Fei Fe-5440a Td1251u

9/13/2019 FEI Model FE-5440A Cesium Beam Primary Frequency Standard
https://freqelec.com/cesium_fe5440a.html Go MAY SEP OCT 👤 ⍰❎

16 captures 15 f 🐦
28 Oct 2015 - 15 Sep 2018 2017 2018 2019 ▾ About this capture
Products
Customer
Service
Cesium Beam Primary Frequency Standard
About Us
A Hybridized High-Reliability Militarized Cesium
Beam Frequency Standard
Military Nomenclature: Master Regulating
Facilities
Clock 0-1824A/U Per MIL-F-28811A
Model FE-5440A, Option FHA
Careers
Features
Easy to use, simple to operate
Extraordinary accuracy and long-term stability
Technical
Literature SC cut quartz crystal to provide stable operation
under dynamic environment
Front panel "C" Field Adjustment
Investor
Relations Hybrid electronics for improved reliability, smaller
size, lower power consumption
Rugged militarized design
News &
May be synchronized externally to any second of
Events
time with a simple control button FE-5440A front and rear view
Fully interchangeable modules
Home
Download FE-5440A PDF Datasheet
Vendors & Function Description Technical Characteristics

Suppliers
The Cesium Beam Primary Frequency Standard is a Accuracy ±3x10-11, maintained over a
rugged, compact, militarized, hybridized, solid-state,
temperature range of –28°C to
Atomic Frequency Standard and Real-Time-Of-Day
+65°C and magnetic fields up to
Clock, which provides
0.2 millitesla (2 gauss) peak.
• Standard and sinusoidal output frequencies of 5 MHz,
Reproducibility ±1x10-11
1 MHz and 100 kHz accurate to ±5x10-12 with a long-
term stablity of ±1x10-11. Settability
±2x10-13
(Frequency)
• Timing information in hours, minutes and seconds, in
Adjustment
the form of a direct real-time LED display Time-Of-Day 6x10-11
Range
Clock, and 1 PPS and PPM outputs for driving external
clocks. Long-Term
±1x10-11
Stability
The Standard owes its exceptional accuracy and stability
to a unique frequency comparison-and-control system, in Short-Term
which a cesium atom (Cs133) is used to stabilize the Stability (5 See Figure 1
output frequency of a precision voltage-controlled quartz MHz)
crystal oscillator using an SC cut crystal to provide Output Voltage 1.0 to 1.5 Vrms into 50 ohms.
improved performance under dynamic environments.
The oscillator output is multiplied and drives the cesium 115 or 230 Vac 10%, single
Voltage
tube. phase. 47 to 100 Hz; or 22 to 30
Required
Vdc.
The signal induces transitions from one hyperfine energy
20 min.from –28°C when
level to another, and those atoms which have undergone Warm-Up Time
operated from ac line
this transition are detected by a hot wire ionizer and
electron multiplier. The output of the electron multiplier is Operating
fed back through an oscillator control circuit to correct Temperature –28°C to + 65°C
the oscillator frequency, maintaining the Standard's Range
accuracy.
Meters and Displays

Major circuit functions and signal levels are metered via
a front panel circuit switch and meter in combination,
providing a convenient means for determining the
operational status of major circuits. Operating status and
alarm LED displays indicate operational status of the
primary and secondary loops.
Compact, Transportable
The Standard may be transported, used as a bench-
mounted unit or rack-mounted by using the four
mounting screws on its front panel.
Three Power Sources

The Standard may be operated from three power
sources:
• First Priority—115 or 230 Vac, 47 to 400 Hz, single-
phase.
• Second Priority—External 22 to 30 Vdc
• Third Priority—Internal battery supply.
Simultaneous connection of ac and dc power source
inputs can be made.
Front panel indicators advise of input power availability,

power source in use, and changing status of the internal
battery supply.
https://web.archive.org/web/20180915191942/https://freqelec.com/cesium_fe5440a.html 1/2
NAVWAR Business Opportunity
NIWC Atlantic - Solicitation
N65236-17-Q-7655 - CESIUM BEAM TUBE FOR MODIFICATION
Close
Subscribe | View Subscribers List
N65236-17-Q-7655
Description
Description: SPAWAR (SSC ATLANTIC) INTENDS TO AWARD A FIRM FIXED PRICE PURCHASE ORDER
ON A SOLE SOURCE BASIS TO MICROSEMI, INC. THE CESIUM BEAM TUBES ARE
NECESSARY FOR PROPER OPERATION OF THE FE-5440A P/N: 21700-6200-4 CESIUM
BEAM FREQUENCY STANDARDS/MASTER REGULATING CLOCKS. THESE FREQUENCY
STANDARDS ARE PRECISION DEVICES AND MICROSEMI IS THE ONLY SOURCE FOR THE
TUBES. IN ORDER TO ENSURE THAT THE TUBES ARE MODIFIED TO THEIR FACTORY
SPECIFIED OPERATING CONDITION, ONLY THE OEM CAN MODIFY THE UNITS FOR THEIR
INTENDED PURPOSES. IF ANY OTHER COMPNAY WAS TO COMPLETE THE
MODIFICATION, IT WOULD CREATE THE RISK THAT THEY WOULD NOT RESTORE THEM
TO THE CORRECT CONDITION IN ORDER TO BE COMPATIBLE WITH FE-5440A P/N
D21700-6200-4 SYSTEM. IF THE GOVERNMENT WAS NOT TO USE THE OEM, IT WOULD
CAUSE THE GOVERNMENT LOST TIME AND ADDITIONAL COSTS WHEN OTHER
COMPANIES ARE UNABLE TO MODIFY THE CESIUM BEAM TUBES.
THIS IS NOT A REQUEST FOR COMPETITIVE QUOTES. IT IS A
NOTICE OF INTENT, HOWEVER, THE GOVERNMENT WILL ACCEPT QUOTES FROM ALL
RESPONSIBLE SMALL BUSINESS SOURCES WITH THE CAPABILITY TO PROVIDE
REQUESTED MODIFICATION OF THE BEAM TUBES AND WILL AWARD A FIRM FIXED
PRICE PURCHASE ORDER RESULTING FROM THIS RFQ TO THE RESPONSIBLE VENDOR
WHOSE OFFER CONFORMS TO THE SOLICITATION AND IS CONSIDERED TO BE MOST
ADVANTAGEOUS TO THE GOVERNMENT. QUOTES ARE DUE BY 18:00 EASTERN
STANDARD TIME ON THURSDAY JULY 27, 2017. NO ELECTRONIC OR HARD COPY RFQ
WILL BE PREPARED OR MADE AVAILABLE FOR DISTRIBUTION. IF QUOTING, PLEASE
SUBMIT QUOTES DIRECTLY TO detra.armstrong@navy.mil. PLEASE STATE YOUR
FEDERAL TAX ID NUMBER, CAGE CODE
AND DUNN AND BRADSTREET NUMBER. PLEASE STATE ANY APPLICABLE SHIPPING
CHARGES FOB DESTINATION AND OR FOB ORIGIN. PLEASE STATE YOUR BEST
DELIVERY ARO. THIS SOLICITAITON IS SET ASIDE FOR SMALL BUSINESSES ONLY. ALL
RESPONDING SMALL BUSINESS VENDORS MUST BE REGISTERED IN SAM PRIOR TO
SUBMITTING INVOICES FOR PAYMENT. INFO TO REGISTER CAN BE FOUND AT
WWW.SAM.GOV. INVOICES MUST BE SUBMITTED ELECTRONICALLY VIA WIDE AREA
WORKFLOW (WAWF). INFO CAN BE FOUND AT WWW.HTTPS://WAWFTRAINING.EB.MIL.
PLEASE QUOTE ON
THE BELOW ATTACHMENT CONSISTING OF REQUESTED ITEMS FOR MODIFICATION. FAR
PART 13 SIMPLITIED ACQUISIITON PROCEDURES APPLY. FAR 13,102(A)(1).FAR 13.105
(B). PLEASE QUOTE ON THE BELOW ATTACHMENT CONSISTING OF REQUESTED ITEMS.
GOVERNMENT FURNISHED MATERIAL WILL BE PROVIDED TO THE CONTRACTOR AND
WILL CONSIT OF MATERIAL IDENTIFIED ON THE BELOW ATTACHMENT AND A
GOVERNMENT FURNISHED PROPERTY FORM WILL ACCOMPANY THE AWARD DOCUMENT.
THE FOLLOWING CLAUSES WILL BE INCORPORATED IN THE AWARD DOCUMENT:
53.245-1 GOVERNMENT PROPERTY
52.245-9 USE AND CHARGES
252.245-7001 TAGGING, LABELING AND MARKING OF GOVERNMENT FURNISHED
PROPERTY
252.245-7002 REPORTING LOSS OF GOVERNMENT PROPERTY
252.245-7003 CONTRACTOR PROPERTY MANAGEMENT SYSTEM ADMINISTRATION
252.245-7004 RE[PRTOMG REUTILIZATION AND DISPOSAL
252.211-7007 REPORTING OF GOVERNMENT FURNISHED PROPERTY
PLEASE QUOTE ON THE BELOW ATTACHMENT CONSISTING OF REQUESTED ITEMS FOR
MODIFICATION.
Show details for Additional Information
Additional Information
Requiring Activity: SSC-C (CHARLESTON)
Technical POC: JERRY LONG

FS Code: 58
View List
NAICS Code: 334419

View List
Type of Action: Other Than Full and Open Competition - Only One Responsible Source of Supply
Recovery Act: No
Estimated Value:
Contact Information
Contracting Officer Name: Detra A Armstrong
Contract Specialist Name: Detra A Armstrong
Contract Specialist Phone Number: Information Unavailable
Contract Specialist E-mail Address: Information Unavailable
Status Information
Status: 4. Evaluation
Current Adjusted Server Time: 3/05/2020 16:28:30 Eastern
Attachments:
RFQMICROSEMI651196Armstrong-Detra-20170722193417.pdf 111 KB
Total 111 KB
Modification Tracking
Created By: Detra A Armstrong
Created Date: 07/23/2017
Last Modified By: System Auto Update
Last Modified Date: 07/27/2017

Performance demonstration of a compact, single optical
frequency Cesium beam clock for space applications
S. Lecomte*, M. Haldimann, R. Ruffieux, P. Berthoud P. Thomann
Observatoire de Neuchâtel Laboratoire Temps-Fréquence, Université de Neuchâtel
Rue de l’Observatoire 58 Rue A.-L. Breguet 1
CH-2000 Neuchâtel, Switzerland CH-2000 Neuchâtel, Switzerland
*Electronic address: steve.lecomte@ne.ch
Observatoire de Neuchâtel has developed a compact optically- distribution and of the 2-fold increase of the useful atoms due
pumped cesium beam frequency standard in the frame of an to optical pumping. It can even compete with the PHM in
ESA-ARTES 5 project. The simplest optical scheme, which is terms of frequency stability. Moreover, due to its inherently
based on a single optical frequency for both preparation and simple design, its manufacturing and its reliability can be
detection processes of atoms, has been chosen to fulfill reliability
strongly improved with respect to the PHM.
constraints of space applications. With the last evolution of our
laboratory demonstrator, we have measured a frequency Observatoire de Neuchâtel (ON) has developed such an
stability of σy=1.14x10-12 τ-1/2, which is compliant with the Optically-pumped Space Cesium Atomic Resonator
Galileo requirement and our frequency stability goal of (OSCAR) prototype in the frame of an ESA-ARTES 5
σy=1x10-12 τ-1/2. Present performance limitations are discussed project. Our goal was to demonstrate a frequency stability of
and further improvements are proposed to possibly increase the σy≤1x10-12 τ-1/2 with a compact atomic resonator and only one
frequency stability. optical frequency. This choice is motivated by space
application prerequisites and has already been discussed [2].
I. INTRODUCTION The best-measured frequency stability with a 1-frequency
Several space applications like navigation systems, scheme is σy=4x10-12 τ-1/2 with a compact laboratory atomic
telecommunications, long-term missions, and scientific resonator and a single laser diode (852 nm, 25-MHz
missions require onboard atomic clocks. The system linewidth) [3]. By also using a compact atomic resonator, but
capability is mainly defined by the atomic clock performance. a more complex optical setup (2-frequency scheme: one
Although standard microwave atomic clock technology is laboratory extended-cavity diode laser and one acousto-optic
well mastered for ground systems, they have to be adapted to modulator), the frequency stability was improved to
the space environment in order to exhibit similar σy=1.4x10-12 τ-1/2 using either a Cs beam [4,5] or a Rb beam
performances with rugged packaging and high system [6].
reliability as well as with strong reductions of mass, volume, In these proceedings, we report on developments of our
and power consumption. compact atomic resonator. First we describe the experimental
The two atomic clocks foreseen to take place onboard the setup (§II), then we present the experimental results and
first generation of Galileo satellites are respectively the discuss the current limitations to the frequency stability
Passive Hydrogen Maser (PHM) and the Rubidium Atomic performances (§III.A and III.B). Finally we present solutions
Frequency Standard (RAFS). While the RAFS is a very to improve the atomic resonator frequency stability in the
compact clock (2.4 litres, 3.4 kg), the PHM is bigger conclusion (§IV).
(26 litres, 18 kg) [1], but exhibits a 5-fold improvement of the II. EXPERIMETAL SETUP
long-term frequency stability σy (< 10-14 for τ > 104 s). These The architecture of the frequency standard is shown in Fig. 1:
two frequency standards being operated in vapour cell the Physics Package (PP) is composed of the Atomic
conditions, the influence of the environment is dramatic Resonator (AR), while the Opto-Electronics Package (OEP)
(frequency temperature coefficient). To overcome the long- contains the Optics and Laser module (OL), the Optics and
term frequency instability of these standards, the atomic- Laser Control module (OLC) and the Atomic Resonator
beam frequency standard is an elegant alternative presently in Control module (ARC).
wide use for GPS and GLONASS. Compared to a
magnetically-deflected Cesium atomic-beam clock, a laser-
pumped resonator has a better short-term stability, owing to
the fact that it makes use of the full atomic velocity
Work sponsored under ESA contract 18040/04/NL/DS
1-4244-0647-1/07/$20.00 ©2007 IEEE 1127

an output power > 15 mW and a spectral linewidth < 2 MHz.
The diverging laser beam is first collimated, then isolated
(40 dB), and finally split into three beams. Neutral density
filters and polarisers properly prepare their respective power
and polarisation before entering the AR by windows. A
homemade external cavity diode laser with a wavelength of
894 nm has also been used in place of the 852-nm DFB laser,
due to the lack of performing DFB laser at this wavelength.
The ARC is composed of the necessary power supplies and
driving electronics for the Cs oven temperature regulation, for
the C-field solenoid current provision and for the ion pump
high voltage supply. The microwave frequency chain (local
oscillator, frequency multiplier and phase modulator) is a
commercial device. It is based on the multiplication of a 10-
MHz quartz oscillator with a step recovery diode. The
frequency locking of the quartz oscillator and of the laser
diode are performed digitally in a single Digital Signal
Processor (DSP). In addition, the C-field amplitude and the
RF interrogation power servo-loops can be implemented
sequentially with the quartz frequency servo loop. Their
different relevant parameters such as modulation amplitudes,
gains, duty cycle and filter topologies can be adapted very
Fig. 1: Block diagram of the Optically-pumped Space Cesium conveniently with this digital electronics. Moreover it offers
Atomic Resonator (OSCAR). The blue module in the middle additional capabilities such as automatic atomic line
is the Physics Package and the orange side modules are parts searching, which however has not yet been implemented.
of the Opto-Electronics Package and of the Atomic Resonator The OLC uses the Cesium beam inside the AR as the
Control module. frequency discriminator (LSU photo-detector). The laser
The AR is fairly standard for an optically-pumped atomic frequency is modulated by its injection current at 10 kHz and
beam resonator. An oven containing 2g of Cesium, fitted with locked on the pumping transition lines Cs D2:44’σ
a multi-channel collimator, produces the atomic beam. It (wavelength ≈852 nm) or Cs D1:34’π (wavelength ≈894 nm)
crosses a first laser beam in the Laser Stabilization Unit by synchronous detection. These hyperfine transitions have
(LSU), which takes advantage of the fluorescence signal to been chosen for their highest population inversion ratio
stabilize the laser frequency. Then, the atomic beam η = 15.5% and η = 14.4% respectively, and for its high
propagates through a second zone to achieve the required fluorescence yield ∆ph = 2.4 ph/at [7]. Note that both
ground state population inversion (Pumping Unit, PU). transitions have nearly the same clock relevant parameters
Subsequently, the atomic beam crosses a short Ramsey cavity (η⋅∆ph).
(12 cm), where the microwave resonance takes place. Finally,
III. EXPERIMENTAL RESULTS
the atomic beam crosses the third laser beam (generated by
the same laser) in the Detection Unit (DU), in which the final Before investigating optically the atomic beam, we have
atomic state is optically probed. In both the LSU and in the calibrated its overall flux with an ionization detector located
DU, the atomic beam fluorescence light is first collected by a downstream of the atomic resonator. Its detection solid angle
specially-machined concave mirror, then converted into is 6x10-5 sr (open surface of 5.5 mm2 at a distance of 301 mm
photo-currents by a Si photodiode (33 mm2) placed under from the oven). The flux has been measured for oven
vacuum, and finally converted into a photo-voltage by a low temperatures from 90°C to 130°C. The measured ion current
noise pre-amplifier outside of the vacuum enclosure. The (proportional to the atomic flux) increases as the calculated
three optical units and the microwave cavity are surrounded Cs vapor density up to 100°C. For higher temperature, the
by a solenoid coil which produces a uniform magnetic C-field atomic flux starts to saturate with respect to the Cs vapor
(up to 100 mG). The AR is maintained under ultra high density, which indicates the transition from molecular flow to
vacuum by Cesium getters (graphite) and by a commercial viscous flow in the collimator. This behavior is similar to the
ion pump. one observed previously [8]. The useful atomic flux for the
The OL has been assembled to operate the atomic resonator atomic resonator (atoms in the mF = 0 hyperfine state having
with the 1-frequency scheme. Note however that it could be contributed to the Ramsey signal) is about 1.5x1010 at/s at the
easily adapted for the 2-frequency scheme, by simply adding oven temperature of 130°C.
an acousto-optic modulator to shift the frequency of the laser The two fluorescence light collection optics with 33-mm2
detection beam. silicon photodiodes convert about 30% of the fluorescence
The optical beam propagates in free space. The single 852- photons into photoelectrons. This collection efficiency is
nm laser source is a Distributed Feed-Back diode (DFB), with presently limited by the size of the photodetectors. It could be
1128
increased by increasing the photodetectors size but at the with the digital electronics, we have measured its frequency
price of more stray light and its associated shot-noise. stability with respect to an active hydrogen maser (Fig. 4).
Birefringent optical elements depolarize the laser beams in The Allan standard deviation extrapolated down to 1s gives at
the PU and the DU [9] for the used D2 transition. A short-term frequency stability of σy=1.51x10-12 τ-1/2. To the
polarization gradient along the atomic beam increases the best of our knowledge, this frequency stability is the best ever
efficiency of the optical pumping in the PU without requiring measured with a compact optically-pumped atomic beam
a strong static magnetic field [10], or a 1D optical molasses frequency standard operated with a single optical frequency
[11]. scheme (single laser diode and no acousto-optic modulator).
The optical power in both the PU and DU was optimized for For long integration time constants (from 4000s),
maximum signal-to-noise ratio (SNR) in order to increase the environmental magnetic and thermal perturbations degrade
frequency stability of the standard. the clock frequency stability. For this laboratory
A. Results obtained with the D2:44’σ transition demonstration, we remind that only one single magnetic
shield is assembled. Moreover, neither C-field nor RF power
The central Ramsey fringe recorded at an oven temperature of
servo loops are operating.
130°C is plotted in Fig. 2. A weak and uniform C-field of
52 mG is applied over all optical units (LSU, PU, and DU)
and over the microwave Ramsey cavity (Fig. 1). The RF
power is adjusted for maximizing the clock signal. The AR
has the following performances: the atomic linewidth is
858 Hz, yielding an atomic quality factor of 107, the Ramsey
central fringe peak-to-valley photo-current is 1681 pA, the σy = 1.51x10-12 τ-1/2
background photo-current is 2.5 nA, and the noise current
density at resonance is 41 fA/Hz1/2 (at a Fourier frequency of
27 Hz, see Fig. 3) yielding a clock SNR of 41’000 Hz1/2 at
clock resonance. Slope∼τ-1/2
4.40E-009
4.20E-009
4.00E-009 Toven=130°C, D2:44'σ

3.80E-009
Photocurrent (A)
3.60E-009
3.40E-009 Fig. 4: Measured frequency stability of OSCAR operated

858 Hz
3.20E-009 with a single wavelength and a 852-nm DFB laser.
3.00E-009
2.80E-009 1681 pA Although the stability of the clock is better than the
2.60E-009 requirement for the Galileo satellite navigation system (σy
2.40E-009
≤3x10-12 τ-1/2), the ultimate frequency stability goal of σy
-1500 -1000 -500 0 500
Frequency detuning (Hz)
1000 1500
≤1x10-12 τ-1/2 has been nearly reached. In fact, the clock noise
is not atomic shot-noise limited. In the following we will
analyze the various noise contributions to the total clock
Fig. 2: Ramsey central fringe recorded with an oven noise and discuss the improvements to be performed in order
temperature of 130°C. to increase the clock SNR and reach the atomic shot-noise
1E-12 limit.
In Fig. 5, we plot the different noise current densities as a

Current noise density (A/Hz )
1/2
function of the Ramsey peak-valley photocurrent. The

different noise measurements and the optimizations have
1E-13 been performed at the Ramsey central fringe maximum. The
corresponding oven temperatures are given for information.
1/2
41 fA/Hz
1E-14
1 10 100
Fourier frequency (Hz)
Fig. 3: Current noise density (measured at Ramsey fringe

peak) recorded with an oven temperature of 130°C.
By operating the atomic resonator with these optimal
parameters, and frequency locking the quartz local oscillator
1129
1E-13 100000
placed under vacuum are sufficiently absorbent, the laser
φ=1.02·10
10
at/s, 120°C
10
φ=1.57·10 at/s, 130°C scattering on the optical windows of the vacuum enclosure
Current noise density (A/Hz )
1/2
9
φ=6.3·10 at/s, 110°C
9
induces a large amount of stray light. New windows of better
φ=3.6·10 at/s, 100°C
9
φ=2.0·10 at/s, 90°C optical quality with efficient anti-reflection (AR) coating will
be mounted and should strongly reduce the amount of stray
SNRRam (Hz )
1E-14 light. By effectively reducing these two major noise
contributions, the total noise of the clock should be close to
10000
1/2
Total noise
Atomic shot-noise
the atomic shot-noise contribution.
Stray light noise
Fluorescence noise
Photodetector noise
B. Results obtained with the D1:34’π transition
SNRRam
1E-15
The characterization of the resonator has been repeated at
1E-10 1E-9 1E-8 894 nm with a C-field of 83 mG. In this case an extended
Ramsey peak-valley photocurrent (A) cavity diode laser with a narrow linewidth was used. By
Fig. 5: Noise budget and overall SNR of OSCAR as function operating the oven at a temperature of 130°C, we obtained a
of the Ramsey peak-valley photocurrent. SNR of 43’000 Hz1/2. This narrower laser (compared to the
DFB laser) allows to injecting less optical power in the
We have identified four noise contributions in the clock detection unit. This reduces the stray light associated noise
signal noise budget: and the laser induced fluorescence noise. The clock SNR is
1. Photodetector noise: it is measured in the dark and is therefore increased since the Ramsey peak-valley is nearly
independent of the Ramsey photocurrent; for the same as with the DFB laser.
nominal operating atomic flux, this contribution is
negligible in the overall noise budget. The frequency stability of 1.14x10-12 τ-1/2 measured against a
2. Stray light noise: it is measured with the nominal hydrogen maser and presented in Fig. 6 is, to our knowledge,
optical power but off-resonance, is proportional to the best stability ever measured for such a resonator. The
the square root of the DC stray light level (shot displayed frequency stability is significantly degraded for
noise), but is independent of the Ramsey integration times τ < 100 s. This is partially due to the C-field
photocurrent; this technical noise contribution is current source which was not actively stabilized (it was not
mainly due to light scattering by the optical the case for the stability presented in Fig. 4). For long
windows mounted on the vacuum enclosure. The integration times τ > 1000 s, environmental perturbations
optical power has been optimized for each oven become visible.
temperature and at the peak of the Ramsey central
fringe.
3. Atomic shot-noise: it is calculated as the square root
of the atomic signal; presently this noise
contribution is largely dominated by the technical σy = 1.14x10-12 τ-1/2
noise.
4. Fluorescence noise: it is computed from the total
measured noise minus the other contributions in
RMS values. This noise contribution appears to be
roughly proportional to the square root of the
Ramsey signal and is related to the residual
frequency noise of the laser converted into Slope∼τ-1/2
amplitude by the atomic beam frequency
discriminator. In order to reduce it, the residual
frequency noise of the laser has to be minimized in
closed loop operation. Presently, our laser
stabilization electronics has a very narrow servo- Fig. 6: Measured frequency stability of OSCAR operated
loop bandwidth (<100 Hz), limited by the with a single wavelength and a 894-nm extended-cavity diode
ADC/DAC stage of our digital electronics. By laser.
increasing it by a factor 100, the amount of
fluorescence noise should be reduced to an IV. CONCLUSIONS AND OUTLOOK
acceptable value. Another alternative, demonstrated
We have reported in these proceedings on the experimental
here, is to optimize the optical power as function of
setup of an optically-pumped cesium beam frequency
the atomic flux. An optimum of the optical power, in
standard. Its concept relies on the simplest optical scheme, in
terms of clock SNR, is found for each atomic flux.
which we use only one optical frequency (one laser, no
AOM). By depolarizing the laser beams (only for the D2
The second major noise contribution arises from the stray
transition) for the optical pumping processes (preparation and
light level and its associated shot-noise. While the light traps
detection), we can use a single, uniform and weak magnetic
1130
C-field for the optical units without trapping atoms in dark Time and Frequency Forum (EFTF), 1994, Munich,
states. A fully digital electronics based on a DSP processor Germany.
has been specially developed for frequency locking both the [4] R. Lutwak et al., “Optically-pumped cesium beam
laser and the quartz local oscillator. This digital electronics frequency standard for GPS-III”, 33rd Annual Precise
also allows to sequentially locking the magnetic C-field and Time and Time Interval (PTTI) Meeting, 2001, Long
the RF injection power in the Ramsey cavity although it has Beach, CA, USA.
not been used in the present measurements. [5] S. Guérandel et al., “Compact cesium beam frequency
We have demonstrated what is, to the best of our knowledge, standard: improvements of the frequency stability
the best ever-measured frequency stability of towards the 10-12 t-1/2 level”, 16th European Time and
σy=1.14x10-12 τ-1/2 with a compact optically-pumped atomic Frequency Forum, 2002, St-Petersburg, Russia.
beam frequency standard operated with a single optical [6] A. Besedina et al., “Preliminary results of investigation
frequency scheme. Although the demonstrated frequency of the high-stable Rudidium atomic beam frequency
stability is sufficient for Galileo, the clock SNR is not yet standard with laser pumping/detection for space”, 20th
limited by the atomic shot-noise. The noise budget has European Time and Frequency Forum, 2006,
identified the two major noise sources: the “fluorescence Braunschweig, Germany.
noise” and the “stray light noise”. While the former could be [7] N. Dimarcq et al., “Comparison of pumping a cesium
reduced by increasing the laser frequency servo loop beam tube with D1 and D2 lines”, J. Appl. Phys., 69 (3),
bandwidth, the later will call for top quality optical windows. p. 1158-1162 (1991).
The marginal fluorescence light collection efficiency could be [8] S. Lecomte et al., “Development of a single-frequency
increased by implementing new and optimize optical optically-pumped cesium beam resonator for space
collectors with possibly larger photodetectors. The results of applications”, 20th European Time and Frequency
this study will be implemented in the design of the Forum, 2006, Braunschweig, Germany.
breadboard Cesium resonator for the second generation of [9] P. Berthoud and P. Thomann, patent number WO
Galileo onboard clocks. This new project is currently ongoing 2006/084829 A1.
with the financial support of ESA within a consortium led by [10] V. Giordano et al., “New design for a high performance
Thalès Electron Devices, France, and with the following optically pumped cesium beam tube”, IEEE
partners: Oerlikon Space AG, Switzerland, SYstème de Transactions on Ultrasonics, Ferroelectrics, and
Référence Temps Espace, France and Observatoire de Frequency Control, 38 (4), p. 350-357 (1991).
Neuchâtel [12, 13]. [11] J. H. Shirley and R. E. Drullinger, “Zeeman coherences
and dark states in optically pumped cesium frequency
REFERENCES standards”, IEEE Prec. Electromagn. Meas., 1994,
[1] F. Droz et al., “The On-board Galileo Clocks: Rubidium Boulder, CO, USA.
Standard and Passive Hydrogen Maser -Current Status [12] S. Guerandel et al., “In-depth analysis of the frequency
and Performance-“, 20th European Time and Frequency analysis of optically pumped cesium beam frequency
Forum, 2006, Braunschweig, Germany. standards”, 21th European Time and Frequency Forum,
[2] P. Berthoud et al., “A feasibility study of an optically- 2007, Geneva, Switzerland, in these proceedings.
pumped cesium beam resonator for space applications”, [13] V. Hermann et al., “OSCC project: a space Cs beam
19th European Time and Frequency Forum, 2005, optically pumped atomic clock for Galileo”, 21th
Besançon, France. European Time and Frequency Forum, 2007, Geneva,
[3] P. Petit et al., “Performance of a 2 dm3 optically pumped Switzerland, in these proceedings.
cesium beam tube: a progress report”, 8th European
1131
3/6/2020 Linked File: fl0072019.php
Bid Protests &

Reports & Testimonies Key Issues About GAO Multimedia
Appropriations Law
B-204483, APR 5, 1982
B-204483, APR 5, 1982

DIGEST: 1. GAO WILL NOT QUESTION CONTRACTING AGENCY'S SOLE-SOURCE AWARD WHERE AGENCY HAS SHOWN THAT TIME WAS OF THE ESSENCE AND ONLY ONE
KNOWN SOURCE COULD MEET ITS NEEDS WITHIN THE TIMEFRAME REQUIRED. 2. CONTRACTING AGENCY SHOULD ENSURE THAT PROTESTER AND OTHER POTENTIAL
SUPPLIERS ARE GIVEN OPPORTUNITY TO HAVE THEIR EQUIPMENT FULLY TESTED AND HAVE OPPORTUNITY TO PARTICIPATE IN COMPETITIVE PROCUREMENT WHICH
AGENCY PROPOSES TO SCHEDULE.
FREQUENCY ELECTRONICS, INC.:
FREQUENCY ELECTRONICS, INC. (FEI), PROTESTS THE AWARD OF A CONTRACT TO HEWLETT-PACKARD COMPANY (HPC) ON A SOLE-SOURCE BASIS UNDER REQUEST
FOR PROPOSALS (RFP) NO. N00039-81-R-0587, ISSUED BY THE NAVAL ELECTRONICS SYSTEMS COMMAND (NAVY).
THE RFP SOLICITED 37 "CESIUM BEAM FREQUENCY STANDARDS" (CBFS) WITH AN OPTION FOR TWO ADDITIONAL UNITS. ACCORDING TO THE NAVY, THE CBFS IS USED
AS THE "PRIMARY TIME AND FREQUENCY REFERENCE FOR THE VERDIN DIGITAL DATA COMMUNICATION SYSTEM - A VERY LOW FREQUENCY/LOW FREQUENCY SHORE-
TO-SHIP AND AIR-TO-SHIP COMMUNICATIONS SYSTEM THAT SUPPORTS THE SECURE COMMAND AND CONTROL REQUIREMENTS OF THE SUBMARINE FORCES AND
SUPPORT ELEMENTS." FEI ARGUES THAT THE NAVY HAD NO BASIS TO PROCURE THESE ITEMS ON A SOLE-SOURCE BASIS SINCE FEI IS ALSO ALLEGEDLY CAPABLE OF
SUPPLYING EQUIPMENT THAT WILL MEET THE NAVY'S MINIMUM NEEDS. FEI WANTS THE NAVY TO CONDUCT A COMPETITIVE PROCUREMENT FOR THESE ITEMS.
WE FIND THAT, UNDER THE CIRCUMSTANCES, THE NAVY WAS JUSTIFIED IN AWARDING A SOLE-SOURCE CONTRACT TO HPC, BUT WE RECOMMEND THAT THE NAVY TAKE
THE STEPS IT HAS OUTLINED IN ITS REPORT ON THE PROTEST TO ENSURE THAT FUTURE PROCUREMENTS ARE CONDUCTED ON A COMPETITIVE BASIS.
HPC IS THE ORIGINAL DEVELOPER AND THE ONLY NAVY-APPROVED PRODUCER OF THE CBFS MODEL WHICH THE NAVY REQUIRES. HPC'S EQUIPMENT WAS CHOSEN
AFTER A TECHNICAL COMPETITION AND HAS BEEN USED BY THE NAVY SINCE 1974. ACCORDING TO THE NAVY, HPC'S EQUIPMENT HAS PASSED THE REQUIRED "SHOCK,
VIBRATION, ENVIRONMENTAL, RELIABILITY, AND MAINTAINABILITY TESTS," WAS SUBJECTED TO FOLLOW-ON TESTING IN JULY 1979, AND MEETS THE REQUIREMENTS OF
THE APPLICABLE MILITARY SPECIFICATION - MIL-F-28811(EC) - DETERMINED BY THE NAVY TO REFLECT ITS MINIMUM NEEDS.
FEI ARGUES THAT ITS CBFS WILL MEET THE NAVY'S NEEDS AND PASS ALL NECESSARY TESTS. ACCORDING TO FEI, IN JANUARY 1981, IT OFFERED TO SUBMIT ITS CBFS
TO THE NAVY FOR TESTING, BUT THE NAVY TURNED DOWN FEI'S REQUEST. FEI POINTS OUT THAT IT HAS SUPPLIED A SIMILAR CBFS TO THE AIR FORCE AND THAT
PRIOR TO THE PRODUCT'S ACCEPTANCE, THIS CBFS WAS FULLY TESTED BY THE AIR FORCE. FEI FURTHER STATES THAT IT IS WILLING TO PROVIDE THE NAVY WITH AN
ADDITIONAL NUMBER OF UNITS IN ORDER TO EXPEDITE THE NAVY'S TESTING PROCESS. FEI ALSO NOTES THAT HPC HAS INFORMED THE NAVY THAT IT WILL SOON
STOP PRODUCING CBFS'S.IN FEI'S OPINION, ALL THESE FACTORS TAKEN TOGETHER INDICATE THAT THE NAVY HAD NO REASONABLE BASIS FOR THE SOLE-SOURCE
AWARD.
PRIOR TO THE SUBMISSION OF ITS PROTEST REPORT TO OUR OFFICE, THE NAVY AWARDED A CONTRACT TO HPC ON THE GROUNDS THAT, DUE TO THE URGENCY OF
THE REQUIREMENT, THE BEST INTERESTS OF THE GOVERNMENT WOULD NOT BE SERVED BY DELAYING THE AWARD UNTIL THE PROTEST WAS RESOLVED. THUS, AN
AWARD TO HPC WAS MADE ON SEPTEMBER 30, 1981.
AS TO THE MERITS OF FEI'S PROTEST, THE NAVY MAINTAINS THAT THE SOLE SOURCE AWARD TO HPC WAS JUSTIFIED SINCE FEI HAS FAILED TO SHOW THAT THIS
DECISION WAS UNREASONABLE. ACCORDING TO THE NAVY, THERE WAS INADEQUATE TIME TO CONDUCT NECESSARY TESTS ON FEI'S EQUIPMENT, DESPITE FEI'S
OFFER TO PROVIDE AN ADDITIONAL NUMBER OF UNITS IN ORDER TO SPEED UP THE TESTING PROCESS. THE NAVY FURTHER NOTES THAT THE MODEL (FE 5440A) FEI
HAS FURNISHED THE AIR FORCE DIFFERS FROM THE MODEL (FE 5440A, OPTION "F") THAT IT INTENDS TO FURNISH THE NAVY. MOREOVER, THE NAVY POINTS OUT THAT
THE TESTS WHICH THE AIR FORCE CONDUCTED ON FEI'S EQUIPMENT ALSO DIFFER FROM THE TESTS REQUIRED BY THE NAVY. FOR EXAMPLE, THE NAVY REQUIRES A
"FIXED- LENGTH RELIABILITY TEST" THAT TOTALS 9,750 HOURS, WHILE UNDER THE AIR FORCE'S FIRST ARTICLE TESTING THE TEST IS A "MEAN-TIME-BETWEEN-
FAILURES" TEST WHICH ONLY REQUIRES TESTING FOR 500 HOURS.
THE NAVY ACKNOWLEDGES THAT HPC INTENDS TO STOP PRODUCING CBFS'S IN THE NEAR FUTURE. BECAUSE OF THIS DEVELOPMENT, THE NAVY STATES THAT IT HAS
FORMULATED A COMPETITIVE PROCUREMENT SCHEDULE WHICH IT PLANS TO IMPLEMENT DURING THE CURRENT FISCAL YEAR FOR APPROXIMATELY 60 TO 80 CBFS'S
REPRESENTING THE NEEDS OF FUTURE YEARS. ACCORDING TO THE NAVY, THERE WILL BE SUFFICIENT TIME UNDER THIS PROPOSED PROCUREMENT FOR ALL
REQUIRED TESTING.
WE HAVE HELD THAT, BECAUSE OF THE REQUIREMENT FOR MAXIMUM PRACTICAL COMPETITION IN THE CONDUCT OF GOVERNMENT PROCUREMENTS, AGENCY
DECISIONS TO PROCURE FROM A SOLE SOURCE MUST BE ADEQUATELY JUSTIFIED AND ARE SUBJECT TO CLOSE SCRUTINY. SUCH DECISIONS, HOWEVER, WILL BE
UPHELD IF THERE IS A REASONABLE OR RATIONAL BASIS FOR THEM. EMI MEDICAL, INC.; PICKER CORPORATION, B-195487, FEBRUARY 6, 1980, 80-1 CPD 96.
WE HAVE ALSO HELD THAT THE DETERMINATION OF THE NEEDS OF THE GOVERNMENT AND THE BEST METHODS OF ACCOMMODATING THOSE NEEDS ARE PRIMARILY
THE RESPONSIBILITY OF THE PROCURING AGENCIES. MORE SPECIFICALLY, WE HAVE RECOGNIZED THAT THE GOVERNMENT PROCUREMENT OFFICIALS, WHO ARE
FAMILIAR WITH THE CONDITIONS UNDER WHICH SUPPLIES, EQUIPMENT, OR SERVICES HAVE BEEN USED IN THE PAST AND HOW THEY ARE TO BE USED IN THE FUTURE,
ARE GENERALLY IN THE BEST POSITION TO KNOW THE GOVERNMENT'S ACTUAL NEEDS. CONSEQUENTLY, WE WILL NOT QUESTION AN AGENCY'S DETERMINATION OF
ITS ACTUAL MINIMUM NEEDS UNLESS THERE IS A CLEAR SHOWING THAT THE DETERMINATION HAS NO REASONABLE BASIS. FENWAL, INC., B-202283, DECEMBER 15,
1981, 81-2 CPD 469.
WITH RESPECT TO THE NEED FOR A CONTRACTOR TO HAVE ITS PRODUCT QUALIFIED BEFORE THAT PRODUCT CAN BE CONSIDERED FOR AN AWARD, WE HAVE
RECOGNIZED THAT THE RESPONSIBILITY FOR THE ESTABLISHMENT OF TESTS AND PROCEDURES NECESSARY TO DETERMINE PRODUCT ACCEPTABILITY IS WITHIN THE
EXPERTISE OF THE CONTRACTING ACTIVITY. TYCO, B-199632, MARCH 24, 1981, 81-1 CPD 220. HOWEVER, AN AGENCY MAY NOT ARBITRARILY REFUSE TO TEST A FIRM'S
PRODUCT, BUT MUST TAKE REASONABLE STEPS TO DETERMINE IF A PROPOSED ALTERNATE WILL SATISFY ITS MINIMUM NEEDS. SEE, E.G., CASTOLEUM CORPORATION,
B-195724, NOVEMBER 29, 1979, 79-2 CPD 381.
AS INDICATED ABOVE, THE NAVY BASED ITS DECISION TO MAKE A SOLE SOURCE AWARD ON THE URGENCY OF THE REQUIREMENT AND THE FACT THAT, SINCE ITS
EQUIPMENT WAS FULLY TESTED, ONLY HPC COULD MEET ITS NEED WITHIN THE REQUIRED TIME. WE HAVE HELD THAT A SOLE-SOURCE AWARD CAN BE JUSTIFIED ON
THE GROUNDS THAT TIME IS OF THE ESSENCE AND ONLY THE KNOWN SOURCE CAN MEET THE REQUIRED TIMEFRAME, PROVIDED THAT THE AGENCY DEMONSTRATES
THAT THESE CIRCUMSTANCES DO IN FACT EXIST. MCDONNELL DOUGLAS CORPORATION, B-202904, AUGUST 18, 1981, 81-2 CPD 154.
https://www.gao.gov/15/fl0072019.php 1/2
3/6/2020 Linked File: fl0072019.php
HERE, THE NAVY HAS PROVIDED US WITH A DETAILED BREAKDOWN OF WHERE EACH OF THE 39 CBFS'S WAS NEEDED. IT ARGUES THAT IF THESE UNITS ARE NOT
DELIVERED WITHIN THE TIMEFRAME OF THE SOLE-SOURCE AWARD, THIS CIRCUMSTANCE WILL "ADVERSELY IMPACT ON VITAL NATIONAL DEFENSE PROGRAMS." FOR
EXAMPLE, "TRIDENT SUBMARINES WILL BE PREVENTED FROM DEPLOYING," SUPPORT FOR THE "GLOBAL POSITIONING SATELLITE" PROGRAM WILL NOT BE PROVIDED
AS REQUIRED, TRAINING IN PROPER OPERATION OF EQUIPMENT WILL BE DELAYED, AND SUBMARINE CONSTRUCTION AND OVERHAUL SCHEDULES WILL ALSO BE
DELAYED.
IN EXPLAINING WHY QUALIFICATION OF FEI'S PRODUCT IN 1981 WOULD HAVE RESULTED IN UNACCEPTABLE DELAY, THE NAVY STATES THAT FEI'S OFFER OF PROVIDING
EXTRA UNITS AS A MEANS OF SPEEDING UP THE TESTING PROCESS WOULD NOT HAVE BEEN MATERIALLY HELPFUL SINCE, REGARDLESS OF THE NUMBER OF UNITS
AVAILABLE FOR THE RELIABILITY TEST, EACH UNIT HAD TO MEET THE MINIMUM TEST TIME OF 3 CALENDAR MONTHS, OR 2,500 TEST HOURS, "WHICH DOES NOT
INCLUDE REPAIR AND RENTAL TIME"; MOREOVER, AN ADDITIONAL 30 DAYS WOULD HAVE BEEN REQUIRED FOR A "MAINTAINABILITY DEMONSTRATION." ALSO, THE NAVY
INSISTS THAT IN ADDITION TO THIS TEST TIME, ADDITIONAL "ADMINISTRATIVE LEAD TIME WOULD HAVE BEEN REQUIRED TO GENERATE TEST PROCEDURE AND FEI
WOULD HAVE HAD TO FABRICATE AND DELIVER THE REQUIRED UNITS WITH SPARE PARTS AND DOCUMENTATION TO SUPPORT THE TESTS."
THE NAVY ALSO INSISTS THAT THE AIR FORCE TESTING OF FEI'S SIMILAR MODEL TO AN "AIRBORNE" SPECIFICATION - MIL-E-5400 - CANNOT BE CONSIDERED TO BE AN
ACCEPTABLE SUBSTITUTE FOR THE TESTING TO THE NAVY'S "SHIPBOARD-USE" SPECIFICATION - MIL-F-28811(EC); MOREOVER, THE NAVY STATES THAT THE "AIR FORCE
HAS EXPERIENCED RELIABILITY PROBLEMS WITH FEI'S EQUIPMENT" AND THAT THE "AIR FORCE IS ATTEMPTING TO RESOLVE THESE DEFICIENCIES."
FINALLY, FEI ALLEGES THAT THE NAVY'S INITIAL PROTEST REPORT INDICATED THAT HPC'S EQUIPMENT HAD NOT BEEN FULLY TESTED UNTIL JULY 1979, EVEN THOUGH
THE EQUIPMENT HAD BEEN USED SINCE 1975. HOWEVER, IN ITS SUPPLEMENTAL PROTEST REPORT, THE NAVY MAINTAINS THAT IT NEVER MEANT TO IMPLY THAT HPC'S
EQUIPMENT WAS USED BEFORE IT WAS FULLY TESTED TO MIL-F- 28811(EC). ACCORDING TO THE NAVY, FIRST ARTICLE TESTING TO THE SPECIFICATION ON HPC'S
EQUIPMENT WAS COMPLETED IN APRIL 1975 AND THE MOST RECENT FOLLOW-ON TESTING ON HPC'S EQUIPMENT WAS CONDUCTED IN JULY 1979. IN VIEW OF THIS
EXPLANATION, WE SEE NO NEED TO CONSIDER THIS ISSUE FURTHER.
IN VIEW OF THE FOREGOING, WE CANNOT QUESTION THE NAVY'S JUSTIFICATION FOR ITS SOLE-SOURCE AWARD TO HPC ON THE GROUNDS THAT TIME WAS OF THE
ESSENCE AND ONLY ONE KNOWN SOURCE - HPC - COULD MEET ITS NEEDS WITHIN THE REQUIRED TIMEFRAME. MCDONNELL DOUGLAS CORPORATION, SUPRA.
SPECIFICALLY, WE CANNOT QUESTION THE NAVY'S POSITION THAT DELIVERY "SCHEDULE CONSTRAINTS FOR CRITICAL NEAR-TERM CBFS UNITS PREVENTED THE
NAVY" - GIVEN THE PROJECTED TIME AND UNCERTAINTIES INVOLVED IN FABRICATING AND IN TESTING FEI'S PRODUCT - FROM ACCEPTING FEI'S JANUARY 1981
QUALIFICATION TEST OFFER.
NEVERTHELESS, WE NOTE THAT THE NAVY PLANS TO IMPLEMENT A COMPETITIVE PROCUREMENT SCHEDULE IN THE NEAR FUTURE WHICH WILL PROVIDE ADEQUATE
TIME FOR TESTING. WE ALSO NOTE THAT ANOTHER FIRM HAS EXPRESSED AN INTEREST IN PARTICIPATING IN ANY FUTURE PROCUREMENT. THEREFORE, THE NAVY
SHOULD ENSURE, AS MUCH AS PRACTICAL, THAT ALL POTENTIAL SOURCES OF CBFS'S ARE NOTIFIED OF THE UPCOMING PROCUREMENT AND ARE GIVEN AN
OPPORTUNITY TO HAVE THEIR EQUIPMENT TESTED IF THEY WISH.
PROTEST DENIED.
https://www.gao.gov/15/fl0072019.php 2/2
SLCET-TR-88-1 (Rev. 8.5.2.2) AD-M001251
Quartz Crystal Resonators and

Oscillators
For Frequency Control and Timing Applications - A Tutorial
March 2004
John R. Vig
US Army Communications-Electronics Research, Development & Engineering Center
Fort Monmouth, NJ, USA
J.Vig@IEEE.org
Approved for public release.
Distribution is unlimited
“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

For Frequency Control and Timing Applications - A TUTORIAL”
Rev. 8.5.1.2, by John R. Vig, July 2001, AD-M001251.
NOTICES
Disclaimer
The findings in this report are not to be construed as an

official Department of the Army position, unless so
designated by other authorized documents.
The citation of trade names and names of manufacturers

in this report is not to be construed as official Government
endorsement or consent or approval of commercial
products or services referenced herein.

Table of Contents
Preface………………………………..……………………….. v
1. Applications and Requirements………………………. 1
2. Quartz Crystal Oscillators………………………………. 2
3. Quartz Crystal Resonators……………………………… 3
4. Oscillator Stability………………………………………… 4
5. Quartz Material Properties……………………………... 5
6. Atomic Frequency Standards…………………………… 6
7. Oscillator Comparison and Specification…………….. 7
8. Time and Timekeeping…………………………………. 8
9. Related Devices and Applications……………………… 9
10. FCS Proceedings Ordering, Website, and Index………….. 10
iii

Preface
Why This Tutorial?
“Everything should be made as simple as I was frequently asked for “hard-copies”
possible - but not simpler,” said Einstein. The of the slides, so I started organizing, adding
main goal of this “tutorial” is to assist with some text, and filling the gaps in the slide
presenting the most frequently encountered collection. As the collection grew, I began
concepts in frequency control and timing, as receiving favorable comments and requests
simply as possible. for additional copies. Apparently, others, too,
found this collection to be useful. Eventually, I
I have often been called upon to brief assembled this document, the “Tutorial”.
visitors, management, and potential users of
precision oscillators, and have also been This is a work in progress. I plan to
invited to present seminars, tutorials, and include new material, including additional
review papers before university, IEEE, and notes. Comments, corrections, and
other professional groups. In the beginning, I suggestions for future revisions will be
spent a great deal of time preparing these welcome.
presentations. Much of the time was spent on
preparing the slides. As I accumulated more John R. Vig
and more slides, it became easier and easier
to prepare successive presentations.
iv

Notes and References
In the PowerPoint version of this document, notes and references
can be found in the “Notes” of most of the pages. To view the notes,
use the “Notes Page View” icon (near the lower left corner of the
screen), or select “Notes Page” in the View menu. In PowerPoint
2000 (and, presumably, later versions), the notes also appear in the
“Normal view”.
To print a page so that it includes the notes, select Print in the
File menu, and, near the bottom, at “Print what:,” select “Notes
Pages”.
The HTML version can be viewed with a web browser (best
viewed at 1024 x 768 screen size). The notes then appear in the
lower pane on the right.
Many of the references are to IEEE publications that are available
online in the IEEE UFFC-S digital archive, www.ieee-uffc.org/archive
or in IEEE Xplore, http://www.ieee.org/ieeexplore .
IEEE Xplore, <http://www.ieee.org/ieeexplore>, provides full-text access to IEEE transactions,

journals, magazines and conference proceedings published since 1988 and to all current IEEE
standards. Access to tables of contents of IEEE transactions, journals, magazines, conference is free
to all. IEEE members can search and access all IEEE abstract/citation records and more...

CHAPTER 1
Applications and Requirements
Quartz for the National Defense Stockpile, Report of the Committee on Cultured Quartz for the
National Defense Stockpile, National Materials Advisory Board Commission on Engineering and
Technical Systems, National Research Council, NMAB-424, National Academy Press, Washington,
D.C., 1985.
J. R. Vig, "Military Applications of High Accuracy Frequency Standards and Clocks," IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 40, pp. 522-527, 1993.

Electronics Applications of Quartz Crystals
Military & Aerospace Industrial Consumer
Communications Communications Watches & clocks
Navigation Telecommunications Cellular & cordless
IFF Mobile/cellular/portable phones, pagers
Radar radio, telephone & pager Radio & hi-fi equipment
Sensors Aviation Color TV
Guidance systems Marine Cable TV systems
Fuzes Navigation Home computers
Electronic warfare Instrumentation VCR & video camera
Sonobouys Computers CB & amateur radio
Digital systems Toys & games
Research & Metrology CRT displays Pacemakers
Atomic clocks Disk drives Other medical devices
Instruments Modems
Astronomy & geodesy Tagging/identification Automotive
Space tracking Utilities Engine control, stereo,
Celestial navigation Sensors clock
Trip computer, GPS
1-1
Technical Systems, National Research Council, NMAB-424, National Academy Press, Washington,
D.C., 1985.

Frequency Control Device Market
(as of ~2001)
Technology Units Unit price, Worldwide

per year typical market, $/year
9
Quartz Crystal ~ 2 x 10 ~$1 ~$1.2B
($0.1 to 3,000)
Atomic Frequency Standards

(see chapter 6)
Hydrogen maser ~ 10 $200,000 $2M
Cesium beam ~ 500 $50,000 $25M

frequency standard
Rubidium cell ~ 60,000 $2,000 $120M
frequency standard
1-2
The units per year are estimates based on informal surveys of industry leaders. The
numbers are probably accurate to better than a factor of two.

Navigation
Precise time is essential to precise navigation. Historically, navigation has
been a principal motivator in man's search for better clocks. Even in ancient times,
one could measure latitude by observing the stars' positions. However, to determine
longitude, the problem became one of timing. Since the earth makes one revolution
in 24 hours, one can determine longitude form the time difference between local time
(which was determined from the sun's position) and the time at the Greenwich
meridian (which was determined by a clock):
Longitude in degrees = (360 degrees/24 hours) x t in hours.
In 1714, the British government offered a reward of 20,000 pounds to the first
person to produce a clock that allowed the determination of a ship's longitude to 30
nautical miles at the end of a six week voyage (i.e., a clock accuracy of three
seconds per day). The Englishman John Harrison won the competition in 1735 for
his chronometer invention.
Today's electronic navigation systems still require ever greater accuracies. As
electromagnetic waves travel 300 meters per microsecond, e.g., if a vessel's timing
was in error by one millisecond, a navigational error of 300 kilometers would result.
In the Global Positioning System (GPS), atomic clocks in the satellites and quartz
oscillators in the receivers provide nanosecond-level accuracies. The resulting
(worldwide) navigational accuracies are about ten meters (see chapter 8 for further
details about GPS).
1-3
Dava Sobel, Longitude, Walker & Co., New York, 1995

Commercial Two-way Radio
Historically, as the number of users of commercial two-way radios
have grown, channel spacings have been narrowed, and higher-
frequency spectra have had to be allocated to accommodate the
demand. Narrower channel spacings and higher operating frequencies
necessitate tighter frequency tolerances for both the transmitters and the
receivers. In 1940, when only a few thousand commercial broadcast
transmitters were in use, a 500 ppm tolerance was adequate. Today,
the oscillators in the many millions of cellular telephones (which operate
at frequency bands above 800 MHz) must maintain a frequency
tolerance of 2.5 ppm and better. The 896-901 MHz and 935-940 MHz
mobile radio bands require frequency tolerances of 0.1 ppm at the base
station and 1.5 ppm at the mobile station.
The need to accommodate more users will continue to require higher
and higher frequency accuracies. For example, a NASA concept for a
personal satellite communication system would use walkie-talkie-like
hand-held terminals, a 30 GHz uplink, a 20 GHz downlink, and a 10 kHz
channel spacing. The terminals' frequency accuracy requirement is a
few parts in 108.
1-4
R. Kinsman and D. Gunn, "Frequency Control Requirements for 800 MHz Land Mobile
Communication," Proc. 35th Ann. Symp. Frequency Control, pp. 501-510, 1981.

Digital Processing of Analog Signals
The Effect of Timing Jitter
(A)
Analog* A/D
A/D Digital
Digital D/A
D/A Analog
input converter
converter processor
processor converter
converter output
Digital
* e.g., from an antenna output
(B) (C)
V(t)
V(t)
Time ∆V
Analog signal Digitized signal ∆t
1-5
As microprocessor and digital signal processing (DSP) chips become more and more capable,
the digital processing of analog signals, as illustrated in (A) above, becomes more and more
advantageous and feasible. Among the advantages of digital (vs. analog) processing are that, in
digital systems, many functions may be integrated on a chip (e.g., filtering, differentiation, integration,
linearization, modulation, and computation), systems can be easily and inexpensively duplicated and
reprogrammed, and systems do not depend on strict component tolerances.
Before an analog signal can processed, however, the signal must be converted into digital form.
An analog-to-digital (A/D) converter (also abbreviated ADC) samples the analog signal at (usually)
equal intervals of time, and converts the analog signal into a sequence of digitized values (i.e., the
analog signal is sampled, measured, then converted into quantized numerical values), as illustrated in
(B) above.
One of the sources of error in ADCs is jitter, i.e., the uncertainty in the time the signal was
sampled. As shown in (C), an error ∆t in the time of the sampling causes an error ∆V in the measured
value of the signal. The higher the resolution (number of bits) and the speed of the ADC, the smaller
the allowable jitter. At GHz frequencies, some 16 bit ADC clock jitter requirements are a few
femtoseconds.
Phase noise of the oscillator that drives the clock is one of the sources of timing jitter. The
oscillator’s contribution to jitter is the integral of the phase noise, L(f), usually from 10 Hz to ~30 MHz.
J. A. Wepman, “Analog-to-Digital Converters and Their Applications in Radio Receivers,” IEEE

Communications Magazine, pp. 39-45, May 1995.
R. J. Lackey and D. W Upmal, “Speakeasy: The Military Software Radio,” IEEE Communications
Magazine, pp. 56-61, May 1995.

Digital Network Synchronization
• Synchronization plays a critical role in digital telecommunication systems.
It ensures that information transfer is performed with minimal buffer overflow or
underflow events, i.e., with an acceptable level of "slips." Slips cause
problems, e.g., missing lines in FAX transmission, clicks in voice transmission,
loss of encryption key in secure voice transmission, and data retransmission.
• In AT&T's network, for example, timing is distributed down a hierarchy of

nodes. A timing source-receiver relationship is established between pairs of
nodes containing clocks. The clocks are of four types, in four "stratum levels."
Accuracy (Free Running)
Stratum Long Term Per 1st Day Clock Type Number Used
1 1 x 10-11 N.A. GPS W/Two Rb 16
2 1.6 x 10-8 1 x 10-10 Rb Or OCXO ~200
3 4.6 x 10-6 3.7 x 10-7 OCXO Or TCXO 1000’s
4 3.2 x 10-5 N.A. XO ~1 million
1-6
J. E. Abate, E. W. Butterline, R. A. Carley, P. Greendyk, A. M. Montenegro, C. D. Near, S. H.

Richman, and G. P. Zampetti, AT&T's New Approach to the Synchronization of Telecommunication
Networks," IEEE Communications Magazine, pp. 35-45, April 1989.
J. Pan, "Present and Future of Synchronization in the US Telephone Network," IEEE Transactions on
Ultrasonics, Ferroelectrics, and Frequency Control, Vol. UFFC-34, No. 6, pp. 629-638, November
1987.
P. Kartaschoff, “Synchronization in Digital Communications networks,” Proc. IEEE, vol. 79, pp. 906-
914, 1991.

Phase Noise in PLL and PSK Systems
The phase noise of oscillators can lead to erroneous detection of
phase transitions, i.e., to bit errors, when phase shift keyed (PSK) digital
modulation is used. In digital communications, for example, where
8-phase PSK is used, the maximum phase tolerance is ±22.5o, of which
±7.5o is the typical allowable carrier noise contribution. Due to the
statistical nature of phase deviations, if the RMS phase deviation is 1.5o,
for example, the probability of exceeding the ±7.5o phase deviation is
6 X 10-7, which can result in a bit error rate that is significant in some
applications.
Shock and vibration can produce large phase deviations even in

"low noise" oscillators. Moreover, when the frequency of an oscillator is
multiplied by N, the phase deviations are also multiplied by N. For
example, a phase deviation of 10-3 radian at 10 MHz becomes 1 radian
at 10 GHz. Such large phase excursions can be catastrophic to the
performance of systems, e.g., of those which rely on phase locked loops
(PLL) or phase shift keying (PSK). Low noise, acceleration insensitive
oscillators are essential in such applications.
1-7
See the acceleration effects section in chapter 4 for further information about acceleration
induced noise and phase excursions.
J. E. Abate, E. W. Butterline, R. A. Carley, P. Greendyk, A. M. Montenegro, C. D. Near, S. H.

Richman, and G. P. Zampetti, AT&T's New Approach to the Synchronization of Telecommunication
Networks," IEEE Communications Magazine, pp. 35-45, April 1989.
J. Pan, "Present and Future of Synchronization in the US Telephone Network," IEEE Transactions on
1987.
P. Kartaschoff, “Synchronization in Digital Communications networks,” Proc. IEEE, vol. 79, pp. 906-
914, 1991.

Utility Fault Location
Zap!
ta tb
Substation
Substation Substation
Substation
AA BB
Insulator
Sportsman
X
L
When a fault occurs, e.g., when a "sportsman" shoots out an insulator, a disturbance
propagates down the line. The location of the fault can be determined from the differences
in the times of arrival at the nearest substations:
x=1/2[L - c(tb-ta)] = 1/2[L - c∆t]

where x = distance of the fault from substation A, L = A to B line length, c = speed of light,
and ta and tb= time of arrival of disturbance at A and B, respectively.
Fault locator error = xerror=1/2(c∆terror); therefore, if ∆terror ≤ 1 microsecond, then

xerror ≤ 150 meters ≤ 1/2 of high voltage tower spacings, so, the utility company
can send a repair crew directly to the tower that is nearest to the fault.
1-8
Another application in the power industry is during “intermesh,” the period when power load is
transferred between two substations. To avoid a power outage during an intermesh, it is necessary to
precisely measure and adjust the phase (within 3 degrees), and also the voltage amplitude differences
between the substations. Precise timing, GPS for example, can serve as an absolute time reference
for measuring the phases at two widely separated substations. In 1 microsecond, the phase of a 60
Hz waveform advances ~0.2 degree, thus allowing the relative phases to be controlled to well within
the required accuracy.
--------------------------
M. A. Street, "Delivery and Application of Precise Timing for a Traveling Wave Powerline Fault Locator
System," Proc. 22nd Ann. Precise Time and Time Interval (PTTI) Applications and Planning Meeting,
pp. 355-360, 1990, NTIS accession no. N91-25755/0
R. E. Wilson, “Use of Precise Time and Frequency in Power Systems,” Proc. IEEE, vol. 79, pp. 1009-
1018, 1991.
P. Stergiou, “Keeping the Lights On - GPS and Power Grid Intermesh,” GPS World, November 2003.

Space Exploration
Schematic of VBLI ∆t
Technique
Mean
wavelength λ
∆θ ∆θ
Wavefront
Local
Local
Time
Time&& θ(t)
Frequency
Frequency
Standard
Standard
Microwave Microwave
mixer Local mixer
Local
Time
Time&&
Frequency
Frequency
Standard
Standard
Recorder Data tape Data tape Recorder
Recorder Recorder
Correlation
Correlation
c∆ t and
∆θ = and
Integration
Lsin θ Integration
Amplitude Interference
Fringes
λ/L sin θ θ (τ ) Angle
1-9
How does NASA know where a spacecraft is in deep space? The spacecraft's precise range,
velocity and angular position are determined with the aid of highly stable frequency standards. The
range is determined from the propagation time of microwave radiation between an antenna on Earth
and the spacecraft. The velocity is determined from the "doppler," i.e., by comparing the phase of the
incoming carrier signal with that of a reference signal generated from the ground station frequency
standard. The angular position is determined by very long baseline interferometry (VLBI) in which
widely separated stations (in California, Spain and Australia) simultaneously receive signals from the
spacecraft. Differences between times of arrival coupled with knowledge of the baseline vectors
joining the station antennas provide direct geometric determination of the angles between the baseline
vectors and the direction to the spacecraft. Hydrogen masers (see chapter 6) provide the best
stability (~10-15) for the propagation times of interest, which typically range from minutes to hours.
VLBI is also used for high resolution angular measurements in radioastronomy.
J. S. Border & E. R. Kursinski, "Deep Space Tracking and Frequency Standards," Proc. 45th Ann.
Symp. on Frequency Control, pp. 594-607, 1991, IEEE Cat. No. 91CH2965-2.
R. F. C. Vessot, "Applications of Highly Stable Oscillators to Scientific Measurements," Proc. of the

IEEE, pp. 1040-1053, 1991.
W. K. Klemperer, “Long-Baseline Radio Interferometry with Independent Frequency Standards,” Proc.

of the IEEE, vol. 60, pp. 602-609, 1972.

Military Requirements
Military needs are a prime driver of frequency control
technology. Modern military systems require
oscillators/clocks that are:
• Stable over a wide range of parameters (time,
temperature, acceleration, radiation, etc.)
• Low noise
• Low power
• Small size
• Fast warmup
• Low life-cycle cost
1-10

Impacts of Oscillator Technology Improvements
• Higher jamming resistance & improved ability to hide signals
• Improved ability to deny use of systems to unauthorized users
• Longer autonomy period (radio silence interval)
• Fast signal acquisition (net entry)
• Lower power for reduced battery consumption
• Improved spectrum utilization
• Improved surveillance capability (e.g., slow-moving target detection,
bistatic radar)
• Improved missile guidance (e.g., on-board radar vs. ground radar)
• Improved identification-friend-or-foe (IFF) capability
• Improved electronic warfare capability (e.g., emitter location via TOA)
• Lower error rates in digital communications
• Improved navigation capability
• Improved survivability and performance in radiation environment
• Improved survivability and performance in high shock applications
• Longer life, and smaller size, weight, and cost
• Longer recalibration interval (lower logistics costs)
1-11

Spread Spectrum Systems
• In a spread spectrum system, the transmitted signal is spread over a bandwidth that is
much wider than the bandwidth required to transmit the information being sent (e.g., a
voice channel of a few kHz bandwidth is spread over many MHz). This is
accomplished by modulating a carrier signal with the information being sent, using a
wideband pseudonoise (PN) encoding signal. A spread spectrum receiver with the
appropriate PN code can demodulate and extract the information being sent. Those
without the PN code may completely miss the signal, or if they detect the signal, it
appears to them as noise.
• Two of the spread spectrum modulation types are: 1. direct sequence, in which the
carrier is modulated by a digital code sequence, and 2. frequency hopping, in which
the carrier frequency jumps from frequency to frequency, within some predetermined
set, the order of frequencies being determined by a code sequence.
• Transmitter and receiver contain clocks which must be synchronized; e.g., in a
frequency hopping system, the transmitter and receiver must hop to the same
frequency at the same time. The faster the hopping rate, the higher the jamming
resistance, and the more accurate the clocks must be (see the next page for an
example).
• Advantages of spread spectrum systems include the following capabilities: 1. rejection
of intentional and unintentional jamming, 2. low probability of intercept (LPI), 3.
selective addressing, 4. multiple access, and 5. high accuracy navigation and ranging.
1-12
R. C. Dixon, Spread Spectrum Systems, John Wiley and Sons, New York, 1976.
D. L. Schilling, R. L. Pickholtz and L. B. Milstein, "Spread Spectrum Goes Commercial," IEEE

Spectrum, Vol. 27, pp. 40-45, 1990.

Clock for Very Fast Frequency Hopping Radio
Jammer Example
J
Let R1 to R2 = 1 km, R1 to
t1 t2 J =5 km, and J to R2 = 5 km.
Then, since propagation
Radio Radio
delay =3.3 µs/km,
R1
tR R2 t1 = t2 = 16.5 µs,
tR = 3.3 µs, and tm < 30 µs.
To defeat a “perfect” follower Allowed clock error ≈ 0.2 tm
jammer, one needs a hop-rate ≈ 6 µs.
given by:
For a 4 hour resynch interval,
tm < (t1 + t2) - tR
clock accuracy requirement is:
where tm ≈ message duration/hop
≈ 1/hop-rate 4 X 10-10
1-13
With the availability of fast spectrum analyzers and synthesizers, it is possible to jam frequency
hopping systems. If a jammer is fast enough, it can detect the frequency of transmission and tune the
jammer to that frequency well before the radio hops to the next frequency. However, with a good
enough clock, it is possible to defeat such “follower” jamming. As illustrated above, even a "perfect"
follower jammer can be defeated if a good enough clock is available. (A perfect jammer is defined
here as one that can identify the frequency of a received signal, tune a synthesizer to that frequency,
and transmit the jamming signal in zero time.)
Because radio waves travel at the speed of light, the radio-to-jammer-to-radio (R1 to J to R2)
and radio-to-radio (R1 to R2) propagation delays are 3.3 µs per km. Therefore, if the hopping rate is
fast enough for the propagation delay difference to be greater than 1/hop-rate, i.e., if the radios can
hop to the next frequency before the jamming signal reaches the receiver, then the radios are
jamming-proof (for follower jammers). In the example above, the propagation delays t1, t2, and tR
imply that the message duration tm be less than 30 µs. Since the clock accuracies required by
frequency hopping systems are usually 10% to 20% of tm, the allowed clock error is about 6 µs. In a
military environment, such accuracies can be maintained for periods of hours and longer only with
atomic clocks.
A. D. Robertson and F. C. Painter, "Tactical Jamming," Defense Science and Engineering, pp. 20-28,
September 1985.

Clocks and Frequency Hopping C3 Systems
Slow hopping ‹-------------------------------›Good clock
Fast hopping ‹------------------------------› Better clock
Extended radio silence ‹-----------------› Better clock
Extended calibration interval ‹----------› Better clock
Othogonality ‹-------------------------------› Better clock
Interoperability ‹----------------------------› Better clock
1-14
The faster the hopping rate, the higher the jamming resistance, and the more accurate the
clocks must be. For example, for a system with a hopping rate of 1,000 hops per second, the dwell
time at each frequency is 1 millisecond. For such a system to operate properly, the clocks must
remain synchronized to about 100 microseconds.
When several radio nets operate in an area, self-jamming (also called co-site interference) can
be a problem if the nets operate independently of one another, i.e., if the nets are not orthogonal.
Radios of neighboring nets can then occasionally hop to the same frequency at the same time, thus
producing self-jamming. When the nets are orthogonal, i.e., when the neighboring nets are
synchronized and use codes that insure that radios do not hop to the same frequency at the same
time, the radios must not only be synchronized within a net, but also to those of neighboring nets.
This requires an even higher clock accuracy.
The requirement for C3 systems to be interoperable places yet another stringent requirement on
accuracy. For example, when an Army unit calls for air support from an Air Force unit that may be
many hundreds of kilometers away, the clocks in the respective units' radios must be synchronized in
order for the units to be able to communicate. Maintaining synchronization for extended periods
among independent clocks that are widely separated requires very high quality clocks.

Identification-Friend-Or-Foe (IFF)
Air Defense IFF Applications
AWACS
FRIEND OR FOE?
F-16
FAAD
PATRIOT STINGER
1-15
In a modern battle, when the sky is filled with friendly and enemy aircraft, and a variety of
advanced weapons are ready to fire from both ground and airborne platforms, positive identification of
friend and foe is critically important. For example fratricide due to identification errors has been a
major problem in all 20th century wars.
Current IFF systems use an interrogation/response method which employs cryptographically
encoded spread spectrum signals. The interrogation signal received by a friend is supposed to result
in the "correct" code being automatically sent back via a transponder on the friendly platform. The
"correct" code must change frequently to prevent a foe from recording and transmitting that code
("repeat jamming"), thereby appearing as a friend. The code is changed at the end of what is called
the code validity interval (CVI).
The better the clock accuracy, the shorter can be the CVI, the more resistant the system can be
to repeat jamming, and the longer can be the autonomy period for users who cannot resynchronize
their clocks during a mission.

Effect of Noise in Doppler Radar System
Moving
Moving Decorrelated
Transmitter Object Clutter Noise
Transmitter Object
fD
Stationary
Stationary Doppler Signal
Receiver Object
Object
fD
f
• Echo = Doppler-shifted echo from moving target + large "clutter" signal
• (Echo signal) - (reference signal) --› Doppler shifted signal from target
• Phase noise of the local oscillator modulates (decorrelates) the clutter

signal, generates higher frequency clutter components, and thereby
degrades the radar's ability to separate the target signal from the clutter
signal.
1-16
Doppler radars, especially, require low-noise oscillators. The velocity of the target and the radar
frequency are primary determinants of the phase noise requirements (see “Doppler Shifts” later in this
chapter). Slow-moving targets produce small Doppler shifts, therefore, low phase-noise close to the
carrier is required. To detect fast-moving targets, low noise far from the carrier is required. For
example, when using an X-band radar to detect a 4 km/hour target (e.g., a slowly moving vehicle), the
noise 70 Hz from the carrier is the important parameter, whereas to detect supersonic aircraft, the
noise beyond 10 kHz is important.
When a radar is on a stationary platform, the phase noise requirements can usually be met with
commercially available oscillators. A good quartz crystal (bulk acoustic wave, BAW) oscillator can
provide sufficiently low noise close to the carrier, and a good surface acoustic wave (SAW) oscillator
can provide sufficiently low noise far from the carrier. Very far from the carrier, dielectric resonator
oscillators (DRO) can provide lower noise than either BAW or SAW oscillators. A combination of
oscillators can be used to achieve good performance in multiple spectral regions, e.g., a DRO phase
locked to a frequency-multiplied BAW oscillator can provide low noise both close to the carrier and far
from the carrier.
The problem with achieving sufficiently low phase noise occurs when the radar platform vibrates,
as is the case when the platform is an aircraft or a missile. The vibration applies time-dependent
stresses to the resonator in the oscillator which results in modulation of the output frequency (see
Chapter 4). The aircraft's random vibration, thereby, degrades the phase noise, and discrete
frequency vibrations (e.g., due to helicopter blade rotation) produce spectral lines which can result in
false target indications. The degradation in noise spectrum occurs in all types of oscillators (BAW,
SAW, DRO, atomic frequency standards, etc.). A large phase-noise degradation can have
catastrophic effects on radar performance. In a coherent radar, the platform-vibration-induced phase
noise can reduce the probability of detection to zero.
Oscillator phase noise also affects the ability of a radar system to discern small targets located
near large targets. Excess phase noise on local oscillators used to translate radar signals can limit
the ability to detect smaller cross-section targets in both the doppler and range domains. Some radar
systems also require the ability to coherently integrate radar-pulse returns on time scales of seconds
to minutes.
M. I. Skolnik, Introduction to Radar Systems, Second Edition, McGraw-Hill Book Co., 1980.

D. B. Leeson, and G. F. Johnson,"Short-Term Stability for a Doppler Radar: Requirements,
Rev. 8.5.1.2, by John R. Measurements, and Techniques," Proc. of the IEEE, Vol. 54, No. 2, pp. 244-248, Feb. 1966.
Vig, July 2001, AD-M001251.
Bistatic Radar
Conventional (i.e., "monostatic") radar, in which the Illuminator
illuminator and receiver are on the same platform, is vulnerable
to a variety of countermeasures. Bistatic radar, in which the
illuminator and receiver are widely separated, can greatly
reduce the vulnerability to countermeasures such as jamming
and antiradiation weapons, and can increase slow moving
target detection and identification capability via "clutter tuning” Receiver
(receiver maneuvers so that its motion compensates for the
motion of the illuminator; creates zero Doppler shift for the area
being searched). The transmitter can remain far from the battle
area, in a "sanctuary." The receiver can remain "quiet.”
The timing and phase coherence problems can be orders
of magnitude more severe in bistatic than in monostatic
radar, especially when the platforms are moving. The
Target
reference oscillators must remain synchronized and syntonized
during a mission so that the receiver knows when the transmitter emits each pulse, and the phase
variations will be small enough to allow a satisfactory image to be formed. Low noise crystal
oscillators are required for short term stability; atomic frequency standards are often required for
long term stability.
1-17
Similar requirements exist in electronic warfare applications. The ability to locate radio and radar
emitters is important in modern warfare. One method of locating emitters is to measure the time
difference of arrival of the same signal at widely separated locations. Emitter location by means of
this method depends on the availability of highly accurate clocks, and on highly accurate methods of
synchronizing clocks that are widely separated. Since electromagnetic waves travel at the speed of
light, 30 cm per nanosecond, the clocks of emitter locating systems must be kept synchronized to
within nanoseconds in order to locate emitters with high accuracy. (Multipath and the geometrical
arrangement of emitter locators usually results in a dilution of precision.) Without resynchronization,
even the best available militarized atomic clocks can maintain such accuracies for periods of only a
few hours. With the availability of GPS and using the "GPS common view" method of time transfer,
widely separated clocks can be synchronized to better than 10 ns (assuming that GPS is not jammed).
An even more accurate method of synchronization is "two-way time transfer via communication
satellites," which, by means of very small aperture terminals (VSATs) and pseudonoise modems, can
attain subnanosecond time transfer accuracies.
Another important application for low-noise frequency sources is the ELINT (ELectronic
INTelligence) receiver. These receivers are used to search a broad range of frequencies for signals
that may be emitted by a potential adversary. The frequency source must be as noise-free as
possible so as not to obscure weak incoming signals. The frequency source must also be extremely
stable and accurate in order to allow accurate measurement of the incoming signal's characteristics.
N. J. Willis, "Bistatic Radar," in Radar Handbook, M. I. Skolnik, editor, Chapter 25, Mc-Graw-Hill
Publishing Co., 1990.
W. Lewandowski and C. Thomas, "GPS Time Transfer," Proc. IEEE, Vol. 79, pp. 991-1000, July 1991.
G. Lippermeier and R. Vernon, "IFFN: Solving the Identification Riddle," Defense Electronics, pp. 83-
88, 1988.

Doppler Shifts
ile
g Vech
d or Air
40
ft
2 Aircra
ft
w Movin
ic Aircra
Radar Frequency (GHz)
, Groun
30
- Mach
lo
Subson
- Vehicle
25
an or S
20
m/h
-
M
100km/h
2,400 k
700km/h
4km/h -
15
X-Band RADAR
10
0
10 100 1K 10K 100K 1M
Doppler Shift for Target Moving Toward Fixed Radar (Hz)
1-18
Doppler radars require low-phase-noise oscillators. The velocity of the target and the radar
frequency are the primary factors that determine the oscillator noise requirements. For example, to
detect slow-moving targets, the noise close to the carrier must be low.
The Doppler shift* of an object moving towards the observer is given by ∆f/f = 2v/c, where ∆f is
the Doppler frequency shift, v is the velocity of the object, and c is the speed of light.
* Doppler shift example: if v = 4 km/h and f = 10 GHz (e.g., a slow-moving vehicle approaching an
X-band radar), then ∆f = 74 Hz, i.e., an oscillator with low phase noise at 74Hz from the carrier is
necessary in order for a coherent radar system to "see" the vehicle.

CHAPTER 2
Quartz Crystal Oscillators
W. L. Smith, "Precision Oscillators," in E. A. Gerber and A. Ballato, Precision Frequency Control, Vol. 2, pp. 45-
98, Academic Press, 1985.
B. Parzen, Design of Crystal and Other Harmonic Oscillators, John Wiley and Sons, Inc., 1983.
M. E. Frerking, "Temperature Control and Compensation," in E. A. Gerber and A. Ballato, Precision Frequency
Control, Vol. 2, pp. 99-111, Academic Press, 1985.
M. E. Frerking, Crystal Oscillator Design and Temperature Compensation, Van Nostrand Reinhold Co., 1978.
Hewlett-Packard Co., Test & Measurement Application Notes

<http://www.tmo.hp.com/tmo/Notes/English/Oscillators.html>
"Fundamentals of Quartz Oscillators," Hewlett-Packard Application Note AN 200-2, Hewlett-Packard Company.
A. Benjaminson, "Computer-Aided Design of Crystal Oscillators," U. S. Army R & D Technical Report DELET-TR-
84-0386-F, August 1985, AD-B096820;
"Advanced Crystal Oscillator Design," U. S. Army R & D Technical Report SLCET-TR-85-0445-F, January 1988,
AD-B121288;
"Advanced Crystal Oscillator Design," U. S. Army R & D Technical Report SLCET-TR-88-0804-1, February 1989,
AD-B134514;
"Advanced Crystal Oscillator Design," U. S. Army R & D Technical Report SLCET-TR-88-0804-F, December
1991, AD-B163808.

Crystal Oscillator
Tuning
Voltage
Crystal
resonator
Output
Frequency
Amplifier
2-1
Above is a simplified circuit diagram that shows the basic elements of a crystal oscillator (XO).
The amplifier of an XO consists of at least one active device, the necessary biasing networks, and
may include other elements for band limiting, impedance matching, and gain control. The feedback
network consists of the crystal resonator, and may contain other elements, such as a variable
capacitor for tuning.
W. L. Smith, "Precision Oscillators," in E. A. Gerber and A. Ballato, Precision Frequency Control, Vol.
2, pp. 45-98, Academic Press, 1985.
B. Parzen, Design of Crystal and Other Harmonic Oscillators, John Wiley and Sons, Inc., 1983.
M. E. Frerking, "Temperature Control and Compensation," in E. A. Gerber and A. Ballato, Precision

Frequency Control, Vol. 2, pp. 99-111, Academic Press, 1985.
M. E. Frerking, Crystal Oscillator Design and Temperature Compensation, Van Nostrand Reinhold
Company, 1978.
"Fundamentals of Quartz Oscillators," Hewlett-Packard application note AN 200-2, Hewlett-Packard

Company,
<http://www.tmo.hp.com/@@2ZcNpBcQ240oRhrt/tmo/Notes/English/5965-7662E.html>

Oscillation
• At the frequency of oscillation, the closed loop phase shift
= 2nπ.
• When initially energized, the only signal in the circuit is
noise. That component of noise, the frequency of which
satisfies the phase condition for oscillation, is propagated
around the loop with increasing amplitude. The rate of
increase depends on the excess; i.e., small-signal, loop
gain and on the BW of the crystal in the network.
• The amplitude continues to increase until the amplifier gain
is reduced either by nonlinearities of the active elements
("self limiting") or by some automatic level control.
• At steady state, the closed-loop gain = 1.
2-2
See “Decay Time, Linewidth, and Q” in chapter 3 for further information on oscillator startup
time.
In addition to noise, switching on the DC power supply is another oscillation trigger.
W. L. Smith, "Precision Oscillators," in E. A. Gerber and A. Ballato, Precision Frequency Control, Vol.
2, pp. 45-98, Academic Press, 1985.
M. Toki and Y. Tsuzuki, “Analysis of Start-up Characteristics of CMOS Crystal Oscillators,” Proc. 1992
IEEE Frequency Control Symposium, pp. 448-452, 1992.

Oscillation and Stability
• If a phase perturbation ∆φ occurs, the frequency must shift ∆f to maintain the
2nπ phase condition, where ∆f/f=-∆φ/2QL for a series-resonance oscillator,
and QL is loaded Q of the crystal in the network. The "phase slope," dφ/df
is proportional to QL in the vicinity of the series resonance frequency (also
see "Equivalent Circuit" and "Frequency vs. Reactance" in Chapt. 3).
• Most oscillators operate at "parallel resonance," where the reactance vs.

frequency slope, dX/df, i.e., the "stiffness," is inversely proportional to C1,
the motional capacitance of the crystal unit.
• For maximum frequency stability with respect to phase (or reactance)

perturbations in the oscillator loop, the phase slope (or reactance slope) must
be maximum, i.e., C1 should be minimum and QL should be maximum. A
quartz crystal unit's high Q and high stiffness makes it the primary frequency
(and frequency stability) determining element in oscillators.
2-3
The importance of high Q is further discussed in chapter 3, see, especially, “What is Q and Why
is it Important?”.

Tunability and Stability
Making an oscillator tunable over a wide frequency range degrades
its stability because making an oscillator susceptible to intentional tuning
also makes it susceptible to factors that result in unintentional tuning. The
wider the tuning range, the more difficult it is to maintain a high stability.
For example, if an OCXO is designed to have a short term stability of
1 x 10-12 for some averaging time and a tunability of 1 x 10-7, then the
crystal's load reactance must be stable to 1 x 10-5 for that averaging time.
Achieving such stability is difficult because the load reactance is affected
by stray capacitances and inductances, by the stability of the varactor's
capacitance vs. voltage characteristic, and by the stability of the voltage
on the varactor. Moreover, the 1 x 10-5 load reactance stability must be
maintained not only under benign conditions, but also under changing
environmental conditions (temperature, vibration, radiation, etc.).
Whereas a high stability, ovenized 10 MHz voltage controlled
oscillator may have a frequency adjustment range of 5 x 10-7 and an
aging rate of 2 x 10-8 per year, a wide tuning range 10 MHz VCXO may
have a tuning range of 50 ppm and an aging rate of 2 ppm per year.
2-4

Oscillator Acronyms
• XO…………..Crystal Oscillator
• VCXO………Voltage Controlled Crystal Oscillator
• OCXO………Oven Controlled Crystal Oscillator
• TCXO………Temperature Compensated Crystal Oscillator
• TCVCXO..…Temperature Compensated/Voltage Controlled

Crystal Oscillator
• OCVCXO.….Oven Controlled/Voltage Controlled Crystal Oscillator
• MCXO………Microcomputer Compensated Crystal Oscillator
• RbXO……….Rubidium-Crystal Oscillator
2-5
Historically, in the USA, military requirements have been the main impetus for research on
crystal oscillators. The US Army sponsored most of the research, especially in the early days.
According to folklore, the abbreviation XO instead of CO came about because, in the military, CO is
the abbreviation for "commanding officer" and because "crystal" sounds a little like "xtal". (Later,
someone pointed out that XO is the abbreviation for "executive officer" in the military, but by then, XO
was generally accepted as the abbreviation for "crystal oscillator".)

Crystal Oscillator Categories
The three categories, based on the method of dealing with the crystal unit's
frequency vs. temperature (f vs. T) characteristic, are:
• XO, crystal oscillator, does not contain means for reducing the crystal's
f vs. T characteristic (also called PXO-packaged crystal oscillator).
• TCXO, temperature compensated crystal oscillator, in which, e.g., the

output signal from a temperature sensor (e.g., a thermistor) is used to
generate a correction voltage that is applied to a variable reactance (e.g., a
varactor) in the crystal network. The reactance variations compensate for
the crystal's f vs. T characteristic. Analog TCXO's can provide about a 20X
improvement over the crystal's f vs. T variation.
• OCXO, oven controlled crystal oscillator, in which the crystal and other
temperature sensitive components are in a stable oven which is adjusted to
the temperature where the crystal's f vs. T has zero slope. OCXO's can
provide a >1000X improvement over the crystal's f vs. T variation.
2-6

Crystal Oscillator Categories
∆f
Voltage f +10 ppm
Tune
250C
-450C +1000C
Output T
• Crystal Oscillator (XO) -10 ppm

∆f
Temperature
Temperature Compensation
Sensor
Compensation
Network f +1 ppm
Sensor Networkoror -450C
Computer
Computer +1000C
T
XO
XO -1 ppm
• Temperature Compensated (TCXO)

Oven
XO
∆f
Oven XO +1 x 10-8
Oven
-450C
f
control
control Temperature +1000C
Temperature
Sensor T
Sensor
-1 x 10-8
• Oven Controlled (OCXO)
2-7
A wide temperature range XO has a typical f vs. T stability of ~10 to 50 ppm. A TCXO can
reduce that to ~1 ppm. An OCXO can reduce that stability to 1 x 10-8 or better (but at the cost of much
higher power consumption). High-end (SC-cut) OCXOs can stay within 1 x 10-10 over a wide
temperature range.

Hierarchy of Oscillators
Oscillator Type* Accuracy** Typical Applications
• Crystal oscillator (XO) 10-5 to 10-4 Computer timing
• Temperature compensated 10-6 Frequency control in tactical
crystal oscillator (TCXO) radios
• Microcomputer compensated 10-8 to 10-7 Spread spectrum system clock
crystal oscillator (MCXO)
• Oven controlled crystal 10-8 (with 10-10
Navigation system clock &
frequency standard, MTI radar
oscillator (OCXO) per g option)
• Small atomic frequency 10-9 C3 satellite terminals, bistatic,
& multistatic radar
standard (Rb, RbXO)
• High performance atomic 10-12 to 10-11 Strategic C3, EW
standard (Cs)
* Sizes range from <5cm3 for clock oscillators to > 30 liters for Cs standards
Costs range from <$5 for clock oscillators to > $50,000 for Cs standards.
** Including environmental effects (e.g., -40oC to +75oC) and one year of
aging.
2-8
See also chapter 7 for more detailed comparisons of various oscillators.

Oscillator Circuit Types
Of the numerous oscillator circuit types, three of the more common ones, the Pierce, the Colpitts and
the Clapp, consist of the same circuit except that the rf ground points are at different locations. The
Butler and modified Butler are also similar to each other; in each, the emitter current is the crystal
current. The gate oscillator is a Pierce-type that uses a logic gate plus a resistor in place of the
transistor in the Pierce oscillator. (Some gate oscillators use more than one gate).
b b ∈ ∈
c c
∈ c b
Pierce Colpitts Clapp
c b
∈ c
b ∈
Modified
Butler Butler Gate
2-9
The choice of oscillator circuit type depends on factors such as the desired frequency stability,
input voltage and power, output power and waveform, tunability, design complexity, cost and the
crystal unit's characteristics.
In the Pierce family, the ground point location has a profound effect on the performance. The
Pierce configuration is generally superior to the others, e.g., with respect to the effects of stray
reactances and biasing resistors, which appear mostly across the capacitors in the circuit rather than
the crystal unit. It is one of the most widely used circuits for high stability oscillators. In the Colpitts
configuration, a larger part of the strays appears across the crystal, and the biasing resistors are also
across the crystal, which can degrade performance. The Clapp is seldom used because, since the
collector is tied directly to the crystal, it is difficult to apply a dc voltage to the collector without
introducing losses or spurious oscillations. (See the references for more details.)
The Pierce family usually operates at "parallel resonance" (see "Resonator Frequency vs.
Reactance" in Chapt. 3), although it can be designed to operate at series resonance by connecting an
inductor in series with the crystal. The Butler family usually operates at (or near) series resonance.
The Pierce can be designed to operate with the crystal current above or below the emitter current.
Gate oscillators are common in digital systems when high stability is not a major consideration.
A. Benjaminson, "Computer-Aided Design of Crystal Oscillators," U. S. Army R & D Technical Report

DELET-TR-84- 0386-F, August 1985, AD-B096820; "Advanced Crystal Oscillator Design," U. S.
Army R & D Technical Report SLCET-TR-85-0445-F, January 1988, AD-B121288; "Advanced Crystal
Oscillator Design," U. S. Army R & D Technical Report SLCET-TR-88-0804-1, February 1989, AD-
B134514; "Advanced Crystal Oscillator Design," U. S. Army R & D Technical Report SLCET-TR-88-
0804-F, December 1991, AD-B163808.
J. P. Buchanan, Handbook of Piezoelectric Crystals for Radio Equipment Designers, WADC Technical
Report 56-156, October 1956 (692 pages), available from NTIS, AD 110448.

OCXO Block Diagram
ϕ
Output
Oven
Each of the three main parts of an OCXO, i.e., the crystal, the sustaining
circuit, and the oven, contribute to instabilities. The various instabilities
are discussed in the rest of chapter 3 and in chapter 4.
2-10
F. L. Walls & J. R. Vig, "Fundamental Limits on the Frequency Stabilities of Quartz Crystal
Oscillators," IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 42, pp.
576-589, July 1995.
F. L. Walls & J.-J. Gagnepain, "Environmental Sensitivities of Quartz Oscillators, IEEE Trans.
Ultrasonics, Ferroelectrics and Frequency Control, vol. 39, pp. 241-249, 1992.

Oscillator Instabilities - General Expression
−1/2
∆f ∆f 1 ⎡ ⎛ 2ff QL ⎞ 2 ⎤
≈ + ⎢1 + ⎜ ⎟ ⎥ dφ(ff )
foscillator fresonator 2QL ⎢⎣ ⎝ f ⎠ ⎥⎦
where QL = loaded Q of the resonator, and dφ(ff) is a small
change in loop phase at offset frequency ff away from carrier
frequency f. Systematic phase changes and phase noise within
the loop can originate in either the resonator or the sustaining
circuits. Maximizing QL helps to reduce the effects of noise and
environmentally induced changes in the sustaining electronics.
In a properly designed oscillator, the short-term instabilities are
determined by the resonator at offset frequencies smaller than
the resonator’s half-bandwidth, and by the sustaining circuit and
the amount of power delivered from the loop for larger offsets.
2-11
576-589, July 1995.

Instabilities due to Sustaining Circuit
• Load reactance change - adding a load capacitance to a crystal
changes the frequency by
C1
δ f ≡ ∆f ≅
f 2(C0 + CL )
∆ (δ f ) C1
then, ≅−
2(C0 + CL )
2
∆ CL
• Example: If C0 = 5 pF, C1 = 14fF and CL = 20pF, then a ∆CL = 10 fF

(= 5 X 10-4) causes ≈1 X 10-7 frequency change, and a CL aging of
10 ppm per day causes 2 X 10-9 per day of oscillator aging.
• Drive level changes: Typically 10-8 per ma2 for a 10 MHz 3rd SC-cut.
• DC bias on the crystal also contributes to oscillator aging.
2-12
See “Frequency vs. Drive Level” in Chapter 4.

A DC voltage on the resonator can be a significant contributor to aging. To minimize this aging
mechanism, the sustaining circuit must be designed so as to not apply a DC voltage to the resonator.
This may be accomplished by, e.g., by placing a capacitor in series and a few MΩ resistor in parallel
with the resonator.
R.L. Filler, J A. Kosinski, V.J. Rosati and J.R. Vig, “Aging Studies On Quartz Crystal Resonators And
Oscillators,” Proc. 38th Ann. Symp. On Frequency Control, pp. 225-231, 1984.

Oscillator Instabilities - Tuned Circuits
Many oscillators contain tuned circuits - to suppress unwanted
modes, as matching circuits, and as filters. The effects of
small changes in the tuned circuit's inductance and
capacitance is given by:
⎛ ⎞
∆f dφ (ff ) ⎜ 1 ⎟⎛ Q ⎞⎛ dC
⎜
dL ⎞
≈ ≈⎜ ⎟⎜ c ⎟ c + c⎟
f oscillator 2QL ⎜⎜ 1 + 2ff ⎟⎟⎝ Q ⎠⎜⎝ Cc L ⎟
c ⎠
⎝ BW ⎠
where BW is the bandwidth of the filter, ff is the frequency
offset of the center frequency of the filter from the carrier
frequency, QL is the loaded Q of the resonator, and Qc, Lc and
Cc are the tuned circuit's Q, inductance and capacitance,
respectively.
2-13
576-589, July 1995.

Oscillator Instabilities - Circuit Noise
Flicker PM noise in the sustaining circuit causes flicker FM
contribution to the oscillator output frequency given by:
f2
Losc (ff ) = Lckt (1Hz ) 3 2
4ff QL
and
1
σ y (τ) = ln2 Lckt (1Hz )
QL
where ff is the frequency offset from the carrier frequency f, QLis the
loaded Q of the resonator in the circuit, Lckt (1Hz) is the flicker PM
noise at ff = 1Hz, and τ is any measurement time in the flicker floor
range. For QL = 106 and Lckt (1Hz) = -140dBc/Hz, σy(τ) = 8.3 x 10-14.
( Lckt (1Hz) = -155dBc/Hz has been achieved.)
2-14
576-589, July 1995.
E. S. Ferre-Pikal, F. L. Walls & C. W. Nelson, “Design Criteria for BJT Amplifiers with Low 1/f AM and
PM Noise,” Proc. 1995 IEEE Int'l Frequency Control Symposium, pp. 305-313, 1995.

Oscillator Instabilities - External Load
If the external load changes, there is a change in the amplitude
or phase of the signal reflected back into the oscillator. The
portion of that signal which reaches the oscillating loop changes
the oscillation phase, and hence the frequency by
∆f dφ(ff ) ⎛ 1 ⎞⎛ Γ − 1 ⎞
≈ ≈⎜ ⎟⎜ ⎟(sinθ ) isolation
f oscillator 2Q ⎝ 2Q ⎠⎝ Γ + 1⎠
where Γ is the VSWR of the load, and θ is the phase angle of
the reflected wave; e.g., if Q ~ 106, and isolation ~40 dB
(i.e., ~10-4), then the worst case (100% reflection) pulling is
~5 x 10-9. A VSWR of 2 reduces the maximum pulling by only
a factor of 3. The problem of load pulling becomes worse at
higher frequencies, because both the Q and the isolation are
lower.
2-15
576-589, July 1995.

Oscillator Outputs
Most users require a sine wave, a TTL-compatible, a CMOS-
compatible, or an ECL-compatible output. The latter three can be
simply generated from a sine wave. The four output types are
illustrated below, with the dashed lines representing the supply
voltage inputs, and the bold solid lines, the outputs. (There is no
“standard” input voltage for sine wave oscillators. The input
voltages for CMOS typically range from 1V to 10V.)
+15V
+10V
+5V
0V
-5V
Sine TTL CMOS ECL
2-16

Resonator Self-Temperature Sensing
fβ (Hz)
fβ ≡ 3f1 - f3
172300
dfβ
= − 14 Hz/oC
dT
171300
-35 -15 5 25 45 65 85
Temperature (oC)
170300
2-17
As is shown in chapter 4, see “Effects of Harmonics on f vs. T,” the f vs. T of the fundamental
mode of a resonator is different from that of the third and higher overtones. This fact is exploited for
“self-temperature sensing” in the microcomputer compensated crystal oscillator (MCXO). The
fundamental (f1) and third overtone (f3) frequencies are excited simultaneously (“dual mode”
excitation) and a beat frequency fβ is generated such that fβ = 3f1 - f3 (or fβ = f1 - f3/3). The fβ is a
monotonic and nearly linear function of temperature, as is shown above for a 10 MHz 3rd overtone
(3.3. MHz fundamental mode) SC-cut resonator. This resonator was 14 mm in diameter, plano-
convex, and had a 3 diopter contour.
The fβ is a measure of the resonator’s temperature exactly where the resonator is vibrating,
thereby eliminating the need for a thermometer other than the resonator. Because the SC-cut is
thermal transient compensated, the thermal transient effects are also eliminated, as are the effects of
temperature gradients between the thermometer and the resonator.
For temperature compensation purposes, the fβ vs. T need not be used; the calibration can
consist of f vs. fβ only. The role of a thermometer during calibration is then only to insure that the
specified temperature range is covered.
See also “Mode Spectrograph of an SC-cut” and the page that follows it in Chapter 3.
S. Schodowski, "Resonator Self-Temperature-Sensing Using a Dual-Harmonic-Mode Crystal

Oscillator," Proc. 43rd Annual Symposium on Frequency Control, pp. 2-7, 1989, IEEE Catalog No.
89CH2690-6.
R. Filler and J. Vig, "Resonators for the Microcomputer-Compensated Crystal Oscillator," Proc. 43rd
Annual Symposium on Frequency Control, pp. 8-15, 1989, IEEE Catalog No. 89CH2690-6.

Thermometric Beat Frequency Generation
DUAL MODE
OSCILLATOR f1
X3
X3
M=1
MULTIPLIER
MULTIPLIER
LOW f = 3f1 - f3
Mixer LOWPASS
PASS β
FILTER
FILTER
M=3
f3
2-18
The microcomputer compensated crystal oscillator (MCXO) uses a high-stability 10 MHz SC-cut
quartz resonator and a dual mode oscillator which simultaneously excites the fundamental and third
overtone modes of the resonator. The beat frequency may be generated either by multiplying the
fundamental mode frequency by three and subtracting from it the third overtone frequency, as shown
above, or by dividing the third overtone frequency by three, in which case the beat frequency is fβ = f1 -
f3/3. The beat frequency is a monotonic and nearly linear function of temperature, as is shown on the
previous page. It provides a high precision, digital measure of the vibrating region’s temperature,
thereby eliminating the need for an external thermometer.

89CH2690-6.

Microcomputer Compensated Crystal Oscillator
(MCXO)
f1
Dual-
Dual-
mode
x3
x3
mode
XO
XO
µcom-
µcom-
Reciprocal
Reciprocal Correction
Correction
fβ
puter
puter f0
f3 Counter
Counter Circuit
Circuit
N1 N2
Mixer
2-19
The fβ is used to gate a reciprocal counter that uses the fundamental mode frequency as the
time base. The counter’s output is a number N1 which varies with temperature. The microcomputer,
in which f1 vs. fβ calibration information specific to each resonator is stored, solves an equation and
outputs a number N2 which is used to correct for the variations of f1 with temperature.
Two correction methods have been used. In one, the SC-cut resonator is made to have a
frequency that is slightly above the output frequency fo at all temperatures, and pulse deletion is used
to obtain an fo that is stable over the temperature range. In the other method, the SC-cut resonator’s
frequency is slightly below the output frequency fo at all temperatures, and and a correction frequency
(generated by means of a direct digital synthesizer) is added to obtain an fo that is stable over the
temperature range. The two methods are explained in more detail on the following pages.

89CH2690-6.

MCXO Frequency Summing Method
Block Diagram
VCXO 10 MHz
PHASE- output
3rd OVERTONE f3 = 10 MHz -
LOCKED
fd
CRYSTAL LOOP
DUAL-MODE Divide by
OSCILLATOR 3 fd
F Divide
f1 by
FUNDAMENTAL
MODE TClock 2500
DIRECT
Mixer DIGITAL
SYNTHESIZER T F
fb
Clock
N2 Divide 1 PPS
Clock MICRO- by output
COUNTER COMPUTER
N1 out 4000
NON-VOLATILE
MEMORY
T = Timing Mode
F = Frequency
Mode
2-20
In the frequency summing method, the direct digital synthesizer (DDS) generates a correction
frequency fd, based on N2, such that f3 + fd = 10 MHz at all temperatures. The phase locked loop
locks the VCXO to this precise 10 MHz.
In the “frequency mode,” the 1 PPS output is derived by division from 10 MHz. In the power
conserving “timing mode,” the 1 PPS is generated directly from f3 driving the DDS, and by using a
different calibration equation. The PLL and portions of the digital circuitry are turned off, the
microprocessor goes to “sleep” between corrections, and the time between corrections is increased to
reduce the power requirement.
A. Benjaminson and S. Stallings, "A Microcomputer-Compensated Crystal Oscillator Using Dual-Mode

Resonator," Proc. 43rd Annual Symposium on Frequency Control, pp. 20-26, 1989, IEEE Catalog No.
89CH2690-6.
A. Benjaminson and B. Rose, "Performance Tests on an MCXO Combining ASIC and Hybrid
Construction," Proc. 45th Annual Symposium on Frequency Control, pp. 393-397, IEEE pub. no.
91CH2965-2, 1991.
E. Jackson, H. Phillips, B. E. Rose, "The Micro-computer Compensated Crystal Oscillator - A Progress

Report," Proc. 1996 IEEE Int'l Frequency Control Symposium, pp. 687-692, IEEE pub. no.
96CH35935, 1996.

MCXO - Pulse Deletion Method
Digital
circuitry (ASIC)
fc output
Dual f
Dualmode
mode β output
oscillator
oscillator Pulse
Counter Pulse
Counter eliminator f0
eliminator
corrected
output
SC-cut crystal
for timing
Frequency
Frequency
evaluator
evaluator
&&correction
correction Microprocessor
determination
determination circuitry
Microcomputer compensated crystal oscillator (MCXO) block diagram - pulse deletion method.
2-21
In the pulse deletion method, the SC-cut resonator’s frequency is slightly above the output
frequency, fo. For example, if fo is 10 MHz, then the SC-cut resonator is made to have a frequency
slightly above 10 MHz at all temperatures over the design temperature range. The dual-mode
oscillator provides output signals at two frequencies, one of which, fβ, is the resonator temperature
indicator. The signals are processed by the microcomputer which, from fβ, determines the necessary
correction to fc and then subtracts the required number of pulses from fc to obtain the corrected output
fo. Fractions of pulses that cannot be subtracted within the update interval (~ 1 s) are used as a carry,
so that the long-term average is within the ±2 x 10-8 design accuracy. Correction data in the PROM
are unique to each crystal and are obtained from a precise fc vs. fβ calibration over the temperature
range. The corrected output signal fo can be divided down to produce a 1 pps time reference, or can
be used directly to drive a clock. Due to the objectionable noise characteristics created by the pulse
deletion process, additional signal processing is necessary to provide a useful RF output for frequency
control applications. This can be accomplished by, for example, imparting the MCXO’s frequency
accuracy to a low-noise, low-cost voltage controlled crystal oscillator (VCXO) via locking the VCXO
frequency to fo.
M. Bloch, M. Meirs and J. Ho, "The Microcomputer Compensated Crystal Oscillator (MCXO)," Proc.
43rd Annual Symposium on Frequency Control, pp. 16-19, 1989, IEEE Catalog No. 89CH2690-6.

MCXO - TCXO Resonator Comparison
Parameter MCXO TCXO
Cut, overtone SC-cut, 3rd AT-cut, fund.
Angle-of-cut tolerance Loose Tight
Blank f and plating tolerance Loose Tight
Activity dip incidence Low Significant
Hysteresis (-550C to +850C) 10-9 to 10-8 10-7 to 10-6
Aging per year 10-8 to 10-7 10-7 to 10-6
2-22
The table shows a comparison between resonators for the MCXO and for precision analog
TCXO. Not only is the angle-of-cut tolerance looser for MCXO resonators, but so are the blank
frequency and plating tolerances. In both MCXO implementations, pulse deletion and frequency-
summing, the resonator target frequency during plating is chosen to ensure that the resonator
frequency is either above or below the nominal clock frequency at all temperatures. Therefore, there
is no need to specify a tight plating tolerance. In fact, no frequency adjustment should be necessary if
the rough plating is reasonably well controlled.
The hysteresis and aging are lower for MCXO resonators than for TCXO resonators because
third-overtone SC-cut resonators are inherently more stable than the fundamental-mode AT-cut
resonators normally used in wide-temperature-range TCXO. Another possible factor may be that the
interface between the rough plating and the fine plating can be eliminated in MCXO resonators.
Hysteresis is the major limitation on the f vs. T stability that is achievable with the MCXO.
Although the majority of resonators so far have exhibited hysteresis at the low 10-8 level, some have
been in the 10-9 range. This shows that MCXO's with an f vs. T stability in the 10-9 range is a
reasonable goal for the future, but, for that goal to be met, further research is needed to gain a better
understanding of the mechanisms responsible for hysteresis.
Y. Kim, B. Rose, T. Schuyler & J.R. Vig, “Hysteresis Measurements of 20 MHz Third Overtone SC-cut
MCXO Resonators,” Proc. 1998 IEEE Int’l Frequency Control Symp., pp. 126-129, 1998.

Opto-Electronic Oscillator (OEO)
Bias
"Pump Laser" Optical out
Piezoelectric
RF driving port fiber stretcher
Filter
Electrical
output
RF coupler Optical
Fiber
Electrical
injection RF Amplifier
Photodetector Optical fiber
Optical Optical Electrical

Injection coupler transmission
line
2-23
The OEO utilizes the transmission characteristics of a modulator together with a fiber optic
delay line to convert light energy into spectrally pure rf/microwave reference signals. The OEO’s
schematic diagram is shown above. Light from a laser is introduced into an electrooptical (E/O)
modulator, the output of which is passed through a long fiber optic link, and detected with a
photodetector. The output of the photodetector is amplified, filtered and fed back to the electrical port
of the modulator. This configuration supports self-sustained oscillations at a frequency determined by
the fiber delay length, bias setting of the modulator, and the bandpass characteristics of the filter. It
also provides for both electrical and optical outputs.
The noise of an OEO has been measured to be -140 dBc/Hz at 10 kHz from a 10 GHz carrier.
This is the highest spectral purity demonstrated by an open loop oscillator in this frequency range (as
of 1999). How stable an OEO can be with respect to other parameters, such as temperature,
acceleration and humidity, is being investigated.
S. Yao and L. Maleki, “New Results with the Opto-electronic Oscillators (OEO),” Proc. 1996 IEEE Int’l
Frequency Control Symposium, pp. 1219-1222, 1996.
X. S. Yao and L. Maleki, "Optoelectronic Oscillator for Photonic Systems," IEEE Journal of Quantum
Electronics, 32, 1141, 1996.
X. S. Yao, L. Maleki and L. Davis, ”Coupled Opto-electronic Oscillators,” Proc. 1998 IEEE Int’l

CHAPTER 3
Quartz Crystal Resonators
General References
A. Ballato, "Piezoelectric Resonators," in B. Parzen, Design of Crystal and Other Harmonic

Oscillators, pp. 66-122, John Wiley and Sons, Inc., 1983.
V. E. Bottom, Introduction to Quartz Crystal Unit Design, Van Nostrand Reinhold Company, 1982.
R. A. Heising, Quartz Crystals for Electrical Circuits, D. Van Nostrand Co., 1946.
T. R. Meeker, “Theory and Properties of Piezoelectric Resonators and Waves,” in E. A. Gerber and A.
Ballato, Precision Frequency Control, Vol. 1, pp. 48-118, Academic Press, 1985.
J. A. Kusters, "Resonator and Device Technology," in E. A. Gerber and A. Ballato, Precision

E. Hafner, "Resonator and Device Measurements," in E. A. Gerber and A. Ballato, Precision

Frequency Control, Vol. 2, pp.1-44, Academic Press, 1985.
J. C. Brice, "Crystals for Quartz Resonators," Rev. of Modern Physics, Vol. 57, pp. 105-146, 1985.
J. R. Vig & F. L. Walls, "Fundamental Limits on the Frequency Instabilities of Quartz Crystal
Oscillators," Proc. 1994 IEEE Int'l Frequency Control Symposium, pp. 506-523, 1994; also, in IEEE
Trans. on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 42, pp. 576-589, July 1995.
J. R. Vig, "Introduction to Quartz Frequency Standards," Army Research Lab. Tech. Rep't SLCET-TR-
92-1 (Rev. 1), October 1992, AD-A256373; also in the Tutorials from the Twenty-third Annual Precise
Time
For Frequency Control and Timingand Time Interval
Applications (PTTI) Applications and Planning Meeting, Pasadena, CA, pp. 1-49,
- A TUTORIAL”
Rev. 8.5.1.2, by John R. December
Vig, July 2001,1991, AD-A254745, and at <http://www.ieee.org/uffc/fc>.
AD-M001251.
Why Quartz?
Quartz is the only material known that possesses the following
combination of properties:
• Piezoelectric ("pressure-electric"; piezein = to press, in Greek)
• Zero temperature coefficient cuts exist
• Stress compensated cut exists
• Low loss (i.e., high Q)
• Easy to process; low solubility in everything, under "normal" conditions,

except the fluoride and hot alkali etchants; hard but not brittle
• Abundant in nature; easy to grow in large quantities, at low cost, and

with relatively high purity and perfection. Of the man-grown single
crystals, quartz, at ~3,000 tons per year, is second only to silicon in
quantity grown (3 to 4 times as much Si is grown annually, as of 1997).
3-1

The Piezoelectric Effect
Y Y
_ _
_ _ + +
+ +
_ _
_ _ + +
+ +
_ _ _
_ +
+ + +
_ _ _
_ _ _ +
+ + + +
+
_ _ _ _ - _
+ + X _ + +
+ _ + • • + X
_ _ _ _
+ _ _
+ + + +
_ _ +
+ _ _
+ +
+
_ _
+ + _ _
+ +
_ _
+ + _ _
+ +
Undeformed lattice Strained lattice
The piezoelectric effect provides a coupling between the mechanical

properties of a piezoelectric crystal and an electrical circuit.
3-2
The direct piezoelectric effect was discovered by the Curie brothers in 1880. They showed that
when a weight was placed on a quartz crystal, charges appeared on the crystal surface; the
magnitude of the charge was proportional to the weight. In 1881, the converse piezoelectric effect
was illustrated; when a voltage was applied to the crystal, the crystal deformed due to the lattice
strains caused by the effect. The strain reversed when the voltage was reversed. The piezoelectric
effect can, thereby, provide a coupling between an electrical circuit and the mechanical properties of
the crystal. Under the proper conditions, a “good” piezoelectric resonator can stabilize the frequency
of an oscillator circuit.
Of the 32 crystal classes, 20 exhibit the piezoelectric effect (but only a few of these are useful).
Piezoelectric crystals lack a center of symmetry. When a force deforms the lattice, the centers of
gravity of the positive and negative charges in the crystal can be separated so as to produce surface
charges. The figure shows one example (from Kelvin’s qualitative model) of the effect in quartz. Each
silicon atom is represented by a plus, and each oxygen atom by a minus. When a strain is applied so
as to elongate the crystal along the Y-axis, there are net movements of negative charges to the left
and positive charges to the right (along the X-axis).
When a crystal has a center of symmetry, i.e., when the properties of the crystal are the same in
both directions along any line in the crystal, no piezoelectric effect can occur. Electrostriction,
however, exists in all dielectric solids. It is a deformation quadratic in the applied electric field
(whereas, piezoelectricity is a linear effect; reversal of the electric field reverses the mechanical
deformation.) Biased electrostriction, where small electric field variations are superimposed on a
constant component, is phenomenologically equivalent to linear piezoelectricity; this artifice may be
used with nonpiezoelectric crystals such as silicon, but the coupling depends upon the bias, and is
often small.
R. A. Heising, Quartz Crystals for Electrical Cicuits - Their Design and Manufacture, D. Van Nostrand
Co., New York, pp. 16-20, 1946.

The Piezoelectric Effect in Quartz
Z
FIELD along:
STRAIN
X Y Z
X √
EXTENSIONAL
along:
Y √
X
Z
X √
SHEAR Y √ Y
about:
Z √
In quartz, the five strain components shown may be generated by an electric field.
The modes shown on the next page may be excited by suitably placed and shaped
electrodes. The shear strain about the Z-axis produced by the Y-component of the
field is used in the rotated Y-cut family, including the AT, BT, and ST-cuts.
3-3
Piezoelectricity is a linear effect. Reversal of the electric field reverses the strain, i.e., the
mechanical deformation.
The electromechanical (also called piezoelectric) coupling factor k is an important characteristic
of a piezoelectric material; k is between zero and one and is dimensionless, e.g., k = 8.8% for AT-cut
quartz, and k = 4.99% for SC-cut quartz. It is a measure of the efficacy of piezoelectric transduction,
and it is a determinant of important device characteristics such as filter bandwidth, insertion loss, and
the location and spacings of resonators' critical frequencies (e.g., the series resonance to
antiresonance frequency spacing).
----------------------------
W. G. Cady, Piezoelectricity, Dover Publications, Inc., New York, 1964.
A. Ballato, Piezoelectricity: Venerable Effects, Modern Thrusts, Army Research Laboratory

Technical Report ARL-TR-70, August 1994. Available from NTIS, and among the “Review Papers” at
www.ieee-uffc.org/fc
G. S. Kino, Acoustic Waves: Devices, Imaging, and Analog Signal Processing, Prentice-Hall, Inc.,
1964. This book is available on-line to IEEE UFFC-S members at www.ieee-uffc.org/archive

Modes of Motion
Flexure Mode Extensional Mode Face Shear Mode
Thickness Shear Fundamental Mode Third Overtone

Mode Thickness Shear Thickness Shear
3-4
Shown above are the bulk acoustic wave (BAW) modes of motion. For example, AT-cut and SC-
cut resonators vibrate in the thickness shear mode. Above 100 MHz, overtone units that operate at a
selected harmonic mode of vibration are often used (e.g., third overtone or 5th overtone). Higher than
100 MHz fundamental mode units can be manufactured by, e.g., chemical polishing (diffusion
controlled wet etching), plasma etching, and ion milling techniques. Below 1 MHz, tuning forks, X-Y
and NT bars (flexure mode), +5° X-cuts (extensional mode), or CT-cut and DT-cut units (face shear
mode) can be used. Tuning forks have become the dominant type of low-frequency units due to their
small size and low cost (see “Quartz Resonators for Wristwatches” and the following pages later in
this chapter).
The velocities of acoustic waves in solids are typically ~3,000 m/s (~10-5 times the velocity of
light). For the shear waves in AT-cut quartz, for example, the velocity of propagation in the thickness
direction is 3,320 m/s; the fundamental mode frequency ~ v/2h, where v is the acoustic wave velocity
and h is the plate thickness. (The thickness of the plate is one half the wavelength.)

Motion Of A Thickness Shear Crystal
CLICK ON FIGURE
TO START MOTION
Note that only the area near the center moves. The edges are inactive in a properly designed
resonator.
This illustration and animation is “borrowed” from “Quartz Crystals vs. Their Environment: Time
Bases or Sensors? (Keeping the World on Time and Your Tanks Full of Gas),” Errol P. EerNisse,
2000 UFFC Society Distinguished Lecture. (Animation requires Funddemo.avi.)

Resonator Vibration Amplitude Distribution
Metallic
electrodes
Resonator
plate substrate
(the “blank”)
Conventional resonator geometry

and amplitude distribution, u
3-5
In an ideal resonator, the amplitude of vibration falls off approximately exponentially outside the
electrodes. In a properly designed resonator, a negligible amount of energy is lost to the mounting and
bonding structure, i.e., the edges must be inactive in order for the resonator to be able to possess a
high Q. The displacement of a point on the resonator surface is proportional to the drive current. At the
typical drive currents used in (e.g., 10 MHz) thickness shear resonators, the peak displacement is a few
atomic spacings.
The peak acceleration of a point on the surface is often more than a million ‘g’s. To show this, if
the displacement u = uo sin ωt, then, the acceleration = d2u/dt2 = - ω2 uo sin ωt, and the peak
acceleration = - ω2 uo. If we assume that uo = two lattice spacings ~1 x 10-9 m, then, at 10 MHz, ω2 uo =
(2π x 107)2 (10-9) ~ 106 g.
B. Capelle, J. Detaint, A. Zarka, Y. Zheng and J. Schwartzel, “Mode Shape Analysis Techniques Using
Synchrotron X-ray Topography,” Proc. 44th Ann. Symp. On Frequency Control, pp. 416-423, 1990.
J. S. Yang and H. F. Tiersten, “An Analysis of Contoured Quartz Resonators with Beveled Cylindrical
Edges,” Proc. 1995 Int’l Frequency Control Symp., pp. 727-739, 1995.
H. F. Tiersten and D. S. Stevens, “The Evaluationof the Coefficients of Nonlinear Resonance for SC-cut
Quartz Resonators,” Proc. 39th Annual Symposium on Frequency Control, pp. 325-332, 1985, IEEE
Catalog No. 85CH2186-5.

Resonant Vibrations of a Quartz Plate
0 db.
3555
3200
MHZ
Response -10 db.
3507
-20
3742
3383
-30 db.
3652
3642
3802
3852
3707
3256
-40 db.
3200 3400 3600 3800

Frequency, in kHz
X-ray topographs (21•0 plane) of various modes excited during a frequency
scan of a fundamental mode, circular, AT-cut resonator. The first peak, at
3.2 MHz, is the main mode; all others are unwanted modes. Dark areas
correspond to high amplitudes of displacement.
3-6
Resonators of finite dimensions have complicated mode specra. “Unwanted” modes occur
above the main resonance, as is shown above. These unwanted (also called “spurious”) modes are
especially troublesome in filter applications. In oscillator applications, controlling the unwanted modes
is not as critical because, as long as the unwanted modes are at least 10 dB below the main mode,
the oscillation will be at the main mode. Contouring, and “energy trapping” rules are used to minimize
the unwanted modes. The energy trapping rules consist of certain relationships between the
electrode and plate dimensions.
As Shockley, et. al pointed out, “energy trapping” and the concepts of “cutoff frequency” and
exponentially decaying waves in piezoelectric resonators are similar to the well-know such
phenomena in microwave and optical cavities (related to total internal reflection).
Mindlin has been quoted as having said about spurious modes “they aren’t spurious; they are
supposed to be there…the theory predicts them!”
-------------------------
K. Haruta and W. J. Spencer, "X-ray Diffraction Study of Vibrational Modes," Proc. 20th Annual
Symposium on Frequency Control, pp. 1-13, April 1966, AD-800523. Copies available from National
Technical Information Service.
W. J. Spencer, "Observation of Resonant Vibrations and Defect Structure in Single Crystals by

X-ray Diffraction Topography," in Physical Acoustics, Vol. V, W. P. Mason and R. N. Thurston, Eds.,
Academic Press, New York, 1968.
B. Capelle, J. Detaint, A. Zarka, Y. Zheng and J. Schwartzel, “Mode Shape Analysis Techniques
Using Synchrotron X-ray Topography,” Proc. 44th Ann. Symp. On Frequency Control, pp. 416-423,
1990.
W. Shockley, D. R. Curran & D. J. Koneval, “Energy Trapping and Related Studies of Multiple
Electrode Filter Crystals,” Proc. 17th Ann. Symp. On Frequency Control, pp. 88-124, 1963.
W. Shockley, D. R. Curran & D. J. Koneval, “Trapped Energy Modes in Quartz Crystal Filters,” J.
Acoustical Soc. Amer., vol. 41, pp. 981-993, 1967.

Overtone Response of a Quartz Crystal
jX
Spurious
Reactance
responses Spurious
Spurious
responses
responses
0
Frequency
5th overtone
3rd overtone
-jX
Fundamental mode
3-7
The above figure illustrates the mode spectrum of a quartz resonator, showing the fundamental
mode, third overtone, fifth overtone, and some of the spurious responses (“spurs”), i.e., unwanted
modes. In oscillator applications, the oscillator usually selects the strongest mode. Some of the
unwanted modes have steep frequency vs. temperature characteristics. Occasionally, as the
temperature changes, at a certain temperature, the frequency of an unwanted mode coincides with
the oscillator frequency, which causes an “activity dip” (see next page, and “Activity Dips” in chapter
4). At the activity dip, excitation of the unwanted mode results in extra energy dissipation in the
resonator, which results in a decrease in the Q, an increase in the equivalent series resistance, and a
change in the frequency of the oscillator. When the resistance increase is sufficiently large, the
oscillation stops, i.e., the oscillator fails. When the temperature changes away from the activity dip
temperature, the oscillation restarts.
Unwanted modes can be controlled by proper design and fabrication methods. Maintaining the
correct relationships among electrode and resonator plate dimensions (i.e., applying energy trapping
rules), and maintaining the parallelism between the major faces of the resonator plate, can minimize
the unwanted modes.
W. Shockley, D. R. Curran & D. J. Koneval, “Trapped Energy Modes in Quartz Crystal Filters,” J.
Acoustical Soc. Amer., vol. 41, pp. 981-993, 1967.
G.K. Guttwein, T. J. Lukaszek, and A. D. Ballato, “Practical Consequences of Modal Parameter

Control in Crystal Resonators,” Proc. 21st Ann. Symp. On frequency Control, pp. 115-137, 1967.

Unwanted Modes vs. Temperature
(3 MHz rectangular AT-cut resonator, 22 X 27 X 0.552 mm)
Activity dips occur where the f vs. T curves of unwanted modes intersect
the f vs. T curve of the wanted mode. Such activity dips are highly
sensitive to drive level and load reactance.
3-8
The above “mode chart” shows the frequency vs. temperature characteristics of the unwanted
modes of an AT-cut quartz resonator. The temperature coefficients range up to -91 ppm/oC. Proper
design can minimize the unwanted modes. For a given design, the unwanted responses are also a
function of the drive level.
H. Fukuyo, H. Yoshie, and M. Nakazawa, "The Unwanted Responses of the Crystal Oscillator
Controlled by AT-cut Plate," Proc. 21st Annual Symposium on Frequency Control, pp. 402-419, April
1967, AD-659792.

Mathematical Description of a Quartz Resonator
• In piezoelectric materials, electrical current and voltage are coupled to elastic displacement and stress:
{T} = [c] {S} - [e] {E}

{D} = [e] {S} + [∈] {E}
where {T} = stress tensor, [c] = elastic stiffness matrix, {S} = strain tensor, [e] = piezoelectric matrix
{E} = electric field vector, {D} = electric displacement vector, and [∈] = is the dielectric matrix
• For a linear piezoelectric material • Elasto-electric matrix for quartz

T1 c11 c12 c13 c14 c15 c16 −e11 −e21 −e31 S1 S1 S2 S3 S4 S5 S6 -E1 -E2 -E3
T2 c21 c22 c23 c24 c25 c26 −e12 −e22 −e32 S2
T1
T3 c31 c32 c33 c34 c35 c36 −e13 −e23 −e33 S3 et
T4 c41 c42 c43 c44 c45 c46 −e14 −e24 −e34 S4 T2
T5 = c51 c52 c53 c54 c55 c56 −e15 −e25 −e35 S5 T3
T6 c61 c62 c63 c64 c65 c66 −e16 −e26 −e36 S6 T4
D1 e11 e12 e13 e14 e15 e16 ∈11 ∈12 ∈13 E1 T5
D2 e21 e22 e23 e24 e25 e26 ∈21 ∈22 ∈23 E2
D3 e31 e32 e33 e34 e35 e36 ∈31 ∈32 ∈33 T6 CE X
E3
D1
where D2 6
T1 = T11 S1 = S11
T2 = T22 S2 = S22
D3 e ∈S 2
2
T3 = T33 S3 = S33 10
LINES JOIN NUMERICAL EQUALITIES
T4 = T23 S4 = 2S23 EXCEPT FOR COMPLETE RECIPROCITY
ACROSS PRINCIPAL DIAGONAL
T5 = T13 S5 = 2S13 INDICATES NEGATIVE OF
T6 = T12 S6 = 2S12 INDICATES TWICE THE NUMERICAL
EQUALITIES
X INDICATES 1/2 (c11 - c12)
3-9
cij in the matrix on the left ≡ cc in the box at the right, and ∈ ij in the matrix ≡ ∈ s at the right
-----------------
ANSI/IEEE Std 176-1987 IEEE Standard on Piezoelectricity
A. Ballato, “Piezoelectricity: Old Effect, New Thrusts,” IEEE Trans. UFFC, v.42, pp. 916-926, Sept.
1995

Mathematical Description - Continued
• Number of independent non-zero constants depend on crystal symmetry. For quartz (trigonal, class 32),
there are 10 independent linear constants - 6 elastic, 2 piezoelectric and 2 dielectric. "Constants” depend
on temperature, stress, coordinate system, etc.
• To describe the behavior of a resonator, the differential equations for Newton's law of motion for a
continuum, and for Maxwell's equation* must be solved, with the proper electrical and mechanical
boundary conditions at the plate surfaces. ∂Tij ∂Di
(F = ma ⇒ i ;
=ρu ∇ ⋅D = 0 ⇒ = 0,
∂xj ∂xi
∂φ 1 ∂ui ∂u j
Ei =− ; S = 2 ( ∂x j +
∂xi ) ; etc.)
∂x i ij
• Equations are very "messy" - they have never been solved in closed form for physically realizable three-
dimensional resonators. Nearly all theoretical work has used approximations.
• Some of the most important resonator phenomena (e.g., acceleration sensitivity) are due to nonlinear
effects. Quartz has numerous higher order constants, e.g., 14 third-order and 23 fourth-order elastic
constants, as well as 16 third-order piezoelectric coefficients are known; nonlinear equations are extremely
messy.
* Magnetic field effects are generally negligible; quartz is diamagnetic, however, magnetic fields can
affect the mounting structure and electrodes.
3-10

Infinite Plate Thickness Shear Resonator
n c ij
fn = , n = 1, 3, 5...
2h ρ
Where fn = resonant frequency of n-th harmonic
h = plate thickness
ρ = density
cij = elastic modulus associated with the elastic wave
being propagated
d(log fn ) 1 dfn − 1 dh 1 dρ 1 dc ij
Tf = = = − +
dT fn dT h dT 2ρ dT 2c ij dT
where Tf is the linear temperature coefficient of frequency. The temperature
coefficient of cij is negative for most materials (i.e., “springs” become “softer”
as T increases). The coefficients for quartz can be +, - or zero (see next page).
3-11
The velocity of propagation, v, of a wave in the thickness direction of a thin plate, and the
resonant frequency, fn, of an infinite plate vibrating in a thickness mode are
v=√(cij/ρ), and fn = n/2h√(cij/ρ)
respectively, where cij is the elastic stiffness associated with the wave being propagated, ρ is the
density of the plate, and 2h is the thickness. The frequency can also be expressed (taking the
logarithm of both sides) as
log fn = log(n/2) - log h + 1/2 (log cij - log ρ)
Differentiating this equation gives the above expression for Tf.
Quartz expands when heated; dh/dT is positive along all directions, and dρ/dT is negative (see
“Thermal Expansion Coefficients of Quartz” in chapter 4). The temperature coefficient of cij also varies
with direction. It is fortunate that, in quartz, directions exist such that the temperature coefficient of cij
balances out the effects of the thermal expansion coefficients, i.e., zero temperature coefficient cuts
exist in quartz.

Quartz is Highly Anisotropic
z The properties of quartz vary greatly with crystallographic direction.
For example, when a quartz sphere is etched deeply in HF, the
sphere takes on a triangular shape when viewed along the Z-axis, and
a lenticular shape when viewed along the Y-axis. The etching rate is
more than 100 times faster along the fastest etching rate direction (the
Z-direction) than along the slowest direction (the slow-X-direction).
z The thermal expansion coefficient is 7.8 x 10-6/°C along the Z-

direction, and 14.3 x 10-6/°C perpendicular to the Z-direction; the
temperature coefficient of density is, therefore, -36.4 x 10-6/°C.
z The temperature coefficients of the elastic constants range from

-3300 x 10-6/°C (for C12) to +164 x 10-6/°C (for C66).
z For the proper angles of cut, the sum of the first two terms in Tf on the
previous page is cancelled by the third term, i.e., temperature
compensated cuts exist in quartz. (See next page.)
3-12
See chapter 5, especially “Deeply Dissolved Quartz Sphere,” for additional information about
quartz and its anisotropy.

Zero
Zero Temperature
Temperature Coefficient Quartz Cuts
Cuts
90 o
60 o
AT FC
FC IT
IT
30o
LC
LC SC
SC
θ 0
zz -30 o SBTC
SBTC
BT
-60 o
-90 o
0o 10oo 20oo
20 30oo
30
θθ φ
DDoo The AT,
AT, FC,
FC, IT,
IT,SC,
SC,BT,
BT,and
andSBTC-cuts
SBTC-cuts are
are some
some
llyy
ngg d
SSiin ttaatteed
uubb
RRoo llyy
θ of the cuts
cuts on
on the
thelocus
locusofofzero
zerotemperature
temperature
oo
RR tt ttaatte coefficient cuts. The
coefficient cuts. TheLC
LCisisaa“linear
“linearcoefficient”
coefficient”
CCuu CCuut edd
t
y cut that
that has
has been
beenused
usedininaaquartz
quartzthermometer.
thermometer.
φφ Y-cut: ≈≈ +90
+90 ppm/00C
C
(thickness-shear mode)
(thickness-shear mode)
xx xl X-cut: ≈≈ -20
-20 ppm/00C
C
(extensional mode)
(extensional mode)
3-13
------------------------
The locus of zero-temperature-coefficient cuts in quartz is shown above. The cuts usually have two-letter
A. Ballato, “Doubly rotated thickness mode plate vibrators,” in Physical Acoustics, Vol. XIII,
names, where the "T" in the name indicates a temperature-compensated cut; for instance, the AT-cut was the first
pp.temperature-compensated
115-181, Academic Press, 1977. The FC, IT, BT, and SBTC-cuts are other cuts along the zero-
cut discovered.
temperature coefficient locus. These cuts were studied in the past (before the discovery of the SC-cut) for some
special properties, but are rarely used today. The highest-stability crystal oscillators employ SC-cut crystal units.
The X, Y, and Z directions have been chosen to make the description of properties as simple as possible. The Z-
axis is an axis of threefold symmetry in quartz; in other words, the physical properties repeat every 120° as the
crystal is rotated about the Z-axis.
------------------------------------
A. Ballato, "Doubly Rotated Thickness Mode Plate Vibrators," in Physical Acoustics, Vol. XIII, W. P. Mason and R.
N. Thurston, Eds., pp. 115-181, Academic Press, New York, 1977.
J. P. Buchanan, Handbook of Piezoelectric Crystals for Radio Equipment Designers, WADC Technical Report 56-
156, October 1956 (692 pages), available from NTIS, AD 110448.
D. L. Hammond, C. A. Adams and P. Schmidt, "A Linear, Quartz Crystal, Temperature Sensing Element," ISA
Transactions, vol. 4, pp. 349-354, 1965.
M. Valdois, B. K. Sinha, and J. J. Boy, “Experimental Verification of Stress Compensation in the SBTC-Cut,” IEEE
Trans. on Ultrasonics, Ferroelectrics and Frequency Control, vol. 36, pp. 643-651, 1989.

Comparison of SC and AT-cuts
• Advantages of the SC-cut
• Thermal transient compensated (allows faster warmup OCXO)
• Static and dynamic f vs. T allow higher stability OCXO and MCXO
• Better f vs. T repeatability allows higher stability OCXO and MCXO
• Far fewer activity dips
• Lower drive level sensitivity
• Planar stress compensated; lower ∆f due to edge forces and bending
• Lower sensitivity to radiation
• Higher capacitance ratio (less ∆f for oscillator reactance changes)
• Higher Q for fundamental mode resonators of similar geometry
• Less sensitive to plate geometry - can use wide range of contours
• Disadvantage of the SC-cut : More difficult to manufacture for OCXO (but is

easier to manufacture for MCXO than is an AT-cut for precision TCXO)
• Other Significant Differences

• B-mode is excited in the SC-cut, although not necessarily in LFR's
• The SC-cut is sensitive to electric fields (which can be used for
compensation)
3-14
See the next chapter for specific comparisons between AT-cut and SC-cut resonators, e.g., see
“Warmup of AT- and SC-cut Resonators”.
A. Ballato, "Doubly Rotated Thickness Mode Plate Vibrators," Physical Acoustics, Vol. XIII, pp. 115-
181, Academic Press, Inc., 1977.
J. A. Kusters, "The SC Cut Crystal - An Overview," Proc. IEEE Ultrasonics Symposium, pp. 402-409,
1981.

Mode Spectrograph of an SC-cut
1.10
0
1.0 a-mode: quasi-longitudinal mode

b-mode: fast quasi-shear mode
1.88
c-mode: slow quasi-shear mode
-10
Attenuation
3.30
-20
3.0
5.65
5.50
-30 5.0
c(1) b(1) a(1) c(3) b(3) c(5) b(5) a(3)

-40
0 1 2 3 4 5 6
Normalized Frequency (referenced to the fundamental c-mode)
3-15
In the above mode spectrograph of an SC-cut resonator, the Nth overtone of mode m is labeled
m(N) (after Ballato). The frequency of each resonance is normalized to that of the fundamental c-
mode. The modal attenuations vary with resonator design. In the above example, the b-modes’
strengths are greater than the c-modes’ (which is often the case). Therefore, the oscillator must
include mode selection circuitry in order to insure that the oscillator operates on the desired c-mode.

J. A. Kusters, M. C. Fischer, J. G. Leach, "Dual mode operation of temperature and stress

compensated crystals," Proc. of the 32nd Annual Frequency Control Symposium, 1978. pp. 389-397.
R. Bourquin, J.J. Boy, B. Dulmet, “SC-cut Resonator With Reduction of b-mode Electrical Response,”
Proc. 1997 IEEE Int’l Frequency Control Symposium, pp. 704-709.

SC- cut f vs. T for b-mode and c-mode
400
200
FREQUENCY DEVIATION (PPM) 0 10 20 Temperature (OC)

0
30 40 50 60 70
-200 c-Mode (Slow Shear)
-400
b-Mode (Fast Shear)

-600
-25.5 ppm/oC
-800
-1000
-1200
3-16
Shown above are frequency vs. temperature characteristics of the b-mode and c-mode of a 10
MHz 3rd overtone SC-cut resonator. The b-mode’s f vs. T is monotonic, with a slope of about -25.5
ppm per degree C. The c-mode is temperature compensated; its exact f vs. T characteristic depends
on the angles of cut, overtone, contour, etc. The b-mode’s frequency is about 9.4% higher than the c-
mode’s (but it can range from ~9% to 10%, depending on design).
Both modes can be excited simultaneously, without the modes interfering with each other. The
b-mode frequency can be used to instantaneously indicate the temperature of the resonator’s active
volume. Dual mode excitation of the b-mode and c-mode frequencies has been used for temperature
compensation. In later work, however, dual mode excitation of the fundamental and third overtone c-
modes was shown to be superior - see “Resonator Self-temperature Sensing” and subsequent pages
in Chapter 2.
J. A. Kusters, M. C. Fischer, J. G. Leach, "Dual mode operation of temperature and stress

compensated crystals," Proc. of the 32nd Annual Frequency Control Symposium, 1978. pp. 389-397.

B and C Modes Of A Thickness Shear Crystal
C MODE B MODE
CLICK ON FIGURES
TO START MOTION
3-17
Note that the c-mode and b-mode displacements are orthogonal.
This illustration and animation is “borrowed” from “Quartz Crystals vs. Their Environment: Time
Bases or Sensors? (Keeping the World on Time and Your Tanks Full of Gas),” Errol P. EerNisse,
2000 UFFC Society Distinguished Lecture. (Animations require Discbfull.avi and diskcmode.avi.)

Singly Rotated and Doubly Rotated Cuts’
Vibrational Displacements
Z
Singly rotated resonator
gl y
Sin ated Do
Ro
t u
Ro bly
θ
t
Cu ta
Cu ted
t
Y Doubly rotated resonator
ϕ
X
X’
3-18
In a vibrating doubly rotated resonator (θ ≈ 35° and ϕ > 0°), the

displacement is partly out of the plane of the plate, as is illustrated. By controlling the ϕ−
angle, one may control the ratio of in-plane to out-of-plane displacements.
A sensor for sensing the properties of a fluid has been proposed based on
this. In a fluid, the out-of-plane component of the displacement propagates a damped
compressional wave, while the in-plane component propagates a damped shear wave. The
frequency changes of doubly rotated resonators have been measured in glycerol solutions of
a variety of concentrations. At each concentration, the frequency change was found to
increase with increasing πηι−angle.
--------------------------
Y. Kim, J. R. Vig and A. Ballato, “Sensing the properties of liquids with doubly rotated resonators,”
Proc. 1998 IEEE Int’l Frequency Control Symp., 1998, pp. 660-666.

Resistance vs. Electrode Thickness
AT-cut; f1=12 MHz; polished surfaces; evaporated 1.2 cm (0.490”) diameter silver electrodes
60
5th
40
RS (Ohms)
20 3rd
Fundamental
0
10 100 1000
−∆f(kHz) [fundamental mode]
3-19
After the electrode deposition starts, the atoms on the quartz surface form
islands through which no current flows. As the deposition continues, the islands grow.
Eventually, the islands touch, form larger islands, then further coalesce to form a continuous
film. At this point, Rs is large due to the high resistivities of the ultrathin films. As the film
thickness (the “plateback”) increases, the film resistivity decreases. Eventually, the Q of the
resonator rather than the film resistance determines Rs. As the film becomes thick, losses in
the film decrease the Q and increase the Rs.
The “Quartz Resonator Handbook…” referenced below suggests minimum and maximum
platebacks for fundamental mode, 3rd overtone and 5th overtone resonators, and for Al, Ag
and Au electrodes (p. 115, Table 5).
The figure above is adapted from the “Quartz Resonator Handbook, Manufacturing
Guide for AT Type Units,” R.E. Bennett, Union Thermoelectric Division, p, 113, Figure 51,
1960. A copy of this publication can be found at http://www.ieee-uffc.org/archive (available to
IEEE UFFC Society members).
C. A. Neugebauer, “Condensation, Nucleation, and Growth of Thin Films,“ in Handbook of Thin Film Technology,
L. I. Maissel & R. Glang editors, Chapt. 8, McGraw Hill Book Co., 1970.

Resonator Packaging
Two-point Mount Package Three- and Four-point Mount Package
Quartz Electrodes Quartz

blank blank
Bonding
Cover area Bonding
area
Mounting Cover
clips
Mounting
Seal Base clips
Pins Seal Pins

Base
Top view of cover
3-20
The above illustration shows the typical packages used for AT-cut and SC-cut resonators. Such
packages are available in both metal and glass constructions. Other package types include surface
mount, ceramic and metal flatpacks, and the cylindrical metal packages used for tuning fork
resonators - see “Watch Crystal” later in this chapter.
J. A. Kusters, “Resonator and Device Technology," in E. A. Gerber and A. Ballato, Precision


Equivalent Circuits
Spring
C
Mass L
Dashpot R
3-21
The mechanically vibrating system and the circuit shown above are "equivalent," because each can
be described by the same differential equation. The mass, spring and damping element (i.e., the
dashpot) correspond to the inductor, capacitor and resistor. The driving force corresponds to the voltage,
the displacement of the mass to the charge on the capacitor, and the velocity to the current.
A crystal resonator is a mechanically vibrating system that is linked, via the piezoelectric effect, to
the electrical world. In the (simplified) equivalent circuit (of one mode of vibration) of a resonator, on the
next page, C0 is called the “shunt” capacitance. It is the capacitance due to the electrodes on the crystal
plate (plus the stray capacitances due to the crystal enclosure). The R1, L1, C1 portion of the circuit is the
"motional arm" which arises from the mechanical vibrations of the crystal. The C0 to C1 ratio is a
measure of the interconversion between electrical and mechanical energy stored in the crystal, i.e., of the
piezoelectric coupling factor, k, and C1 is a measure of the crystal's "stiffness," i.e., its tunability - see the
equation under the equivalent circuit on the next page. When a dc voltage is applied to the electrodes of
a resonator, the C0/C1 is also a measure of the ratio of electrical energy stored in the capacitor formed by
the electrodes to the energy stored elastically in the crystal due to the lattice strains produced by the
piezoelectric effect. The C0/C1 is also a measure of the antiresonance-resonance frequency separation.
(Let r = C0/C1, then fA - fR≈fR/2r, and 2r = (πN/2k)2, where N = 1,3,5...is the overtone number.)
Some of the numerous advantages of quartz crystal resonator over a tank circuit built from discrete
R's, C's and L's are that the crystal is far stiffer and has a far higher Q than what could be built from
normal discrete components. For example, a 5 MHz fundamental mode AT-cut crystal may have C1 =
0.01 pF, L1 = 0.1 H, R1 = 5 Ω, and Q = 106. A 0.01pF capacitor is not available, since the leads attached
to such a capacitor would alone probably contribute more than 0.01 pF. Similarly, a 0.1 H inductor would
be physically large, it would need to include a large number of turns, and would need to be
superconducting in order to have a ≤ 5 Ω resistance.
E. Hafner, "Resonator and Device Measurements," in E. A. Gerber and A. Ballato, Precision Frequency
Control, Vol. 2, pp.1-44, Academic Press, 1985.

Equivalent Circuit of a Resonator
Symbol for crystal unit CL

C0
CL
C1 L1 R1
∆f C1
{
1. Voltage control (VCXO)
≈ →
fS 2(C0 + CL )
2. Temperature compensation
(TCXO)
3-22
The oscillator designer treats the crystal unit as a circuit component and just deals with the
crystal unit's equivalent circuit. Shown above is a simple equivalent circuit of a single-mode quartz
resonator. A resonator is a mechanically vibrating system that is linked, via the piezoelectric effect, to
the electrical world. A load capacitor CL is shown in series with the crystal. C0, called the "shunt"
capacitance, is the capacitance due to the electrodes on the crystal plate plus the stray capacitances
due to the crystal enclosure. The R1, L1, C1 portion of the circuit is the "motional arm" which arises
from the mechanical vibrations of the crystal.
The C0 to C1 ratio is a measure of the interconversion between electrical and mechanical
energy stored in the crystal, i.e., of the piezoelectric coupling factor, k. C0/C1 increases with the
square of the overtone number; the relationship of C0/C1 to k and N is 2C0/C1 = [πN/2k]2, where N is
the overtone number. When a dc voltage is applied to the electrodes of a resonator, the capacitance
ratio C0/C1 is also a measure of the ratio of electrical energy stored in the capacitor formed by the
electrodes to the energy stored elastically in the crystal due to the lattice strains produced by the
piezoelectric effect. The C0/C1 is also inversely proportional to the antiresonance-resonance frequency
separation (i.e., the pole-zero spacing - see the page after next) which is an especially important
parameter in filter applications. The slope of the reactance vs. frequency curve near fs is inversely
proportional to C1, i.e., ∆X/(∆f/f) ~ 1/πfC1 near fs, where X is the reactance. C1 is, therefore, a
measure of the crystal's "stiffness," i.e., its tunability. For a simple RLC circuit, the width of the
resonance curve is inversely proportional to the quality factor Q, but in a crystal oscillator, the situation
is complicated by the presence of C0 and by the fact that the operating Q is lower than the resonator
Q. For a quartz resonator, Q = (2πfsC1R1)-1.
When the load capacitor is connected in series with the crystal, the frequency of operation of
the oscillator is increased by a ∆f, where ∆f is given by the equation below the equivalent circuit. A
variable load capacitor can thus be used to vary the frequency of the resonator-capacitor combination,
which may be applied in, e.g., a VCXO or TCXO.
----------------------------
V. E. Bottom, Introduction to Quartz Crystal Unit Design, Chapters 6 and 7, Van Nostrand Reinhold
Company, 1982.

Crystal Oscillator f vs. T Compensation
Uncompensated
frequency
Frequency / Voltage
Compensating Compensated
voltage frequency
on varactor CL of TCXO
3-23
When the load capacitor is connected in series with the crystal, the frequency of operation of
the oscillator is increased by a ∆f', where ∆f' is given by the equation on the previous page. When an
inductor is connected in series with the crystal, the frequency of operation is decreased. The ability to
change the frequency of operation by adding or changing a reactance allows for compensation of the
frequency versus temperature variations of crystal units in TCXOs, and for tuning the output frequency
of voltage controlled crystal oscillators (VCXO). In both, the frequency can be changed, e.g., by
changing the voltage on a varactor.
Other means of temperature compensation include the use of a temperature sensitive
reactance element such that the variations of the reactance with temperature compensate for the f vs.
T variations of the resonator, and the use digital compensation techniques. The microcomputer
compensated crystal oscillator (MCXO), which uses a high-accuracy digital compensation technique,
is discussed in chapter 2.
M. E. Frerking, “Temperature Control and Compensation,” in E. A. Gerber and A. Ballato, Precision

M. E. Frerking, Crystal Oscillator Design and Temperature Compensation, Van Nostrand Reinhold
Company, 1978.
S. Okano, T. Mitsuoka & T. Ohshima, “Direct-temperature Compensated Crystal Oscillator for

Advanced VHF/UHF Radio Communication Systems,” Proc. 34th Ann. Frequency Control
Symposium, pp. 488-497, 1980.

Resonator Reactance vs. Frequency
Area of usual
+ operation in an
oscillator
Reactance
Resonance, fr Antiresonance, fa
0 Frequency
1
2πfC0
3-24
At frequencies far from the resonance frequency, the resonator is a simple parallel plate
capacitor having a capacitance Co ≅ kεo(A/t), where A is the area of the electrodes, t is the thickness
of the plate, k is the dielectric constant and εo is the permittivity of free space. The reactance is zero at
resonance, and it is maximum at the antiresonance frequency. The antiresonance to resonance
separation, fa - fr ≈ fr/2r’ – r’/2Q2, where r’ = Co/C1.
In most oscillator circuits, the frequency is in the region shown, where the resonator’s reactance
is inductive. An adjustable capacitance in series (or parallel) with the resonator can then used to
adjust the frequency of oscillation.
The reactance vs. fractional frequency slope, ∆X/(∆f/f), is a measure of the resonator’s
“stiffness,” i.e., the amount the resonator’s frequency changes with load capacitance. The stiffer the
resonator, the less the resonator’s frequency changes with a change in load capacitance. Near fS,
∆X/(∆f/f) = 1/(πfSC1). Overtone resonators are stiffer than fundamental mode units because the C1 of
overtone resonators is smaller than the C1 of fundamental mode units.


V. E. Bottom, Introduction to Quartz Crystal Unit Design, Chapters 6 and 7, Van Nostrand
Reinhold Company, 1982.

Equivalent Circuit Parameter Relationships
n: Overtone number
A C C0: Static capacitance
C0 ≅ ε r≡ 0 C1: Motional capacitance
t C1 C1n: C1 of n-th overtone
L1: Motional inductance
L1n: L1 of n-th overtone
1 1 fs R1: Motional resistance
fs =
2π L1C1
fa − fs ≅ R1n: R1 of n-th overtone
2r ε: Dielectric permittivity of quartz
≈40 pF/m (average)
A: Electrode area
1 t: Plate thickness
Q=
1 ω L1 −
2π fSR1C1 ϕ=
ω C1 r:
r’:
Capacitance ratio
f1/fn
R1 fs: Series resonance frequency ≈fR
fa: Antiresonance frequency
Q; Quality factor
dϕ 360 Q τ1: Motional time constant
τ 1 = R1C1 ≅ 10 −14 s ≅ ω: Angular frequency = 2πf
df π fs ϕ: Phase angle of the impedance
k; Piezoelectric coupling factor
=8.8% for AT-cut, 4.99% for SC
r' C11 n3L11 n3R11 ⎛π n ⎞
2
C1n ≈ L1n ≈ R1n ≈ 2r = ⎜ ⎟
n3 r '3 r' ⎝ 2k ⎠
3-25
In the last row, the r’ is the ratio between the fundamental mode and the n-th overtone
frequencies. For example, when comparing the parameters of a 10 MHz fundamental mode and a 10
MHz 3rd overtone (of similar design), r’ = 1, and when comparing the parameters of a 10 MHz
fundamental mode resonator when excited at the fundamental mode, and when excited at the 3rd
overtone frequency, i.e., at ~30 MHz, r’ = 3.

What is Q and Why is it Important?
Energy stored during a cycle
Q≡2π
Energy dissipated per cycle
Q is proportional to the decay-time, and is inversely

proportional to the linewidth of resonance (see next page).
• The higher the Q, the higher the frequency stability and

accuracy capability of a resonator (i.e., high Q is a
necessary but not a sufficient condition). If, e.g., Q = 106,
then 10-10 accuracy requires ability to determine center of
resonance curve to 0.01% of the linewidth, and stability (for
some averaging time) of 10-12 requires ability to stay near
peak of resonance curve to 10-6 of linewidth.
• Phase noise close to the carrier has an especially strong

dependence on Q (L(f) ∝ 1/Q4).
3-26
See the next page for other definitions of Q, and see chapter 5 for additional information about
the Q of quartz resonators. When the signal is decaying, as shown on the next page, the energies in
the definition above are averaged over the cycle. Close to the carrier, a factor of two difference in Q
results in a factor of 16 difference in phase noise.
IEEE Std 100-1996, The IEEE Standard Dictionary of Electrical and Electronics Terms,
http://shop.ieee.org/store/
H. Hellwig, "Frequency Standards and Clocks: A Tutorial Introduction," NBS Technical Note 616,
1977, Time and Frequency Division, NIST, Boulder, CO 80303.
V. B. Braginsky, V. P. Mitrofanov & V. I. Panov, Systems with Small Dissipation, The University of
Chicago Press, 1985.
E. I. Green, "The Story of Q," American Scientist, pp. 584-595, 1955.

Decay Time, Linewidth, and Q
Decaying oscillation
of a resonator
Oscillation
1 1
= of maximum intensity
e 2.7
TIME
Exciting
pulse ends
Max.
intensity
1
BW ≅
td πt d
Maximum intensity Resonance
ν behavior of
BW
Q = o ≅ νo π t d a resonator
½ Maximum intensity
BW
ν0 FREQUENCY
3-27
In addition to the definition on the previous page, equivalent definitions of Q are shown above.
Q is the frequency divided by the bandwidth of resonance, and it also determines the rate at which a
signal decays after the vibration excitation stops - the higher the Q, the narrower the bandwidth and
the longer it takes for the excitation to decay. Q is proportional to the time it takes for the signal to
decay to 1/e of the amplitude of vibration prior to the cessation of excitation. This relationship is used
in one method (sometimes referred to as the “logarithmic decrement” method) of measuring Q.
The relationship between Q and decay time is also relevant to oscillator startup time. When an
oscillator is turned on, it takes a finite amount of time for the oscillation to build up. The oscillator’s
startup time depends on the loaded Q of the resonator in the sustaining circuit, and the loop gain of
the circuit.
1977, Time and Frequency Division, NIST, Boulder, CO 80303.

Factors that Determine Resonator Q
The maximum Q of a resonator can be expressed as:
1
Q max = ,
2π fτ
where f is the frequency in Hz, and τ is an empirically determined “motional
time constant” in seconds, which varies with the angles of cut and the mode
of vibration. For example, τ = 1 x 10-14s for the AT-cut's c-mode (Qmax = 3.2
million at 5 MHz), τ = 9.9 x 10-15s for the SC-cut's c-mode, and τ = 4.9 x 10-15s
for the BT-cut's b-mode.
Other factors which affect the Q of a resonator include:
z Overtone z Blank geometry (contour,

z Surface finish dimensional ratios)
z Material impurities and defects z Drive level
z Mounting stresses z Gases inside the enclosure
z Bonding stresses (pressure, type of gas)
z Temperature z Interfering modes
z Electrode geometry and type z Ionizing radiation
3-28
The motional time constant τ = R1C1 (= 10-14 s for an AT-cut) is related to the internal friction, i.e.,
to the attenuation of an acoustic wave as the wave travels in a resonator. It is independent of
frequency.
See also “Ions in Quartz…” and “Internal Friction in Quartz” at the end of Chapter 5.


Resonator Fabrication Steps
DESIGN GROW
SWEEP CUT LAP ROUND
RESONATORS QUARTZ
ETCH
ORIENT ANGLE X-RAY
CLEAN (CHEMICAL CONTOUR
IN MASK CORRECT ORIENT
POLISH)
DEPOSIT PREPARE
MOUNT BOND INSPECT CLEAN
CONTACTS ENCLOSURE
TEST
FREQUENCY FINAL
SEAL PLATE BAKE
ADJUST CLEAN
OSCILLATOR
3-29
In the manufacturing of typical quartz resonators, wafers are cut from a quartz crystal bar into
plates (“blanks”), along precisely controlled directions with respect to the crystallographic axes. The
cutting is usually done with a slurry saw. This saw consists of a set of ~100 thin, stretched metal
bands moving back and forth across multiple quartz bars in a flood of abrasive slurry. The properties
of the resonator depend strongly on the angles of cut (see chapter 4). The angles of cut are
determined by x-ray diffraction. After shaping to required dimensions, lapping, etching, (polishing) and
cleaning, metal electrodes are applied to the plates which are mounted into holder structures (see
“Resonator Packaging” earlier in this chapter). The plates are bonded to the mounting clips of the
holder with, e.g., a silver-filled epoxy or polyimide. The assembly, called a crystal unit (or crystal or
resonator) is hermetically sealed.
The above flow chart shows the major steps in the fabrication of a resonator. Although all the
steps can affect the stability, the steps shown in the box are usually the most important with respect to
the long term stability of the resonator. Ideally, these final steps should be performed in an ultrahigh
vacuum. Ultrahigh vacuum baking immediately before hermetic sealing is highly desirable in order to
produce low aging, especially if the resonator is exposed to air after frequency adjustment. (See
“Aging Mechanisms” in chapter 4.)
Quartz growing, sweeping and etching/chemical-polishing are discussed in chapter 5.
W.L. Bond, Crystal Technology, John Wiley & Sons, New York, 1976.


X-ray Orientation of Crystal Plates
Shielding
Monochromator
crystal
Detector
X-ray beam
Crystal under test
Copper target
X-ray source
Goniometer
Double-crystal x-ray diffraction system
3-30
-
The frequency vs. temperature characteristics of resonators depend on the angles of cut of the
quartz plate with respect to the crystallographic axes (see the section of Chapter 4 starting with
“Quartz Wristwatch Accuracy vs. Temperature”). In some applications, seconds of arc accuracies are
required. Because of imperfections in the cutting techniques (and in the quartz), the angles of cut of
each blank must be measured, the blanks must be sorted, and, if necessary, angle-corrected in order
to achieve the required angle-of-cut precision.
Double-crystal X-ray diffraction is generally used to measure the angles of cut - via measuring
the angle between the major surface of a blank and a specified set of atomic planes. X-rays are
reflected from atomic planes in a crystal in accordance with Bragg’s law: nλ = 2d sin θ, where n is an
integer, λ is the wavelength of the reflected X-rays, d is the distance between the reflecting atomic
planes, and θ is the “Bragg angle,” the angle at which the peak of the reflection occurs. In most X-ray
orientation systems, the Kα radiation from a copper target is used because the wavelength of this
radiation is near the typical atomic spacings.
In the above drawing, the monochromator crystal serves to collimate the X-rays (which allows a
more accurate determination of the Bragg angle than is possible with a single-crystal orientation
system), and the goniometer allows variation of the angle of incidence of the X-rays and the
determination of the angle at which the reflection is maximum. When a laser is used to define the
plane of the blank, a measurement precision of ~2 seconds is possible.
------------------------------------
W.L. Bond, Crystal Technology, John Wiley & Sons, New York, 1976.
R. A. Heising, Quartz Crystals for Electrical Circuits, D. Van Nostrand Co., 1946.
V. E. Bottom, Introduction to Quartz Crystal Unit Design, Van Nostrand Reinhold Company, Chapter
11, 1982.
J.R. Vig, “A High Precision Laser Assisted X-Ray Goniometer For Circular Plates,” Proc. 29th Ann.
Symposium on Frequency Control, pp. 240-247, 1975.

Contamination Control
Contamination control is essential during the fabrication of
resonators because contamination can adversely affect:
• Stability (see chapter 4)
- aging
- hysteresis
- retrace
- noise
- nonlinearities and resistance anomalies (high starting
resistance, second-level of drive, intermodulation in filters)
- frequency jumps?
• Manufacturing yields
• Reliability
3-31
-
To illustrate the importance of contamination control, consider an AT-cut resonator. The lattice
spacing is ~0.52 nm, and the frequency (f)-thickness (t) product is 1.66 MHz-mm. Therefore, a 10 MHz
3rd overtone (~3 MHz fundamental mode) AT-cut quartz plate is about 1 million atomic layers thick.
As f ∝1/t, it follows that ∆f/f = -∆t/t. Therefore, for t = 106 atomic layers, an aging rate of 1 x 10-10
per day requires that ∆t/t < 10-4 atomic layers per day.
This is not easily achieved because in a vacuum, if all the molecules striking a surface stick:
• At 10-6 torr, a monolayer forms in 1 second; 10-4 monolayer in 10-4 s
• Even at 10-9 torr, a monolayer forms in 1000 s (= 16 min); 10-4 monolayer forms in 0.1 s.
Therefore, it is essential that surface contamination be removed from the resonator and enclosure
surfaces prior to hermetically sealing the resonators. Transfer of contamination to and from the
resonator’s surfaces is one of the important causes of the instabilities listed above.
Surface contamination can also leads to poor or nonuniform adhesion of the electrodes and
bonding agents, which can lead to yield and reliability problems. Particulate contamination can lead to
high starting resistance (i.e., high resistance at low drive levels), second-level of drive effects, and
intermodulation in crystal filters. Unfiltered factory air typically contains 2 million to 10 million > 0.5 µm
particles per m3, and a slowly walking person emits > 1 million such particles per minute.
-----------------------------------
Handbook of Semiconductor Wafer Cleaning Technology, W. Kern editor, Noyes Publications, Mill Rd. at Grand
Ave., Park Ridge, NJ 07656, U.S.A., 1992
Treatise on Clean Surface Technology, edited by K. L. Mittal, Plenum Press, 1987.
Environmental Control in Electronic Manufacturing, edited by P. W. Morrison, Van Nostrand Reinhold, 1973.
Handbook of Thin Film Technology, L. I. Maissel & R. Glang editors, Chapt. 2, McGraw Hill Book Co., 1970.
Contamination Control Handbook, NASA Publication SP-5076, 1969; NTIS accession no. N70-13566.

Crystal Enclosure Contamination
The enclosure and sealing process can have important
influences on resonator stability.
• A monolayer of adsorbed contamination contains ~ 1015

molecules/cm2 (on a smooth surface)
• An enclosure at 10-7 torr contains ~109 gaseous
molecules/cm3
Therefore:
In a 1 cm3 enclosure that has a monolayer of contamination
on its inside surfaces, there are ~106 times more adsorbed
molecules than gaseous molecules when the enclosure is sealed
at 10-7 torr. The desorption and adsorption of such adsorbed
molecules leads to aging, hysteresis, noise, etc.
3-32
Enclosure Outgassing Example: Let the average outgassing rate of an enclosure during the
1st day after it is hermetically sealed = 2 x 10-9 torr liter/sec/cm2, let the enclosure volume = 1cm3, let
the enclosure area = 5cm2, and let the sealing pressure = 0. Then, at the end of the first day after
sealing, the pressure inside the enclosure will have risen to (2 x 10-9) (103) (86,000) (5) torr = 0.86 torr.
The number of molecules in the enclosure is then
(0.86/760)*(6.02 x 1023) /22,400 = 3.0 x 1016 molecules.
If a blank diameter is 14mm and the lattice spacing is 5 x 10-8 cm, then number of lattice sites on
blank is
3.1 cm2/(5 x 10-8 cm)2 = 1.2 x 1015 lattice sites.
Therefore there are 3.0 x 1016/1.2 x 1015 = 25 outgassed molecules per lattice site.
If the enclosure had been vacuum baked at, e.g., > 250 0C just before sealing, then the outgassing
rate would have been about 1000 times lower (at 20 0C), and the aging and retrace characteristics
would have been substantially better.
Desorption rate (the time it takes before a molecule desorbs from a surface) depends
exponentially on temperature. A contaminant molecule that did not desorb during a > 250 0C bake is
unlikely to desorb at the normal operating temperatures of a resonator (i.e., the molecule will have a
long lifetime on the surface).

What is an “f-squared”?
It is standard practice to express the thickness removed by lapping, etching and polishing,
and the mass added by the electrodes, in terms of frequency change, ∆f, in units of “f2”, where
the ∆f is in kHz and f is in MHz. For example, etching a 10MHz AT-cut plate 1f2 means that a
thickness is removed that produces ∆f= 100 kHz; and etching a 30 MHz plate 1f2 means that
the ∆f= 900 kHz. In both cases, ∆f=1f2 produces the same thickness change.
To understand this, consider that for a thickness-shear resonator (AT-cut, SC-cut, etc.)
N
f=
t
where f is the fundamental mode frequency, t is the thickness of the resonator plate and N is
the frequency constant (1661 MHz•µm for an AT-cut, and 1797 MHz•µm for a SC-cut’s c-
mode). Therefore,
∆f ∆t
=−
f t
and,
∆f
∆t = −N
f2
So, for example, ∆f = 1f2 corresponds to the same thickness removal for all frequencies.
For an AT-cut, ∆t=1.661 µm of quartz (=0.83 µm per side) per f2. An important advantage of
using units of f2 is that frequency changes can be measured much more accurately than
thickness changes. The reason for expressing ∆f in kHz and f in MHz is that by doing so, the
numbers of f2 are typically in the range of 0.1 to 10, rather than some very small numbers.
3-33

Milestones in Quartz Technology
1880 Piezoelectric effect discovered by Jacques and Pierre Curie
1905 First hydrothermal growth of quartz in a laboratory - by G. Spezia
1917 First application of piezoelectric effect, in sonar
1918 First use of piezoelectric crystal in an oscillator
1926 First quartz crystal controlled broadcast station
1927 First temperature compensated quartz cut discovered
1927 First quartz crystal clock built
1934 First practical temp. compensated cut, the AT-cut, developed
1949 Contoured, high-Q, high stability AT-cuts developed
1956 First commercially grown cultured quartz available
1956 First TCXO described
1972 Miniature quartz tuning fork developed; quartz watches available
1974 The SC-cut (and TS/TTC-cut) predicted; verified in 1976
1982 First MCXO with dual c-mode self-temperature sensing
3-34
The Frequency Control website <http://www.ieee.org/uffc/fc> contains a collection of historical

papers on both quartz and atomic frequency control technologies.
W. A. Marrison, "The Evolution of the Quartz Crystal Clock," Bell System Technical Journal, Vol. 27,
pp. 510-588, 1948.
F. D. Lewis, "Frequency and Time Standards," Proc. of the Institute of Radio Engineers, pp. 1046-
1068, Sept. 1955.
V. E. Bottom, "A History of the Quartz Crystal Industry in the USA," Proc. 35th Annual Symposium on
Frequency Control, pp. 3-12, 1981.

Quartz Resonators for Wristwatches
Requirements:
• Small size
• Low power dissipation (including the oscillator)
• Low cost
• High stability (temperature, aging, shock,

attitude)
These requirements can be met with 32,768 Hz quartz

tuning forks
3-35
E. Momosaki, “A Brief Review of Progress in Quartz Tuning Fork Resonators,” Proc. 1997 IEEE Int'l

Why 32,768 Hz?
32,768 = 215 32,768
16,384
• In an analog watch, a stepping motor receives 8,192
one impulse per second which advances the 4,096
2,048
second hand by 6o, i.e., 1/60th of a circle,
1,024
every second. 512
256
• Dividing 32,768 Hz by two 15 times results 128
in 1 Hz. 64
32
• The 32,768 Hz is a compromise among size, 16
power requirement (i.e., battery life) and 8
stability. 4
2
1
3-36

Quartz Tuning Fork
Z
Y
X
a) natural faces and crystallographic axes of quartz

Z
Y’
0~50
Y
arm
base
X
b) crystallographic orientation of tuning fork c) vibration mode of tuning fork
3-37
See also “Quartz Wristwatch Accuracy vs. Temperature” in Chapter 4, and “Tuning Fork
Resonator Sensors” in Chapter 9.
E. Momosaki, “A Brief Review of Progress in Quartz Tuning Fork Resonators,” Proc. 1997 IEEE Int'l
J. H. Staudte, “Subminiature Quartz Tuning Fork Resonator,” Proc. 27th Ann. Symp. on Frequency
Control, pp. 50-54, 1973.

Watch Crystal
3-38
More than 109 tuning fork resonators are produced annually - mostly for wristwatches. In spite
of the numerous steps and the precision required to make such a resonator, the wholesale price of a
tuning fork resonator (in 1999) was ~$0.1.

Lateral Field Resonator
Lateral Field Thickness Field

In lateral field resonators (LFR): 1. the electrodes are absent from the
regions of greatest motion, and 2. varying the orientation of the gap between
the electrodes varies certain important resonator properties. LFRs can also be
made with electrodes on only one major face. Advantages of LFR are:
• Ability to eliminate undesired modes, e.g., the b-mode in SC-cuts
• Potentially higher Q (less damping due to electrodes and mode traps)
• Potentially higher stability (less electrode and mode trap effects, smaller C1)
3-39
The LFR above shows electrodes on both major faces of the quartz plate. LFRs can also be
made with electrodes on only one major face (e.g., only on the top).
A. W. Warner, "Measurement of Plano-Convex SC Quartz Blanks Using Lateral Field Excitation,"

Proc. 42nd AnnualSymposium on Frequency Control, pp. 202-204, 1988, AD-A217275.

Electrodeless (BVA) Resonator
D2
D1
Quartz
bridge
Side view of BVA2 Side and top views of

resonator construction center plate C
3-40
BVA resonators are designed primarily to minimize stresses due to the mounting structure and the
instabilities due to the electrodes (see “Stresses on a Quartz Resonator Plate” and subsequent pages in
the next chapter). The BVA2 structure shown above consists of three quartz plates of the same angles
of cut and azimuthal orientation. The resonator plate is the plate “C” in the center. The active part of
plate C is separated from the outer inactive part near the edges by means of cutouts, and the active
area is connected to the inactive area by means of small quartz bridges. The outside plates D1 and D2
contact the C plate only outside the cutouts. The electrodes, deposited onto the center areas of D1 and
D2, are separated from the active area of the resonator by means of small, 5 µm to 50 µm, gaps.
The best resonator short term stability measurements reported to date have been obtained with
BVA resonators, i.e., σy(τ) of parts in 1014 at the flicker floor (see chapter 4).
R. J. Besson, “A New ‘Electrodeless’ Resonator Design,” Proc. 31st Ann. Symp. on Frequency Control,”
pp. 147-152, 1977
R. J. Besson, J-J Boy, M. M. Maurey, “BVA Resonators and Oscillators: A Review. Relation with Space
Requirements and Quartz Material Characterization,” Proc. 1995 IEEE Int’l Frequency Control
Symposium, pp. 590-599, 1995

CHAPTER 4
Oscillator Stability
E. A. Gerber and A. Ballato, Precision Frequency Control, Academic Press, 1985
P. Kartaschoff, Frequency and Time, Academic Press, 1978.

Oscillators, John Wiley & Sons, 1983.
Characterization of Clocks and Oscillators, edited by D. B. Sullivan, et al., NIST Technical Note 1337,
March 1990. Time & Frequency Div., NIST, 325 Broadway, Boulder, CO 80303.
J. R. Vig & A. Ballato, "Frequency Control Devices," in Ultrasonic Instruments & Devices II, edited by
R.N. Thurston, A. D. Pierce, vol. ed. E. Papadakis, Physical Acoustics Vol. XXIV, Academic Press, pp.
209-273, 1999.
T. R. Meeker, and J. R. Vig, "The Aging of Bulk Acoustic Wave Resonators, Oscillators and Filters,"
Proc. 45th Ann. Symp. on Frequency Control 1991, IEEE Cat. No. 91CH2965-2.
Oscillators," Proc. 1994 IEEE Int'l Frequency Control Symposium, pp. 506-523, 1994.

The Units of Stability in Perspective
• What is one part in 1010 ? (As in 1 x 10-10/day aging.)

• ~1/2 cm out of the circumference of the earth.
• ~1/4 second per human lifetime (of ~80 years).
• Power received on earth from a GPS satellite, -160 dBW, is
as “bright” as a flashlight in Los Angeles would look in New
York City, ~5000 km away (neglecting earth’s curvature).
• What is -170 dB? (As in -170 dBc/Hz phase noise.)
• -170 dB = 1 part in 1017 ≈ thickness of a sheet
of paper out of the total distance traveled by all
the cars in the world in a day.
4-1
The human mind is limited in its ability to understand very small and very large numbers. Above
is an attempt to make the small numbers used in the frequency and time field a bit more
understandable.
GPS analogy is courtesy of Raymond Filler, March 2004.

Accuracy, Precision, and Stability
Precise but Not accurate and Accurate but Accurate and

not accurate not precise not precise precise
f f f f
Time Time Time Time

Accurate
Stable but Not stable and Stable and
(on the average)
not accurate not accurate accurate
but not stable
4-2
The terms accuracy, stability, and precision are often used in describing an oscillator's quality.
Above is an illustration of the meanings of these terms for a marksman and for a frequency source.
(For the marksman, each bullet hole's distance to the center of the target is the "measurement.")
Accuracy is the extent to which a given measurement, or the average of a set of measurements for
one sample, agrees with the definition of the quantity being measured. It is the degree of
"correctness" of a quantity. Reproducibility is the ability of a single frequency standard to produce
the same frequency, without adjustment, each time it is put into operation. From the user's point of
view, once a frequency standard is calibrated, reproducibility confers the same advantages as
accuracy. Stability describes the amount something changes as a function of parameters such as
time, temperature, shock, and the like. Precision is the extent to which a given set of measurements
of one sample agrees with the mean of the set. (A related meaning of the term is used as a descriptor
of the quality of an instrument, as in a "precision instrument." In that context, the meaning is usually
defined as accurate and precise, although a precision instrument can also be inaccurate and precise,
in which case the instrument needs to be calibrated.)
The military specification for crystal oscillators, MIL-PRF-55310D*, defines “Overall Frequency
Accuracy” as “6.4.33 Overall frequency accuracy. The maximum permissible frequency deviation of
the oscillator frequency from the assigned nominal value due to all combinations of specified operating
and nonoperating parameters within a specified period of time. In the general case, overall accuracy
of an oscillator is the sum of the absolute values assigned to the following:
a. The initial frequency-temperature accuracy (see 6.4.24).
b. Frequency-tolerances due to supply voltage changes (see 6.4.17) and other environmental effects
(see 6.4.12).
Total frequency change from an initial value due to frequency aging (see 6.4.11) at a specified
temperature.”
The International System (SI) of units for time and frequency (the second and Hz, respectively)
are obtained in laboratories using very accurate frequency standards called primary standards. A
primary standard operates at a frequency calculable in terms of the SI definition of the second**: "the
duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two
hyperfine levels of the ground state of the cesium atom 133”.
* MIL-PRF-55310, http://www.dscc.dla.mil/Programs/MilSpec/ListDocs.asp?BasicDoc=MIL-PRF-
55310
Influences on Oscillator Frequency
z Time
• Short term (noise)
• Intermediate term (e.g., due to oven fluctuations)
• Long term (aging)
z Temperature
• Static frequency vs. temperature
• Dynamic frequency vs. temperature (warmup, thermal shock)
• Thermal history ("hysteresis," "retrace")
z Acceleration
• Gravity (2g tipover) • Acoustic noise
• Vibration • Shock
z Ionizing radiation
• Steady state • Photons (X-rays, γ-rays)
• Pulsed • Particles (neutrons, protons, electrons)
z Other
• Power supply voltage • Humidity • Magnetic field
• Atmospheric pressure (altitude) • Load impedance
4-3
Many factors influence the frequency stability of an oscillator. Changes in the environment can
cause especially large instabilities. For example, orders of magnitude (tens of dBs) changes can be
observed when the phase noise of an oscillator is measured in a quiet laboratory environment, and in
a vibrating environment, such as a moving vehicle.

Idealized Frequency-Time-Influence Behavior
∆f
X 10 8 Oscillator
f Turn Off 2-g
Temperature Vibration Shock Radiation
3 &
Step Tipover
Turn On
Off
2
1 Aging
-1
On
-2
Short-Term
Instability
-3
t0 t1 t2 t3 t4 t5 t6 t7 t8 Time
4-4
Shown above are the major types of oscillator frequency instabilities. The pages that follow
show each of the changes, and some others, in more detail.

Aging and Short-Term Stability
Short-term instability
(Noise)
30
25
∆f/f (ppm)
20
15
10
5 10 15 20 25 Time (days)
4-5
The difference between aging and short-term instability, i.e., noise, is illustrated above. One is a
systematic effect that is observed over long periods of time (days to years), whereas the other is
random, observed over periods that are typically measured in fractions of a second to minutes. Over
periods of hours, a combination of systematic and random effects are usually observed. The
frequency vs. time characteristics over such periods often appear to be random walk (at least some of
which is usually environmentally caused).
"Aging" and "drift" have occasionally been used interchangeably in the frequency control
literature. However, in 1990, recognizing the "need for common terminology for the unambiguous
specification and description of frequency and time standard systems," the CCIR adopted a glossary
of terms and definitions. According to this glossary, aging is "the systematic change in frequency with
time due to internal changes in the oscillator." Added to the definition is: "Note - It is the frequency
change with time when factors external to the oscillator (environment, power supply, etc.) are kept
constant." Drift is defined as "the systematic change in frequency with time of an oscillator." Drift is
due to a combination of factors, i.e., it due to aging plus changes in the environment and other factors
external to the oscillator. Aging is what one specifies and what one measures during oscillator
evaluation. Drift is what one observes in an application. For example, the drift of an oscillator in a
spacecraft is due to (the algebraic sum of) aging and frequency changes due to radiation, temperature
changes in the spacecraft, and power supply changes.
CCIR Recommendation No. 686, “[TF.686-1] Glossary,” CCIR 17th Plenary Assembly, Vol. VII,
"Standard Frequency and Time Signals (Study Group 7)," 1990. Consultative Committee on
International Radio (CCIR); copies available from: International Telecommunications Union, General
Secretariat - Sales Section, Place des Nations, CH-1211 Geneva, Switzerland.
http://www.itu.int/itudoc/itu-r/rec/tf/

Aging Mechanisms
z Mass transfer due to contamination
Since f ∝ 1/t, ∆f/f = -∆t/t; e.g., f5MHz ≈ 106 molecular layers,
therefore, 1 quartz-equivalent monolayer ⇒ ∆f/f ≈ 1 ppm
z Stress relief in the resonator's: mounting and bonding structure,

electrodes, and in the quartz (?)
z Other effects
{ Quartz outgassing
{ Diffusion effects
{ Chemical reaction effects
{ Pressure changes in resonator enclosure (leaks and outgassing)
{ Oscillator circuit aging (load reactance and drive level changes)
{ Electric field changes (doubly rotated crystals only)
{ Oven-control circuitry aging
4-6
The main causes of aging appear to be mass transfer to or from the resonator surfaces (due to
adsorption and desorption of contamination) and stress relief within the mounting structure or at the
interface between the quartz and the electrodes. Many aging measurements have been reported, but
few have included a detailed scientific or statistical study of the aging processes. Our understanding
of resonator aging processes is often based on indirect evidence.
Because the frequency of a thickness shear resonator, such as an AT-cut or SC-cut, is inversely
proportional to the thickness of the crystal plate, and because, for example, a 5-MHz 3rd overtone
plate is on the order of 1 million atomic layers thick, the adsorption or desorption of contamination
equivalent to the mass of one atomic layer of quartz changes the frequency by about 1 ppm. In
general, if contamination equal in mass to 1½ monolayers of quartz is adsorbed or desorbed from the
surfaces, the resulting frequency change in parts per million is equal to the resonator's fundamental
mode frequency in megahertz. Therefore, in order to achieve low-aging, crystal units must be
fabricated and hermetically sealed in ultraclean, ultrahigh vacuum environments, and hermetically
sealed into packages that are capable of maintaining the clean environment for long periods.
Stress (and stress relief) effects are discussed on the pages that follow (see, e.g., “Mounting
Force Induced Frequency Change”).
Proc. 45th Ann. Symp. on Frequency Control 1991, IEEE Cat. No. 91CH2965-2. This paper is also
available at <http://www.ieee.org/uffc/fc> .

Typical Aging Behaviors
A(t) = 5 ln(0.5t+1)
Time
∆f/f
A(t) +B(t)
B(t) = -35 ln(0.006t+1)
4-7
Aging can be positive or negative. Occasionally, a reversal in aging direction is observed. Typical
(computer simulated) aging behaviors are illustrated above. The curve showing the reversal is the
sum of the other two curves. The simplest proposed aging model showing a reversal consists of two
simultaneously acting aging mechanisms, with different directions and decay times. In the example
above, initially, A(t) dominates. It decays faster than B(t), therefore, eventually, the aging mechanism
described by B(t) dominates the aging.
The aging rate of an oscillator is highest when it is first turned on. Since the aging rate during the
first few days to weeks is generally significantly higher than during subsequent intervals, the early part
of the aging curve is sometimes referred to as "initial aging" or "stabilization period." At a constant
temperature, aging usually has an approximately logarithmic dependence on time. After a long
period, the aging rate often becomes approximately constant (i.e., linear frequency vs. time). When
the temperature of a crystal unit is changed, e.g., when an OCXO is turned off, and turned on at a
later time, a new aging cycle usually starts (see “Retrace”).
Proc. 45th Ann. Symp. on Frequency Control 1991, IEEE Cat. No. 91CH2965-2.

Stresses on a Quartz Resonator Plate
Causes:
• Thermal expansion coefficient differences
• Bonding materials changing dimensions upon solidifying/curing
• Residual stresses due to clip forming and welding operations,
sealing
• Intrinsic stresses in electrodes
• Nonuniform growth, impurities & other defects during quartz
growing
• Surface damage due to cutting, lapping and (mechanical) polishing
Effects:
• In-plane diametric forces
• Tangential (torsional) forces, especially in 3 and 4-point mounts
• Bending (flexural) forces, e.g., due to clip misalignment and
electrode stresses
• Localized stresses in the quartz lattice due to dislocations,
inclusions, other impurities, and surface damage
4-8
There are a variety of stresses built, inadvertently, into a newly manufactured resonator. The
resonator experiences additional stresses due to changes in its environment, as is discussed later in
chapter 4.
It is the changes in the stresses, and the changes produced by the stresses that cause
frequency instabilities. There exists evidence that, on a microscopic level, stress relief is not a
continuous process. It can occur in bursts that can, possibly, contribute to noise and frequency jumps.

Thermal Expansion Coefficients of Quartz
Thermal Expansion Coefficient, α, of AT-cut Quartz, 10-6/0K

14 ZZl
13.71
Ψ XXl
13
Radial
12
11.63
11
α (Thickness) = 11.64
10
Tangential
9.56
9
00 100 200 300 400 500 600 700 800 900
Orientation, Ψ, With Respect To XXl
4-9
The thermal expansion coefficient of single crystal quartz varies with direction, as shown above*.
This makes it virtually impossible to avoid stresses due to thermal expansion coefficient differences -
e.g., at interfaces between the quartz plate and its electrodes and the mounting structure.
Electrodeless designs, such as the “BVA” resonator, and the “all quartz package,” attempt to
eliminate these stresses by mounting the quartz resonator plate on a single crystal quartz structure of
the same orientation. However, it is not the stresses but the changes in the stresses that cause
instabilities, so, in other designs, compliant, spring-like mounting structures are used to minimize
instabilities due to stress relief.
*Provided by Arthur Ballato, U.S. Army CECOM, private communication, circa 1978
R.J. Besson, “A New Electrodeless Resonator Design,” Proc. 31st Ann. Symp. On Frequency Control,
pp. 147-152, 1977.
T.E. Parker, J. Callerame and G.R. Montress, “A New All Quartz Package For SAW Devices,” Proc.
39th Ann. Symp. on Frequency Control, pp. 519-525, 1985.

Force-Frequency Coefficient
30
* 10-15 m • s / N AT-cut quartz
25
20
15 Z’
F
10 Ψ
Kf (Ψ) X’
5 F
-5
-10
∆f
= KF
(Force ) (Frequency constant )
f (Diameter ) (Thickness )
-15 0
0 100 200 300 400
Ψ 500 600 700 800 900
4-10
A force applied to a resonator causes lattice deformation that results in a frequency change.
The above curve shows the force-frequency coefficient, KF, as a function of the direction of the applied
force. An AT-cut quartz plate was squeezed, as shown, by a diametrically applied force. The
frequency changes were measured as a function of direction of the force with respect to the x’-axis,
and KF was calculated using the expression shown above. The zero crossing is at Ψ = 61o. However,
the the force-frequency coefficient, and it’s zero crossing, also depend on temperature. For both the
AT-cut and the SC-cut, the zero crossing is at a few degrees smaller Ψ angle when, e.g., measured at
80oC vs. 25oC.
A. Ballato, E. P. EerNisse, and T. Lukaszek, "The Force--Frequency Effect in Doubly Rotated Quartz
Resonators," Proc. 31st Annual Symposium on Frequency Control, pp. 8-16, 1977, AD-A088221.

Strains Due To Mounting Clips
X-ray topograph of an AT-cut, two-point mounted resonator. The topograph

shows the lattice deformation due to the stresses caused by the mounting clips.
4-11
R. A. Young, R. B. Belser, A. L. Bennett, W. H. Hicklin, J. C. Meaders, and C. E. Wagner, "Special X-

ray Studies of Quartz Frequency Control Units," Proc. 19th Annual Symposium on Frequency Control,
pp. 23-41, 1965, AD-471229.

Strains Due To Bonding Cements
(a) (b)
X-ray topographs showing lattice distortions caused by bonding cements; (a) Bakelite
cement - expanded upon curing, (b) DuPont 5504 cement - shrank upon curing
4-12
Bonding agents such as silver-filled epoxies and silver-filled polyimides are commonly used to
bond quartz plates to the mounting clips. In addition to producing stresses, such bonding agents also
absorb water, and can produce aging due to stress relief, and the outgassing of organic materials and
water.
R. B. Belser and W. H. Hicklin, "Aging Characteristics of Quartz Resonators With Comments on the
Effects of Radiation," Proc. 17th Annual Symposium on Frequency Control, pp. 127-175, 1963,
AD423381.
J. A. Kusters, "Resonator and Device Technology," in E. A. Gerber and A. Ballato, Precision

R. L. Filler*. J. M. Frank**. R. D. Peters, and J. R. Vig, “Polyimide Bonded Resonators” Proc. 32nd
Ann. Symp. on Frequency Control, pp. 290-298, 1972.
R. J. Byrne, “Thermocompression Bonding To Quartz Crystals,” Proc. 26th Ann. Symp. on Frequency
Control, pp. 71-77, 1972.

Mounting Force Induced Frequency Change
The force-frequency coefficient, KF (ψ), is defined by Z’
∆f
= KF
(Force) (Frequency − constant ) F
f (Diameter ) (Thickness) Ψ
X’
Maximum KF (AT-cut) = 24.5 x 10-15 m-s/N at ψ = 0o
F
Maximum KF (SC-cut) = 14.7 x 10-15 m-s/N at ψ = 44o
As an example, consider a 5 MHz 3rd overtone, 14 mm diameter resonator.

Assuming the presence of diametrical forces only, (1 gram = 9.81 x 10-3
newtons),
⎛ ∆f ⎞
⎝ f ⎠Max
{
⎜ ⎟ = 2.9 x 10 -8 per gram for an AT-cut resonator
-8
1.7 x 10 per gram for an SC-cut resonator
⎛ ∆f ⎞
⎜ ⎟ = 0 at ψ = 61o for an AT-cut resonator, and at ψ = 82o for an
⎝ f ⎠Min SC-cut.
4-13
The mounting structure can produce not only in-plane diametric forces but also tangential
(torsional) forces, especially in 3 and 4-point mounts, and bending (flexural) forces, e.g., due to clip
misalignment and electrode stresses. These stresses produce frequency shifts, and the changes in
these stresses result in frequency aging.
A. Ballato, E. P. EerNisse, and T. Lukaszek, "The Force--Frequency Effect in Doubly Rotated

Quartz Resonators," Proc. 31st Annual Symposium on Frequency Control, pp. 8-16, 1977, AD-
A088221.
E. D. Fletcher and A. J. Douglas, “A Comparison Of The Effects Of Bending Moments On The

Vibrations Of AT And SC (or TTC) Cuts Of Quartz,” Proc. 33rd Annual Symposium on
Frequency Control, pp. 346-350, 1979.

Bonding Strains Induced Frequency Changes
Apparent angle shift (minutes)

6’
• Blank No. 7
• Z’
5’
• Blank No. 8
4’
Ψ X’
3’
•
2’
•
1’
•
0’
•
-1’
• •
-2’
300 600 900

Bonding orientation, Ψ
When 22 MHz fundamental mode AT-cut resonators were reprocessed so as to vary the
bonding orientations, the frequency vs. temperature characteristics of the resonators changed
as if the angles of cut had been changed. The resonator blanks were 6.4 mm in diameter
plano-plano, and were bonded to low-stress mounting clips by nickel electrobonding.
4-14
The mounting and bonding stresses can affect the f vs. T of resonators. The experimental
results shown above demonstrate that the same quartz plate can exhibit a range of f vs. T
characteristics depending on the orientation of the bonding stresses. (Thicker plates and more
compliant bonding agents produce smaller apparent angle shifts.)
R. L. Filler, and J. R. Vig, "The Effect of Bonding on the Frequency vs. Temperature Characteristics of
AT-cut Resonators," Proc. 30th Annual Symposium on Frequency Control, pp. 264-268, 1976, AD-
046089.

Bending Force vs. Frequency Change
AT-cut resonator SC-cut resonator
fo = 10Mz fo = 10Mz
30 5gf 5gf
• •
• • •
• • • +10
• • •
Frequency Change (Hz)

• • •
20
• • •
•
• •• • •• • ••
•
• • •
• • • •• • • • •• • • •• •
• • 60 120 180 240 • 300 • • 360
• • •• •
• • • •• •
10 • • • Azimuth angle ψ (degrees)
•
• ••
-10
0 60 120 180 240 300 360

Azimuth angle ψ (degrees)
Frequency change for symmetrical Frequency change for symmetrical
bending, AT-cut crystal. bending, SC-cut crystal.
4-15
The above shows experimental results of the frequency change vs. the azimuthal angle for
symmetrical bending about a diameter. The resonators were 12 mm diameter, AT-cut and SC-cut,
plano-plano. A constant force of 5 grams was applied, the measurements were repeated as the
resonator was rotated, and the frequency changes were plotted as a function of azimuthal angle Ψ.
At all azimuthal angles, the SC-cut is less sensitive to bending than the AT-cut when the force is
applied normal to the plane of the resonator, as it was in these experiments. For the AT-cut, the
frequency change vs. angle has no zero crossing. The integral of frequency changes from 0o to 360o
is much smaller for the SC-cut than for the AT-cut (even though there was a significant measurement
uncertainty in the SC-cut results). In addition, when the force was varied from zero to ~15 grams, the
frequency change vs. applied force was much more linear for the SC-cut than the AT-cut.
E. D. Fletcher and A. J. Douglas, “A Comparison Of The Effects Of Bending Moments On The

Vibrations Of AT And SC (or TTC) Cuts Of Quartz,” Proc. 33rd Annual Symposium on Frequency
Control, pp. 346-350, 1979.

Short Term Instability (Noise)
Stable Frequency (Ideal Oscillator)
Φ(t)
1
V
-1
T1 T2 T3
Time
V(t) = V0 sin(2πν0t) Φ(t) = 2πν0t
Unstable Frequency (Real Oscillator)

1 Φ(t)
V -1
T1 T2 T3
Time
V(t) =[V0 + ε(t)] sin[2πν0t + φ(t)] Φ(t) = 2πν0t + φ(t)
1 d Φ(t ) 1 dφ(t )
Instantaneous frequency, ν(t ) = = ν0 +
2π d t 2π d t
V(t) = Oscillator output voltage, V0 = Nominal peak voltage amplitude
ε(t) = Amplitude noise, ν0 = Nominal (or "carrier") frequency
Φ(t) = Instantaneous phase, and φ(t) = Deviation of phase from nominal (i.e., the ideal)
4-16
The output voltage of an ideal oscillator would be a perfect sine wave. The outputs of all real
oscillators deviate from a perfect sine wave due to noise. The amplitude deviation is represented by
ε(t), and the phase deviation by φ(t). As frequency is the rate of change of phase, the frequency
deviation is ν(t) - ν0 = [1/(2π)][dφ(t)/dt].
See the next page for another illustration of the amplitude, phase and frequency instabilities.
S. R. Stein, "Frequency and Time - Their Measurement and Characterization," in E. A. Gerber and A.

Instantaneous Output Voltage of an Oscillator
Amplitude
instability
- Voltage +
Phase
instability
0
Frequency
instability
Time
4-17
As shown on the previous page, the instantaneous output voltage of a precision oscillator can
be expressed as
V(t) = (Vo + ε(t)) sin(2πνot + φ(t)),
where Vo is the nominal peak voltage amplitude, ε(t) is the deviation from the nominal amplitude, νo is
the nominal frequency, and φ(t) is the phase deviation from the nominal phase 2πνot. The figure
illustrates a signal with amplitude, phase and frequency instabilities. Fluctuations in the peak value of
the signal is the amplitude instability. Fluctuations in the zero crossings of the voltage is the phase
instability. Frequency instability is the fluctuations in the period of the voltage.
In the signal shown, the frequency components of the noise are higher than the carrier
frequency. This is for illustration purposes only. In general, the measures of stability apply to the
frequency components of amplitude, phase and frequency instabilities which are lower in frequency
than the carrier frequency.
The above figure was provided by Prof. Eva Ferre-Pikal, Univ. of Wyoming, 1999.
IEEE Standard 1139-1999

Impacts of Oscillator Noise
• Limits the ability to determine the current state and the
predictability of oscillators
• Limits syntonization and synchronization accuracy
• Limits receivers' useful dynamic range, channel spacing, and
selectivity; can limit jamming resistance
• Limits radar performance (especially Doppler radar's)
• Causes timing errors [~τσy(τ )]
• Causes bit errors in digital communication systems
• Limits number of communication system users, as noise from
transmitters interfere with receivers in nearby channels
• Limits navigation accuracy
• Limits ability to lock to narrow-linewidth resonances
• Can cause loss of lock; can limit acquisition/reacquisition
capability in phase-locked-loop systems
4-15
Impacts of Oscillator
Noise

Time Domain - Frequency Domain
A
f
(a)
Amplitude - Time
Amplitude - Frequency
t
(b) (c)
A(t) A(f)
4-18
Shown in (a) is the amplitude vs. frequency vs. time plot of two sine waves. An oscilloscope
can display such signals in the time domain, e.g., the voltage output vs. time of the two sine waves,
and their sum, is shown in (b) above. A spectrum analyzer can display voltage or power as a function
of frequency, i.e., it can be used to analyze signals in the frequency domain, as in (c) above. In the
frequency domain, signals are separated into their frequency components and the power level at each
frequency is displayed.
An ideal sine wave appears as a spectral line of zero bandwidth in the frequency domain. Real
sine wave outputs are always noisy, so the spectral lines have a finite bandwidth.
Noise is usually present over a wide band of frequencies. The total power (or voltage)
measured by a spectrum analyzer depends on the resolution bandwidth used. The unit of phase noise
is dBc/Hz (dB below the carrier per Hz of bandwidth). Reporting phase noise measurement results
must include the bandwidth and the carrier frequency.
“Spectrum Analysis….Spectrum Analyzer Basics,” Hewlett-Packard Aplication Note 150, April, 1974.
“Spectrum Analysis….Noise Measurements,” Hewlett-Packard Aplication Note 150-4, April, 1974.

<http://www.tmo.hp.com/@@gR622hcQNYU06mWM/tmo/Notes/English/NotesSubjects.html>

Causes of Short Term Instabilities
z Johnson noise (thermally induced charge fluctuations, i.e., "thermal emf” in
resistive elements)
z Phonon scattering by defects & quantum fluctuations (related to Q)
z Noise due to oscillator circuitry (active and passive components)
z Temperature fluctuations- thermal transient effects

- activity dips at oven set-point
z Random vibration
z Fluctuations in the number of adsorbed molecules
z Stress relief, fluctuations at interfaces (quartz, electrode, mount, bond)
z Shot noise in atomic frequency standards
z ???
4-19
The irregular, random dancing motions of tiny solid particles suspended in a liquid were
observed in the 18th century, and the phenomenon was later described by Robert Brown. The effect,
now called Brownian motion, was explained by Einstein. It was theoretically predicted and
experimentally verified by J. B. Johnson (at the output of an audio amplifier) that the Brownian motion
of electrons would lead to random voltages at the terminals of a resistor.
Johnson noise, also called thermal noise, is the basic limit in the amplification of weak signals,
and it also limits the noise of crystal oscillators, due to the equivalent series resistance of the
resonator, and due to the resistances of other components in the sustaining circuit. According to the
equipartition law of Maxwell and Boltzmann, under thermal equilibrium, the energy associated with a
degree of freedom of a system is kT/2 = 2 x 10-21 J per degree of freedom. It can be shown that a
pure resistance R is a white thermal noise source with available power = kT = 4 x 10-21 W/Hz = -204
dBW/Hz = -174 dBm/Hz (dBm is dB referenced to one mW).
W. R. Bennett, Electrical Noise, McGraw-Hill Book Co., New York, 1960.
P. H. Handel, “Quantum 1/f Quartz Resonator Theory Versus Experiment,” Proc. 1999 IEEE
Frequency Control Symp.
Y.-K. Yong and J. R. Vig, “Resonator Surface Contamination - A Cause of Frequency Fluctuations?,”
IEEE Tr. Ultrason. Ferroelec. Freq. Contr., vol. UFFC-36, No. 4, pp. 452-458, July 1989.
R. L. Filler, “The Acceleration Sensitivity of Quartz Crystal Oscillators: A Review,” IEEE Tr. Ultrason.
Ferroelec. Freq. Contr., vol. UFFC-35, pp. 297-305, 1988.
Short-Term Stability Measures
Measure Symbol
Two-sample deviation, also called “Allan deviation” σy(τ)*
Spectral density of phase deviations Sφ(f)
Spectral density of fractional frequency deviations Sy(f)
Phase noise L(f)*
* Most frequently found on oscillator specification sheets
f2Sφ(f) = ν2Sy(f); L(f) ≡ ½ [Sφ(f)] (per IEEE Std. 1139),
and 2 ∞
σ 2y ( τ) =
(πντ)2 ∫
0
S φ (f)sin4 ( πfτ)df
Where τ = averaging time, ν = carrier frequency, and f = offset or

Fourier frequency, or “frequency from the carrier”.
4-20
IEEE Standard 1139 is the standard for characterizing measurements of frequency, phase, and
amplitude instabilities. The standard measure for characterizing phase and frequency instabilities in
the frequency domain is L(f), defined as one half of the double-sideband spectral density of phase
fluctuations. When expressed in decibels, the units of L(f) are dBc/Hz (dB below the carrier in a 1 Hz
bandwidth). A device is to be characterized by a plot of L(f) versus offset frequency f. In some
applications, providing L(f) versus discrete values of offset frequency is sufficient.
The standard measure for characterizing amplitude instability in the frequency domain is one
half of the double-sideband spectral density of the fractional amplitude fluctuations, 1/2 Sa(f). When
expressed in decibels, the units of Sa(f) are dBc/Hz.
In the time domain, the standard measure of frequency and phase instabilities is the fully
overlapped Allan deviation σy(τ) - see the next few pages. A device shall be characterized by a plot of
σy(τ) versus sampling time τ. In some cases, providing discrete values of σy(τ) versus τ is sufficient.
The measurement system bandwidth and the total measurement time shall be indicated.
IEEE Standard 1139-1999
Characterization of Clocks and Oscillators, edited by D. B. Sullivan, et al., NIST Technical Note 1337,
March 1990. Time & Frequency Div., NIST, 325 Broadway, Boulder, CO 80303.
Ballato, Precision Frequency Control, Vol. 2, pp. 191-232, Academic Press, 1985

Allan Deviation
Also called two-sample deviation, or square-root of the "Allan
variance," it is the standard method of describing the short term
stability of oscillators in the time domain. It is denoted by σy(τ),
1
where σ 2y ( τ ) = < ( y k +1 - y k )2 > .
2
∆f
The fractional frequencies, y = are measured over a time
f
interval, τ; (yk+1 - yk) are the differences between pairs of
successive measurements of y, and, ideally, < > denotes a time
average of an infinite number of (yk+1 - yk)2. A good estimate
can be obtained by a limited number, m, of measurements
(m≥100). σy(τ) generally denotes σ2y ( τ, m), i.e.,
∑ 21 (yk +1 − yk )
1 m
σ 2
y
(τ) = σ ( τ,m) =
2
y
2
j
m j=1
4-21
At long averaging times, especially when the averaging time is a substantial fraction of the
record length, “TOTALDEV” statistics yield better results than the two-sample deviation (see the Howe
reference below).
------------------------
D. W. Allan, "Time and Frequency (Time-Domain) Characterization, Estimation, and Prediction of
Precision Clocks and Oscillators," IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency
Control, Vol. UFFC-34, No. 6, pp. 647-654, November 1987.
“Characterization of Clocks and Oscillators,” edited by D.B. Sullivan, D.W. Allan, D.A. Howe, F.L.
Walls, NIST Technical Note 1337, March 1990.
D. A. Howe, “Total Variance Explained,” Proc. 1999 IEEE Int’l Frequency Control Symp., Joint
Meeting EFTF-IEEE IFCS, pp. 1093-1099.

Why σy(τ)?
z Classical variance: σ 2yi =
1
m -1
∑ ( )
yi − y 2,
diverges for some commonly observed noise

processes, such as random walk, i.e., the variance
increases with increasing number of data points.
z Allan variance:
• Converges for all noise processes observed in precision
oscillators.
• Has straightforward relationship to power law spectral
density types.
• Is easy to compute.
• Is faster and more accurate in estimating noise
processes than the Fast Fourier Transform.
4-22

Frequency Noise and σy(τ)
3 X 10-11 0.1 s averaging time
∆f
0
f
-3 X 10-11 100 s
3 X 10-11
1.0 s averaging time
∆f
0
f 100 s
-11
-3 X 10
σy(τ) 10 -10
10-11
10-12
0.01 0.1 1 10 100
Averaging time, τ, s
4-23
“The noise” is a function of the averaging time (also called “measurement time” or “tau”), as is
illustrated above. For the same oscillator, the fluctuations in the frequency vs. time plot measured
with a 0.1 second averaging time are larger than when measured with a 1 second averaging time.
Also shown are the corresponding Allan deviations.
At short averaging times, the longer the averaging time, the lower the noise, up to the “flicker
floor,” i.e., for certain noise processes (see the next four pages), the hills and valleys in the frequency
vs. time data average out. Longer averaging does not help when the dominant noise process is flicker
of frequency. At the flicker floor, the Allan deviation is independent of averaging time. At longer
averaging times, the Allan deviation increases because the dominant noise process is random walk of
frequency, for which the longer the averaging time, the larger the Allan deviation.

Time Domain Stability
Aging* and
random walk
Frequency noise of frequency
σy(τ)
1s 1m 1h Sample time τ
Short-term Long-term
stability stability
*For σy(τ) to be a proper measure of random frequency fluctuations,

aging must be properly subtracted from the data at long τ’s.
4-24

Power Law Dependence of σy(τ)
τ-1
σy(τ)
τ-1
τ-1 τ1/2 to τ1
τ0
White Flicker White Flicker Random

Noise type: phase phase freq. freq. walk freq.
Below the flicker of frequency noise (i.e., the “flicker floor”) region, crystal
oscillators typically show τ-1 (white phase noise) dependence. Atomic
standards show τ-1/2 (white frequency noise) dependence down to about the
servo-loop time constant, and τ-1 dependence at less than that time constant.
Typical τ’s at the start of flicker floors are: 1s for a crystal oscillator, 103s for a
Rb standard and 105s for a Cs standard. At large τ’s, random walk of
frequency and aging dominate.
4-25

Pictures of Noise
Plot of z(t) vs. t Sz(f) = hαfα Noise name
α=0 White
α = -1 Flicker
α = -2 Random
walk
α = -3
Plots show fluctuations of a quantity z(t), which can be,e.g., the output of a counter (∆f vs. t)
or of a phase detector (φ[t] vs. t). The plots show simulated time-domain behaviors
corresponding to the most common (power-law) spectral densities; hα is an amplitude
coefficient. Note: since S∆f = f 2Sφ, e.g. white frequency noise and random walk of phase are
equivalent.
4-26
David W. Allan, “The Measurement of Frequency and Frequency Stability of Precision Oscillators,”
David W. Allan, NBS Technical Note 669, May 1975.
R. F. Voss, "1/f (Flicker) Noise: A Brief Review," Proc. 33rd Annual Symposium on Frequency
Control, pp. 40-46, 1979, AD-A213544.

Spectral Densities
V(t) = [V0 + ε (t) ] sin [ 2πν0 t + φ(t)]

In the frequency domain, due to the phase deviation, φ(t), some of
the power is at frequencies other than ν0. The stabilities are
characterized by "spectral densities." The spectral density, SV(f), the
mean-square voltage <V2(t)> in a unit bandwidth centered at f, is not a
good measure of frequency stability because both ε(t) and φ(t) contribute
to it, and because it is not uniquely related to frequency fluctuations
(although ε(t) is often negligible in precision frequency sources.)
The spectral densities of phase and fractional-frequency fluctuations,
Sφ(f) and Sy(f), respectively, are used to measure the stabilities in the
frequency domain. The spectral density Sg(f) of a quantity g(t) is the
mean square value of g(t) in a unit bandwidth centered at f. Moreover,
the RMS value of g2 in bandwidth BW is given by g RMS 2
(t) = ∫ S g(f)d f .
BW
4-27
The Fourier frequency is a fictitious frequency used in Fourier analysis of a signal. Zero Fourier
frequency corresponds to the carrier, and a negative Fourier frequency refers to the region below the
carrier. The integral of the spectral density over all Fourier frequencies from minus infinity to infinity is
the mean-square value of the quantity. The spectral density of phase-noise Sφ(f') is important
because it is directly related to the performance of oscillators in RF signal processing applications. Up
until 1988, the single-sideband (SSB) noise power per Hz to total signal power ratio was often
specified for oscillators instead of the phase spectral density. This ratio has been designated L(f). In
IEEE standard 1139-1988 (current version is 1139-1999), the definition of L(f) was changed to one-
half Sφ(f'). When defined this way, L(f) is equal to the SSB noise-to-signal ratio only as long as the
integrated phase-noise from f' to infinity is small compared to one rad2.
The phase spectral density (phase noise) depends on carrier frequency. When the signal from
an oscillator is multiplied by n in a noiseless multiplier, the frequency modulation (FM) sidebands
increase in power by n2, as does the spectral density of phase. Consequently, it is important to state
the oscillator frequency together with the phase noise.

Mixer Functions
V1 = A 1sin(ω1t + ϕ1 ) V1V2
Filter
Filter V0
V2 = A2sin (ω2t + ϕ2 )
Trigonometric identities: sin(x)sin(y) = ½cos(x-y) - ½cos(x+y)
cos(x±π/2) = sin(x)
Let ω1 = ω 2 ;Φ1 ≡ ω1t + φ 1, and Φ2 ≡ ω 2 t + φ 2 . Then the mixer can become :
• Phase detector: When Φ1 = Φ2 + π / 2 and A 1 = A 2 = 1, then

1 1
V0 = sin(φ 1 − φ 2 ) = (φ 1 − φ 2 ) for small φ ' s
2 2
• AM detector: When A 2 = 1 and the filter is a low − pass filter, then
1 1
V0 = A 1cos(φ 1 − φ 2 ); if φ 1 ≈ φ 2 , then V0 ≈
A1
2 2
• Frequency multiplier: When V1 = V2 and the filter is bandpass at 2ω1
1 2
then, V0 = A 1 cos(2ω1t + 2φ 1 ) ⇒ doubles the frequency and phase error.
2
4-28
A mixer is a versatile device in frequency metrology. As shown above, the input to the mixer are
two voltages, and the output is the product of those voltages. Through the application of elementary
trigonometry, a mixer can be made into a phase detector, an amplitude detector, and a frequency
multiplier.
P. Kartaschoff, Frequency and Time, section 3.4 “Frequency Multiplication and Division,” Academic
Press, 1978.

Phase Detector
DUT
V(t) Sφ(f)
~ LPF
fO VO(t) Vφ(t)
∆Φ = 900
Low-Noise
~ VR(t) Amplifier Spectrum Analyzer
Reference
Quadrature
Monitor*
* Or phase-locked loop
4-29
The device under test (DUT) and a reference source, at the same frequency and in phase
quadrature (i.e., 900 out of phase), are input to a double-balanced mixer. Then,
V0(t) = V(t) VR(t) = Kcos[φ(t) - φR(t) + π/2] + Kcos[2π(ν+ νR)t + …].
The low-pass filter (LPF) eliminates the second cosine term. Then,
for φR(t) << φ(t) << π/2,
Vφ(t) = Kφ(t),
i.e., the phase detector converts phase fluctuations to voltage fluctuations.

Phase Noise Measurement
RF
RFSource
Source
V (t ) = Vo sin[2π0 t + Φ(t )]
Phase
PhaseDetector
Detector
VVφ(t)
(t)==kΦ(t)
kΦ(t)
φ
Vφ(t)
RF
Voltmeter
Oscilloscope Spectrum Analyzer
Φ(t) ΦRMS(t) in BW of meter Sφ(f) vs. f
4-30
An oscilloscope can display the phase vs. time of a signal. An RF voltmeter can measure the
RMS phase fluctuations within the bandwidth of the instrument. A spectrum analyzer can display the
phase noise as a function of offset frequency from the carrier frequency.
W. P. Robins, Phase Noise in Signal Sources (Theory and Applications), IEE Tele-communications
Series 9, 1982.

Frequency - Phase - Time Relationships
1 dφ(t ) t
ν(t ) = ν 0 + = " instantane ous" frequency; φ(t ) = φ0 + ∫ 2π[ν(t' ) − ν 0 ]dt'

2π dt 0
•
ν(t ) − ν 0 φ(t )
y (t ) ≡ = = normalized frequency; φRMS
2
= ∫ S φ (f )dt
ν0 2πν 0
2
φRMS
2
⎛ν ⎞
S φ (f ) = = ⎜ 0 ⎟ S y (f ); L (f ) ≡ 1/2 Sφ (f ), per IEEE Standard 1139 − 1988
BW ⎝ f ⎠
( )
∞
2
σ 2y (τ) = 1/2 < y k +1 − y k S (f ) sin 4 (πfτ ) df
2
(πν 0 τ)2 ∫0 φ
>=
The five common power-law noise processes in precision oscillators are:

S y (f ) = h2 f 2 + h1f + h0 + h −1f −1 + h −2 f −2
(White PM) (Flicker PM) (White FM) (Flicker FM) (Random-walk FM)
φ(t )
t
Time deviation = x(t ) = ∫ y (t' )dt' =
o
2πν
4-31

Sφ(f) to SSB Power Ratio Relationship
Consider the “simple” case of sinusoidal phase modulation at frequency fm. Then,
φ(t) = φo(t)sin(2πfmt), and V(t) = Vocos[2πfct + φ(t)] = Vocos[2πfct + φ0(t)sin(πfmt)],
where φo(t)= peak phase excursion, and fc=carrier frequency. Cosine of a sine
function suggests a Bessel function expansion of V(t) into its components at
various frequencies via the identities:
cos(X + Y ) = cosX cosY − sinX sinY
cosXcosY = 1/2[cos(X + Y ) + cos(X − Y )]
− sinXsinY = [cos(X + Y ) − cos(X − Y )]
cos(BsinX ) = J0 (B) + 2 ∑ J2n (B )cos(2nX )

∞
sin(BsinX ) = 2 ∑ J2n +1 (B )sin[(2n + 1)X]
n =0
After some messy algebra, SV(f) and Sφ(f) are as shown on the next page. Then,
V0 J1 [Φ(fm )]
2 2
SSB Power Ratio at fm = ∞
V0 J0 [Φ(fm )] + 2∑ Ji [Φ(fm )]
2 2 2
if Φ(fm ) << 1, then J0 = 1, J1 = 1/2Φ(fm ), Jn = 0 for n > 1, and

i =1
Φ2 (fm ) S φ (fm )
SSB Power Ratio = L (fm ) = =
4 2
4-32
According to IEEE Standard 1139-1999, the phase noise, L(f), is defined as
L(f) ≡ ½Sφ(f)
which is equal to the SSB power ratio only when <φ2(t)> = the integral of Sφ(f) from f = 0 to f = ∞, is
less than about 0.1 rad2. Sφ(f) can always be measured unambiguously, whereas the SSB power ratio
(the pre-1988 definition of phase noise) diverges close to the carrier.
The above analysis and the graph on the next page were provided by Raymond L. Filler, U.S.
Army LABCOM, 1989.

Sφ(f), Sv(f) and L (f)
Sφ (f )
Φ(t ) = Φ(fm )cos(2πfm t ) Φ2
(fm )
2
0 fm f
V (t ) = V0cos[2πfC t + Φ(fm )]
2 2
SV(f) V0 J0
[Φ(fm )]
2 2 2
V0 J1
[Φ(fm )]
2 2 2
V0 J2
[Φ(fm )]
2 2 2
V0 J3
[Φ(fm )]
2
fC-3fm fC-2fm fC-fm fC fC+fm fC+2fm fC+3fm f
V0 J1 [Φ(fm )] Sφ(fm )
2 2
SSB Power Ratio = ≅ L (fm ) ≡

V0 J0 [Φ(fm )] + 2∑ J 2 ⎡Φ⎛⎜ f ⎞⎟⎤
∞
2 2 2
⎢
i=1 i ⎣ ⎝
m ⎠⎥⎦
4-33

Types of Phase Noise
40 dB/decade (ff-4)
L(ff) Random walk of frequency
30 dB/decade (ff-3)
Flicker of frequency
20 dB/decade (ff-2)
White frequency; Random walk of phase
10 dB/decade (ff-1)
Flicker of phase
0 dB/decade (ff0)
White phase
ff
~BW of resonator Offset frequency
(also, Fourier frequency,
sideband frequency,
or modulation frequency)
4-34
Phase noise is important in many applications, but, the offset frequencies where the noise is
important differs from application to application. For example, in many communication systems, the
noise far from the carrier is important (e.g., to minimize the interference between a strong transmitted
signal and a weak signal received on a neighboring channel), whereas in Doppler radar systems
designed for detection of slowly moving targets, noise close to the carrier is important (see in Chapter
1, “Effect of Noise in Doppler Radar System,” and “Doppler Shifts”).

Noise in Crystal Oscillators
z The resonator is the primary noise source close to the carrier; the oscillator sustaining
circuitry is the primary source far from the carrier.
z Frequency multiplication by N increases the phase noise by N2 (i.e., by 20log N, in dB's).
z Vibration-induced "noise" dominates all other sources of noise in many applications

(see acceleration effects section, later).
z Close to the carrier (within BW of resonator), Sy(f) varies as 1/f, Sφ(f) as 1/f3, where f =
offset from carrier frequency, ν. Sφ(f) also varies as 1/Q4, where Q = unloaded Q. Since
Qmaxν = const., Sφ(f) ∝ ν4. (Qmaxν)BAW = 1.6 x 1013 Hz; (Qmaxν)SAW = 1.05 x 1013 Hz.
z In the time domain, noise floor is σy(τ) ≥ (2.0 x 10-7)Q-1 ≈ 1.2 x 10-20ν, ν in Hz. In the
regions where σy(τ) varies as τ-1 and τ-1/2 (τ-1/2 occurs in atomic frequency standards),
σy(τ) ∝ (QSR)-1, where SR is the signal-to-noise ratio; i.e., the higher the Q and the signal-
to-noise ratio, the better the short term stability (and the phase noise far from the carrier,
in the frequency domain).
z It is the loaded Q of the resonator that affects the noise when the oscillator sustaining
circuitry is a significant noise source.
z Noise floor is limited by Johnson noise; noise power, kT = -174 dBm/Hz at 290°K.
z Higher signal level improves the noise floor but not the close-in noise. (In fact, high drive
levels generally degrade the close-in noise, for reasons that are not fully understood.)
z Low noise SAW vs. low noise BAW multiplied up: BAW is lower noise at f < ~1 kHz,
SAW is lower noise at f > ~1 kHz; can phase lock the two to get the best of both.
4-35
BAW = bulk-acoustic wave resonator (e.g., an AT-cut or

SC-cut)
SAW = surface acoustic wave oscillator (e.g., an ST-cut)

Low-Noise SAW and BAW Multiplied to 10 GHz
(in a nonvibrating environment)
0
BAW = bulk-acoustic wave
-20 oscillator
SAW = surface acoustic
-40
wave oscillator
L(f) in dBc/Hz
-60
-80
BAW
-100 5 MHz x 2000
-120
BAW
-140 100 MHz x 100
BAW is SAW is SAW
-160 lower noise lower noise 500 MHz x 20
200 5500
10-1 100 101 102 103 104 105 106
Offset frequency in Hz
4-36
When low noise is required in the microwave (or higher) frequency range, SAW oscillators and
dielectric resonator oscillators (DROs) are sometimes used. When compared with multiplied-up (bulk-
acoustic-wave) quartz oscillators, these oscillators can provide lower noise far from the carrier at the
expense of poorer noise close to the carrier, poorer aging, and poorer temperature stability. SAW
oscillators and DROs can provide lower noise far from the carrier because these devices can be
operated at higher drive levels, thereby providing higher signal-to-noise ratios, and because the
devices operate at higher frequencies, thereby minimizing the "20 log N" losses due to frequency
multiplication by N. L(f) = -185 dBc/Hz noise floor (at 400 kHz from the carrier) has been achieved
with a 400 MHz SAW oscillator. Of course, as is the case for high-frequency bulk-wave oscillators,
such noise floors are realizable only in environments that are free of vibrations at the offset
frequencies of interest.
Shown above are the phase noises (with no vibration) of three low-noise oscillators multiplied
to 10 GHz: a 5 MHz bulk-acoustic-wave (BAW) oscillator, a 100 MHz BAW oscillator and a 500 MHz
SAW oscillator. The SAW oscillator is better at far from the carrier; the BAW oscillators are better
close to the carrier.
As shown on the next page, in the presence of vibration, it is the vibration induced phase noise
that dominates the phase noise (except very far from the carrier). The device that has the lower
acceleration sensitivity will have the lower phase noise, independent of the effects of frequency
multiplication.
G.K. Montress & T.E. Parker,”Design and Performance of an Extremely Low Noise Surface Acoustic
Wave Oscillator”, Proc. 1994 IEEE Int’l Frequency Control Symposium, pp. 365-373, 1994.

Low-Noise SAW and BAW Multiplied to 10 GHz
(in a vibrating environment)
0
Vibration induced phase noise dominates the phase
noise of both (whichever has lower acceleration
-20 sensitivity will have lower phase noise; currently,
BAW can provide lower sensitivity than SAW.)
Illustration assumes 1 x 10-9/g acceleration
-40
sensitivity for both BAW and SAW, and 0.01
g2 / Hz random vibration power spectral
L(f) in dBc / Hz
-60 density at all vibration frequencies
-80
5 MHz x 2000
BA BAW
-100 W
an
dS
AW
-120
-140 100 MHz x 100
-160 500 MHz x 20
10-1 100 101 102 103 104 105 106
Offset frequency in Hz
4-37
The solid line shows the phase noises under vibration for both BAW and SAW oscillators. The
dotted lines show the phase noises without vibration. In the presence of vibration, it is the
acceleration sensitivity that determines the phase noise. As is illustrated above, if the BAW and SAW
oscillators have equal acceleration sensitivities, then, after multiplication to 10 GHz, they will have
equal phase noise levels (up to about 10 kHz from the carrier).
Acceleration sensitivity is discussed later in this chapter. See, especially, the notes under
“Vibration-Induced Sidebands - After Frequency Multiplication”.
R. L. Filler, "The Acceleration Sensitivity of Quartz Crystal Oscillators: A Review," IEEE Transactions
on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 35, No. 3, pp. 297-305, May 1988.

Effects of Frequency Multiplication
fi ≡ fin fo ≡ fout = Mfi
∆ fi ∆ fo = M∆ fi
∆ fi ∆ fo ∆ fi
≡y =
fi fo fi
∆φ i ∆ φ o = M∆φ i
fi x M = f o L (f )o = L (f )i + 20 log M
L (f )i
Sφ (f )i Sφ (f )o = M2S φ (f )i
Noiseless
Multiplier
S y (f )i S y (f )o = S y (f )i
σ y (τ)i σ y (τ )o = σ y (τ )i
∆f
Note that y = , Sy(f), and σy(τ) are unaffected by frequency multiplication.
f 4-38

TCXO Noise
The short term stabilities of TCXOs are temperature (T) dependent, and are generally
worse than those of OCXOs, for the following reasons:
• The slope of the TCXO crystal’s frequency (f) vs. T varies with T. For example,
the f vs. T slope may be near zero at ~20oC, but it will be ~1ppm/oC at the T extremes. T
fluctuations will cause small f fluctuations at laboratory ambient T’s, so the stability can be
good there, but millidegree fluctuations will cause ~10-9 f fluctuations at the T extremes. The
TCXO’s f vs. T slopes also vary with T; the zeros and maxima can be at any T, and the
maximum slopes can be on the order of 1 ppm/oC.
• AT-cut crystals’ thermal transient sensitivity makes the effects of T fluctuations

depend not only on the T but also on the rate of change of T (whereas the SC-cut crystals
typically used in precision OCXOs are insensitive to thermal transients). Under changing T
conditions, the T gradient between the T sensor (thermistor) and the crystal will aggravate the
problems.
• TCXOs typically use fundamental mode AT-cut crystals which have lower Q and
larger C1 than the crystals typically used in OCXOs. The lower Q makes the crystals inherently
noisier, and the larger C1 makes the oscillators more susceptible to circuitry noise.
• AT-cut crystals’ f vs. T often exhibit activity dips (see “Activity Dips” later in this
chapter). At the T’s where the dips occur, the f vs. T slope can be very high, so the noise due
to T fluctuations will also be very high, e.g., 100x degradation of σy(τ) and 30 dB degradation of
phase noise are possible. Activity dips can occur at any T.
4-39

Quartz Wristwatch Accuracy vs. Temperature
Temperature coefficient of frequency = -0.035 ppm/0C2
Time Error per Day (seconds)

0
10
20
-550C -100C +280C +490C +850C

Military Winter Wrist Desert Military
“Cold” Temp. “Hot”
4-40
The frequencies of crystal (as well as all other) oscillators are sensitive to temperature.
Illustrated above is the effect for the oscillators used in quartz watches.
Quartz wristwatches are sufficiently accurate, usually, while worn as intended, i.e., on the wrist
for ~16 h and off the wrist for ~8 h each day. The accuracies degrade when the watch is off the wrist
for extended periods. The further the storage temperature is from the optimum temperature, the faster
the watch loses time. At temperature extremes, e.g., in a freezer at - 55oC, or at the temperature of
boiling water, wristwatches lose about 20 s per day.
The angle of cut of the resonator used in wristwatches is such that the zero temperature
coefficient is at ~250C. This has been found to provide the highest probability of accuracy, based on
the typical durations and temperatures while the watch is on the wrist and while it is off the wrist.

Frequency vs. Temperature Characteristics
f (LTP)
Frequency
Inflection Point
f (UTP)
Lower Temperature Upper

Turnover Turnover
Point (LTP) Point (UTP)
4-41
Shown above is the frequency vs. temperature (f vs. T) characteristic that is typical of e.g., AT-
cut and SC-cut resonators. The upper and lower turnover points (UTP and LTP) are the points where
the f vs. T has zero slope, i.e., df(T)/dT = 0. The corresponding frequencies are the upper and lower
turnover frequencies, respectively.
The inflection point is where the curvature of the f vs. T changes from convex to concave, i.e.,
where d2f(T)/dT2 = 0. The inflection temperatures are ≈ 260C for AT-cuts, and ≈ 960C for SC-cuts.
The values of the inflection temperature, and the dependence of the f vs. T on the angles of cut
depend not only on the angles of cut but also on the resonator’s design. For example, whereas a 3rd
overtone SC-cut’s inflection temperature is ~96oC, a fundamental mode SC-cut’s is ~105oC.
“Quartz Resonator Handbook - Manufacturing Guide for AT-Type Units,” edited by R. E. Bennett,
prepared for the US Dep’t of the Army, pp. 77-103, 1960, AD-274031.
J. A. Kusters, "The SC Cut Crystal - An Overview," Proc. 1981 IEEE Ultrasonics Symposium, pp. 402-
409.

Resonator f vs. T Determining Factors
z Primary: Angles of cut
z Secondary:
• Overtone
• Blank geometry (contour, dimensional ratios)
• Material impurities and strains
• Mounting & bonding stresses (magnitude and
direction)
• Electrodes (size, shape, thickness, density, stress)
• Drive level
• Interfering modes
• Load reactance (value & temperature coefficient)
• Temperature rate of change
• Thermal history
• Ionizing radiation
4-42
See the next page for the f vs. T vs. angle-of-cut family of curves for the AT-cut (when all
secondary effects are constant).

Frequency-Temperature vs. Angle-of-Cut, AT-cut
Z
25
AT-cut BT-cut
49o
∆θ 35¼o
20 R r R
8’ m m Y
R r R -1’
15 7’
0’
10 6’ Z
Y-bar quartz 1’
5’
f (ppm)
5
4’ 2’
0
3’
3’
∆f
-5 2’
4’
-10 1’
5’
0’
-15 θ = 35o 20’ + ∆θ, ϕ = 0
for 5th overtone AT-cut 6’
-1’
-20 θ = 35 12.5’+ ∆θ, ϕ = 0 for
o
fundamental mode plano-plano AT-cut 7’

8’
-25 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90
Temperature (oC)
4-43
The inset in the above illustration shows how an AT-cut plate and a BT-cut plate are related to
the directions in a quartz bar. The illustration shows how the AT-cut’s f vs. T characteristics change
as the θ angle of cut (see “Zero Temperature Coefficient Quartz Cuts” in chapter 3) is varied in one
minute of arc increments.
The ∆θ = 0 curve shows the f vs. T characteristic at the “reference angle,” which is θ = 35o12.5’
for a fundamental mode AT-cut resonator. At the reference angle, the turnover points coincide with
the inflection point. The reference angle varies with design, i.e., with overtone, plate contour, etc.

Desired f vs. T of SC-cut Resonator
for OCXO Applications
20
Frequency Offset (ppm) 15
10
5
-5
-10
Frequency remains within ± 1 ppm
over a ± 250C range about Ti
-15
-20
20 40 60 80 100 120 140 160
Temperature (0C)
4-44
When the turnover temperatures coincide with the inflection temperature of an SC-cut resonator,
the f vs. T slope remains small over a wide temperature range, as is shown above. Such an f vs. T
characteristic is ideal for OCXO applications, but such resonators are difficult to manufacture.
However, with angle-correction or recontouring, resonators with close to the ideal f vs. T
characteristics can be manufactured.
J. R. Vig, W. Washington, and R. L. Filler, "Adjusting the Frequency vs. Temperature Characteristics
of SC-cut Resonators by Contouring," Proc. 35th Ann. Symp. on Frequency Control, pp. 104-109,
1981, AD-A110870.

OCXO Oven’s Effect on Stability
Frequency
TURNOVER
POINT OVEN SET POINT
TURNOVER
POINT
OVEN
OFFSET
2∆ To
Typical f vs. T characteristic
OVEN CYCLING RANGE
for AT and SC-cut resonators
Temperature
Oven Parameters vs. Stability for SC-cut Oscillator
Assuming Ti - TLTP = 100C
Ti - TLTP = Oven Cycling Range

100C (millidegrees)
10 1 0.1 0.01
100 4 x 10-12 4 x 10-13 4 x 10-14 4 x 10-15
(millidegrees)
Oven Offset
10 6 x 10-13 4 x 10-14 4 x 10-15 4 x 10-16

1 2 x 10-13 6 x 10-15 4 x 10-16 4 x 10-17
0.1 2 x 10-13 2 x 10-15 6 x 10-17 4 x 10-18
0 2 x 10-13 2 x 10-15 2 x 10-17 2 x 10-19
A comparative table for AT and other non-thermal-transient compensated cuts of oscillators
would not be meaningful because the dynamic f vs. T effects would generally dominate the
static f vs. T effects.
4-45
The f vs. T stability of an OCXO depends on the static and dynamic f vs. T characteristics of the
resonator, the design temperature range of the OCXO, the stability of the oven and of the components
in the sustaining circuitry, and the accuracy with which the oven is set to the turnover temperature of
the resonator. The table shows the theoretically achievable f vs. T of an OCXO as functions of oven
offset from turnover temperature, and oven stability. The temperature coefficients of components in
the oven and sustaining circuitry make the theoretical values in the table difficult to approach.
The oven stability depends on the temperature range outside the OCXO and the thermal gain of
the oven. The thermal gain is defined as the external to internal temperature excursion ratio. For
example, if during an external temperature excursion from -40oC to +60oC the temperature inside the
oven changes by 0.1oC, the thermal gain is 103. For precise temperature control, the oven
temperature is typically ~15oC above the maximum operating temperature, e.g., in an OCXO designed
for -40oC to +60oC operation, the oven temperature is maintained at +75oC.
The thermal transient effect makes small oven offsets more difficult and time consuming to
achieve with AT-cut resonators than with SC-cuts (see “Warmup of AT- and SC-cut Resonators” two
pages forward in this chapter for a discussion and illustration of the thermal transient effect).
Using SC-cut resonators and high thermal gain ovens, OCXO stabilities of ~10-10 over a -40oC
to +75oC temperature range have been achieved. High thermal gains can be achieved with a single
oven, as described in the Walls reference, or, with double ovens (i.e., an oven in an oven).
A. Ballato, and J. R. Vig, "Static and Dynamic Frequency-Temperature Behavior of Singly and Doubly
Rotated, Oven-Controlled Quartz Resonators," Proc. 32nd Annual Symposium on Frequency Control,
pp. 180-188, 1978, AD-A955718.
F. L. Walls, "Analysis of High Performance Compensated Thermal Enclosures," Proc. 41st Ann.
Symp. on Frequency Control, pp. 439-443, 1987.
Warren L. Smith & T. E. Parker, “Precision Oscillators," in E. A. Gerber and A. Ballato, Precision Frequency
Marvin E. Frerking, “Temperature Conrol and Compensation,” in E. A. Gerber and A. Ballato, Precision Frequency
Oven Stability Limits
• Thermal gains of 105 has been achieved with a feed-forward
compensation technique (i.e., measure outside T of case &
adjust setpoint of the thermistor to anticipate and
compensate), and with double ovens. For example, with
a 105 gain, if outside ∆T = 100oC, inside ∆T = 1 mK.
• Stability of a good amplifier ~1µK/K
• Stability of thermistors ~1mK/year to 100mK/year
• Noise < 1µK (Johnson noise in thermistor + amplifier noise
+ shot noise in the bridge current)
• Quantum limit of temperature fluctuations ~ 1nK
• Optimum oven design can provide very high f vs. T stability
4-46
When power is applied to a frequency standard, it takes a finite amount of time before the
equilibrium frequency stability is reached. The figure on the next page illustrates the warmup of two
OCXOs. The warmup time of an oscillator is a function of the thermal properties of the resonator, the
oscillator and oven control circuits, the oven construction, the input power, and the oscillator's
temperature prior to turn-on. Typical warmup time specifications of OCXOs (e.g., from a 0°C start)
range from 3 minutes to 10 minutes. Even TCXOs, MCXOs, and simple XOs take a few seconds to
"warm up," although these are not ovenized. The reasons for the finite warmup, i.e., the stabilization
periods, are that it takes a finite amount of time for the signal to build up in any high-Q circuit, and the
few tens of milliwatts of power which are dissipated in these oscillators can change the thermal
conditions within the oscillators.
Double oven oscillators (in production) have achieved f vs. T stabilities of a few parts in 1010
over a wide temperature range.
F. L. Walls, "Analysis of High Performance Compensated Thermal Enclosures," Proc. 41st Ann.

Warmup of AT- and SC-cut Resonators
Fractional Frequency Deviation From Turnover Frequency

10-3
10-4
10-5
dT
{
Deviation from static f vs. t = ~
a ,
10-6 dt
where, for example,
~
a ≈-2 x 10-7 s/K2
10-7
for a typical AT-cut resonator
10-8
0
3 6 9 12 Time (min) 15
-10-8
Oven Warmup Time
-10-7
-10-6
4-47
Changing the temperature surrounding a crystal unit produces thermal gradients when, for
example, heat flows to or from the active area of the resonator plate through the mounting clips. The
static f vs. T characteristic is modified by the thermal-transient effect resulting from the thermal-
gradient-induced stresses. When an OCXO is turned on, there can be a significant thermal-transient
effect. The above figure shows what happens to the frequency output of two OCXOs, each containing
an oven that reaches the equilibrium temperature in six minutes. One oven contains an AT-cut, the
other, an SC-cut crystal. Thermal gradients in the AT-cut produce a large frequency undershoot that
anneals out several minutes after the oven reaches equilibrium. The SC-cut crystal, being "stress-
compensated" and thereby insensitive to such thermal-transient-induced stresses, reaches the
equilibrium frequency as soon as the oven stabilizes.
In addition to extending the warmup time of OCXOs, when crystals other than SC-cuts are used,
the thermal-transient effect makes it much more difficult to adjust the temperature of OCXO ovens to
the desired turnover points, and the OCXO frequencies are much more sensitive to oven-temperature
fluctuations.
The testing and compensation accuracies of TCXOs are also adversely affected by the thermal-
transient effect. As the temperature is changed, the thermal-transient effect distorts the static f vs. T
characteristic, which leads to apparent hysteresis (see “Apparent Hysteresis” later in this chapter).
The faster the temperature is changed, the larger is the contribution of the thermal-transient effect to
the f vs. T performance.
A. Ballato, and J. R. Vig, "Static and Dynamic Frequency-Temperature Behavior of Singly and
Doubly Rotated, Oven-Controlled Quartz Resonators," Proc. 32nd Annual Symposium on
Frequency Control, pp. 180-188, 1978, AD-A955718.

TCXO Thermal Hysteresis
1.0
Fractional Frequency Error (ppm)

0.5
0.0
-25 -5 15 35 55 75
Temperature (0C)
-0.5
-1.0 TCXO = Temperature Compensated Crystal Oscillator
4-48
The f vs. T characteristics of crystal oscillators do not repeat exactly upon temperature cycling.
The lack of repeatability in temperature-compensated crystal oscillators (TCXOs), called "thermal
hysteresis," is illustrated above, showing that the f vs. T characteristic upon increasing temperature
differs from the characteristic upon decreasing temperature.
Hysteresis is defined as the difference between the up-cycle and the down-cycle f vs. T
characteristics, and is quantified by the value of the difference at the temperature where the difference
is maximum. Hysteresis is determined during at least one complete quasistatic temperature cycle
between specified temperature limits.
Hysteresis is the major factor limiting the stability achievable with TCXOs. It is especially so in
the microcomputer compensated crystal oscillator (MCXO - see chapter 2) because, in principle, the
digital compensation method used in the MCXO would be capable of compensating for the f vs. T
variations to arbitrary accuracy (up to the noise limit) if the f vs. T characteristics could be described
by single-valued functions.
Typical values of hysteresis in TCXOs range from 1 ppm to 0.1 ppm when the temperature-
cycling ranges are 0°C to 60°C, and -55°C to +85°C. Hysteresis of less than 1 x 10-8 has been
observed in a some SC-cut (MCXO) resonators, but the typical MCXO resonator hysteresis is a few
parts in 108.
J. A. Kusters and J. R. Vig, "Thermal Hysteresis in Quartz Resonators - A Review," Proc. 44th Annual
Symposium on Frequency Control, pp. 165-175, 1990, IEEE Catalog No. 90CH2818-3. This paper is
also available at <http://www.ieee.org/uffc/fc> .

Apparent Hysteresis
45
40
Normalized frequency change (ppm)

35
30
25
20
15
10
0
-15
25
75
85
-55
-45
-35
-5
15
35
45
-25
55
65
Temperature (C)
4-49
Shown above is the -55oC to +85oC to -55oC frequency vs. temperature characteristic of an SC-
cut resonator when the temperature was measured with a quartz thermometer external to the
resonator. When, during the same temperature cycle, the temperature was measured by means of
the self-temperature sensing method, no hysteresis could be detected when observed on the scale
shown above (for a discussion of the self-temperature sensing method, see “Effects of Harmonics on
f vs. T” later in this chapter, and “Resonator Self-Temperature Sensing” in chapter 2).
The apparent hysteresis shown above is due to the thermal lag between the resonator and
thermometer during the temperature cycle (the rate of change of temperature was ~0.25oC/min). The
left to right shift between the two curves is an indicator that one is observing apparent, rather than real
hysteresis. Real hysteresis usually shifts the curves vertically, i.e., in real hysteresis, at the same
temperature, there is a frequency difference between the two curves. In apparent hysteresis, the
thermal gradients dominate the f vs. T results.
The apparent hysteresis would have been much greater if the resonator had been, e.g., an AT-
cut, because of the thermal transient effect (see “Warmup of AT- and SC-cut Resonators,” earlier in
this chapter).
R. L. Filler, “"Measurement and Analysis of Thermal Hysteresis in Resonators and TCXO's" Proc.
42nd Ann. Symp. On Frequency Control, pp. 380-388, 1988.
J. A. Kusters and J. R. Vig, "Thermal Hysteresis in Quartz Resonators - A Review," Proc. 44th Annual
Symposium on Frequency Control, pp. 165-175, 1990, IEEE Catalog No. 90CH2818-3.

OCXO Retrace
15
14 days
10
OVEN
5 OFF
OVEN ON (a)
-9
f X 10
15
∆f
14 days
10
OSCILLATOR
5
OFF
OSCILLATOR ON (b)
0
In (a), the oscillator was kept on continuously while the oven was cycled
off and on. In (b), the oven was kept on continuously while the oscillator
was cycled off and on.
4-50
The lack of repeatability of the f vs. T characteristics of oven controlled crystal oscillators
(OCXOs), called "retrace," is illustrated above. Retrace is defined as the nonrepeatability of the f vs. T
characteristic, usually at the oven temperature of an OCXO, upon on-off cycling under specified
conditions. Retrace is a function of the storage temperature during the off period.
The OCXO retrace example above shows that upon restarting the oscillator after a 14 day off-
period, the frequency was about 7x10-9 lower than what it was just before turn-off, and that the aging
rate had increased significantly upon the restart. About a month elapsed before the pre-turn-off aging
rate was reached again. (Figure shows ∆f/f in parts in 109 vs. time in days.)
Retrace limits the accuracies achievable with OCXOs in applications where the OCXO is on-off
cycled. Typical OCXO retrace specifications, after a 24 hour off period at about 25°C, range from 2 x
10-8 to 1 x 10-9. Low-temperature storage during the off period, and extending the off period, often
make the retrace worse.
The causes of hysteresis and retrace are not well understood. The experimental evidence to
date is inconclusive. The mechanisms that can cause these effects include strain changes in the
resonator’s mounting structure, changes in the quartz, oscillator circuitry changes, contamination
redistribution in the crystal enclosure, and apparent hysteresis or retrace due to thermal gradients.
W. L. Smith, and W. J. Spencer, "Quartz Crystal Controlled Oscillators," Final Report, U.S. Army
Contract DA36-039 SC-85373, Report No. 25335-H, 15 March 1963, AD-419717.
R. A. Sykes, W. L. Smith, and W. J. Spencer, "Studies on High Precision Resonators," Proc. 17th
Annual Symposium on Frequency Control, pp. 4-27, 1963, AD423381. Proc. copies available from
NTIS.
J. A. Kusters and J. R. Vig, "Thermal Hysteresis in Quartz Resonators - A Review," Proc. 44th Annual Symposium
on Frequency Control, pp. 165-175, 1990, IEEE Catalog No. 90CH2818-3. This paper is also available at
<http://www.ieee.org/uffc/fc>

TCXO Trim Effect
2
-6 ppm aging adjustment
(ppm)
15
55
-25
35
75
-5
T (0C)
∆f
f
-1
+6 ppm aging adjustment
In TCXO’s, temperature sensitive reactances are used to compensate for f vs. T

variations. A variable reactance is also used to compensate for TCXO aging. The effect
of the adjustment for aging on f vs. T stability is the “trim effect”. Curves show f vs. T
stability of a “0.5 ppm TCXO,” at zero trim and at ±6 ppm trim. (Curves have been
vertically displaced for clarity.)
4-51
An important effect in TCXOs is the interaction between the frequency adjustment during
calibration and the f vs. T stability. This phenomenon is called the trim effect. In TCXOs,
temperature-dependent reactance variations are used to compensate for the crystal's f vs. T
variations. During calibration, the crystal's load reactance is varied to compensate for the TCXO's
aging. Since the frequency versus reactance relationship is nonlinear (see next page), the
capacitance change during calibration moves the operating point on the frequency versus reactance
curve to a point where the slope of the curve is different, which changes the compensation (i.e.,
compensating for aging changes the f vs. T stability). The next page shows how, for the same
compensating CL vs. T, the compensating f vs. T changes when the operating point is moved to a
different CL. Shown above are test results for a “0.5 ppm TCXO” that had a ±6 ppm frequency-
adjustment range (to allow for aging compensation for the life of the device). When delivered, this
TCXO met its 0.5 ppm f vs. T specification; however, when the frequency was adjusted ±6 ppm during
testing, the f vs. T performance degraded significantly. The 0.5 ppm TCXO was shown to be a 2 ppm
TCXO.
In specifying a TCXO, it is important to require that the f vs. T stability include the hysteresis
and trim effects.
R. L. Filler, V. J. Rosati, S. S. Schodowski, and J. R. Vig, "Specification and Measurement of the

Frequency Versus Temperature Characteristics of Crystal Oscillators," Proc. 43rd Annual Symposium
on Frequency Control, pp. 253-255, 1989, IEEE Catalog No. 89CH2690-6.

Why the Trim Effect?
∆f C1
∆f ≈
fS 2(C0 + CL )
fS
Compensated f vs. T
CL
Compensating CL vs. T
4-52
The same variation of Cl with T causes a different f vs. T compensation when the value of CL is
changed.

Effects of Load Capacitance on f vs. T
12
* 10-6 SC-cut
∆f
0
f
-4
-8
r = Co/C1 = 746
α = 0.130
-12
-50 200 450 700 950 1200 1450 1700 1950
T DEGREES CELSIUS
4-53
A load capacitor, CL, changes not only the frequency (see chapter 3), but also the frequency vs.
temperature characteristic. Shown above are the f vs. T characteristics of the same resonator with
and without a CL. The CL rotates the f vs. T curve as if the angle of cut had been lowered (because
the combined effect of the resonator and CL lowers the effect of piezoelectricity, as explained in the
reference below). So, the CL raises the frequency at all temperatures (the curve with fL has been
vertically displaced for clarity), and it also rotates the f vs. T to a lower apparent angle of cut, i.e., it
reduces the turning-point-to-turning-point frequencies and temperatures.
The temperature coefficient of CL can greatly amplify the f vs. T rotation.
A. Ballato, "Frequency-Temperature-Load Capacitance Behavior of Resonators for TCXO

Application," IEEE Trans. Sonics Ultrasonics., SU-25, pp.185-191, July 1978.

Effects of Harmonics on f vs. T
50
40
30
20 ∞
∆f 10
5
f 0 3
(ppm)-10
1
-20 AT-cut
M
Reference angle-of-cut (θ) is about
-30
8 minutes higher for the overtone modes.
-40 (for the overtone modes of the SC-cut,
the reference θ-angle-of-cut is about 30
-50 minutes higher)
-100 -80 -60 -40 -20 -0 20 40 60 80 ∆T, 0C
4-54
Shown above are the frequency vs. temperature characteristics of a resonator excited on the
fundamental mode, third overtone, and fifth and higher overtones. The f vs. T of the fundamental
mode is different from that of the same resonator’s third and higher overtones. When excited on the
third overtone, the f vs. T is only slightly different from that of the 5th overtone (because the higher the
overtone, the less the effects of piezoelectricity, as is explained in the Ballato reference). The rotation
of the f vs. T with overtone is as if the angle of cut had been lowered. For example, for an AT-cut, the
third overtone’s f vs. T is the same as that of a fundamental mode resonator the angle of cut of which
is eight minutes lower.
When the f vs. T characteristics are described by polynomials, it is found that the change
between the fundamental and 3rd and higher overtones are due almost entirely to a change in the first
order temperature coefficient, i.e., the linear terms of the polynomials. This fact is exploited in the
microcomputer compensated crystal oscillator (MCXO - see the MCXO discussions in chapter 2). In
the MCXO, the fundamental (f1) and third overtone (f3) frequencies are excited simultaneously (“dual
mode” excitation) and a beat frequency fβ is generated such that fβ = 3f1 - f3 (or = f1 - f3/3). The fβ is
then a monotonic and nearly linear function of temperature. The resonator can, thereby, become its
own thermometer (“self-temperature sensing”). The fβ senses the resonator’s temperature exactly
where the resonator is vibrating, thereby eliminating temperature gradient effects, and, because an
SC-cut is used, thermal transient effects are also eliminated.
A. Ballato, and T. Lukaszek, "Higher Order Temperature Coefficients of Frequency of Mass-Loaded

Piezoelectric Crystal Plates," Proc. 29th Ann. Symp. on Frequency Control, pp. 10-25, 1975, AD-
A017466. Proc. copies available from NTIS.

Amplitude - Frequency Effect
Normalized current amplitude
10 µ W 100 µ W 400 µ W 4000 µ W

10 -6
Frequency
At high drive levels, resonance curves become asymmetric
due to the nonlinearities of quartz.
4-55
When the frequency is swept through resonance, as the driving voltage is increased, the
resonance curve bends over due to the nonlinear constants of quartz. The peak of each curve is the
resonance frequency. The AT-cut, illustrated above, bends towards higher frequencies, i.e., it
behaves as a hard spring (a hard spring’s stiffness increases with increasing displacement). Some
other cuts behave as soft springs; the resonance curve bends towards lower frequencies. The locus
of the maxima of the resonance curves varies as the square of the current, ∆f/f = aI2, where a is
typically in the range of 0.02/A2 to 0.2/A2; a depends on resonator design - angles of cut, overtone,
plate contour, etc.
At high drive levels, the amplitude vs. frequency curves can be triple valued functions, but only
the highest and lowest values are accessible experimentally (upon increasing and decreasing
voltages, respectively). The current vs. frequency exhibits discontinuities as the driving voltage (or
frequency) is increased or decreased around the resonance peak.
D. Hammond, C. Adams, and L. Cutler, "Precision Crystal Units," Proc. 17th Annual Symposium on
Frequency Control, pp. 215-232, 1963, AD-423381.
R. L. Filler, "The Amplitude-Frequency Effect in SC-Cut Resonators," Proc. 39th Annual Symposium
H. F. Tiersten and D. S. Stevens, “The Evaluationof the Coefficients of Nonlinear Resonance for SC-
cut Quartz Resonators,” Proc. 39th Annual Symposium on Frequency Control, pp. 325-332, 1985,
IEEE Catalog No. 85CH2186-5.
J.J. Gagnepain and R. Besson, “Nonlinear Effects in Piezoelectric Quartz Crystals,” in Physical
Acoustics, Vol. XI, pp. 245-288, W.P. Mason & R.N. Thurston, editors, Academic Press, 1975

Frequency vs. Drive Level
Frequency Change (parts in 109)

80
5 MHz AT
60
3 diopter
40 10 MHz SC
2 diopter
10 MHz SC
20
1 diopter
10 MHz SC
0
-20
10 MHz BT
100 200 300 400 500 600 700
Crystal Current (microamperes)
4-56
Frequency varies as the current amplitude squared; f = f0 (1 + aI2), where a is a constant that is
a function of the design, i.e., the appropriate nonlinear elastic constant for the resonator type used,
and I is the current.
Useful drive levels are limited at the high end by the nonlinearities of quartz, and at the low end
by noise (or defects, as discussed on the next page).
J. A. Kusters, "The SC Cut Crystal - An Overview," Proc. 1981 IEEE Ultrasonics Symposium, pp. 402-
409.
J.J. Gagnepain and R. Besson, “Nonlinear Effects in Piezoelectric Quartz Crystals,” in Physical
Acoustics, Vol. XI, pp. 245-288, W.P. Mason & R.N. Thurston, editors, Academic Press, 1975
R. L. Filler, "The Amplitude-Frequency Effect in SC-Cut Resonators," Proc. 39th Annual Symposium
D. Hammond, C. Adams, and L. Cutler, "Precision Crystal Units," Proc. 17th Annual Symposium on
Frequency Control, pp. 215-232, 1963, AD-423381.

Drive Level vs. Resistance
Anomalous Normal Drive level
starting operating effects
resistance range
Resistance R1
10-3 10-2 10-1 1 10 100

IX (mA)
4-57
The high resistance at low drive currents, and the “second level of drive” effect shown on the
next page, are troublesome because they can appear and cause failure well after the resonator has
passed the final manufacturing tests. The effect can prevent oscillator start-up when the oscillator
circuit’s gain is insufficient. In some resonators, the effect can be “cured” by applying a high drive
level, however, such cures are often not permanent. The effect can reappear long after the high drive
level was applied. (Increasing the drive level increases the amplitude of vibration at the resonator’s
surface, which can remove surface contamination. However, when the high drive is applied to a
hermetically sealed resonator, the contamination stays within the enclosure and can redeposit at a
later time.)
Removing loose particles and other surface contaminants during processing, before the
resonator is sealed into its enclosure, e.g., by properly etching and cleaning the resonators, can
significantly reduce the incidence of this effect. Imperfections in the electrodes, such as scratches,
blisters and irregular adhesion can also produce the effect.
Properly made resonators show no resistance increase at low drive levels - down to at least 10-
5 amperes of drive current*.
------------------------------
* R. Smythe, Piezo Technology, Inc., private communication, April 1999.
M. Bernstein, “Increased Crystal Unit Resistance at Oscillator Noise Levels,” Proc. 21st Annual
J. E. Knowles, "On the Origin of the 'Second Level of Drive' Effect in Quartz Oscillators," Proc. 29th
Annual Symposium on Frequency Control, pp. 230-236, 1975, AD-A017466.
E. P. EerNisse, “An Analysis of the Drive Level Sensitivity in Thickness Shear Quartz Resonators,”
Proc. 1996 Int’l Frequency Control Symposium, pp. 346-356, 1996, IEEE Cat. No. 96CH35935.
E.P EerNisse, E. Benes, M. Schmid, “The Role of Localized Rotational Imbalance in Drive Level
Dependence Phenomena,” Pro. 2002 IEEE Int’l Frequency Control Symp., pp. 2-7, 2002.

Second Level of Drive Effect
Activity (current)
O A D Drive level (voltage)
4-58
A “good” crystal will follow the path OABCBAO without hysteresis. A “bad” crystal will follow the
path OADBCBAO: hence the term “second level of drive”. Upon increasing the drive again, there is a
tendency for the magnitude of the effect to decrease, but in a very irregular and irreproducible
manner. The effect is usually due to particulate contamination, loose electrodes, or other surface
defects.
M. Bernstein, “Increased Crystal Unit Resistance at Oscillator Noise Levels,” Proc. 21st Annual
J. E. Knowles, "On the Origin of the 'Second Level of Drive' Effect in Quartz Oscillators," Proc. 29th
Annual Symposium on Frequency Control, pp. 230-236, 1975, AD-A017466.
E. P. EerNisse, “An Analysis of the Drive Level Sensitivity in Thickness Shear Quartz Resonators,”
Proc. 1996 Int’l Frequency Control Symposium, pp. 346-356, 1996, IEEE Cat. No. 96CH35935
E.P EerNisse, E. Benes, M. Schmid, “The Role of Localized Rotational Imbalance in Drive
Level Dependence Phenomena,” Pro. 2002 IEEE Int’l Frequency Control Symp., pp. 2-7,
2002.

Activity Dips
fL2
∆f fL1
Frequency
10 X10-6
f
fR
RL2
Resistance
RL1
R1
-40 -20 0 20 40 60 80 100

Temperature (0C)
Activity dips in the f vs. T and R vs. T when operated with and without load
capacitors. Dip temperatures are a function of CL, which indicates that the dip is
caused by a mode (probably flexure) with a large negative temperature coefficient.
4-59
Anomalies in the f vs. T and R vs. T characteristics, as illustrated above, are called “activity
dips”. The curves labeled fR and R1 are the f vs. T and R vs. T without a load capacitor. The fL and RL
curves are the f vs. T and R vs. T of the same resonator when load capacitors are in series with the
resonator. The load capacitors shift the frequency to higher values; the curves have been vertically
displaced for clarity.
An activity dip can cause intermittent failures. It affects both the frequency and the resistance
(i.e., the Q) of resonators. When the oscillator gain is insufficient, the resistance increase stops the
oscillation. For example, the clock in one satellite stopped periodically a certain time interval after the
satellite entered the earth’s shadow. As the satellite cooled, the oscillator’s temperature reached the
activity dip temperature and the oscillation stopped. Upon further temperature change, the oscillation
resumed.
Even when the resistance increase is not large enough to stop the oscillation, the frequency
change can cause intermittent failures, e.g., it can cause a loss of lock in phase locked systems.
Activity dips are usually caused by interfering modes (e.g., by high overtone flexure modes).
Such activity dips are strongly influenced by the crystal's drive level and load reactance. The activity-
dip temperature is a function of CL because the interfering mode’s frequency usually has a large
temperature coefficient and a C1 that is different from that of the desired mode. When the frequency
of the interfering mode coincides with the frequency of the main mode, energy is lost from the main
mode and an activity dip occurs.
Activity dip like features can also be caused by processing problems, e.g., loose contacts. Such
dips usually change with temperature cycling, rather than with load capacitance.
--------------------------------
A. Ballato and R. Tilton, "Electronic Activity Dip Measurement," IEEE Trans. on Instrumentation and
Measurement, Vol. IM-27, No. 1, pp. 59-65, March 1978; also “Ovenless Activity Dip Tester,” in Proc.
31st Ann. Symp. On Frequency Control, pp. 102-107, 1977.
See also “Overtone Response…” & “Unwanted Modes vs. Temperature” in Chapter 3.
E.P EerNisse, E. Benes, M. Schmid, “The Role of Localized Rotational Imbalance in Drive Level
Dependence Phenomena,” Pro. 2002 IEEE Int’l Frequency Control Symp., pp. 2-7, 2002.

Frequency Jumps
2.0 x 10-11
30 min.
Frequency deviation (ppb)
4.0
3.0
2.0
No. 2
1.0 No. 3
0.0 No. 4
-1.0
0 2 4 6 8 10
Elapsed time (hours)
4-60
When the frequencies of oscillators are observed for long periods, occasional frequency jumps
can be seen. In precision oscillators, the magnitudes of the jumps are typically in the 10-11 to 10-9
range. The jumps can be larger in general purpose units. The jumps occur many times a day in some
oscillators, and much less than once a day in others. The frequency excursions can be positive or
negative. The causes (and cures) are not well understood.
The causes are believed to include the resonator - nearby spurious resonances, stress relief,
changes in surface and electrode irregularities; and noisy active and passive circuit components, such
as the bistable “burst noise” observed in solid state devices and resistors (which look similar to some
of the features in the upper curve above). The effect can depend on resonator drive level. In some
units, frequency jumps can be produced at certain drive levels (but not below or above, as can be
seen in the Koyama reference). Aging affects the incidence. Well aged units show a lower incidence
of jumps than new units (or newly turned on units). Minimizing surface and electrode imperfections
can contribute to lowering the incidence of jumps (e.g., according to hearsay, unetched or lightly
etched crystals exhibit more jumps than deeply etched crystals; and plating and frequency adjustment
in two steps result in more jumps than "one-shot" plating).
Environmental effects can also produce jumps. Magnetic field, pressure, temperature, and
power transients can produce sudden frequency excursions, as can shock and vibration. It is not
unusual, for example, to experience shock and vibration levels of >0.01g in buildings as trucks pass
by, heavy equipment is moved, boxes are dropped, etc. [Note that, for example, 0.02g x 10-9/g = 2 x
10-11.]
-------------------------
K. K. Tuladhar & G. Jenni, “Frequency Jumps on BVA & Other Precision Quartz Crystal Resonators
and Burst-Noise on Overtone Mode High-Frequency Quartz Crystal Resonators,” Proc. 1996 IEEE
Int’l Frequency Control Symp., pp. 339-342, 1996, IEEE Cat. No. 96CH35935.
M. Koyama, “An Experimental Study of Long Term Aging of Quartz Oscillators,” Proc. 1995 IEEE Int’l
Frequency Control Symp., pp. 620-622, 1995, IEEE Cat. No. 95CH35752.
G. E. Moulton, “Burst Noise and 1/F Noise in Quartz Crystals and Oscillators,” Proc. 42nd Ann. Symp.
on Frequency Control, pp. 389-396, 1988.
M. J. Buckingham, Noise in Electronic Devices and Systems, Chapter 7, “Burst Noise,” Ellis Horwood
Ltd., Halsted Press: a div. of John Wiley & Sons, New York, 1983.

Acceleration Affects “Everything”
• Acceleration Force Deformation (strain)
Change in material and device properties - to some level
• Examples:
- Quartz resonator frequency
- Amplifier gain (strain changes semiconductor band structure)
- Laser diode emission frequencies
- Optical properties - fiber index of refraction (acoustooptics)
- Cavity frequencies
- DRO frequency (strain changes dielectric constants)
- Atomic clock frequencies
- Stray reactances
- Clock rates (relativistic effects)
4-61

Acceleration vs. Frequency Change
Z’
∆f
f ±
´
³ µ
± ³
O
Y’
G
² ²
Crystal
plate
´
X’ µ
Supports

Frequency shift is a function of the magnitude and direction of the

acceleration, and is usually linear with magnitude up to at least 50 g’s.
4-62
When a resonator experiences an acceleration, the strains due to the acceleration cause
frequency changes, as shown above. Under vibration, the time varying strains cause time dependent
frequency changes, i.e., the vibration causes frequency modulation, as shown on the pages that
follow.
The relationship between frequency and acceleration can become nonlinear at high
accelerations due to deformation of the mounting structure. The acceleration sensitivity can also be a
function of temperature.
M. Valdois, J. Besson, and J.J. Gagnepain, “Influence of Environment conditions on a Quartz

Resonator,” Proc. 28th Annual Symposium on Frequency control, pp. 19-32, 1794, AD-A011113.
J. R. Vig, C. Audoin, L. S. Cutler, M. M. Driscoll, E. P. EerNisse, R. L. Filler, R. M. Garvey, W. L. Riley,

R. C. Smythe, and R. D. Weglein, "Acceleration, Vibration and Shock Effects - IEEE Standards Project
P1193," Proc. 1992 IEEE Frequency Control Symposium, 763-781, 1992; also, The Effects of
Acceleration on Precision Frequency Sources, U. S. Army Laboratory Command Research and
Development Technical Report SLCET-TR-91-3, March 1991, AD-A235470.

2-g Tipover Test
(∆f vs. attitude about three axes)
Axis 3 10.000 MHz oscillator’s tipover test

∆f
X 10 −9 (f(max) - f(min))/2 = 1.889x10-09 (ccw)
(f(max) - f(min))/2 = 1.863x10-09 (cw)
f delta θ = 106.0 deg.
4
2
0
180
225
270
360
135
315
45
90
-2
-4 Axis 1
Axis 2 4 (f(max) - f(min))/2 = 6.841x10-10 (ccw)
(f(max) - f(min))/2 = 6.896x10-10 (cw)
2 delta θ = 150.0 deg.
135
180
225
270
360
90
315
45
-2
-4 Axis 2
(f(max) - f(min))/2 = 1.882x10-09 (ccw)
4 (f(max) - f(min))/2 = 1.859x10-09 (cw)
delta θ = 16.0 deg.
2
Axis 1 0
315
360
135
180
225
270
45
90
-2
g -4
4-63
When an oscillator is rotated 1800 about a horizontal axis, the scalar product of the gravitational
field and the unit vector normal to the initial “top” of the oscillator changes from -1g to +1g, i.e., by
2g. A simple “2g tipover” test is sometimes used to test an oscillator’s acceleration sensitivity. Above
is actual data of the fractional frequency shifts of an oscillator when the oscillator was rotated about
three mutually perpendicular axes in the earth’s gravitational field. For each curve, the axis of rotation
was horizontal. The sinusoidal shape of each curve is a consequence of the scalar product being
proportional to the cosine of the angle between the acceleration-sensitivity vector (see later) and the
acceleration due to gravity.
The 2g tipover test must not be used indiscriminately because many oscillators exhibit irregular
variations of frequency with attitude when tested in small increments of angle rather than with just
simple 180o rotations. Irregularities can be caused by, for example, temperature changes due to air
convection in the oscillator, and tiny movements of components, circuit boards and wires. When the
frequency vs. attitude behavior is nonsinusoidal, the results of a simple 2g tipover test can be highly
misleading.
--------------------------
R. L. Filler, "The Acceleration Sensitivity of Quartz Crystal Oscillators: A Review," IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 35, No. 3, pp. 297-305, May
1988.
J. R. Vig, C. Audoin, L. S. Cutler, M. M. Driscoll, E. P. EerNisse, R. L. Filler, R. M. Garvey, W. L.
Riley, R. C. Smythe, and R. D. Weglein, "Acceleration, Vibration and Shock Effects - IEEE Standards
Project P1193," Proc. 1992 IEEE Frequency Control Symposium, 763-781, 1992; also, The Effects of
IEEE Standard 1193-1994, “IEEE Guide for Measurement of Environmental Sensitivities of
Standard Frequency Generators”.

Sinusoidal Vibration Modulated Frequency
t=0
Acceleration
f0 - ∆f f0 + ∆f
Time
π Time
t=
f0 - ∆f f0 + ∆f 2fv
π
t=
f0 - ∆f f0 + ∆f fv
Voltage
3π
t=
f0 - ∆f f0 + ∆f 2fv
Time
2π
t=
f0 - ∆f f0 + ∆f fv
4-64
The above illustration shows how a time varying stress due to a time varying (sinusoidal) acceleration
changes the frequency of a resonator, i.e., how the vibration frequency modulates the oscillator’s
frequency.

Acceleration Levels and Effects
Environment Acceleration ∆f
typical levels*, in g’s x10-11, for 1x10-9/g oscillator
Buildings**, quiesent 0.02 rms 2
Tractor-trailer (3-80 Hz) 0.2 peak 20
Armored personnel carrier 0.5 to 3 rms 50 to 300
Ship - calm seas 0.02 to 0.1 peak 2 to 10
Ship - rough seas 0.8 peak 80
Propeller aircraft 0.3 to 5 rms 30 to 500
Helicopter 0.1 to 7 rms 10 to 700
Jet aircraft 0.02 to 2 rms 2 to 200
Missile - boost phase 15 peak 1,500
Railroads 0.1 to 1 peak 10 to 100
* Levels at the oscillator depend on how and where the oscillator is mounted
Platform resonances can greatly amplify the acceleration levels.
** Building vibrations can have significant effects on noise measurements
4-65
Acceleration (vibration) is present nearly everywhere. When an accelerometer is attached to a

workbench for a day in a typical “quiet laboratory,” 10-3 to 10-2 g peak levels can usually be measured.
Buildings vibrate, and these vibrations can affect noise measurements. Even “low” levels of vibration
can produce significant errors while evaluating low-noise oscillators.
In applications where the platform vibrates, such as in an aircraft, the vibration induced phase
noise degradation can cause severe problems - see, for example, “Random-Vibration-Induced Phase
Noise,” and “Phase Noise Degradation Due to Vibration,” in the pages that follow.
The above table was compiled by Arthur Ballato based on AMC Pamphlet 706-117, Engineering
Design Handbook, “Environmental Series, Part Three, Induced Environmental Factors,” HQ US Army
Materiel Command, Alexandria, VA 22333, USA, January 1976, chapters 4, 5, and 6.

Acceleration Sensitivity Vector
Axis 3
Γ = γ1î + γ2 ĵ + γ3k̂
γ3
Γ = γ12 + γ22 + γ32
Γ
γ1
γ2
Axis 1 Axis 2
4-66
Acceleration-sensitivity is a vector, i.e., the acceleration-induced frequency shift is maximum when

the acceleration is along the acceleration-sensitivity vector; ∆f = Γ • Α
It has been shown, empirically, that the acceleration sensitivity of a quartz crystal oscillator is a
vector quantity. The frequency of an accelerating oscillator is a maximum when the acceleration is
parallel to the acceleration-sensitivity vector. The frequency shift is zero for any acceleration in the
plane normal to the acceleration-sensitivity vector, and it is negative when the acceleration is
antiparallel to the acceleration-sensitivity vector.
Typical values of |Γ| for precision crystal oscillators are in the range of 10-9 per g to 10-10 per g.
|Γ| is independent of acceleration amplitude for the commonly encountered acceleration levels (i.e., at
least up to 20g); however, high acceleration levels can result in changes, e.g., in the mounting
structure, that can lead to Γ being a function of acceleration. Γ can also be a function of temperature.
The magnitude of the acceleration sensitivity is the vector sum (square-root of the sum of the
squares) of the sensitivities along three mutually perpendicular axes.


Vibration-Induced Allan Deviation Degradation
σy (τ)
10-9
−12
γA τν
σy (τ) = 10
+
τ π τ
-10
10
−12
σy (τ) = 10
10-11
τ
10-12
0.001 0.01 τ(sec) 0.1 1
Example shown: fv = 20, Hz A = 1.0 g along Γ, Γ = 1 x 10-9/g
4-67
Vibration modulates the frequency and, thereby, degrades the short-term stability. The typical
degradation due to sinusoidal vibration varies with averaging time, as shown. Since a full sine wave
averages to zero, the degradation is zero for averaging times that are integer multiples of the period of
vibration. The peaks occur at averaging times that are odd multiples of half the period of vibration.
The σy(τ) due to a single-frequency vibration is:
σy (τ) = (Γ•A/π)(τv/τ) sin2 [π(τ/τv)],
where τv is the period of vibration, τ is the measurement averaging time, Γ is the acceleration
sensitivity vector, and A is the acceleration.


Vibration-Induced Phase Excursion
The phase of a vibration modulated signal is
⎛∆ f ⎞
φ(t ) = 2πf0 t + ⎜⎜ ⎟⎟sin(2πfv t )
⎝ fv ⎠
When the oscillator is subjected to a sinusoidal vibration, the peak
phase excursion is
∆ φ peak =
(
∆ f Γ • A f0
=
)
fv fv
Example: if a 10 MHz, 1 x 10-9/g oscillator is subjected to a 10 Hz
sinusoidal vibration of amplitude 1g, the peak vibration-induced phase
excursion is 1 x 10-3 radian. If this oscillator is used as the reference
oscillator in a 10 GHz radar system, the peak phase excursion at 10GHz
will be 1 radian. Such a large phase excursion can be catastrophic to
the performance of many systems, such as those which employ phase
locked loops (PLL) or phase shift keying (PSK).
4-68


Vibration-Induced Sidebands
0
NOTE: the “sidebands” are spectral
lines at ±fV from the carrier frequency
(where fV = vibration frequency). The -10 L(f)
lines are broadened because of the finite
bandwidth of the spectrum analyzer. -20
-30
10g amplitude @ 100 Hz

-40
Γ = 1.4 x 10-9 per g
-50
-60
-70
-80
-90
-100
-100
200
-250
-200
-150
100
250
150
-50
50
0 f
4-69
As shown in the references, for small modulation index β, i.e., β ≡ ∆f/fv = (Γ•A)fo/fv<0.1, sinusoidal
vibration produces spectral lines at ± fv from the carrier, where fv is the vibration frequency. For an
ideal sine wave, the “sidebands” are spectral lines (i.e., delta functions) not spectral densities. Most of
the power is in the carrier, a small amount is in the first spectral line pair, and the higher order spectral
lines are negligible. On a spectrum analyzer, the spectral lines appear to be sidebands - due to the
finite bandwidth of the spectrum analyzer.


Vibration-Induced Sidebands
After Frequency Multiplication
L(f) 0 Each frequency multiplication

-10
by 10 increases the sidebands
by 20 dB.
-20
-30
-40
10X
-50
-60
-70
1X
-80
-90
-100
250
-250
-200
-150
100
-100
150
200
-50
50
0
f
4-70
Upon frequency multiplication by a factor N, the vibrational frequency fv is unaffected, as it is an

external influence. The peak frequency change due to vibration, ∆f, however, becomes
∆f = (Γ•A) Nf0.
The modulation index β is therefore increased by the factor N. Expressed in decibels, frequency
multiplication by a factor N increases the phase noise by 20 log N. When exposed to the same
vibration, the relationship between the vibration-induced phase noise of two oscillators with the same
vibration sensitivity and different carrier frequencies is
LB(f) = LA(f) +20 log (fB/fA),
where LA(f) is the sideband level, in dBc/Hz (or dBc for sinusoidal vibration), of the oscillator at
frequency fA and LB(f) is the sideband level of the oscillator at frequency fB. For the same acceleration
sensitivity, vibration frequency and output frequency, the sidebands are identical, whether the output
frequency is obtained by multiplication from a lower frequency or by direct generation at the higher
frequency. For example, when a 2 x 10-9/g sensitivity 5 MHz oscillator’s frequency is multiplied by a
factor of 315 to generate a frequency of 1575 MHz, its output will contain vibration-induced sidebands
which are identical to those of a 1575 MHz SAW oscillator that has the same 2 x 10-9/g sensitivity.


Sine Vibration-Induced Phase Noise
Sinusoidal vibration produces spectral lines at ±fv from the
carrier, where fv is the vibration frequency.
⎛ Γ • Af0 ⎞
L' (fv ) = 20 log ⎜⎜ ⎟⎟
⎝ 2f v ⎠
e.g., if Γ = 1 x 10-9/g and f0 = 10 MHz, then even if the
oscillator is completely noise free at rest, the phase “noise”
i.e., the spectral lines, due solely to a sine vibration level of
1g will be;
Vibr. freq., fv, in Hz L’(fv), in dBc

1 -46
10 -66
100 -86
1,000 -106
10,000 -126
4-71


Random Vibration-Induced Phase Noise
Random vibration’s contribution to phase noise is given by:
⎛ Γ • Af0 ⎞
L (f ) = 20 log ⎜⎜ where lAl = [(2)(PSD )]
1
⎟⎟, 2
⎝ 2 f ⎠
e.g., if Γ = 1 x 10-9/g and f0 = 10 MHz, then even if the
oscillator is completely noise free at rest, the phase “noise”
i.e., the spectral lines, due solely to a vibration of power
spectral density, PSD = 0.1 g2/Hz will be:
Offset freq., f, in Hz L’(f), in dBc/Hz

1 -53
10 -73
100 -93
1,000 -113
10,000 -133
4-72


Random-Vibration-Induced Phase Noise
Phase noise under vibration is for Γ = 1 x 10-9 per g and f = 10 MHz
L(
-70 f) u
nd
e
PSD (g2/Hz)
.07
-80 r th
er .04
-90
an
do
m
-100 vib
rat
ion
L (f) (dBc)
5 300 1K 2K
-110 s ho Frequency (Hz)
L( wn
-120 f) Typical aircraft
wi 45 dB
tho
-130 ut random vibration
vib
ra envelope
-140 tio
n
-150
-160
5 300 1K 2K
4-73
Shown above is a typical aircraft random vibration specification (power spectral density vs.
vibration frequency is shown in the upper right portion of the figure) and the resulting vibration-induced
phase noise degradation. The degradation of 45 dB can produce severe system performance
degradation.

Acceleration Sensitivity vs. Vibration Frequency
10-8
Vibration Sensitivity (/g)

10-9
mit
range li
z er d ynamic
m analy
10-10 Spectru
100 200 300 400 500 1000

Vibration Frequency (Hz)
4-74
The acceleration sensitivity, Γ, can be calculated from the vibration induced sidebands. In an
ideal oscillator, Γ vs. fv would be constant, but real oscillators exhibit resonances. In the above
example, the resonance at 424 Hz resulted in a 17-fold increase in Γ. The preferred test method
includes measurement of Γ at multiple vibration frequencies in order to reveal resonances.
MIL-PRF-55310 requires that “Measurements shall be made at least 5 equally spaced
acceleration levels between 20 percent of the maximum and the maximum specified.”

The complete MIL-PRF-55310 is available on the Internet via a link from <http:\\www.ieee.org/uffc/fc>,
or directly, from <http://www.dscc.dla.mil/Programs/MilSpec/ListDocs.asp?BasicDoc=MIL-PRF-
55310>
Copies of MIL-PRF-55310 are also available by mail from: Military Specifications and Standards,
Bldg. 4D, 700 Robbins Avenue, Philadelphia, PA 19111-5094, USA. Customer Service telephone:
(215) 697-2667/2179; Telephone Order Entry System (requires a touch tone telephone and a
customer number) (215) 697-1187 thru 1195.

Acceleration Sensitivity of Quartz Resonators
Resonator acceleration sensitivities range from the low parts in 1010 per g for
the best commercially available SC-cuts, to parts in 107 per g for tuning-fork-type
watch crystals. When a wide range of resonators were examined: AT, BT, FC, IT,
SC, AK, and GT-cuts; 5 MHz 5th overtones to 500 MHz fundamental mode inverted
mesa resonators; resonators made of natural quartz, cultured quartz, and swept
cultured quartz; numerous geometries and mounting configurations (including
rectangular AT-cuts); nearly all of the results were within a factor of three of 1x10-9
per g. On the other hand, the fact that a few resonators have been found to have
sensitivities of less than 1x10-10 per g indicates that the observed acceleration
sensitivities are not due to any inherent natural limitations.
Theoretical and experimental evidence indicates that the major variables yet
to be controlled properly are the mode shape and location (i.e., the amplitude of
vibration distribution), and the strain distribution associated with the mode of
vibration. Theoretically, when the mounting is completely symmetrical with respect
to the mode shape, the acceleration sensitivity can be zero, but tiny changes from
this ideal condition can cause a significant sensitivity. Until the acceleration
sensitivity problem is solved, acceleration compensation and vibration isolation can
provide lower than 1x10-10 per g, for a limited range of vibration frequencies, and at
a cost.
4-75
J. T. Stewart, P. Morley, & D. S. Stevens, “Theoretical and Experimental Results for the Acceleration
Sensitivity of Rectangular Crystal resonators,” Proc. 1999 IEEE Int’l Frequency Control Symposium,
1999.
B. J. Lwo & H. F. Tiersten, "Calculation of the Optimal Clip Dimensioning to Minimize the Influence of
Fabrication Imperfections on the Acceleration Sensitivity of SC-cut Quartz Resonators...," Proc. IEEE
Int'l Frequency Control Symp., pp. 165-171, 1994.
E. P. EerNisse, R. W. Ward and O. L. Wood, "Acceleration-Induced Frequency Shifts in Quartz

Resonators," Proc. 43rd Ann. Symposium on Frequency Control, pp. 388-395, 1989, IEEE Cat. No.
89CH2690-6

Phase Noise Degradation Due to Vibration
• Data shown is for a 10 MHz, Impacts on Radar Performance
2 x 10-9 per g oscillator • Lower probability of detection
• Lower probability of identification
• Radar spec. shown is for a coherent • Shorter range
radar (e.g., SOTAS) • False targets
-50
“Good’ oscillator on
vibrating platform (1g)
Required to
“see” 4km/hr
target 53 dB -100 “Good’
oscillator
at rest
dBc/Hz
Radar oscillator
specification
-150
100K 10K 1K 100 10 1 1 10 100 1K 10K 100K

OFFSET FROM CARRIER (Hz)
4-76
At X-band, for example, low phase noise 70 Hz from the carrier is required to detect slowly
moving (4 km/h) objects. Shown above is an example of the phase noise degradation due to vibration
- showing phase noise vs. offset frequency from the carrier. The phase noise degradation was 53dB at
70 Hz from the carrier. In the example on the next page, a degradation > ~25 dB causes the radar to
have a zero probability of detection for 4 km/hr objects. Therefore, with the degradation shown above,
the radar was blind to slow moving objects.

Coherent Radar Probability of Detection
To “see” 4 km/h targets, low phase noise 70 Hz from the
100 carrier is required. Shown is the probability of detection of
4 km/h targets vs. the phase noise 70 Hz from the carrier
Probability of Detection (%)

of a 10 MHz reference oscillator. (After multiplication to 10
GHz the phase noise will be at least 60 dB higher.) The
phase noise due to platform vibration, e.g., on an aircraft,
80 reduces the probability of detection of slow-moving targets
to zero.
60
40
20
Low Noise High Noise

-140 -135 -130 -125 -120 -115 -110
Phase Noise (dBc/Hz)
at 70 Hz from carrier, for 4 km/h targets
4-77
Shown above is the probability of detection vs. reference oscillator phase noise for an X-band
(~10 GHz) coherent radar system. To detect a slowly moving (4km/h) object with a 90% probability
of detection (as was the requirement for one developmental airborne system), the phase noise
requirement had to be -130 dBc/Hz at 70 Hz from the carrier frequency. On a quiet, stationary
platform, the phase noise requirement is well within the state of the art. However, on an airborne
platform, the vibration of the aircraft degrades the phase noise (see previous page) to the point
where the probability of detection is zero. To meet the requirements while in an aircraft, the
oscillator’s acceleration sensitivity had to be << 1x 10-10 per g.

Vibration Isolation
Limitations
1
Transmissibility
Region of • Poor at low frequencies

Amplification
• Adds size, weight and cost
Region of • Ineffective for acoustic

Isolation
noise
0.2 1 2
Forcing Freq./Resonant Freq.
4-78
A simple vibration isolation system is itself a resonant structure. It can be effective at high
frequencies (along one direction), but it amplifies the vibration at, and below its resonant frequency.
Moreover, the isolation system’s dimensions must accommodate large displacements at low
frequencies and high accelerations.
For sinusoidal vibration, the vibration displacement d = dosin 2πft, and the acceleration a = -
do(2πf)2 sin 2πft, where do is the peak displacement and f is the vibration frequency. Therefore, do =
0.50 G/f2 meters, peak-to-peak, where G is the acceleration in units of g. For example, the peak-to-
peak displacement at 1 Hz and 1 g is 0.5 meters.
Acoustic noise can be especially troublesome in certain applications. For example, when an
extremely low noise oscillator was required in an aircraft radar application, after the system designers
built a three-level vibration isolation system to isolate the oscillator from the vibration of the aircraft,
they discovered that the isolation system failed to deliver the expected phase noise of the oscillator
because the isolation system failed to deal with the acoustic noise in the aircraft; i.e., the isolation
system was effective in isolating the oscillator from the vibrations of the airframe, but it was ineffective
in blocking the intense sound waves that impinged on the oscillator.


Vibration Compensation
Stimulus Response Compensated Oscillator
f lt
-9 /Vo
10
OSC. 7x ACC AMP OSC.
ACC
DC Voltage on Crystal 5 MHz fund. SC

Vibration Compensated
Oscillator
ACC = accelerometer
fv OSC.
Response to Vibration
AC Voltage on Crystal fO - fV fO fO - fV
fv OSC.
fO - fV fO fO - fV
Crystal Being Vibrated fO - fV fO fO - fV
4-79
The frequency of a doubly rotated resonator, such as the SC-cut, varies linearly with a voltage
applied to the resonator’s electrodes. (The frequency of a singly rotated, i.e., φ = 0, resonator, such
as the AT-cut, does not vary with voltage.) For example, the voltage sensitivity of a 5 MHz
fundamental mode SC-cut resonator is 7 x 10-9 per volt, as shown in the upper middle above.
Therefore, when an AC voltage of frequency fv is applied to the electrodes, sidebands (i.e.,
spectral lines) at ± fv from the carrier frequency fo are generated, as shown above, in the middle. By
adjusting the magnitude of the AC voltage, the sidebands can be made identical to the sidebands
produced by sinusoidal vibration at frequency fv. When the vibration and AC voltage are applied
simultaneously, the amplitude and phase of the AC voltage can be adjusted so as to cancel the
vibration induced sidebands.
A 60 dB suppression of the sidebands, at a single frequency, has been demonstrated. An
accelerometer sensed the vibration, and the output signal from the accelerometer was applied to the
resonator’s electrodes after calibration of the output’s phase and amplitude. Away from the frequency
of optimum suppression, the suppression degraded due to frequency dependent phase shifts in the
(simple) circuitry used, and mechanical resonances in the setup.
V. J. Rosati and R. L. Filler, "Reduction of the Effects of Vibration on SC-Cut Quartz Crystal
Oscillators," Proc. 35th Annual Symposium on Frequency Control, pp. 117-121, 1981, AD-A110870

Vibration Sensitivity Measurement System
Controller
Controller
Plotter
Plotteror
or Spectrum
Spectrum
Printer
Printer Analyzer
Analyzer
Signal
Signal
Generator
Generator
Frequency
Frequency Synthesizer
Synthesizer fVf
Multiplier
Multiplier V
(Local
(LocalOscillator)
Oscillator)
(x10)
(x10)
Accelerometer
Accelerometer
Vibration
Vibration
Test
TestOscillator
Oscillator Power Level
Level
Power
Shake Amplifier Controller
ShakeTable
Table Amplifier Controller
4-80
One method of measuring Γ is illustrated above. From the vibration induced sidebands, one can
calculate Γ via the equation
L(f) = 20 log (Γ•Af0/2fv)
from which it follows that
Γi = (2fv/Aif0)10B, where B = L i(fv)/20
where Γi and Ai are the components of the acceleration sensitivity vector and of the acceleration,
respectively, in the (unit vector) i direction. Measurements along three mutually perpendicular axes
characterize Γ, which becomes
Γ = Γii+ ΓI j+ Γkk
The local oscillator is used to mix the carrier frequency down to the range of the spectrum analyzer. If
the local oscillator is not modulated, the relative sideband levels are unchanged by mixing. The
frequency multiplier is used to overcome dynamic range limitations of the spectrum analyzer, using the
"20 log N" enhancement discussed previously. The measured sideband levels must be adjusted for the
multiplication factor prior to insertion into the above equation. (The equation is valid only if β < 0.1. )
The illustration was provided by R.L. Filler, private communication, circa 1988


Shock
∆f
X 10 8
The frequency excursion during a shock is due to f
3
the resonator’s stress sensitivity. The magnitude of
the excursion is a function of resonator design, and
2
of the shock induced stresses on the resonator
(resonances in the mounting structure will Shock
amplify the stresses.) The permanent frequency 1
offset can be due to: shock induced stress

changes, the removal of (particulate) contamination 0
from the resonator surfaces, and changes in the

oscillator circuitry. Survival under shock is primarily -1
a function of resonator surface imperfections.
Chemical-polishing-produced scratch-free resonators -2
have survived shocks up to 36,000 g in air gun tests,
and have survived the shocks due to being fired from -3
a 155 mm howitzer (16,000 g, 12 ms duration).
tO t1
4-81
See chapter 5 for further information about the etching and chemical polishing of quartz crystals.
The shock testing of a frequency source generally consists of measuring the frequency or phase
of the source before and after exposing the device to the specified shock. The phase deviation
resulting from the shock (which is the time integral of the fractional frequency change) can provide
useful information about the frequency excursion during the shock (including the possible cessation of
operation).
R. L. Filler, L. J. Keres, T. M. Snowden, and J. R. Vig, "Ceramic Flatpack Enclosed AT and SC-cut
Resonators," Proc. 1980 IEEE Ultrasonic Symp., pp. 819-824, 1980.
J. R. Vig, J. W. LeBus, and R. L. Filler, "Chemically Polished Quartz," Proc. of the 31st Ann. Symp. On
Frequency Control, pp. 131-143, 1977

Radiation-Induced Frequency Shifts
fO fO = original, preirradiation
frequency
∆fSS
Frequency
fSS = steady-state frequency
(0.2 to 24 hours after
fSS exposure)
∆fSS = steady-state frequency

ft offset
fT = frequency at time t
t0 t
Time
10-11 for natural quartz (and R increase can stop the oscillation)
∆fSS/rad* = { 10-12 for cultured quartz
10-13 for swept cultured quartz
* for a 1 megarad dose (the coefficients are dose dependent)
Idealized frequency vs. time behavior for a quartz resonator following a pulse
of ionizing radiation.
4-82
The above illustration shows a crystal oscillator's idealized frequency response due to a pulse of
ionizing radiation. The response consists of two parts. Initially, there is a transient frequency change
that is due primarily to the thermal-transient effect caused by the sudden deposition of energy into the
resonator. This effect is a manifestation of the dynamic f vs. T effect discussed earlier (see “Warmup
of AT- and SC-cut Resonators”). The transient effect is absent in SC-cut resonators made of high
purity quartz.
In the second part of the response, after steady state is reached, there is a permanent frequency
offset that is a function of the radiation dose and the nature of the crystal unit. The frequency change
versus dose is nonlinear, the change per rad being much larger at low doses than at large doses. At
doses above 1 kilorad (SiO2), the rate of frequency change with dose is quartz-impurity-defect
dependent. For example, at a 1 megarad dose, the frequency change can be as large as 10 ppm
when the crystal unit is made from natural quartz; it is typically 1 to a few ppm when the crystal is
made from cultured quartz, and it can be as small as 0.02 ppm when the crystal is made from swept
cultured quartz.
“The Effects of Ionizing and Particle Radiation on Precision Frequency Sources (Proposal for IEEE
Standards Project P1193),” Proc. 1992 IEEE Frequency Control Symposium, pp. 798-806, 1992.
1193-1994 IEEE Guide for Measurement of Environmental Sensitivities of Standard Frequency

Generators.
King, J. C., and Koehler, D. R., “Radiation Effects on Resonators” in: Precision Frequency Control,
Vol. 2 (E. A. Gerber and A. Ballato, eds.), Academic Press, New York, pp. 147-159, 1985.

Effects of Repeated Radiations
10 2. Reirradiation (after 2.5 x 104 rad)
Fractional Frequency, ppb 3. Reirradiation (after >106 rad)
-10
Five irradiations; responses during the 4th and 5th
irradiations repeated the results of the 3rd. At
least 2 days elapsed between successive
irradiations.
-20
Initial slopes:
1st: -1 x 10-9/rad
-30 1. Initial irradiation 2nd: +1 x 10-11/rad
3rd: +3 x 10-12/rad
4th: +3 x 10-12/rad
5th: +5 x 10-12/rad
-40
10 102 103 104 105 106
Dose, rad(SiO2)
4-83
In the experiments described in the reference, irradiation with X-rays and gamma rays from a
cobalt-60 source gave similar results, thus confirming that a “rad is a rad”. The initial irradiations
resulted in negative shifts, subsequent irradiations resulted in smaller shifts. The frequency changes
vs. time for hours to days after the irradiations significantly exceeded the pre-irradiation aging rates.
Higher doses generally resulted in higher increases in aging rates.
L. J. Palkuti and Q. T. Truong, "An X-Ray Irradiation System for Total-Dose Testing of Quartz
Resonators," Proc. 38th Annual Symposium on Frequency Control, pp. 55-62, 1984, AD-A217381.

Radiation Induced ∆f vs. Dose and Quartz Type
10 MeV electrons,
5 MHz 5th overtone
50 AT-cut resonators

30
Z-growth
cultured
10
0
-10 Swept Z-growth
cultured
-30
-50
Natural
104 5 105 5 106 5 107

Reds (Si)
4-84
The steady-state frequency shifts due to ionizing radiation are due to radiation caused changes
at impurity defects. The defect of major concern in quartz is the substitutional Al3+ defect with its
associated interstitial charge compensator, which can be an H+, Li+, or Na+ ion, or a hole. This defect
substitutes for a Si4+ in the quartz lattice (see chapter 5). Radiation can result in a change in the
position of weakly bound compensators, which changes the elastic constants of quartz and thereby
leads to a frequency change. The movement of ions also results in a decrease in the crystal's Q, i.e.,
in an increase in the crystal's equivalent series resistance, especially upon exposure to a pulse of
ionizing radiation. If the oscillator's gain margin is insufficient, the increased resistance can stop the
oscillation for periods lasting many seconds.
B. R. Capone, A. Kahan, R. N. Brown, and J. R. Buckmelter, "Quartz Crystal Radiation Effects," IEEE
Trans. Nuclear Sci., NS-17, pp. 217-221, 1970.
J. C. King and D. R. Koehler, Radiation Effects on Resonators. In: Precision Frequency Control, Vol. 2
(E. A. Gerber and A. Ballato, eds.), Academic Press, New York, pp. 147-159, 1985.
J. J. Suter, et al., “The Effects of Ionizing and Particle Radiation on Precision Frequency Sources,”
Proc. 1992 IEEE Frequency Control Symposium, IEEE Cat. No. 92CH3083-3, 1992.
J. C. King and H. H. Sander, “Rapid Annealing of Frequency Change in High Frequency Crystal
Resonators Following Pulsed X-irradiation at Room Temperature,” Proc. 27th Ann. Symp. Frequency
Control, pp. 117-119, NTIS Accession No. AD-771042, 1973.
R. E. Paradysz, and W. L. Smith, “Crystal Controlled Oscillators for Radiation Environments,” Proc.
27th Ann. Symp. Frequency Control, pp. 120-123, NTIS Accession No. AD-771042, 1973.

Annealing of Radiation Induced f Changes
-∆f
40 ∆fS = 41 Hz
Frequency change, Hz
x x
x x
x X T= 4330K (1600C)
x x
X T= 4540K
x
x X T= 4680K
x
20
x
X T= 4880K (2150C)
T = 5130K(2400C)
0 x x
100 200 300
Annealing time, minutes
• For a 4 MHz AT-cut resonator, X-ray dose of 6 x 106 rads produced ∆f = 41 Hz.
• Activiation energies were calculated from the temperature dependence of the
annealing curves. The experimental results can be reproduced by two
processes, with activation energies E1 = 0.3 ± 0.1 eV and E2 = 1.3 ± 0.3eV.
• Annealing was complete in less than 3 hours at > 2400C.
4-85
P. Freymuth and G. Sauerbrey, "Ausheilung von Bestrahlungsschaden in Quarz- Schwingkristallen,"

J. Phys. Chem. Solids, Pergamon Press, Vol. 24, pp. 151-155, 1963.

Transient ∆f After a Pulse of γ Radiation
X
4
X
X X
0 X
X
X
X
-4 X
∆f/f (pp 108)
X X XX
X
X
X X
X
-8 X
X X
X
X X
X
X X
-12 X
-16
Experimental data, dose = 1.3 x 104 rads, SC-cut
X Experimental data, dose = 2.3 x 104 rads, AT-cut

-20
Model Calculation: AT-cut

-24
0.1 1.0 10 100 Time 1000
(seconds after event)
4-86
J. C. King and D. R. Koehler, "Radiation Effects on Resonators," in E. A. Gerber and A. Ballato,

Precision Frequency Control, Vol. 1, pp. 147-159, Academic Press, 1985.
J. C. King and H. H. Sander, “Rapid Annealing of Frequency Change in High Frequency Crystal
Resonators Following Pulsed X-irradiation at Room Temperature,” Proc. 27th Ann. Symp. Frequency
Control, pp. 117-119, NTIS Accession No. AD-771042, 1973.

Effects of Flash X-rays on RS
32 MHz AT-cut resonators 14
80
13
N-4 12
70
N-4 (Natural) 4.5 x 104 R
Value of Q-1 x 106

11
C-7 (Unswept synthetic) 4 x 104 R
RS in Ohms
60
C-22 (Unswept synthetic) 3.5 x 104 R 10
S-25 Swept synthetic) 4 x 104 R
C-7 S-26 Swept synthetic) 4 x 104 R 9
50
8
7
40
C-22
6
(S-25)
30 5
(S-26)
Preirradiation value RS
0.001 0.01 0.1 1.0 10 100 1000

Time following exposure (seconds)
The curves show the series resonance resistance, RS, vs. time following a
4 x 104 rad pulse. Resonators made of swept quartz show no change in RS
from the earliest measurement time (1 ms) after exposure, at room
temperature. Large increases in RS (i.e., large decrease in the Q) will stop the
oscillation.
4-87
Ionizing radiation changes a crystal oscillator's frequency primarily because of changes the
radiation produces in the crystal unit (also see chapter 5). Under certain conditions, the radiation will
also produce an increase in the crystal unit's equivalent series resistance. The resistance increase
can be large enough to stop the oscillation when the oscillator is not radiation hardened.
A high level pulse of ionizing radiation will produce photocurrents in the circuit which result in a
momentary cessation of oscillation, independent of the type of quartz used in the resonator. In
oscillators using properly designed oscillator circuitry and resonators made of swept quartz, the
oscillator recovers within 15 µs after exposure.
The energy band gap of quartz is ~9 eV; it is an insulator, however, a pulse of ionizing radiation
(x-rays, γ-rays, high energy particles) creates electrons and holes, and these result in a momentary
conductivity that lasts ~5 to 30 ns after the pulse. In addition, the radiation induced electrons and
holes lead to a freeing of interstitial H+, Li+ and Na+ which results in additional conductivity. The
conductivity results in losses, i.e., a drop in the resonator’s Q.
J. C. King and H. H. Sander, "Transient Change in Q and Frequency of AT-Cut Quartz Resonators
Following Exposure to Pulse X-Rays," IEEE Trans. Nucl. Sci., NS-20, pp. 117-125, 1973.

Frequency Change due to Neutrons
1000
X106
900 5 MHz
AT-cut
800
∆f
Frequency Deviation, f
700
600
Slope = 0.7 x 10-21/n/cm2
500
400
300
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Fast Neutron Exposure (nvt) x1017
4-88
When a fast neutron hurtles into a crystal lattice and collides with an atom, it is scattered like a
billiard ball. A single such neutron can produce numerous vacancies, interstitials, and broken
interatomic bonds (see next page). The effect of this "displacement damage" on oscillator frequency
is dependent primarily upon the neutron fluence (and not on the type of quartz). The frequency of
oscillation increases nearly linearly with neutron fluence at rates of: 8 x 10-21 neutrons per square
centimeter (n/cm2) at a fluence range of 1010 to 1012 n/cm2, 5 x 10-21/n/cm2 at 1012 to 1013 n/cm2, and
0.7 x 10-21/n/cm2 at 1017 to 1018 n/cm2.
J. C. King and D. B. Fraser, "Effects of Reactor Irradiation on Thickness Shear Crystal Resonators,"
Proc. 16th Annual Symposium on Frequency Control, pp. 8-31, 1962, AD-285086.
T. M. Flanagan & T. F. Wrobel, “Radiation Effects in Swep-Synthetic Quartz,” IEEE Trans. Nuclear
Science, vol. NS-16, pp. 130-137, Dec. 1969
W. Primak, “Extrusion of Quartz on Ion Bombardment: Further Evidence for Radiation-Induced Stress
Relaxation of the Silica Network,” Phys. Rev. B, vol. 14, pp. 4679-4686, 1976.

Neutron Damage
(1) (2) (3) (4)
A fast neutron can displace about 50 to 100 atoms before it comes

to rest. Most of the damage is done by the recoiling atoms. Net result
is that each neutron can cause numerous vacancies and interstitials.
4-89
Shown above is the typical sequence of events when an energetic neutron strikes a crystal. In
(1), a neutron has arrived and has dislodged a single atom. The (2), (3) and (4) drawings show how
the process builds up and ends, with both the neutron and the recoiling atoms dislodging further
atoms. The end result is a number of vacant sites and interstitial atoms.
F. Seitz and E. P. Wigner, “The Effects of Radiation on Solids,” Scientific American, pp. 76-84, August
1956.

Summary - Steady-State Radiation Results
z Dose vs. frequency change is nonlinear; frequency change per rad is larger at low doses.
z At doses > 1 kRad, frequency change is quartz-impurity dependent. The ionizing radiation
produces electron-hole pairs; the holes are trapped by the impurity Al sites while the
compensating cation (e.g., Li+ or Na+) is released. The freed cations are loosely trapped along
the optic axis. The lattice near the Al is altered, the elastic constant is changed; therefore, the
frequency shifts. Ge impurity is also troublesome.
z At a 1 MRad dose, frequency change ranges from pp 1011 per rad for natural quartz to pp 1014
per rad for high quality swept quartz.
z Frequency change is negative for natural quartz; it can be positive or negative for cultured and
swept cultured quartz.
z Frequency change saturates at doses >> 106 rads.
z Q degrades upon irradiation if the quartz contains a high concentration of alkali impurities; Q of
resonators made of properly swept cultured quartz is unaffected.
z High dose radiation can also rotate the f vs. T characteristic.
z Frequency change anneals at T > 240°C in less than 3 hours.
z Preconditioning (e.g., with doses > 105 rads) reduces the high dose radiation sensitivities upon
subsequent irradiations.
z At < 100 rad, frequency change is not well understood. Radiation induced stress relief & surface
effects (adsorption, desorption, dissociation, polymerization and charging) may be factors.
4-90
See also “Ions in Quartz - Simplified Model,” “Aluminum Associated Defects,” and “Sweeping” in
Chapter 5.
T. M. Flanagan, R. E. Leadon and D. L. Shannon, "Evaluation of Mechanisms for Low-Dose

Frequency Shifts in Crystal Oscillators," Proc. 40th Ann. Symp. on Frequency Control, pp. 127-133,
1986, IEEE Cat. No. 86CH2330-9
J. C. King and D. R. Koehler, Radiation Effects on Resonators. In: Precision Frequency Control, Vol. 2
(E. A. Gerber and A. Ballato, eds.), Academic Press, New York, pp. 147-159, 1985.

Summary - Pulse Irradiation Results
z For applications requiring circuits hardened to pulse irradiation, quartz resonators are the
least tolerant element in properly designed oscillator circuits.
z Resonators made of unswept quartz or natural quartz can experience a large increase in Rs
following a pulse of radiation. The radiation pulse can stop the oscillation.
z Natural, cultured, and swept cultured AT-cut quartz resonators experience an initial
negative frequency shift immediately after exposure to a pulse of X-rays (e.g., 104 to
105 Rad of flash X-rays), ∆f/f is as large as -3ppm at 0.02sec after burst of 1012 Rad/sec.
z Transient f offset anneals as t-1/2; the nonthermal-transient part of the f offset is probably
due to the diffusion and retrapping of hydrogen at the Al3+ trap.
z Resonators made of properly swept quartz experience a negligibly small change in Rs

when subjected to pulsed ionizing radiation (therefore, the oscillator circuit does not require
a large reserve of gain margin).
z SC-cut quartz resonators made of properly swept high Q quartz do not exhibit transient
frequency offsets following a pulse of ionizing radiation.
z Crystal oscillators will stop oscillating during an intense pulse of ionizing radiation because
of the large prompt photoconductivity in quartz and in the transistors comprising the
oscillator circuit. Oscillation will start up within 15µsec after a burst if swept quartz is used
in the resonator and the oscillator circuit is properly designed for the radiation environment.
4-91

Summary - Neutron Irradiation Results
z When a fast neutron (~MeV energy) hurtles into a crystal lattice and collides with an
atom, it is scattered like a billiard ball. The recoiling atom, having an energy (~104
to 106 eV) that is much greater than its binding energy in the lattice, leaves behind
a vacancy and, as it travels through the lattice, it displaces and ionizes other atoms.
A single fast neutron can thereby produce numerous vacancies, interstitials, and
broken interatomic bonds. Neutron damage thus changes both the elastic
constants and the density of quartz. Of the fast neutrons that impinge on a
resonator, most pass through without any collisions, i.e., without any effects on the
resonator. The small fraction of neutrons that collide with atoms in the lattice cause
the damage.
z Frequency increases approximately linearly with fluence. For AT- and SC-cut
resonators, the slopes range from +0.7 x 10-21/n/cm2, at very high fluences (1017 to
1018n/cm2) to 5 x 10-21/n/cm2 at 1012 to 1013n/cm2, and 8 x 10-21/n/cm2at 1010 to
1012n/cm2. Sensitivity probably depends somewhat on the quartz defect density
and on the neutron energy distribution. (Thermonuclear neutrons cause more
damage than reactor neutrons.)
z Neutron irradiation also rotates the frequency vs. temperature characteristic.
z When a heavily neutron irradiated sample was baked at 500°C for six days, 90% of
the neutron-induced frequency shift was removed (but the 10% remaining was still
93 ppm).
4-92

Other Effects on Stability
z Electric field - affects doubly-rotated resonators; e.g., a voltage on the electrodes of a 5 MHz
fundamental mode SC-cut resonator results in a ∆f/f = 7 x 10-9 per volt. The voltage can also
cause sweeping, which can affect the frequency (of all cuts), even at normal operating
temperatures.
z Magnetic field - quartz is diamagnetic, however, magnetic fields can induce Eddy currents, and
will affect magnetic materials in the resonator package and the oscillator circuitry. Induced ac
voltages can affect varactors, AGC circuits and power supplies. Typical frequency change of a
"good" quartz oscillator is <<10-10 per gauss.
z Ambient pressure (altitude) - deformation of resonator and oscillator packages, and change in
heat transfer conditions affect the frequency.
z Humidity - can affect the oscillator circuitry, and the oscillator's thermal properties, e.g.,
moisture absorbed by organics can affect dielectric constants.
z Power supply voltage, and load impedance - affect the oscillator circuitry, and indirectly, the
resonator's drive level and load reactance. A change in load impedance changes the amplitude
or phase of the signal reflected into the oscillator loop, which changes the phase (and
frequency) of the oscillation. The effects can be minimized by using a (low noise) voltage
regulator and buffer amplifier.
z Gas permeation - stability can be affected by excessive levels of atmospheric hydrogen and
helium diffusing into "hermetically sealed" metal and glass enclosures (e.g., hydrogen diffusion
through nickel resonator enclosures, and helium diffusion through glass Rb standard bulbs).
4-93
Ambient pressure change (as during an altitude change) can change a crystal oscillator's frequency if the
pressure change produces a deformation of the crystal unit's or the oscillator's enclosure (thus changing stray
capacitances and stresses). The pressure change can also affect the frequency indirectly through a change in
heat-transfer conditions inside the oscillator. Humidity changes can also affect the heat-transfer conditions. In
addition, moisture in the atmosphere will condense on surfaces when the temperature falls below the dew point,
and can permeate materials such as epoxies and polyimides, and thereby affect the properties (e.g.,
conductivities and dielectric constants) of the oscillator circuitry. The frequency of a properly designed crystal
oscillator changes less than 5 x 10-9 when the environment changes from one atmosphere of air to a vacuum.
The medium and long term stability of some oscillators can be improved by controlling the pressure and humidity
around the oscillators.
Electric fields can change the frequency of a crystal unit. An ideal AT-cut is not affected by a dc voltage on
the crystal electrodes, but "doubly rotated cuts," such as the SC-cut, are affected. For example, the frequency of
a 5-MHz fundamental mode SC-cut crystal changes 7 x 10-9 per volt. Direct-current voltages on the electrodes
can also cause sweeping, which can affect the frequencies of all cuts.
Power-supply and load-impedance changes affect the oscillator circuitry and, indirectly, the crystal's drive
level and load reactance. A change in load impedance changes the amplitude or phase of the signal reflected
into the oscillator loop, which changes the phase (and frequency) of the oscillation. The effects can be minimized
through voltage regulation and the use of buffer amplifiers. The frequency of a "good" crystal oscillator changes
less than 5 x 10-10 for a 10% change in load impedance. The typical sensitivity of a high-quality crystal oscillator to
power-supply voltage changes is 5 x 10-11/V.
Gas permeation under conditions where there is an abnormally high concentration of hydrogen or helium in
the atmosphere can lead to anomalous aging rates. For example, hydrogen can permeate into "hermetically"
sealed crystal units in metal enclosures, and helium can permeate through the walls of glass-enclosed crystal
units, and through the glass of glass-to-metal seals.
---------------------------
J. R. Vig & F. L. Walls, "Fundamental Limits on the Frequency Instabilities of Quartz Crystal Oscillators,"
Proc. 1994 IEEE Int'l Frequency Control Symposium, pp. 506-523, 1994
IEEE Standard 1193-1994 “IEEE Guide for Measurement of Environmental Sensitivities of Standard
Frequency Generators”.
R. Brendel, “Influence Of a Magnetic Field on Quartz Crystal Resonators,” IEEE Transactions on
Ultrasonics, Ferroelectrics, and Frequency Control, vol. 43 no. 5, pp. 818-831, September 1996. This paper is also
available at <http://www.ieee.org/uffc/fc> .

Interactions Among Influences
In attempting to measure the effect of a single influence, one often encounters
interfering influences, the presence of which may or may not be obvious.
Measurement Interfering Influence
Resonator aging ∆T due to oven T (i.e., thermistor) aging

∆ drive level due to osc. circuit aging
Short term stability Vibration
Vibration sensitivity Induced voltages due to magnetic fields
2-g tipover sensitivity ∆T due to convection inside oven
Resonator f vs. T Thermal transient effect, humidity

(static) T-coefficient of load reactances
Radiation sensitivity ∆T, thermal transient effect, aging
4-94
The various influences on frequency stability can interact in ways that lead to erroneous test
results if the interfering influence is not recognized during testing. For example, building vibrations
can interfere with the measurement of short-term stability. Vibration levels of 10-3 g to 10-2 g are
commonly present in buildings. Therefore, if an oscillator's acceleration sensitivity is 1 x 10-9/g, then
the building vibrations alone can contribute short-term instabilities at the 10-12 to 10-11 level.
The 2-g tipover test is often used to measure the acceleration sensitivity of crystal oscillators.
Thermal effects can interfere with this test because, when an oscillator is turned upside down, the
thermal gradients inside the oven can vary due to changes in convection currents [6]. Other examples
of interfering influences include temperature and drive-level changes interfering with aging tests;
induced voltages due to magnetic fields interfering with vibration-sensitivity tests; and the
thermal-transient effect, humidity changes, and the effect of load-reactance temperature coefficient
interfering with the measurement of crystal units' static f vs. T characteristics.
An important effect in TCXOs is the interaction between the frequency adjustment during
calibration and the f vs. T stability [71]. This phenomenon is called the trim effect.
IEEE Standard 1193-1994 “IEEE Guide for Measurement of Environmental Sensitivities of Standard
Frequency Generators”.
Series of papers on IEEE Standard 1193 in the Proc. 1992 IEEE Frequency Control Symposium, pp.
762-830, 1992.
S. R. Stein and J. R. Vig, Frequency Standards for Communications, U. S. Army Laboratory

Command Research and Development Technical Report SLCET-TR-91-2, January 1991, AD-
A231990

CHAPTER 5
Quartz Material Properties
Technical Systems, National Research Council, NMAB - 424, National Academy Press, Washington,
D.C., 1985.
R. W. Ward, “The Constants of Alpha Quartz,” Proc. 38th Ann. Symposium on Frequency Control, pp.
22-31, 1984.
L. E. Halliburton, J. J. Martin and D. R. Koehler, “Properties of Piezoelectric Materials,” in E. A. Gerber

and A. Ballato, Precision Frequency Control, Vol. 1, pp. 1-45, Academic Press, 1985.
C. Frondel, The System of Mineralogy, Vol. III, “Silica Minerals”, John Wiley and Sons, Inc., New York,
1962.
J. C. Brice, “Crystals for Quartz Resonators,” Reviews of Modern Physics, Vol. 57, No. 1, pp. 105-146,
January 1985.
J. A. Weil, “A Review of Electron Spin Spectroscopy and Its Application to the Study of Paramagnetic
Defects in Crystalline Quartz,” Physics and Chemistry of Minerals, vol. 10, pp. 149-165, 1984.

Hydrothermal Growth of Quartz
• The autoclave is filled to some

Cover predetermined factor with water plus
mineralizer (NaOH or Na2CO3).
Closure • The baffle localizes the temperature gradient

area so that each zone is nearly isothermal.
• The seeds are thin slices of (usually)

Autoclave Z-cut single crystals.
Growth
zone, T1 Seeds
• The nutrient consists of small (~2½ to 4 cm)
pieces of single-crystal quartz (“lascas”).
Baffle
• The temperatures and pressures are
Nutrient typically about 3500C and 800 to 2,000
Solute- atmospheres; T2 - T1 is typically 40C to 100C.
dissolving
nutrient
zone, T2
• The nutrient dissolves slowly (30 to 260 days
Nutrient per run), diffuses to the growth zone, and
deposits onto the seeds.
T2 > T1
5-1
Prior to ~1956, the material used for quartz resonators was natural quartz, i.e., mined quartz.
Today, it is “cultured quartz,” i.e., quartz grown in factories. Although this quartz is often referred to as
“synthetic quartz,” nobody has yet found a way to synthesize single crystal quartz directly from silicon
and oxygen. Large quartz bars (typically ~15 cm long) of uniform size and shape are grown from
small, irregularly shaped pieces of quartz (called “lascas” by the culturing process described above.
So, strictly speaking, the quartz is “cultured quartz”.
Quartz is a common material in the earth’s crust (e.g., sand is mostly quartz), however, the high
purity crystals needed for quartz growing are not so common. Most of the nutrient materials used by
quartz growers are mined in Brazil and the USA (near Jessieville, Arkansas).
The autoclave is a long, thick-walled ~25 to 100 cm inner diameter steel tube that can withstand
the high temperatures and pressures of the growth process.
The anisotropy of quartz is discussed on the next page, and in chapter 3, where it is pointed out
that the highest etching rate direction is the Z-direction. Similarly, during quartz growing, the Z-
direction is the fastest direction of growth.
R. A. Laudise and R. L. Barns, “Perfection of Quartz and Its Connection to Crystal Growth,” IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 35, No. 3, pp. 277-287, May
1988, IEEE Catalog 88CH2588-2.

Deeply Dissolved Quartz Sphere
Anisotropic Etching
Z
Y
+X
+X
Looking along Y-axis Looking along Z-axis
5-2
Polished quartz spheres, when deeply etched in concentrated HF, dissolve in a highly
anisotropic manner. The partially dissolved spheres become “triangular, lenticular,” as shown above -
the shape is triangular when observed along the Z-axis, and lenticular when observed along the Y-
axis. The etching rate along the fastest etching direction, the Z-direction, is nearly 1000 times faster
than the rate along the slowest direction, the -X direction.
A good review of the early etching studies can be found in C. Frondel, The System of Mineralogy, Vol.
III, “Silica Minerals”, John Wiley and Sons, Inc., New York, 1962.
R. W. Ward, “Etching of Quartz Crystal Spheres,” Proc. 1993 IEEE Int’l Frequency Control
Symposium, pp. 390-396, 1993.

Etching & Chemical Polishing
Diffusion Controlled Etching:
Etchant Must:
1. Diffuse to Surface
2. Be Adsorbed
3. React Chemically Lapped surface
Reaction Products Must:

4. Be Desorbed
5. Diffuse Away
Chemically polished surface
5-3
The etching of a crystal may be limited by any of the five steps shown above. When the etching
rate is reaction limited (i.e., when there is a plentiful supply of etchant molecules at the surface), the
morpholgy of the etched surface is determined primarily by the properties of the material being etched.
Reaction limited etching of an anisotropic material usually results in a rough, faceted surface.
However, when the etching is diffusion limited, i.e., the inherent rate at which a reaction can take
place is higher than the rate of diffusion of etchant molecules to the surface, a depleted surface layer
of etchant molecules exist, outside which the etchant concentration is uniform, and inside which the
concentration decreases to near zero at the surface. When starting with a rough, e.g., a lapped,
surface consisting of hills and valleys, as illustrated in the upper right above, the probability of an
etchant molecule diffusing to the top of a hill will be greater than the probability of it diffusing to the
bottom of a valley. The hills will, therefore, be etched faster than the valleys, and the surface will
become smoother as the etching progresses, i.e., the surface will become “chemically polished”.
Chemically polished surfaces are not perfectly flat. They are microscopically undulating, but
atomically smooth, as illustrated in the lower right above.
J. R. Vig, J. W. LeBus and R. L. Filler, "Chemically Polished Quartz," Proc. 31st Annual Symposium
on Frequency Control, pp. 131-143, 1977, AD-A088221.
J. R. Vig, R. J. Brandmayr and R. L. Filler, “Etching Studies On Singly And Doubly Rotated Quartz
Plates,” Proc. 33rd Annual Symposium on Frequency Control, pp. 131-143, 1979, AD-A213544.
M. Deleuze, et. al, “Controlled Dissolution Applied to Berlinite and Quartz Materials,” Proc. 1993 IEEE
Int’l Frequency Control Symposium, pp. 381-389, 1993.

Left-Handed and Right-Handed Quartz
X X
Y
Y
r r
z z
S
X r z r z S
Z X
Z
Y m m
m m Y
z r z r
Left-Handed Right-Handed
5-4
Quartz is an enantiomorphic crystal, i.e., both right-handed and left-handed crystals exist. The
crystal structures of the two forms are mirror images. Neither can be made to look like the other by
rotation. Both kinds could be equally useful, however, resonator manufacturing processes have been
standardized on right-handed quartz.
Quartz has trigonal structure. The ideal crystal is a hexagonal prism with six cap faces at each
end. The prism faces are called the m-faces, and the major cap faces are called the r-faces; they are
also called the major rhomb faces. The z-faces are called the minor rhomb faces.
The Z-axis is also called the “optic axis,” and the “c-axis”. It is an axis of three-fold symmetry,
i.e., all the physical properties repeat every 120o as the crystal is rotated about the Z-axis. The
polarization of a beam of plane polarized light passed through quartz along the Z-axis will be rotated
by the crystal. The polarization is rotated clockwise in right handed quartz, as seen by an observer
looking through the quartz towards the light source. It is rotated counterclockwise by left handed
quartz. This “optical rotation” ability is used in optical instruments. (It has no significance in frequency
control applications.)
1962.

The Quartz Lattice
Si Si
Z
O
O
144.2o
109o
Y
O Si
Si Si
5-5
For pure α-quartz, the lattice constant a = 0.4913 nm and c/a = 1.10013. Impurities increase the
a lattice constant, and decrease the axial ratio c/a. Small differences between natural quartz and
cultured quartz, and between cultured quartz and swept cultured quartz have been measured.
Pressure also affects the lattice constants.
The average Si-O bond length is 0.1607 nm, the O-Si-O bond angles vary from 108.3o to 110.7o
and the Si-O-Si bond angle is 144.2o.
J. D. Jorgensen, “Compression Mechanisms in α-Quartz Structures - SiO2 and GeO2,” J. Appl. Phys.
Vol 49, pp. 5473-78, 1978.
J. C. Brice, The Lattice Constants of a-Quartz," J. Materials Science, vol. 15, pp. 161-167, 1980.
R. W. Ward, “The Constants of Alpha Quartz,” Proc. 38th Ann. Symposium on Frequency Control, pp.
22-31, 1984.

Quartz Properties’ Effects on Device Properties
Quartz Property Device and Device-Fabrication Property
Q Oscillator short-term stability, phase noise

close to carrier, long-term stability, filter loss
Purity (Al, Fe, Li, Na, K, Radiation hardness, susceptibility to twinning,

-OH, H2O) optical characteristics
Crystalline Perfection, Sweepability, etchability for chem. polishing

Strains and photolithographic processing, optical
properties, strength, aging(?), hysteresis(?)
Inclusions High-temperature processing and applications,

resonator Q, optical characteristics, etchability
5-6

Ions in Quartz - Simplified Model
a
A) B)
= Oxygen
Axis of channel
H
H
= Si4+
Al Al
C) D) E)
Na
Li K
Al Al Al
0.143 eV
0.2 eV
0.089 eV
0.055 eV
5-7
The above diagrams show qualitative models of the possible positions of H+ and alkali ions in the
channels of the quartz lattice, and the corresponding trends in the potential energy curves. The arrows
show the main channels in the quartz lattice - along the z-axis. It takes much less force to move ions
along this channel than along other directions.
The H+ ion is strongly bound to the O ions; A) and B) show two possible models of H+ in quartz
(due to Kats). As the OH bond is very strong, it is unlikely for a H+ to move along the channel at normal
temperatures. C) to E) show the positions of alkali ions in the channels and the corresponding
potentials. There is a potential well of ~1 eV depth around the Al3+ ion. Superimposed on that are a
series of low potential barriers along the channel that are shown above. In the Li+ case, only one kind of
transition is likely. In the Na+ and K+ cases, two different kinds of transitions are likely.
When the Q of a resonator is measured as a function of temperature, defects such as the Al-alkali
centers cause acoustic losses as the thermally activated alkalis couple to the oscillating stress field;
e.g., the Al-Na+ center causes an acoustic loss peak at 50 K in 5 MHz 5th overtone AT-cut resonators.
Above room temperature, the interstitial alkalis are thermally liberated and diffuse along the z-axis which
results in Q losses that increase exponentially with temperature. Such Q losses, and the Al-Na+ loss
peak are not present in swept crystals (see “Sweeping” later in this chapter).
Ionizing radiation creates electron-hole pairs that make a transient electrical conductivity increase.
It also liberates ions and allows the ions to move from one potential well to another, thereby changing
the elastic constants. A pulse of radiation thereby causes both a transient and steady-state change in
the resonator’s frequency (see radiation effects discussion in chapter 4).
J. M. Stevels and J. Volger, “Further Experimental Investigations on the Dielectric Losses of

Quartz Crystals in Relation to Their Imperfections,” Philips Res. Reports. 17, pp. 283-314, 1962.
A. Kats, “Hydrogen in Alpha-quartz, Philips Res. Reports. 17, pp. 133-195, 201-279, 1962.
L. E. Halliburton, J. J. Martin & D. R. Koehler, “Properties of Piezoelectric Materials,” in E. A.

Gerber and A. Ballato, Precision Frequency Control, Vol. 1, pp. 1-45, Academic Press, 1985.

Aluminum Associated Defects
Al-OH center Al-M+ center
Ox
Ox
Ox
Ox
Al3+
M+ Al3+
H+ Interstitial
Ox Alkali
OH molecule Ox
Ox
Ox
Al-hole center Ox
Ox
Al3+
Ox
h+
Hole trapped in nonbonding

oxygen p orbital
5-8
Al3+ ions readily substitute for Si4+ ions in quartz. When such substitution occurs, a charge
compensator is needed for charge neutrality. Four compensators are known in quartz: H+, Li+, Na+
and a hole trapped at an oxygen ion. Above is a schematic representation of the aluminum
associated centers: Al-OH- center, Al-M+ center (where M is an interstitial alkali, either Li+, or Na+),
and Al-hole center. The Al-OH- center is formed when an interstitial proton bonds to an oxygen ion.
The aluminum related centers are related to acoustic losses and radiation induced frequency
shifts in quartz resonators. These effects can be reduced by an electrodiffusion process called
“sweeping,” a process that removes the interstitial cations (H+, Li+, Na+) from the quartz lattice - see
next two pages.
The Al-hole center consists of a hole, i.e., a missing electron, trapped in a nonbonding p orbital
of an oxygen ion located near a substitutional Al. These are so weakly bound (0.03 eV) that, at room
temperature, the hole is rapidly jumping among the four oxygens surrounding the Al. Al-hole centers
affect the optical absorption (smoky coloration) of quartz.
Oxygen vacancy centers, called E’ centers, are another class of point defects. These centers
affect the ultraviolet absorption of quartz.
L. E. Halliburton, J. J. Martin & D. R. Koehler, “Properties of Piezoelectric Materials,” in E. A. Gerber

and A. Ballato, Precision Frequency Control, Vol. 1, pp. 1-45, Academic Press, 1985.
L. E. Halliburton, M. E. Markes, and J. J. Martin, “Point Defects in Synthetic Quartz: A Survey of

Spectroscopic Results With Application to Quality Assurance,” Proc. 34th Annual Symposium on
Frequency Control, pp. 1-8, 1980, AD-A213670. Proc. copies available from NTIS.
L. E. Halliburton, N. Koumvakalis, M. E. Markes, and J. J. Martin, “Radiation Effects in Crystalline

SiO2: The Role of Aluminum,” J. Appl. Phys., Vol. 52, pp. 3565-3574, 1981.

Sweeping
Oven
T = 500OC Thermometer
Ammeter
Z Cr-Au
E = 1000 V/cm High voltage power

Quartz bar
supply
0.5 µa/cm2
Time
5-9
Sweeping is a purification process which removes certain impurities from the quartz and thereby
improves the radiation hardness and etching properties of quartz crystals. It is an electric-field driven,
solid-state diffusion process that is performed at an elevated temperature. As illustrated above, the
major steps of a typical sweeping process consist of applying electrodes to the Z-surfaces of a
lumbered quartz bar, heating the bar slowly to 5000C, applying a voltage to the electrodes such that
the electric field along the Z-direction is about 1 kV/cm, monitoring the current through the bar (as the
sweeping progresses, the current decreases), and, after the current decays to some constant value,
cooling the bar slowly to room temperature, then removing the voltage.
Under the influences of high electric field and high temperature, the positive impurity ions, such
as Li+ and Na+, diffuse to the cathode and are removed when the electrodes are removed in
subsequent processing. In addition to improving radiation hardness, sweeping also greatly reduces
the number of etch channels that are produced when quartz is etched.
Quartz has large channels along the c-axis (also called Z-axis). The interstitial ions can readily
migrate along these channels. The sweeping rate is much lower when the electric field is applied
along other directions.
J. J. Martin, “Electrodiffusion (Sweeping) of Ions in Quartz,” IEEE Transactions on Ultrasonics,

Ferroelectrics, and Frequency Control, Vol. 35, No. 3, pp. 288-296, May 1988, IEEE Catalog
88CH2588-2.
J. G. Gualtieri, “Sweeping Quartz Crystals,” Proc. 1989 IEEE Ultrasonics Symposium, pp. 381-391,
1989.

Quartz Quality Indicators
• Infrared absorption coefficient*
• Etch-channel density *
• Etch-pit density
• Inclusion density *
• Acoustic attenuation
• Impurity analysis
• X-ray topography
• UV absorption
• Birefringence along the optic axis
• Thermal shock induced fracture
• Electron spin resonance
• ???
* EIA Standard 477-1 contains standard test method for this quantity
5-10
The first four tests listed above are the most commonly used quality indicators.
G. R. Johnson and R. A. Irvine, “Etch Channels in Single Crystal Cultured Quartz,” Proc. 41st Annual
Symposium on Frequency Control, pp. 175-182, 1987, IEEE Catalog No. 87CH2427-3.
J. R. Vig, J. W. LeBus and R. L. Filler, "Chemically Polished Quartz," Proc. 31st Annual Symposium
on Frequency Control, pp. 131-143, 1977, AD-A088221. Proc. copies available from NTIS.

Infrared Absorption of Quartz
2.5 3.0 3.5 4.0
100
3500 E parallel to Z
80 E parallel to X
Transmission (%)
3410
60 3585
40
3200
3300
20
0
4000 3500 3000 2500
Wave number (cm-1)
5-11
The importance of Q, and the factors that influence it, are discussed in chapter 3. The Q of a
resonator is due to a combination of factors, one of which is the quartz material. The material can limit
a resonator’s Q, but the contribution of the material cannot be determined from measuring the Q of a
resonator.
Infrared absorption by the quartz material used to make a resonator correlates well with the
maximum Q achievable with that resonator, at a particular frequency; 5 MHz is the standard frequency
used for correlations. The figure above shows the transmission spectra of a 20 mm thick sample, with
the IR incident along two perpendicular directions, X and Z.
“Infrared Q” measurements, per EIA Standard 477-1, are routinely used by quartz growers and
users as an indicator of quartz quality (see next page).
J. C. Brice and A. M. Cole, "The Characterization of Synthetic Quartz by Using Infrared Absorption,"
Proc. 32nd Annual Symposium on Frequency Control, pp. 1-10, 1978.
B. Sawyer, "Quality Indications From Infrared Absorption Measurements on Cultured Quartz," IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. UFFC-34, No. 5, pp. 558-
565, September 1987.
D. M. Dodd and D. B. Fraser, “Infra-red Studies of the Variation of H-bonded OH in Synthetic α-

Quartz,” J. Phys. Chem. Solids, vol. 26, pp. 673-686, 1965.

Infrared Absorption Coefficient
One of the factors that determine the maximum achievable resonator Q is the
OH content of the quartz. Infrared absorption measurements are routinely used
to measure the intensities of the temperature-broadened OH defect bands.
The infrared absorption coefficient α is defined by EIA Standard 477-1 as
A (3500 cm-1) - A (3800 cm-1)

α= Y-cut thickness in cm
where the A’s are the logarithm (base 10) of the fraction of the incident beam
absorbed at the wave numbers in the parentheses.
Grade α, in cm-1 Approx. max. Q*
A 0.03 3.0
B 0.045 2.2
C 0.060 1.8
D 0.12 1.0
E 0.25 0.5
* In millions, at 5 MHz (α is a quality indicator for unswept quartz only).
5-12
EIA Standard 477-1, available from Electronic Industries Alliance, 2500 Wilson Blvd., Arlington, VA
22201-3834, USA,
Tel: +1 703 907 7500, <http://www.eia.org>

Quartz Twinning
>>>>>
>>>
>>>
>
∧∧ ∧ ∧∧ ∧∧∧∧∧ ∧
>> >>>
> >>> ∧∧∧ ∧∧∧ ∧ ∧ ∧ ∧ ∧∧ ∧∧
>>> ∧∧∧∧ ∧∧ ∧∧∧∧∧∧
∧
>> >
> > >>> >> ∧ ∧∧ ∧∧ ∧∧ ∧ ∧∧
∧∧ ∧∧
∧∧∧ ∧∧
> >>>> >> ∧ ∧ ∧ ∧ ∧∧ ∧∧ ∧
> > >> ∧ ∧∧ ∧
∧∧∧ ∧∧ ∧∧∧∧
> > > >> ∧∧ ∧ ∧
∧ ∧ ∧ ∧ ∧∧ ∧∧ ∧∧ ∧ ∧ ∧∧∧ ∧
∧∧ ∧∧ ∧∧ ∧∧ ∧ ∧
> >>>> > >>>> >> >
> > > ∧ ∧∧∧ ∧∧ ∧∧ ∧ ∧
>>> >> >>>>> >> >>>>>
> > > >> >>>>
>>>>>>> >>>> >> >> >
>> > > >> > >> > >>
>>>> >
>>>>>> >>>>>
> > >>>>> >>
∧ ∧
>> > > ∧∧ ∧∧∧∧ ∧∧∧∧∧
Electrical twinning Optical Twinning
• The X-axes of quartz, the electrical axes, are parallel to the line bisecting
adjacent prism faces; the +X-direction is positive upon extension due to
tension.
• Electric twinning (also called Dauphiné twinning) consists of localized reversal
of the X-axes. It usually consists of irregular patches, with irregular
boundaries. It can be produced artificially by inversion from high quartz,
thermal shock, high local pressure (even at room temperature), and by an
intense electric field.
• In right-handed quartz, the plane of polarization is rotated clockwise as seen
by looking toward the light source; in left handed, it is counterclockwise.
Optically twinned (also called Brazil twinned) quartz contains both left and
right-handed quartz. Boundaries between optical twins are usually straight.
• Etching can reveal both kinds of twinning.
5-13
Natural quartz often consists of partly right handed and partly left handed quartz. Such crystals
are said to be optically twinned. Optical twins are growth twins, i.e., if absent from a crystal when the
crystal is grown, they cannot be induced subsequently without breaking silicon-oxygen bonds.
Electrical twins, on the other hand, can be produced in quartz by the application of mechanical and
thermal stress. The higher the temperature, the less stress is required to produce this type of
twinning. The atomic shifts necessary to produce electrical twinning (~0.03 nm) are less than one-
tenth of the lattice spacings; no breaking of bonds takes place.
Twinned plates are not useful for frequency control applications. It is, therefore, necessary to
avoid high stresses during fabrication, especially at elevated temperatures. Processes such as the
thermocompression bonding of mounting clips to the edges of quartz plates can readily produce
twinning, if not done carefully. Twinning has also produced failures in quartz pressure transducers
used at high pressures and temperatures, e.g., in oil wells.
It has been observed that it takes less stress to produce twinning in very high Q, high purity
quartz than in lower quality quartz. One company specifies both a minimum and a maximum material
Q; a minimum to ensure that the material does not limit the overall electrical Q performance, and a
maximum to minimize the incidence of twinning*.
1962.
T. L. Anderson, R. E. Newnham and L. E. Cross, "Coercive Stress for Ferrobielastic Twinning in

Quartz," Proc. 31st Annual Symposium on Frequency Control, pp. 171-177, 1977, AD-A088221.
Jack Kusters, Hewlett-Packard Co., private communication circa 1985, confirmed in 1999.

Twinning - Axial Relationships
- Z + - Z +
r + r - r + Z -
r r
Z Z r r
- Electrical -
+ + + - Optical + -
(Dauphine) (Brazil)
r r - Z + Z Z
Z Z Z Z
+ Z - r
+ r -
r - + r
- r + Z Z - r +
+ + Combined - -
r r
Z - Z
Z +
- r +
The diagrams illustrate the relationship between the axial system

and hand of twinned crystals. The arrows indicate the hand.
5-14
T. L. Anderson, R. E. Newnham & L. E. Cross, "Coercive Stress for Ferrobielastic Twinning in Quartz,"
Proc. 31st Annual Symposium on Frequency Control, pp. 171-177, 1977, AD-A088221.
Processing Related Twinning References:
D. A. Cocuzzi & J. W. Laughner, “Effects of Surface Abrasion And Impurity Levels on Stress-induced
Dauphine Twinning in Alpha Quartz,” Proc. 43rd Ann. Symp. On Frequency Control, Pp. 617-622,
1989.
E. P. EerNisse & R. W. Ward, “Quartz Resonator Sensors in Extreme Environments,” Proc. 45th Ann.
J.J. Boy & P.L. Guzzo, “Quartz Crystal Twinning Under Mechanical Stress: Experimental
Measurements,” Proc. 1996 IEEE Frequency Control Symp., pp. 155-160, 1996.
T. Uno, “Temperature Compensation of Thickness Shear Mode Resonators Formed on Artificial

Twinned Quartz Plates,” Proc. 1996 IEEE Frequency Control Symp., pp. 526-531, 1996.

Quartz Lattice and Twinning
Z-axis projection showing electric (Dauphiné)
twins separated by a twin wall of one unit cell
thickness. The numbers in the atoms are
atom heights expressed in units of percent of
a unit cell height. The atom shifts during
twinning involve motions of <0.03 nm.
Domain
Wall
79
00
21
88 88
45 33 45
Silicon
67 55 67
12
79 79
00 00
Oxygen 21 21
88
33 45
67
5-15
T. L. Anderson, R. E. Newnham and L. E. Cross, "Coercive Stress for Ferrobielastic Twinning in

Quartz," Proc. 31st Annual Symposium on Frequency Control, pp. 171-177, 1977, AD-A088221.
Proc. copies available from NTIS.

Quartz Inversion
• Quartz undergoes a high-low inversion (α - β transformation) at

5730C. (It is 5730C at 1 atm on rising temperature; it can be 10 to 20C
lower on falling temperature.)
• Bond angles between adjoining (SiO4) tetrahedra change at the

inversion. Whereas low-quartz (α-quartz) is trigonal, high quartz
(β -quartz) is hexagonal. Both forms are piezoelectric.
• An abrupt change in nearly all physical properties takes place at the

inversion point; volume increases by 0.86% during inversion from
low to high quartz. The changes are reversible, although Dauphiné
twinning is usually acquired upon cooling through the inversion
point.
• Inversion temperature decreases with increasing Al and alkali

content, increases with Ge content, and increases 10C for each 40
atm increase in hydrostatic pressure.
5-16
1962.

Phase Diagram of Silica (SiO2)
12
Stishovite
10
P(gpa)
Coesite
6
4 Liquid
High or
2 Low or β-quartz
α-quartz
Tridymite Cristobalite
0
0 500 1000 1500 2000 2500
T(oC)
5-17
Shown above is the phase diagram for single crystal silica, SiO2. The Y-axis shows pressure
in gigapascals. The temperature of phase transition between α-quartz and β-quartz increases
with increasing pressure. The α-β transition temperature is 573oC at one atmosphere pressure.
It increases to about 600oC at the typical pressures used for growing quartz in autoclaves.
In frequency control applications, only α-quartz is used.
-
------------------------------------
R. B. Susman, The Phases of Silica, Rutgers Univ. Press, New Brunswick, NJ 1965.
C. Frondel, The System of Mineralogy, Vol. III, “Silica Minerals”, John Wiley and Sons, Inc.,
New York, 1962.

Internal Friction (i.e., the Q) of Quartz
100
60
40
Most probable internal
Value of Q, in millions
20
friction curve for quartz;
excluding mounting losses
10
8
6
4
90 mm
30 mm
2
Diameter of
1
shaped quartz 15 mm
0.8
plates, in
0.6
vacuum
0.4
Flat quartz
0.2
plates, in
air
0.1
0.1 0.2 0.4 0.6 1.0 2 4 6 8 10 20 40 60 100
Frequency in MHz
Empirically determined Q vs. frequency curves indicate that the maximum achievable
Q times the frequency is a constant, e.g., 16 million for AT-cut resonators, when f is in MHz.
5-18
See chapter 3 for further information about Q. Also see “Ions in Quartz - Simplified Model”
earlier in chapter 5 for information about acoustic losses due to aluminum-alkali centers.
A. W. Warner, "Design and Performance of Ultraprecise 2.5-mc Quartz Crystal Units," Bell System
Technical Journal, pp. 1193-1217, Sept. 1960; also in "An Ultraprecise Standard of Frequency," W. L.
Smith, final report for U. S. Army contract DA 36-039 sc-73078, p. 13, 1 Dec. 1960, AD-253034.

Langasite and Its Isomorphs
La3Ga5SiO14 Langasite (LGS)

La3Ga5.5Nb0.5O14 Langanite (LGN)
La3Ga5.5Ta0.5O14 Langatate (LGT)
• Lower acoustic attenuation than quartz (higher Qf than
AT- or SC-cut quartz)
• No phase transition (melts at ~1,400 oC vs. phase transition
at 573 oC for quartz)
• Higher piezoelectric coupling than quartz
• Thicker than quartz at the same frequency
• Temperature-compensated
5-19
Ever since quartz became the material of choice for crystal oscillators (in the 1920s),
researchers have been looking for materials that are even better than quartz. Many materials have
looked promising, e.g., berlinite, lithium tetraborate, and gallium phosphate, but nothing has equaled
quartz.
Langasite (also called LGS) and its isomorphs, langanite (LGN), langatate (LGT), etc. look highly
promising. The Qf product of LGT resonators has been shown to be double that of quartz. Some of
the possible consequences of the improved properties of these materials are as follows:
• Higher Q at the same frequency can lead to lower phase noise close to the carrier (where L(f) ∝
1/Q4)
• Absence of phase transition and high melting point allows higher temperature processing - which
can lead to cleaner, lower aging and lower hysteresis resonators, and to sensors capable of high-
temperature operation
• Higher piezoelectric coupling allows higher overtones; possibly lower hysteresis in higher overtone
resonators; greater tunability in VCXOs; and, in filter applications, wider bandwidth, lower impedance,
and higher frequency operation
• Larger thickness at the same frequency implies less deformation under acceleration and hence,
possibly, lower acceleration sensitivity.
B.V. Mill, Y.V. Pisarevsky, E.L. Belokoneva, “Synthesis, Growth and Some Properties of Single
Crystals with Ca3Ga2Ge4O14 Structure,” Proc. 1999 IEEE Int’l Frequency Control Symposium, 1999
Joint Meeting EFTF-IEEE IFCS, pp. 829-834, 1999.
B. Chai, et al., “Growth and Evaluation of Large Size LGS (La3Ga5SiO14), LGN (La3Ga5.5Nb0.5O14), &
LGT (La3Ga5.5Ta0.5O14) Single Crystals,” Proc. 1998 IEEE Int’l Frequency Control Symposium, pp.
748-760, 1998.
R.C. Smythe, et al., “Langasite, langanite, and langatate bulk-wave Y-cut resonators;”
Ultrasonics, Ferroelectrics and Frequency Control, IEEE Transactions on , Volume: 47
Issue: 2 , pp. 355 -360, March 2000.

CHAPTER 6
Atomic Frequency Standards*
* There are two important reasons for including this chapter: 1. atomic frequency standards are
one of the most important applications of precision quartz oscillators, and 2. those who study or
use crystal oscillators ought to be aware of what is available in case they need an oscillator with
better long-term stability than what crystal oscillators can provide.
General References
1977, Time and Frequency Division, NIST, 325 Broadway, Boulder, Colorado, 80303.
H. Hellwig, "Microwave Frequency and Time Standards," in E. A. Gerber and A. Ballato, Precision
H. Hellwig, "Microwave Time and Frequency Standards," Radio Science, Vol. 14, No. 4, pp. 561-572,
July-August 1979.
F. G. Major, The Quantum Beat - The Physical Principles of Atomic Clocks, Springer-Verlag, 1998.
J. Vanier and C. Audoin, The Quantum Physics of Atomic Frequency Standards, ISBN 0-85274-434-
X, Adam Hilger, 1978.
S. R. Stein and J. R. Vig, "Frequency Standards for Communications," U. S. Army Laboratory

Command Research and Development Technical Report SLCET-TR-91-2 (Rev. 1), October 1991,
AD-A243211. This report is a reprint of a chapter "Communications Frequency Standards," in The
Froehlich/Kent Encyclopedia of Telecommunications, Vol. 3, pp. 445-500, Marcel Dekker, Inc., 1992.
L. L. Lewis, “An Introduction to Frequency Standards,” Proc. IEEE, vol. 79, pp. 927-935, 1991.
N. F. Ramsey, Molecular Beams, Oxford University Press, 1956.
P. Forman, "Atomichron: The Atomic Clock from Concept to Commercial Product," Proc. of the IEEE,
Vol. 73, No. 7, pp. 1181-1204, July 1985.
Several review papers, including three on the environmental sensitivities of atomic frequency
standards, are contained in the Proc. 22nd Ann. Precise Time and Time Interval (PTTI) Applications
and Planning Meeting, NASA Conference Publ. 3116, 1990; AD-A239372.
Proceedings of the IEEE, Special Issue on Time and Frequency, J. Jespersen & D. W. Hanson, ed's.,
Vol. 79, No.
“QUARTZ CRYSTAL RESONATORS AND7, OSCILLATORS
July 1991.
Precision Frequency Standards
• Quartz crystal resonator-based (f ~ 5 MHz, Q ~ 106)

• Atomic resonator-based
Rubidium cell (f0 = 6.8 GHz, Q ~ 107)
Cesium beam (f0 = 9.2 GHz, Q ~ 108)
Hydrogen maser (f0 = 1.4 GHz, Q ~ 109)
Trapped ions (f0 > 10 GHz, Q > 1011)
Cesium fountain (f0 = 9.2 GHz, Q ~ 5 x 1011)
6-1
A high Q is necessary (but not sufficient) for high frequency stability - see Chapter 3 for a
discussion of Q. The higher the Q, the higher the frequency stability and accuracy capability of a
resonator. If, e.g., Q = 106, then 10-10 accuracy requires the ability to determine the center of the
resonance curve to 0.01% of the linewidth, and stability (for some averaging time) of 10-12 requires the
ability to stay near the peak of the resonance curve to 10-6 of linewidth.
A high Q is not sufficient for high stability because a high Q resonator may, for example, have a
poor temperature stability. Sapphire resonators, for example, can have a very high Q, but their poor
temperature stability prevents their use in clocks.
The Q, or line width of an atomic transition is determined by the observation time. The atomic
resonance Qs listed above are typical values. Laser cooling of atoms can significantly extend the
observation time and Q (see “Laser Cooling of Atoms” later in this chapter. Laser cooling is necessary
to achieve a Cs fountain).

Atomic Frequency Standard Basic Concepts
When an atomic system changes energy from an exited state to a
lower energy state, a photon is emitted. The photon frequency ν is given
by Planck’s law E2 − E1
ν =
h
where E2 and E1 are the energies of the upper and lower states,
respectively, and h is Planck’s constant. An atomic frequency standard
produces an output signal the frequency of which is determined by this
intrinsic frequency rather than by the properties of a solid object and how it
is fabricated (as it is in quartz oscillators).
The properties of isolated atoms at rest, and in free space, would not
change with space and time. Therefore, the frequency of an ideal atomic
standard would not change with time or with changes in the environment.
Unfortunately, in real atomic frequency standards: 1) the atoms are
moving at thermal velocities, 2) the atoms are not isolated but experience
collisions and electric and magnetic fields, and 3) some of the components
needed for producing and observing the atomic transitions contribute to
instabilities.
6-2
Atomic frequency standards must be understood in terms of the concepts of quantum

mechanics. The properties of simple atomic systems cannot assume arbitrary values. For example,
the energies of the bound states of an atomic system are constrained to discrete values called energy
levels. When an atomic system changes energy from an excited state to a state with lower energy, it
emits a quantity of electromagnetic energy called a photon, the frequency of which is determined by
the energy difference between the two states, in accordance with Planck’s law, shown above.
Atomic systems can be isolated from unwanted perturbations, which result in small sensitivities
to temperature, pressure, and other environmental conditions. The low level of interaction also results
in extremely sharp resonance features, and reduces errors due to imperfections in the electronics. All
atoms of an element are identical, and atomic properties are time invariant, which makes it possible to
build very stable devices.
Atomic frequency standards are categorized in several ways; most often, they are referred to
by the type of atom: hydrogen, rubidium, or cesium. Actually, these three devices are based on the
same type of atomic interaction, but there are great practical differences in their implementation.
Some atomic frequency standards, called oscillators, are active, in which case the output signal is
derived from the radiation emitted by the atom. Others are passive; the atoms are then employed as
a discriminator to measure and control the frequency of an electronic oscillator, such as a quartz
oscillator. The third classification follows the method of interaction. In atomic beams, the atoms are
observed "on the fly"; they pass through the interaction region and are not used again. In contrast,
storage devices contain some type of cell that holds the atoms to be observed indefinitely (ideally).

AD-A243211. This report is a reprint of a chapter, "Communications Frequency Standards," in The

Hydrogen-Like Atoms
3 MF =
2
Electron 1
spin and 0
Nucleus 2 -1
dipole Electron
Closed
1
electronic
shell N F=2
S
∆W 2 3 4 X
S F=1
Electron -1
Nuclear N
spin and MF =
dipole -2 -2
-1
0
-3 1
Hyperfine structure of 87Rb, with nuclear spin I=3/2,
Hydrogen-like (or alkali) ν0=∆W/h=6,834,682,605 Hz and X=[(-µJ/J) +(µI/I)]H0/∆W
atoms calibrated in units of 2.44 x 103 Oe.
6-3
All commercial atomic frequency standards are based on hyperfine transitions of one of three
hydrogen-like atoms, rubidium, cesium and hydrogen.
The energy levels of an atom are generally classified according to their physical origin. For
example, the principal levels of an atom are associated with the radius of the "orbit" of an electron about
the nucleus. These levels have the largest atomic energy separations. The principal energy levels are
subdivided as a result of the quantization of the angular momentum of the atom. The angular momentum
due to the motion of a particle, such as an electron, is called orbital angular momentum. Even when their
motion is such that there is no orbital angular momentum, atomic particles may possess an intrinsic
angular momentum or spin and a proportional intrinsic magnetic moment. The principal levels are first
divided according to the shape of the electron "orbits." Still finer division occurs as a consequence of the
particular orientation of the electron's spin and the spin of the nucleus.
The photons emitted when atoms change states among the principal energy levels are usually in the
infrared and higher energy regions of the electromagnetic spectrum. The frequencies of these energetic
photons are too high for practical electronic devices. Atomic frequency standards are feasible because of
the splitting of the ground state of the atom. Next lower, in terms of energy, is the fine structure of the
atom, which results from the interaction of the spin of the electron with the magnetic field due to the
motion of the electron through the nuclear electric field. This structure is thousands of times smaller than
the separation of the principal energy levels. Laboratory atomic frequency standards based on fine
structure in calcium and magnesium have been built, but the fundamental frequencies of the atomic
transitions are higher than 600 GHz, which is difficult to synthesize.
A finer energy splitting than the spin-orbit coupling is produced by the interaction of the electron and
nuclear spins; this is called the hyperfine structure. The ground state of a hydrogen-like atom (e.g., H, Li,
Na, K, Rb, Cs, and singly ionized Be) has a single unpaired electron in a symmetric orbit. In this case,
there is no orbital angular momentum and no fine structure. The energy splitting due to the intrinsic
magnetic moments of the electron and the nucleus can be a million times smaller than the separation of
the principal energy levels. The transition frequencies are convenient: 1.4 GHz for hydrogen, 6.8 GHz for
rubidium, and 9.2 GHz for cesium.

Atomic Frequency Standard*
Block Diagram
Multiplier
Multiplier Quartz
Quartz
Atomic
Atomic Crystal
Crystal
Resonator
Resonator Feedback
Feedback Oscillator
Oscillator
5 MHz
Output
* Passive microwave atomic standard (e.g., commercial Rb and Cs standards)
6-4
Atomic resonators are inherently noisy due to the discreet nature of atomic transitions. The
short term stabilities, σy(τ) vs. τ, vary as the square-root of the measurement interval, i.e., as τ1/2, for
short intervals. This is due to the statistics of counting atomic transitions; σy(τ) varies as the square-
root of the number of transitions. Crystal oscillators are less noisy at small τ. Therefore, in all
commercial atomic standards, the atomic resonator frequency is generated from the crystal oscillator’s
frequency (by frequency multiplication or frequency synthesis), and the crystal oscillator frequency is
locked to the frequency of the atomic resonator with a servo loop time constant that is selected to
provide optimum performance for the intended application. Of the many atomic transitions available,
the ones selected are those which are least sensitive to environmental effects and which can be
conveniently locked to the VCXO.
The atomic standard behaves as the crystal oscillator for measurement times shorter than the
time constant (which, for example, is typically 100 ms to 500 ms for a Rb standard, longer in Cs
standards), and it behaves as an atomic oscillator for measurement times longer than the time
constant.
Since all atomic frequency standards derive their output signal from quartz oscillators, the
performance of the atomic standards is significantly affected by the capabilities of the crystal
oscillators. In particular, the short-term frequency stability, the vibration sensitivity, the radiation pulse
sensitivity, and the sensitivity to thermal transients depend on the performance of the crystal oscillator.
The atomic resonator’s superior long term stability and lower sensitivity to environmental changes is
used to “servo out” the crystal oscillator’s aging and some of the crystal oscillator’s environmental
sensitivities.
H. Hellwig, "Microwave Frequency and Time Standards," in E. A. Gerber and A. Ballato,

Precision Frequency Control, Vol. 2, pp. 113-176, Academic Press, 1985.

Generalized Microwave Atomic
Resonator
Prepare
PrepareAtomic
Atomic Apply
Apply Detect
DetectAtomic
Atomic B
State
State Microwaves
Microwaves State
StateChange
Change
hν0
Tune
TuneMicrowave
MicrowaveFrequency
Frequency A
For
ForMaximum
MaximumState
StateChange
Change
6-5
• Let A and B be two possible energy states of an atom, separated by energy hνo; then νo
is the frequency of the electromagnetic radiation required to convert the atoms from A to
B, or from B to A; νo is in the microwave range for all currently manufactured atomic
standards.
• Population difference between energy states, when hνo << kT, is near zero. Therefore, in
a natural ensemble of atoms, when νo is applied, about half the atoms absorb hνo and
half emit hνo; the net effect is zero.
• A nonthermal distribution is prepared, i.e., one of the states is "selected," by optical excita-
tion from one of the levels to a third level or by magnetic deflection of an atomic beam.
• Microwave energy is absorbed in the process of converting the selected atoms to the
other energy state, e.g., from A to B. Thus, the applied microwave frequency can be
"locked" to the frequency corresponding to the atomic transition.
The microwave signals that interrogate the atoms are generally modulated at audio frequencies.
Phase sensitive detection of the atomic signal is used to adjust the frequency of the crystal oscillator
to the frequency that produces the maximum atomic signal.

Atomic Resonator Concepts
• The energy levels used are due to the spin-spin interaction between the atomic nucleus and
the outer electron in the ground state (2S1/2) of the atom; i.e., the ground state hyperfine
transitions.
• Nearly all atomic standards use Rb or Cs atoms; nuclear spins I = 3/2 and 7/2, respectively.
• Energy levels split into 2(I ± 1/2)+1 sublevels in a magnetic field; the "clock transition" is the
transition between the least magnetic-field-sensitive sublevels. A constant magnetic field, the
"C-field," is applied to minimize the probability of the more magnetic-field-sensitive transitions.
• Magnetic shielding is used to reduce external magnetic fields (e.g., the earth's) at least
100-fold.
• The Heisenberg uncertainty principle limits the achievable accuracy: ∆E∆t ≥ h/2π, E = hν,
therefore, ∆ν∆t ≥1, and, long observation time → small frequency uncertainty.
• Resonance linewidth (i.e., 1/Q) is inversely proportional to coherent observation time ∆t; ∆t is
limited by: 1.) when atom enters and leaves the apparatus, and 2.) when the atom stops
oscillating due to collisions with other atoms or with container walls (collisions disturb atom's
electronic structure).
• In microwave atomic standards, as atoms move with respect to the microwave source,
resonance frequency is shifted due to the Doppler effect (k•v); velocity distribution results in
"Doppler broadening"; the second-order Doppler shift (1/2 v2/c2) is due to relativistic time
dilation.
6-6

Rubidium Cell Frequency Standard
Energy level diagrams of 85Rb and 87Rb
F=2
F=3
F=2
363 MHz 52P1/2 816 MHz
F=1
795 nm 795 nm
F=3 F=2
3.045 GHz
52S1/2 6.834,682,608 GHz

F=2
F=1
85Rb 87Rb
6-7
Referring back to “Generalized Atomic Resonator” earlier in this chapter, the A and B levels for a
rubidium standard are the energy levels between the 87Rb F = 1 and F = 2 hyperfine levels of the 52S1/2
ground states. The difference in the energies of those two levels correspond to 6.834,682,608 GHz,
which is the atomic resonance frequency utilized in rubidium standards. Initially, about an equal number
of atoms are in the two ground states. A nonthermal distribution can be prepared by optical pumping to a
third level, the F = 1 excited state of 87Rb, as follows.
A 87Rb discharge lamp emits wavelengths that can pump both the A and B levels to the F = 1
excited state. However, due to the natural coincidence of optical resonance frequencies between 85Rb
and 87Rb, both of which correspond to a 795 nm wavelength, a 85Rb filter cell can filter the wavelength
corresponding to the transition from the F = 2 ground state of 87Rb. Assuming a perfect filter, the filtered
light pumps atoms only from the F = 1 ground state to the excited state, not from the F = 2 ground state.
The 87Rb atoms remain in the excited state for a short time only; they emit a photon and return (with about
equal probability) to either the A or B ground states. If they return to A, they will be pumped again to the
excited state, if they return to B, they will not be pumped. Eventually, all the atoms end up in the level B
(F = 2 ground) state. Light from the discharge lamp then passes through the absorption cell without
attenuation.
If 6.834,682,608 GHz microwave energy is applied to the atoms, the atoms in level B return to level
A (via a spin-flip; see “Hydrogen-Like Atoms,” earlier in this chapter), light is absorbed as the level A
atoms are optically pumped, and this light attenuation can be sensed with a photodetector. The
microwave frequency can, thereby, be controlled by the dip in the photodetector output, which occurs only
when the microwave frequency is at the 6.834,682,608 GHz atomic resonance frequency. When the
microwave frequency starts to drift, the change in photodetector output brings the frequency back to the
proper frequency, i.e., the microwave frequency is locked to the atomic resonance frequency.
The apparatus that can make all this happen is shown schematically on the next page.

Rubidium Cell Frequency Standard
Atomic resonator schematic diagram
Magnetic shield
“C-Field”
Absorption
cell
85Rb
87Rb Rb-87
+ buffer Light + buffer
lamp gas gas
Photo Detector
Filter cell output
Cell
Cavity
Frequency
input
6.834,685 GHz
rf Power supplies C-field
lamp for lamp, filter power
exciter and absorption supply
cell thermostats
6-8
Light from the 87Rb (rf discharge) lamp passes through the 85Rb filter cell and into the absorption
cell, which contains 87Rb gas plus a buffer gas. The 87Rb lamp emits wavelengths corresponding to
both the 87Rb F = 1 and F = 2 transitions. The 85Rb filter cell absorbs more of the F=2 transition light.
The light which passes through the filter cell is absorbed by the 87Rb F=1 state, the excited atoms
relax to both the F=1 and F=2 states, but the F=1 states are excited again; the F=2 state is
overpopulated; the applied 6.8 GHz microwave converts F=2 back to F=1, which provides more atoms
to absorb light. The microwave at the correct resonance frequency causes increased light absorption,
i.e., a (< 1%) dip, in the light detected by the photocell. The microwave frequency is locked to
photocell detection dip, thus the atomic transition frequency controls the microwave frequency, i.e.,
the frequency of the crystal oscillator.
The absorption cell contains Rb gas at ~10-6 torr and an inert buffer gas at ~1 torr. The Rb atom
oscillation lifetime is limited by collisions to ~10-2 s; the atomic resonance linewidth ~100 Hz; Q ~5 x
107. The buffer gas, a mixture of positive (e.g., N2) and negative (e.g., Ar) pressure-shift gases,
provides zero temperature coefficient at some temperature in the operating temperature range, and
confines Rb atoms to a small region to reduce wall-collisions and first order Doppler effects.

Cs Hyperfine Energy Levels
(F, mF)
(4,4)
(4,3)
(4,2)
(4,1)
Energy (Frequency) (GHz)

9.2 (4,0)
(4,-1)
(4,-2)
(4,-3)
(4,-4)
9.192,631,770 GHz
(3,-3)
(3,-2)
(3,-1)
0 (3,0)
(3,1)
(3,2)
(3,3)
Magnetic Field HO Energy states
at H = HO
6-9
Shown are the magnetic field dependence of the hyperfine energy levels in the ground state of
the cesium atom (nine in the upper state, seven in the lower). The magnetic field is plotted up to the
value HO. The solid arrow represents the “clock” transition; the dashed arrows depict the magnetic-
field-sensitive (Zeeman) transitions. F is the hyperfine quantum number, and mF is the magnetic
quantum number of the atom. The atomic resonance utilized is at 9,192,631,770 Hz - by
definition (of the second*), which corresponds to the (3,0) to (4,0) hyperfine transition, called the clock
transition. Referring back to “Generalized Atomic Resonator” earlier in this chapter, the (3,0) and (4,0)
levels are the A and B levels.
This (3,0) to (4,0) transition has a small quadratic dependence on magnetic field. The C-field
must be stable and uniform; high degree of shielding is required for ±1x10-13/gauss magnetic field
sensitivity (e.g., one laboratory Cs standard uses a triple magnetic shield).
It would be desirable to operate at zero magnetic field - all transitions would then behave as a
single transition, the signal would be 7X larger, but that would require < 10-8 gauss for errors < 1 x 10-
12. This is not feasible; a C-field must be applied. A 0.06 gauss C-field separates the sublevels by 40
kHz, and the the (3,0) and (4,0) levels, the level with the minimum magnetic field sensitivity are utilized
in making a Cs frequency standard. The way these levels are utilized is shown schematically on the
next two pages.
See “Magnetic Field Sensitivities of Atomic Clocks” later in this chapter.
* See “The Second” in chapter 8.

Cesium-Beam Frequency Standard
Atomic state selection Cs atom detection
R
NO SIGNAL C TO
TE
DE
S NO SIGNAL
Cs VAPOR, CONTAINING AN EQUAL

KIND 1 - ATOMS
AMOUNT OF THE TWO
(LOWER STATE) STATE SELECTED
KINDS OF Cs ATOMS
ATOMIC BEAM N
MICROWAVE
CAVITY MAGNET
S
R
MICROWAVE SIGNAL
C TO
(OF ATOMIC RESONANCE TE
DE
FREQUENCY)
ATOMIC BEAM MAXIMUM
N S SIGNAL
ATOMIC BEAM
MAGNET
SOURCE (STATE SELECTOR)
KIND 2 - ATOMS STATE SELECTED
N
VACUUM CHAMBER (UPPER STATE) ATOMIC BEAM
MICROWAVE
MAGNET
CAVITY
6-10
A Cs vapor is generated in the oven, the atoms are collimated, and the beam of Cs atoms are
directed to pass through a strongly diverging field of the “A” magnet, the “state-selector” magnet. The
force on an atom of magnetic moment µi in a magnetic field B is
Fi = -µi (∇•B)
Therefore, the atoms are deflected by amounts that depend on their magnetic moments, i.e., their
energy states. The atoms in the (3,0) state are deflected in a different direction than those in the (4,0)
state. In this manner, the two types of atoms can be physically separated. The (3,0) and (4,0) levels
are the A and B levels, referring back to “Generalized Atomic Resonator” earlier in this chapter,
The state selecting magnet "selects" one of the two atomic levels. The applied microwave at the
atomic resonance frequency causes a state change (a spin-flip; see “Hydrogen-Like Atoms,” earlier in
this chapter); the second magnet deflects those atoms to the detector which have undergone the state
change. The magnets' peak field is ~10 kgauss.
The atom detector is a ribbon or wire (e.g., W or Pt) at ~ 900°C. The Cs atoms are ionized, the
ions are collected, the current is amplified and fed back into feedback network. In this way, the
microwave frequency is locked to the frequency of maximum ion current, thus the atomic transition
frequency controls the microwave frequency, i.e., the frequency of the crystal oscillator.
Much less than 1% of the Cs atoms reach the detector in conventional Cs standards (hence
optical pumping's advantage - see “Optically Pumped Cs Standard” later in this chapter.)

Cesium-Beam Frequency Standard
Cs atomic resonator schematic diagram
DC C-FIELD
MAGNETIC SHIELD POWER SUPPLY
HOT WIRE
“C-FIELD” IONIZER
B-MAGNET
GETTER
Cs-BEAM
A-MAGNET CAVITY GETTER

ION
COLLECTOR
DETECTOR PUMP
VACUUM SIGNAL
ENVELOPE FREQUENCY
OVEN INPUT
OVEN PUMP
HEATER 9,192,631,770 Hz DETECTOR
HEATER POWER
POWER SUPPLY POWER
POWER SUPPLY
SUPPLY
SUPPLY
6-11
The A magnet selects one of the states, say the kind 2 state, which is deflected downward by
the magnets A and B. (Atoms in the kind 1 state would be deflected upward.) As long as the atoms
selected remain in the kind 2 state, none of them reach the hot-wire ionizer detector. However, if
microwaves of frequency 9.192,631,770 GHz are applied to the Cs beam, the atoms that absorb this
frequency undergo a transition and become a kind 1 atom. The B magnet deflects these atoms
upward to the ionizer. The electrical current generated in the ionizer is proportional to the number of
atoms that make the microwave induced transitions. Thus the microwave frequency can be locked to
the value that produces the maximum current.
The oven is at ~100°C, the Cs pressure in the oven is ~10-3 torr, the cavity is at ~10-9 torr; the
typical average atom speed is 100 m/s; the typical cavity length in commercial standards is 10 to 20
cm; the interaction time is ~1 to 2 x 10-3 s; the linewidth is ~0.5 to 1 kHz; the Q ~ 107; in standard
laboratories, the cavity length is ~4 meters and the Q ~ 108.
Cs standards are more accurate than Rb standards because the Cs atoms pass through a high-
vacuum region without collisions with buffer gas molecules or walls which cause frequency instabilities
in Rb standards. Instead of confining the atoms to a small cell, as is done in Rb standards, the Cs
atoms travel through a relatively long microwave cavity. Cs standards also use the Ramsey
separated field method which further narrows the atomic resonance linewidth - see the references for
details.

Atomic Hydrogen Energy Levels
(F, mF)
(1, +1)
(1, 0)
(1, -1)
W
1.42040…GHz
(0, 0)
H’
Ground state energy levels of atomic hydrogen as a function of magnetic field H’.
6-12
The above figure shows the relevant energy levels of a hydrogen atom in a magnetic field. The
hydrogen maser (H-maser) operates on the nearly field independent (at low fields) hyperfine transition
between the (1,0) and (0,0) states. The energy difference between these two states corresponds to a
1,420,405,752 Hz frequency (21 cm wavelength).
There are two types of hydrogen masers, active and passive. In the active maser, the 1.42 GHz
output signal is obtained directly from stimulated radiation by the H atoms (“maser” was originally an
abbreviation for microwave amplification by stimulated emission of radiation). In the passive H-maser,
as in all passive atomic standards, a microwave oscillator stimulates the desired atomic transitions
and a control loop provides feedback to tune a quartz oscillator to the frequency that maximizes the
transition rate.

Passive H-Maser Schematic Diagram
Teflon coated
storage bulb
Microwave
cavity Microwave
Microwave output
input
State
selector
Hydrogen
atoms
6-13
In a H-maser, molecular hydrogen is dissociated by an electrical discharge, and the resulting atomic
hydrogen is collimated into a beam. As is done in the Cs standard, an inhomogeneous magnetic field is
used to separate, i.e., “state-select,” atomic states. The maser cavity is coated with Teflon to minimize
atomic perturbations due to collisions with the cavity wall.
A frequency modulated 1.42…GHz microwave signal is passed through the cavity to interrogate the
atoms. As the microwaves induce atomic transitions, the atoms emit 1.42…GHz radiation. In active
masers, this 1.42…GHz is detected and is used directly. In passive standards, the applied modulated
microwave signal interacts with the atomic resonance line’s frequency. A phase shift of the carrier
frequency results, as well as an AM signal at the modulation frequency which is proportional to the offset of
the carrier frequency from the atomic resonance frequency. The AM is detected and is used to tune a quartz
oscillator so that the center of the carrier frequency coincides with the atomic resonance frequency.
Because of their superior resonance line Q (~109), H-masers have short term stabilities that are
superior to those of Cs and Rb standards. Stabilities of <10-15 at the noise floor and <10-14 at one day have
been reported. H-masers are the principal clocks used for radio astronomy, where the clocks must stay
coherent to a small fraction of a radio frequency cycle during the observation period (minutes to hours). The
large size of most hydrogen masers (typical active hydrogen masers are ten times the volume of
commercial Cs frequency standards) results from the 1.4 GHz microwave cavity and the surrounding layers
of magnetic shielding and temperature-stabilizing ovens. Hydrogen masers are not primary standards; they
typically exhibit frequency aging of 1 to 10 X 10-12 per year.
H. Hellwig, "Microwave Frequency and Time Standards," in E. A. Gerber and A. Ballato, Precision Frequency
S. R. Stein and J. R. Vig, "Frequency Standards for Communications," U. S. Army Laboratory Command Research
and Development Technical Report SLCET-TR-91-2 (Rev. 1), October 1991, AD-A243211. This report is a reprint of a
chapter "Communications Frequency Standards," in The Froehlich/Kent Encyclopedia of Telecommunications, Vol. 3,
pp. 445-500, Marcel Dekker, Inc., 1992.
T. E. Parker, “Environmental Factors and Hydrogen Maser Frequency Stability,” IEEE Trans. UFFC, vol. 46, pp.
745-751, 1999.

Atomic Resonator Instabilities
• Noise - due to the circuitry, crystal resonator, and atomic resonator. (See next
page.)
• Cavity pulling - microwave cavity is also a resonator; atoms and cavity behave
as two coupled oscillators; effect can be minimized by tuning the cavity to the
atomic resonance frequency, and by maximizing the atomic resonance Q to cavity
Q ratio.
• Collisions - cause frequency shifts and shortening of oscillation duration.
• Doppler effects - 1st order is classical, can be minimized by design; 2nd

order is relativistic; can be minimized by slowing the atoms via laser cooling - see
“Laser Cooling of Atoms” later in this chapter.
• Magnetic field - this is the only influence that directly affects the atomic
resonance frequency.
• Microwave spectrum - asymmetric frequency distribution causes frequency

pulling; can be made negligible through proper design.
• Environmental effects - magnetic field changes, temperature changes,

vibration, shock, radiation, atmospheric pressure changes, and He
permeation into Rb bulbs.
6-14
Frequency Control, Vol. 2, pp. 113-176, Academic Press, 1985

Noise in Atomic Frequency Standards
If the time constant for the atomic-to-crystal servo-loop is to, then
at τ < to, the crystal oscillator determines σy(τ), i.e., σy (τ) ~ τ-1. From
τ > to to the τ where the "flicker floor" begins, variations in the atomic
beam intensity (shot-noise) determine σy(τ), and σy(τ) ~ (iτ)-1/2, where
i = number of signal events per second. Shot noise within the
feedback loop shows up as white frequency noise (random walk of
phase). Shot noise is generally present in any electronic device
(vacuum tube, transistor, photodetector, etc.) where discrete particles
(electrons, atoms) move across a potential barrier in a random way.
In commercial standards, to ranges from 0.01 s for a small Rb

standard to 60 s for a high-performance Cs standard. In the regions
where σy(τ) varies as τ -1 and τ -1/2, σy(τ) ∝ (QSR)-1, where SR is the
signal-to-noise ratio, i.e., the higher the Q and the signal-to-noise
ratio, the better the short term stability (and the phase noise far from
the carrier, in the frequency domain).
6-15

Short-Term Stability of a Cs Standard
σy(τ)
10-9
STANDARD TUBE
10-10 τ0 = 1 SECOND
10-11
τ0 = 60 SECONDS*
10-12
OPTION 004 TUBE
10-13 τ0 = 1 SECOND
10-14
10-15 SYSTEM BW = 100 kHz * The 60 s time constant

provides better short-
term stability, but it is
10-16
10-3 10-2 10-1 100 101 102 103 104 105 106 usable only in benign
environments.
AVERAGING TIME τ (SECONDS)
6-16
-Hewlett-Packard 5061B Data Sheet (Pub. 5952-7912D), Hewlett-Packard Co.

Short-Term Stability of a Rb Standard
σy(τ) fL (LOOP BW) =
1
2πτ
fL = 100 Hz (STANDARD)
fL = 100 Hz
OTHERS OPTIONAL
10-9
RUBIDIUM - WORST CASE
10-10
fL = 0.01 Hz
fL = 1 Hz
10-11
VCXO
τ(seconds)
10-12
.001 .01 .1 1 10 100
6-17
Rubidium Atomic Frequency Standard, Model FE-5600A data sheet, Frequency Electronics, Inc., 55
Charles Lindbergh Blvd., Mitchel Field, NY 11553.

Acceleration Sensitivity of Atomic Standards
Multiplier
Multiplier Quartz
Quartz
Atomic
Atomic Crystal
Crystal
Resonator
Resonator Feedback
Feedback Oscillator
Oscillator
5 MHz Output
Let the servo loop time constant = t0, let the atomic standard's Γ = ΓA, and the
crystal oscillator’s (VCXO's) Γ = ΓO. Then,
• For fast acceleration changes (fvib >> 1/2πt0), ΓA = ΓO
• For slow acceleration changes, (fvib << 1/2πt0), ΓA << ΓO
• For fvib ≈ fmod, 2fmod, servo is confused, ΓA ≈ ΓO, plus a frequency offset
• For small fvib, (at Bessel function null), loss of lock, ΓA ≈ ΓO
6-18

P1193," Proc. 1992 IEEE Frequency Control Symposium, 763-781, 1992.

Atomic Standard Acceleration Effects
In Rb cell standards, high acceleration can cause ∆f due to light shift,
power shift, and servo effects:
• Location of molten Rb in the Rb lamp can shift
• Mechanical changes can deflect light beam
• Mechanical changes can cause rf power changes
In Cs beam standards, high acceleration can cause ∆f due to changes

in the atomic trajectory with respect to the tube & microwave cavity
structures:
• Vibration modulates the amplitude of the detected signal.
Worst when fvib = f mod.
• Beam to cavity position change causes cavity phase shift
effects
• Velocity distribution of Cs atoms can change
• Rocking effect can cause ∆f even when fvib < f mod
In H-masers, cavity deformation causes ∆f due to cavity pulling effect
6-19

P1193," Proc. 1992 IEEE Frequency Control Symposium, 763-781, 1992.

Magnetic Field Sensitivities of Atomic Clocks
Clock transition frequency ν = νo + CHHo2, where CH is the quadratic

Zeeman effect coefficient (which varies as 1/νo).
Atom Transition Frequency C-field* Shielding Sensitivity

(milligauss)** Factor* per gauss**
Rb v=6.8 GHz + (574 Hz/G2) Bo 2 250 5k 10-11
Cs v=9.2 GHz + (427 Hz/G2) Bo 2 60 50k 10-13
H v=1.4 GHz + (2750 Hz/G2) Bo 2 0.5 50k 10-13
* Typical values
** 1 gauss = 10-4 Tesla; Tesla is the SI unit of magnetic flux density.
6-20

Crystal’s Influences on Atomic Standard
• Short term stability - for averaging times less than the atomic-to-
crystal servo loop time constant, τL, the crystal oscillator determines
σy(τ).
• Loss of lock - caused by large phase excursions in t < τL (due to

shock, attitude change, vibration, thermal transient, radiation pulse).
At a Rb standard's 6.8 GHz, for a ∆f = 1 x 10-9 in 1s, as in a 2g
tipover in 1s, ∆φ ~ 7π. Control voltage sweeping during
reacquisition attempt can cause the phase and frequency to change
wildly.
• Maintenance or end of life - when crystal oscillator frequency

offset due to aging approaches EFC range (typically ~ 1 to 2 x 10-7).
• Long term stability - noise at second harmonic of modulation f

causes time varying ∆f's; this effect is significant only in the highest
stability (e.g., H and Hg) standards.
6-21

Optically Pumped Cs Standard
Fluorescence Detector
F=5
2
6 P 3/2 F = 4 453 MHz
• F=3
Tuned laser
diode pumps
Oven 852 nm
Spontaneous (~350 THz)
2 decays
Pump Detection 6 S 1/2 F=4
laser(s) laser 9.192631770 GHz
F=3
State selection State detection
Essential Elements of an Optically Pumped
Cesium Beam Standard Atomic Energy Levels
The proper atomic energy levels are populated by optical pumping with
a laser diode. This method provides superior utilization of Cs atoms,
and provides the potential advantages of: higher S/N, longer life, lower
weight, and the possibility of trading off size for accuracy. A miniature
Cs standard of 1 x 10-11 accuracy, and <<1 liter volume, i.e., about
100x higher accuracy than a Rb standard, in a smaller volume (but not
necessarily the same shape factor) seems possible.
6-22
The optical pumping technique manipulates the populations in the hyperfine levels of the ground
state by exciting transitions to higher principal quantum states with infrared, or higher frequency, light.
As shown above, the atoms in one hyperfine level are excited optically to a higher state from which
they decay spontaneously to both ground state hyperfine levels. The population of the hyperfine state
involved in the stimulated transition is rapidly depleted; the population of the second hyperfine level is
enhanced. Microwaves applied at the 9.192…GHz frequency can be locked to the atomic transition
frequency, as in a Cs beam standard.
Optical pumping has both advantages and disadvantages compared to magnetic state selection.
On the positive side, it can be accomplished in a more compact device and it can enhance the number
of atoms in the desired state rather than just rejecting the atoms in the undesired state. On the
negative side are increases in complexity, the difficulty of obtaining laser diodes at the proper
frequency and with suitable stability (as of 1999), and some additional performance-degrading
mechanisms.
Optical pumping can eliminate the need for state-selection magnets, and result in a larger
number of atoms contributing to the signal which results in a superior signal-to-noise ratio. In addition,
the spatial symmetry of the optical pumping reduces certain frequency shifts.
L. L. Lewis, “Miniature Optically Pumped Cesium Standards,” Proc. 45th Ann. Symp. on Frequency
Control, pp. 521-533, 1991.
P. J. Chantry, B. R. McAvoy, J. M. Zomp, and I. Liberman, “Towards A Miniature Laser-pumped

Cesium Cell Frequency Standard,” Proc. 1992 IEEE Int’l Frequency Control Symp., pp. 114-122,
1992.
P. J. Chantry, I. Liberman, et. al, “Miniature Laser Pumped Cesium Cell Atomic Clock Oscillator,”
Proc. 1996 IEEE Int’l Frequency Control Symp., pp. 1002-1010, 1996.

Laser Cooling of Atoms
1 Direction of motion
Light Light
Atom
2 3 4
Direction
of force
6-23
Laser cooling of atoms can create atoms that move very slowly (equivalent to temperatures of
microkelvins). This allows long observation times. The slow speed virtually eliminates Doppler shifts,
and the long observation times allow high accuracy determinations of atomic transition frequencies,
per the Heisenberg uncertainty principle, i.e., ∆E∆t ~ h and E = hν, so ∆ ν ~ 1/∆t. Laser cooling
promises frequency accuracies of parts in 1016. The explanation of laser cooling is as follows. The
numbers correspond to the numbers in the illustration above:
1. Consider two rays of light that bombard an atom. One ray travels in the same direction as the
atom,; the other moves in the opposite direction. The frequency of the light is slightly lower than the
frequency that the atom readily absorbs.
2. From the atom’s perspective, the ray moving in the same direction as the atom is shifted down in
frequency; the other ray is shifted up in frequency.
3. The atom is likely to absorb the high-frequency light but not the low. It is therefore pushed in a
direction opposite its motion and slows down.
4. The emission of the absorbed light pushes the atom in some random direction, but if the process is
repeated many times, the emission exerts no net force.
Chu, Steven, ”Laser Trapping of Neutral Particles," Scientific American, February 1992,
pp. 71-76.

Rubidium - Crystal Oscillator (RbXO)
Rubidium
Low-power
Frequency RbXO
Crystal
Standard Interface
Oscillator
(≈25W @ -550C)
6-24
The RbXO provides “the best of both worlds” - the long term stability of a Rb standard and the
low power of a crystal oscillator. Occasionally, power is applied to the Rb standard for a few minutes.
Upon warmup of the Rb standard, the RbXO interface syntonizes the crystal oscillator and cuts off
power to the Rb standard. When the crystal oscillator is an MCXO (see chapter 2), the MCXO digital
circuit can include the RbXO interface, and the average RbXO power consumption can be less than
100 mW.
J. R. Vig and V. J. Rosati, "The Rubidium-Crystal Oscillator Hybrid Development Program,"

Proceedings of the 16th Annual PTTI Applications and Planning Meeting, pp. 157-165, 1984, NTIS
Accession No. N85-29221/7.
W. J. Riley, Jr. and J. R. Vaccaro, "A Rubidium-Crystal Oscillator (RbXO)," IEEE Transactions on
1987.

RbXO Principle of Operation
Output
OCXO
RF
Rb Reference Sample
Power Source Rb Control
Reference Voltage
Control
Signals Tuning
Memory
OCXO and Tuning

Memory Power Source
6-25
The Rb reference is a miniature Rb frequency standard (RFS) that has been modified to control
an external crystal oscillator. The OCXO includes a digital tuning memory to hold the frequency
control voltage while the Rb reference is off. The OCXO is ON continually. At specified intervals, the
system applies power to the RFS. After the warmup of the RFS (a few minutes), the interface circuits
adjust the frequency of the OCXO to the RFS reference, then shut off the RFS. For portable
applications, the OCXO can be separable from the rest of the RbXO (see the next page) so that one
can operate with minimum size, weight, and power, and with nearly the accuracy of the RFS for the
duration of a mission. An MCXO can be used for even lower power consumption.
1987.

Rubidium Crystal Oscillator
Shown above is one implementation of the RbXO - in which the OCXO is detachable.
1987.

CHAPTER 7
Oscillator Comparisons and
Specifications
A listing of "Specifications and Standards Relevant to Frequency Control," appears in the back pages
of the Proceedings of the IEEE Frequency Control Symposium - see the proceedings ordering
information in chapter 10.
E. Hafner, "Specifications and Standards," in E. A. Gerber and A. Ballato, Precision Frequency



Oscillator Comparison
Quartz Oscillators Atomic Oscillators

TCXO MCXO OCXO Rubidium RbXO Cesium
Accuracy * 2 x 10-6 5 x 10-8 1 x 10-8 5 x 10-10 7 x 10-10 2 x 10-11
(per year)
Aging/Year 5 x 10-7 2 x 10-8 5 x 10-9 2 x 10-10 2 x 10-10 0
Temp. Stab. 5 x 10-7 3 x 10-8 1 x 10-9 3 x 10-10 5 x 10-10 2 x 10-11

(range, 0C) (-55 to +85) (-55 to +85) (-55 to +85) (-55 to +68) (-55 to +85) (-28 to +65)
Stability,σy(τ) 1 x 10-9 3 x 10-10 1 x 10-12 3 x 10-12 5 x 10-12 5 x 10-11

(τ = 1s)
Size 10 30 20-200 200-800 1,000 6,000
(cm3)
Warmup Time 0.03 0.03 4 3 3 20
(min) (to 1 x 10-6) (to 2 x 10-8) (to 1 x 10-8) (to 5 x 10-10) (to 5 x 10-10) (to 2 x 10-11)
Power (W) 0.04 0.04 0.6 20 0.65 30
(at lowest temp.)
Price (~$) 10 - 100 <1,000 200-2,000 2,000-8,000 <10,000 50,000
* Including environmental effects (note that the temperature ranges for Rb

and Cs are narrower than for quartz).
7-1
The specifications in this table are typical for oscillators made for a wide temperature range.
Laboratory standards that are designed to operate in a benign environment can perform better. This
is especially true for laboratory cesium standards, the accuracy of which can be >100x better than
what is shown above.

Clock Accuracy vs. Power Requirement*
(Goal of R&D is to move technologies toward the upper left)
10-12
Cs
Cs 1µs/day
1ms/year
10-10
Rb
Rb
Accuracy
10-8 1ms/day
OCXO 1s/year
OCXO
10-6 TCXO
TCXO
1s/day
10-4 XO
0.001 0.01 0.1 1 10 100
Power (W)
* Accuracy vs, size, and accuracy vs. cost have similar relationships
7-2
Commercially available frequency sources cover an accuracy range of several orders of

magnitude - from the simple XO to the cesium-beam frequency standard. As the accuracy increases,
so does the power requirement, size, and cost. Shown above is the relationship between accuracy
and power requirement. (Note that it is a log-log scale.) Accuracy versus cost would be a similar
relationship, ranging from about $1 for a simple XO to about $40,000 for a cesium standard (1997
prices).

Clock Accuracy vs. Power Requirement*
(Goal of R&D is to move technologies toward the upper left)
10-12
= in production Cs
Cs 1µs/day
Mini-Rb/Cs 1ms/year
10-10 = developmental
Accuracy
RbXO Rb
Rb
10-8 1ms/day
MCXO TMXO 1s/year
OCXO
OCXO
10-6
TCXO
TCXO 1s/day
10-4 XO
XO
0.001 0.01 0.1 1 10 100
Power (W)
* Accuracy vs, size, and accuracy vs. cost have similar relationships
7-3
Shown above is the relationship between accuracy and power requirement for the major classes
of frequency standards, and for some developmental standards. The accuracies shown are for wide
temperature range devices, and include environmental effects and one year of aging. Laboratory
standards, (e.g., the best available cesium standard) have higher accuracies.
Power availability often limits the available options. For example, small, handheld devices are
limited to XO, TCXO or MCXO.
The MCXO and RbXO were discussed earlier. One version of a miniature Cs standard is
discussed at “Optically Pumped Cs Standard,” in chapter 6. Other versions of miniature Rb and Cs
standards, e.g., using coherent population trapping (CPT), promise to allow small, low power atomic
standards in the future. The TMXO was the Tactical Miniature Crystal Oscillator, was an experimental
<20 cm3, low-power (300 mW @-40oC, vacuum insulated) OCXO. Similar size OCXOs are now
available commercially.
J. Messina, D. Bowman, R. Filler, R. Lindenmuth, V. Rosati, S. Schodowski, “Results Of Long Term

Testing Of Tactical Miniature Crystal Oscillators,” Proc. 43rd Annual Symposium on Frequency
Control, pp. 47-50, 1989.

Battery Life vs. Oscillator Power
1000
Operation at –30oC. The oscillators are assumed to
consume ½ of the battery capacity. 1 Year
Batteries (except alkaline) are derated for temperature. 6 Months
Days of Battery Life
100
1 Month
BA AA Li BA BA BA
55 Alk I on 58 50 55
10 67 al (Ce 00
, 93
,
90
,8
,9 ine ll P 12
.2
cm 3 ,8 ho 7 c 621 c 83
cm 3
1 Week
cm 3 ne m3 m3
,2 ), 1
5°C 35
cm 3
1 1 Day
0.1
0.001 0.01 0.1 1 10
RB Std
Rb/Cs
OCXO
XO
TCXO
MCXO
Small
Small
Mini
Oscillator Power (Watts)
7-4
Clock is assumed to run continuously. In most instances, even when the equipment is OFF, the clock continues to
operate.
Curves are approximations and should not be used for design. Although the curves show battery lifetime derated
for temperature, at very low current drains, most of the battery’s energy can be recovered. For example, a D-
sized Li SO2 battery at -40°C retains about 50% of its maximum capacity when providing 0.5A (1.5 W) , 65% of its
capacity when providing 0.1A (300 mW), and 85% of its capacity when providing 0.01A (30 mW).“ Conversely, at
at high current drain, the batteries deliver less than 50% of capacity. When these factors are taken into account,
the curves rise at the left (i.e., at low current drain) and curve downward at the right.
D-cell capacity is equivalent to the Li-Ion cell shown on the chart.
The AA alkaline battery WAS NOT derated for temperature because at –30°C the battery is virtually useless.
BA batteries are lithium/sulfur dioxide type.
The “mini Rb/Cs” power is the goal of a development program; no such atomic standard exists as of 2001.
--------------------------
David Linden, ed., Handbook of Batteries and Fuel Cells, McGraw-Hill Book Co., pp. 11-21 to 11-30, 1984.
The above chart was prepared April, 2001, by Yoonkee Kim and Vince Rosati, using Sigma Plot to generate the
graph, which was imported into Power Point by Vaughn Skidmore.

Short Term Stability Ranges of Various
Frequency Standards
-9
a rtz
-10 Qu
Log (σy(τ))
-11
i um
bi d
-12
Ru
-13
si um
Ce
-14
-15
Hydrogen Maser
-16
-3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
Log (τ), seconds

1 day1 month
7-5
The above graph is based on one prepared and provided by Richard Sydnor, Jet Propulsion
Laboratory, 1989.

Phase Instabilities of Various Frequency
Standards
-30
Ru
-40
bi
-50
di
Ce
um
10*Log(Sφ(f)) -60
si
-70
um
H
-80
yd
r
-90
og
-100
-110
en Quartz
M
as
-120
er
-130
-140
-150
-160
-3 -2 -1 0 1 2 3 4
Log (f)
Typical one-sided spectral density of phase deviation vs. offset
frequency, for various standards, calculated at 5 MHz. L(f) = ½ Sφ
7-6
The above graph is based on one prepared and provided by Richard Sydnor, Jet Propulsion
Laboratory, 1989.

Weaknesses and Wearout Mechanisms
Weaknesses Wearout Mechanisms
Quartz Aging None
Rad hardness
Rubidium Life Rubidium depletion

Power Buffer gas depletion
Weight Glass contaminants
Cesium Life Cesium supply depletion

Power Spent cesium gettering
Weight Ion pump capacity
Cost Electron multiplier
Temp. range
7-7
See also “Why Do Crystal Oscillators Fail?” on the next page.

The wearout mechanisms listed for Rb standards have become less of a problem in recent years
than they were when early GPS satellites failed due to the failure of on-board atomic clocks.
D. Ringer, H. Bethke, and M. Van Melle, “Rubidium and Cesium Frequency Standards - Status and
Performance on the GPS Program,” Proc. 16th Precise Time and Time Interval (PTTI) Applications
and Planning Meeting, pp. 127-141, 1986, NTIS Accession no. N85-29221/7.
M. J. Van Melle, “Cesium and Rubidium Frequency Standards Status and

Performance on the GPS Program,” Proc. 27th Precise Time and Time Interval (PTTI) Applications
and Planning Meeting, pp. 167-180, 1995, NASA Conference Publication 3334.
Lt. R. E. Bower, G. L. Dieter, and M. J. Van Melle,“SVN 20 End-Of-Life Frequency Standard Test
Results,” Proc. 29th Precise Time and Time Interval (PTTI) Applications and Planning Meeting, pp.
181- , 1997.
PTTI Proc. Availability information: <http://tycho.usno.navy.mil/ptti/orderform.html>

Why Do Crystal Oscillators Fail?
Crystal oscillators have no inherent failure mechanisms. Some have operated
for decades without failure. Oscillators do fail (go out of spec.) occasionally for reasons
such as:
• Poor workmanship & quality control - e.g., wires come loose at poor quality solder
joints, leaks into the enclosure, and random failure of components
• Frequency ages to outside the calibration range due to high aging plus insufficient
tuning range
• TCXO frequency vs. temperature characteristic degrades due to aging and the
“trim effect”.
• OCXO frequency vs. temperature characteristic degrades due to shift of oven set
point.
• Oscillation stops, or frequency shifts out of range or becomes noisy at certain
temperatures, due to activity dips
• Oscillation stops or frequency shifts out of range when exposed to ionizing
radiation - due to use of unswept quartz or poor choice of circuit components
• Oscillator noise exceeds specifications due to vibration induced noise
• Crystal breaks under shock due to insufficient surface finish
7-8
Aging, trim effect, oven instability effects, activity dips, vibration-induced noise effects, shock
effects, and other instabilities are discussed in chapter 4.

Oscillator Selection Considerations
• Frequency accuracy or reproducibility requirement
• Recalibration interval
• Environmental extremes
• Power availability - must it operate from batteries?
• Allowable warmup time
• Short term stability (phase noise) requirements
• Size and weight constraints
• Cost to be minimized - acquisition or life cycle cost
7-9
In relation to the last item, i.e., what cost is to be minimized, the initial acquisition cost or the life-
cycle cost, often, the cost of recalibration is far higher than the added cost of an oscillator that can
provide calibration-free life. A better oscillator may also allow simplification of the system's design.
The frequency of the oscillator can be another important consideration, because the choice can
have an impact on both cost and performance. Everything else being equal, an oscillator of standard
frequency, such as 5 MHz or 10 MHz, will cost less than one of an unusual frequency, such as
8.34289 MHz. Moreover, for thickness-shear crystals, such as the AT-cut and SC-cut, the lower the
frequency, the lower the aging. Since at frequencies much below 5 MHz thickness-shear crystals
become too large for economical manufacturing, and since all the highest stability oscillators use
thickness-shear crystals, the highest stability commercially available oscillator's frequency is 5 MHz.
Such oscillators will also have the lowest phase-noise capability close to the carrier. There are also
some excellent 10 MHz oscillators on the market; however, oscillators of much higher frequency than
10 MHz usually have higher aging rates and phase-noise levels close to the carrier than do 5 MHz
oscillators. For lowest phase-noise far from the carrier, where the signal-to-noise ratio in the
sustaining circuit determines the noise level, higher frequency crystals (e.g., 100 MHz) can provide
lower noise because such crystals can tolerate higher drive levels, thereby allowing higher signal
levels.
When resonator size is an issue, higher frequency resonators are generally smaller than lower
frequency resonators (of the same type). For example, a 100 MHz 5th overtone resonator is
substantially smaller than a 5 MHz 5th overtone unit. On the other hand, a 32 kHz watch crystal is
much smaller than either.

Crystal Oscillator Specification: MIL-PRF-55310
MIL-PRF-55310D
15 March 1998
-----------------------
SUPERSEDING
MIL-0-55310C
15 Mar 1994
PERFORMANCE SPECIFICATION
OSCILLATOR, CRYSTAL CONTROLLED
GENERAL SPECIFICATION FOR
This specification is approved for use by all Departments and Agencies of the Department of Defense.
1. SCOPE
1.1 Statement of scope. This specification covers the general requirements for
quartz crystal oscillators used in electronic equipment.
----------------------------------
Full text version is available via a link from <http:\\www.ieee.org/uffc/fc>
7-10
This document is a good guide for preparing oscillator specifications, whether it be an oscillator
for a military or a civilian application, and whether it be a quartz or atomic oscillator.
The complete MIL-PRF-55310 is available on the Internet via a link from <http:\\www.ieee.org/uffc/fc>,
or directly, from <http://www.dscc.dla.mil/Programs/MilSpec/ListDocs.asp?BasicDoc=MIL-PRF-
55310>
Copies of MIL-PRF-55310 are also available by mail from: Military Specifications and Standards,
Bldg. 4D, 700 Robbins Avenue, Philadelphia, PA 19111-5094, USA. Customer Service telephone:
(215) 697-2667/2179; Telephone Order Entry System (requires a touch tone telephone and a
customer number): (215) 697-1187 thru 1195.

CHAPTER 8
Time and Timekeeping
Proceedings of the IEEE, Special Issue on Time and Frequency, J. Jespersen & D. W. Hanson, ed's., Vol.
79, No. 7, July 1991.
P. Kartaschoff, Frequency and Time, Academic Press, 1978.
G. M. R. Winkler, "Timekeeping and Its Applications," Advances in Electronics and Electron Physics, Vol. 44,
pp. 34-97, 1977.
G. Kamas, & M. A. Lombardi, Time & Frequency Users' Manual, NIST Special Publication 559, revised 1990;
Time and Frequency Division, NIST, 325 Broadway, Boulder, Colorado, 80303.
Characterization of Clocks and Oscillators, ed. by D. B. Sullivan, et al., NIST Technical Note 1337, 1990,
Time and Frequency Division, NIST, 325 Broadway, Boulder, Colorado, 80303.
D. W. Allan, "Frequency and Time Coordination, Comparison, and Dissemination," in E. A. Gerber and A.
H. Hellwig, K. M. Evenson and D. J. Wineland, "Time, Frequency and Physical Measurement," Physics
Today, pp. 23-30, December 1978.
“The Science of Timekeeping,” Application Note 1289, Hewlett Packard Co., 1997.
"Fundamentals of Time and Frequency Standards," Application Note 52-1, Hewlett Packard Co., September
1986.
"Timekeeping and Frequency Calibration,” Application Note 52-2, Hewlett Packard Co., November 1986.
HP application notes can be found at: <http://www.tmo.hp.com/tmo/Notes/English/>
“QUARTZ CRYSTAL RESONATORS
R. Morris, Time'sAND OSCILLATORS
Arrows - Scientific Attitudes Toward Time, Simon and Schuster, NY, 1985.
What Is Time?
• "What, then, is time? If no one asks me, I know; if I wish to explain to
him who asks, I know not." --- Saint Augustine, circa 400 A.D.
• The question, both a philosophical and a scientific one, has no entirely
satisfactory answer. "Time is what a clock measures." "It defines the
temporal order of events." "It is an element in the four-dimensional
geometry of space-time.” “It is nature’s way of making sure that
everything doesn't happen at once.”
• Why are there "arrows" of time? The arrows are: entropy,
electromagnetic waves, expansion of the universe, k-meson decay,
and psychological. Does time have a beginning and an end? (Big
bang; no more "events", eventually.)
• The unit of time, the second, is one of the seven base units in the
International System of Units (SI units)*. Since time is the quantity
that can be measured with the highest accuracy, it plays a central role
in metrology.
8-1
None of the fundamental laws of physics distinguish between the forward and backward
directions of time; e.g., the laws of mechanics are valid when time is reversed. If the collision of two
particles is recorded, and the recording is played forward and backward, there is no way to tell the
“correct” direction. The particles follow the laws of physics in either direction. However,
five ways have been found in which the directions of time, the “arrows of time,” can be distinguished.
The most important arrow is the second law of thermodynamics, according to which the past and
future look different; there will be more entropy tomorrow, there was less entropy yesterday. (The
second law is not a fundamental law; it is a statistical law; it cannot be applied to a single particle or a
small number of particles.)
In the laws of physics, there is no “present moment.” The laws deal only with time intervals.
The “flow of time” cannot be measured. “At best, one can say only that time moves onward at the rate
of one second per second, which is about as meaningful as defining the word ‘cat’ by saying, ‘A cat is
a cat.’ “. **
* See “Units of Measurement Having Special Names in the International System of Units (SI)” and the
following page later in this chapter.
** R. Morris, Time's Arrows - Scientific Attitudes Toward Time, Simon and Schuster, NY, 1985.
K. Lippincott, et. al, The Story of Time, Merrell Holberton Publishers in association with National
Maritime Museum (UK), 1999.

Dictionary Definition of “Time”
(From The Random House Dictionary of the English Language ©1987)
tim e (tim ), n ., a d j., u , tim e d , tim ~ in g . — n . 1 . th e s y s te m th e p itc h e r's tim e a t b a t. 2 1. tim e s , u s e d a s a d o e s h to m a rk o r o b s e rv e th e te m p o . c .to p e rfo rm
o f th o s e s e q ue n tial re la tio n s th a t a n y ev e n t h a s to a n y m u ltip lic a tiv e w o rd in p h ra s a l c o m b in a tio n s e x p re s s in g rh y th m ic m o v e m e n ts in u n is o n . 4 1 . k ill tim e , to o c c u p y
o th e r, a s p a s t, p re s e n t, or futu re , in de fin ite a n d c o n - h o w m a n y in s tan c e s o f a q u a n tity o r fa c to r a re ta k e n o n e s e lf w ith s o m e a c tiv ity to m a k e tim e p a s s q u ic k ly :
tin u o u s d u ra tio n re g a rd e d a s th a t in w h ic h e v e n ts s u c - to g e th e r: T w o g o e s in to s ix th re e tim e s , fiv e tim e s W h ile I w a s w a itin g , I k ille d tim e c o u ntin g th e c a rs o n
c e e d o n e a n o th er. 2 . d u ra tio n re g a rd e d a s b e lo n g in g to fa s te r 2 2 . D ra m a . o n e o f th e th re e u n itie s . C f. u n ity th e fre ig h t tra in s . 42 . m a k e tim e, a . to m o v e q u ic k ly ,
th e p res e n t life a s d is tin c t fro m th e life to c o m e or fro m (d e f. 8 ) 23 . P ro s . a u n it o r a g ro u p o f u n its in th e e s p . in a n a tte m p t to re c o v e r lo s t tim e . b . to tra v e l a t a
e te rn ity ; finite d u ra tio n . 3 . (s o m e tim e s c a p .) a s y s te m m e a s u re m e n t o f m e te r. 2 4. M u s ic a . te m p o , re la tiv e p a rtic u la r s p e e d . 4 3 . m a k e tim e w ith , S la n g . to p u rs u e
o r m e th o d o f m e a s urin g o r re c k o n in g th e p a s s a g e o f ra p id ity o f m o v e m e n t. b . th e m e tric a l d u ra tio n of a n o te o r ta k e a s a s e x u a l p a rtn e r. 4 4 . m a n y a tim e , a g a in
tim e : m e a n tim e ; a p p a ren t tim e ; G re e n w ic h T im e . 4 . a o r re s t. c . p ro p e r o r c h a ra c te ris tic te m p o . d . th e g e n e ra l a n d a g a in ; fre q u e n tly : M a n y a tim e th e y d id n 't h a v e
lim ite d p e riod o r in te rv a l, a s b e tw e e n tw o s u c c e s s iv e m o v e m e n t o f a p a rtic u la r k in d o f m u s ic a l c o m p o s itio n e n o u g h to e a t a n d w e n t to b e d h u n g ry . 4 5 . m a rk tim e ,
e v e n ts : a lo n g tim e . 5 . a p a rtic u lar p e riod c o n s id e re d w ith re fe re n c e to its rh y th m , m e tric a l s tru c tu re , a n d a . to s u s p e n d p ro g re s s te m p o ra rily , a s to a w a it
a s d is tin c t fro m o th e r p e rio d s : Y o uth is th e b e s t tim e o f te m p o . e . th e m ov e m e n t o f a d a n c e o r th e lik e to m us ic d e v e lo p m e n ts ; fa il to a d v a n c e . b . M il. to m o v e the fe e t
life . 6. O fte n , tim e s . a . a p e rio d in th e h is to ry o f th e s o a rra ng e d : w a ltz tim e . 2 5 . M il. ra te o f m a rc h in g , a lte rn a te ly a s in m a rc h in g , b u t w ith o u t ad v a nc in g . 4 6 .
w o rld , o r c o n tem p o ra ry w ith th e life o r a c tiv itie s o f a c a lc u la te d o n th e n u m b e r o f p a c e s ta k en p e r m in u te : o n o n e 's o w n tim e , d u rin g o ne 's fre e tim e ; w ith o u t
n o ta b le p e rs o n : p re h is to ric tim e s ; in L in c o ln 's tim e . b . d o u b le tim e ; q u ic k tim e . 2 6 . M a n e g e . e a c h c o m p le te d p a y m e n t: H e w o rk e d o u t m o re effic ie n t p ro d u c tio n
th e p e rio d o r e ra n o w o r p re v io u s ly p re s e n t: a s ig n o f a c tio n o r m o v e m e n t o f th e h o rs e . 2 7 . a g a in s t tim e , in m e th o d s o n h is o w n tim e . 4 7 . o n tim e , a . at th e
th e tim e s ; H o w tim e s h a v e c h a n g e d ! c . a p e rio d a n e ffo rt to fin is h s o m eth in g w ith in a lim ite d p e rio d : W e s p e c ifie d tim e ; p u n c tu a B y . b . to b e p a id fo r w ith in a
c o n s id e re d w ith re fe re n c e to it’s e v e n ts o r p re v a ilin g w ork e d a g a in s t tim e to g e t o u t th e n e w s p ap e r. 2 8 . d e s ig n a te d p e rio d o f tim e , a s in in s ta llm e n ts : M a n y
c o n d itio n s , te n d e n c ie s Id e a s , e tc .: h a rd tim e s ; a tim e o f a h e a d o f tim e , b e fo re th e tim e d u e ; e a rly : T h e b u ild in g p e o p le a re n e v e r o u t o f de b t b ec a u s e th e y b u y
w a r. 7 . a p re s c rib e d or a llo tte d p e rio d , a s o f o ne 's life , w as c o m p le te d a h e a d o f tim e . 2 9 . a t o n e tim e , a . e v e ry th in g o n tim e . 4 8 . o u t o f tim e , n o t in th e p ro p e r
fo r p a y m e n t o f a d e b t, e tc . 8 . th e e n d o f a p re s c rib e d o r o n c e ; in a fo rm e r tim e : A t o n e tim e th e y o w n e d a rh y th m : H is s in g ing w as o u t o f tim e w ith the m u s ic 4 9 .
a llo tte d p e rio d , a s o f o n e's life o r a p re g n a n c y : H is tim e re s ta u ra n t. b . a t th e s a m e tim e ; a t o n c e : T h e y a ll trie d P a s s th e tim e o f d a y . to c o n v e rs e b rie fly w ith o r g re e t
h a d c o m e , b u t th e re w a s n o o n e le ft to m o u rn o v e r to ta lk a t o n e tim e . 3 0 . a t th e s a m e tim e , n e v e rth e le s s ; s o m e o n e : T h e w o m e n w o u ld s to p in th e m a rk e t to
h im . W h e n h e r tim e c a m e , h e r h u s b a n d a c c o m p a n ie d y e t: I'd lik e to try it b u t a t th e s a m e tim e I'm a little p a s s th e tim e o f d a y . 5 0 . ta k e o n e 's tim e , to b e s lo w
h e r to th e d e liv e ry ro o m . 9 . a pe rio d w ith re fere n c e to a fra id 31 . a t tim e s a t in te rv a ls ; o c c a s io n a lly : A t tim e s o r le is u re ly ; d a w d le : S p e e d w a s im p o rta n t h e re , b u t h e
p e rs o n a l e x p e rie n c e o f a s p e c ifie d k in d : to h a v e a g o o d th e c ity be c o m e s in to le ra b le . 32 . b e a t s o m e o n e 's ju s t to o k h is tim e . 5 1 . tim e a fte r tim e . a g a in a n d a g a in ;
tim e ; a h ot tim e in th e o ld to w n to n ig h t. 1 0 . a p e rio d o f tim e . S la n g . to c o m p e te fo r o r w in a p e rs o n b e in g re p e a te d ly ; o fte n : I'v e to ld h im tim e a fte r tim e n o t to
w o rk o f a n e m p lo y e e , o r th e p a y fo r it; w o rk in g h o u rs o r d a te d o r c o u rte d b y a n oth e r; p re v a il o v e r a riv a l: H e s la m th e d o o r. 5 2 . tim e a n d tim e ag a in , re p e a te d ly ;
d a y s o r a n h o u rly or d a ily p a y ra te . 1 1 . In fo rm a l. a te rm a c c u s e d m e , h is o w n b ro th e r, o f try in g to b e a t h is tim e . o fte n : T im e a n d tim e a g a in I w a rn e d h e r to s to p
o f e n fo rc e d d u ty o r im p ris o n m e n t: to s e rv e tim e in th e 3 3 . b e h in d th e tim e s , o ld -fa s h io n e d ; d a ted : T h e s e s m o k in g . A ls o , tim e a n d a g a in . 5 3 . tim e o f llfe ,
a rm y , d o tim e in p ris o n . 1 2 . th e p e rio d n e c e s s a ry fo r o r a ttitu d e s a re b e h in d the tim e s . 3 4 . fo r th e tim e b e in g , (o n e 's ) a g e : A t y o u r tim e o f life y o u m u s t b e c a re fu l n o t
o c c u p ie d b y s o m e th in g : T h e tim e o f th e b a s e b a II g a m e te m p o ra rily ; fo r th e p re s e n t: L e t's fo rg e t a b o u t it fo r to o v e rd o th in g s . 5 4 . tim e o f o n e 's life , In fo rm a l. a n
w a s tw o h o u rs a n d tw o m in u te s . T h e b u s ta k e s to o th e tim e b e in g . 3 5 . fro m tim e to tim e , o n oc c a s io n ; e x tre m e ly e n jo y a b le e x p e rie n c e : T h e y h a d th e tim e o f
m u c h tim e , s o I'll ta k e a p la n e . 1 3 . Ie is u re tim e ; o c c a s io n ally ; a t in te rv a ls : S h e c o m e s to s e e u s fro m th e ir liv e s o n th e ir trip to E u ro p e .. 5 5 . o f. p e rta in ing to ,
s u ffic ien t o r s p a re tim e : to h a v e tim e fo r a v a c a tio n ; I tim e to tim e. 3 6 . g a in tim e , to p o s tp o n e in o rd e r to o r s h o w in g th e p a s s a g e o f tim e 5 6 . (o f a n e x p lo s iv e
h a v e n o tim e to s to p n o w . 1 4 . a p a rtic u la r o r d e fin ite m a k e p re p a ra tio n s o r g a in a n a d v a n ta g e d e la y th e d e v ic e ) c o n ta in ing a c lo c k s o th a t it w ill d e to n a te a t th e
p o in t in tim e, a s in d ic a te d b y a c lo c k : W h a t tim e is it? o u tc o m e o f: H e h o p e d to g a in tim e b y p u ttin g o ff d e s ire d m o m e n t a tim e b o m b 57 . C o m . p a y a b le a t a
1 5 . a p a rtic u la r p a rt o f a y e a r, d a y , e tc .; s e a s o n o r s ig n ing th e p a p e rs fo r a fe w d a y s m o re . 3 7 . in g o o d s ta te d p e rio d o f tim e a fte r p re s e n tm e n t: tim e d ra fts o r
p e rio d : It's tim e fo r lu n c h . 1 6 . a n a p p o in te d , fit, d u e , o r tim e , a . a t th e rig h t tim e ; o n tim e , p u n c tu a lly . b . in a d - n o te s . 5 8 . o f o r p e rta in in g to p u rc h a s e s o n th e
p ro p e r in s ta n t o r p e rio d : A tim e fo r s o w ing ; th e tim e v a n c e of th e rig h t tim e ; e a rly : W e a rriv e d a t th e a p - in s tallm e n t p la n , or w ith p a y m e n t po s tpo n e d . v .t. 5 9 . to
w h e n th e s u n c ro s s e s the m e rid ia n T h ere is A tim e fo r p o in te d s po t in g o o d tim e . 38 . in n o tim e , in a v e ry m e a s u re o r re c o rd th e s p e e d, d u ra tio n, o r ra te o f: to
e v e ry th in g . 1 7 . th e p a rtic u la r p oin t in tim e w h e n a n b rie f tim e ; a lm o s t a t o n c e : W o rk in g to g e th e r, th e y tim e a ra c e . 6 0 . to fix th e d ura tio n o f: T h e p ro c to r tim e d
e v e n t Is s c h ed u le d to tak e p la c e : tra in tim e ; c u rta in c le a n e d th e e ntire h ou s e in n o tim e . 3 9 . in tim e , a . th e te s t a t 1 6 m in u te s . 6 1. to fix th e in terv a l b e tw e e n
tim e . 1 8 . a n in d e fin ite , fre q u e n tly p ro lo n g e d p e rio d o r e a rly e n o u gh : to c o m e in tim e fo r d in n er. b . in th e fu - (a c tio n s , e v e n ts , e tc .): T h e y tim e d th e ir s tro k e s a t s ix
d u ra tio n in th e fu tu re : T im e w ill te ll if w h a t w e h a v e tu re ; e v e n tu a lly : In tim e h e 'll s e e w h a t is rig h t. c . in th e p e r m in u te . 6 2 . to re g ula te (a tra in , c lo c k e tc .) a s to
d o n e h e re to d a y w a s rig h t. 1 9 . th e rig h t o c c a s io n o r c o rre c t rh y th m o r te m p o : T h e re w o u ld a lw a y s b e a t tim e . 6 3 . to a p po in t o r c h o o s e th e m o m e n t o r o c c a s lo n
o p p o rtu n ity : to w a tc h o n e 's tim e . 2 0 . e a c h o c c a s io n o f le a s t o n e c h ild w h o c o u ld n't p la y in tim e w ith th e m u s ic fo r; s c h e d u le : H e tim e d th e a tta c k p erfe c tly . — v .i. 6 4 .
a re c u rrin g a c tio n o r e v e n t: to d o a th in g fiv e tim e s , It's 4 0 . k ee p tim e , a . to re c o rd tim e , a s a w a tc h o r c lo c k to k e e p tim e ; s o u n d o r m o v e in u n is o n . [b e f. 9 0 0 ; (n .)
8-2

The Second
• The SI unit of time is the second (symbol s).
• The second was defined, by international agreement, in October,

1967, at the XIII General Conference of Weights and Measures.
• The second is "the duration of 9,192,631,770 periods of the

radiation corresponding to the transition between the two
hyperfine levels of the ground state of the cesium atom 133."
• Prior to 1967, the unit of time was based on astronomical

observations; the second was defined in terms of ephemeris time,
i.e., as "1/31,556,925.9747 of the tropical year..."
• The unit of frequency is defined as the hertz (symbol Hz). One

hertz equals the repetitive occurrence of one "event" per second.
8-3
More information about the SI units can be found later in this chapter.

Frequency and Time
1
f =
τ
where f = frequency (= number of “events” per unit time), and
τ = period (= time between “events”)
Total number of events
Accumulate d clock time =
Number of events per unit of time
3 rotations of the earth

Example : = 3 days
1 rotation/d ay
Frequency source + counting mechanism → clock
Examples of frequency sources: the rotating earth, pendulum,

quartz crystal oscillator, and atomic frequency standard.
8-4
The essential parts of a clock are a frequency source (oscillator) and a counting device.

Typical Clock System
Counting
Mechanism
Frequency Setting Display or
Source Mechanism Time code
Synchronization
Mechanism
t = t0 + Σ∆τ
Where t is the time output, t0 is the initial setting,
and ∆τ is the time interval being counted.
8-5
A clock may or may not have a display. In many consumer applications, clocks display the time
of day. In many other applications, clocks are used internally only; their output is typically a one-
pulse-per-second (1 pps) or a time code signal which are used for sequencing or time-tagging events
(see “One Pulse-Per-Second Timing Signal” and “BCD Time Code” later in this chapter.

Evolution of Clock Technologies
• Sundials, and continuous flow of:
• Water (clepsydra)
• Sand (hour glass)
• Falling weights, with frictional control of rate
• Vibrating, but non-resonant motion - escapement

mechanisms: falling weight applies torque through train of
wheels; rate control depends on moments of inertia,
friction and torque; period is the time it takes to move from
one angular position to another.
• Resonant control
• Mechanical: pendulum, hairspring and balance wheel
• Mechanical, electrically driven: tuning fork, quartz
resonator
• Atomic and molecular
8-6
H. Tait, Clocks and Watches, British Museum Publications, 1983
W. A. Marrison, “The Evolution of the Quartz Crystal Clock,” The Bell System Technical Journal, Vol.
XXVII, pp. 510-588, 1948. Reprinted at <http://www.ieee.org/uffc/fc>
<http://www.horology.com/>

Progress in Timekeeping
Time Period Clock/Milestone Accuracy Per Day
4th millennium B.C. Day & night divided into 12 equal hours
Up to 1280 A.D. Sundials, water clocks (clepsydrae) ~1 h
~1280 A.D. Mechanical clock invented- assembly time for prayer ~30 to 60 min
was first regular use
14th century Invention of the escapement; clockmaking becomes ~15 to 30 min
a major industry
~1345 Hour divided into minutes and seconds
15th century Clock time used to regulate people’s lives (work hours) ~2 min
16th century Time’s impact on science becomes significant ~1 min
(Galileo times physical events, e.g., free-fall)
1656 First pendulum clock (Huygens) ~100 s
18th century Temperature compensated pendulum clocks 1 to 10 s
19th century Electrically driven free-pendulum clocks 10-2 to 10-1 s
~1910 to 1920 Wrist watches become widely available
1920 to 1934 Electrically driven tuning forks 10-3 to 10-2 s
1921 to present Quartz crystal clocks (and watches. Since ~1971) 10-5 to 10-1 s
1949 to present Atomic clocks 10-9 to 10-4 s
8-7
H. Tait, Clocks and Watches, British Museum Publications, 1983
W. A. Marrison, “The Evolution of the Quartz Crystal Clock,” The Bell System Technical Journal, Vol.
XXVII, pp. 510-588, 1948. Reprinted at <http://www.ieee.org/uffc/fc>
<http://www.horology.com/>

Clock Errors
t t
T(t) = T0 + ∫ R(t)dt + ε(t) = T0 + (R0t + 1/2At2 + …) + ∫ Ei(t)dt +ε(t)

0 0
Where,
T(t) = time difference between two clocks at time t after synchronization
T0 = synchronization error at t = 0
R(t) = the rate (i.e., fractional frequency) difference between the two clocks
under comparison; R(t) = R0 + At + …Ei(t)
ε(t) = error due to random fluctuations = τσy(τ)
R0 = R(t) at t = 0
A = aging term (higher order terms are included if the aging is not linear)
Ei(t) = rate difference due to environmental effects (temperature, etc.)
--------------
Example: If a watch is set to within 0.5 seconds of a time tone (T0 = 0.5 s), and the
watch initially gains 2 s/week (R0 = 2 s/week), and the watch rate ages -0.1 s per
week2, (A = -0.1 s/week2), then after 10 weeks (and assuming Ei(t) = 0):
T (10 weeks) = 0.5 (2 x 10) + 1/2(-0.1 x (10)2) = 15.5 seconds.
8-8
No clock can ever keep perfect time because all oscillators exhibit random and systematic errors
(as discussed in chapters 4 and 6), clocks cannot be set perfectly (e.g., due to noise), and time is a
function of position and motion (relativistic effects are discussed later in this chapter).

Frequency Error vs. Time Error
Frequency
Time Error
Frequency
Time Error
fr fr
t t t t
Time Error
Frequency
Frequency
Time Error
fr
fr
t t t t
1 2 3 1 2 3 1 2 3
1 2 3
fr = reference (i.e., the “correct”) frequency
8-9
ft = fo + Afrt
where ft = f at time t, fo = initial f at time zero, A is the aging (e.g., 5 x 10-10 per day), and fr = the
reference frequency. The time error T(t) is given by the equation two pages back. Neglecting the
noise term,
T(t) = T0 + ([fo- fr]/fr] t + 1/2At2)
The plot of To vs. t is a parabola (for finite A), for which the vertical displacement is To. The above
illustration shows the relationship between frequency and time errors, for positive and negative
constant frequency errors, and for positive and negative aging.
"Timekeeping and Frequency Calibration," Hewlett-Packard Co. Application Note 52-1

<http://www.tmo.hp.com/tmo/Notes/English/>

Clock Error vs. Resynchronization Interval
25 Aging/Day = 5 x 10-10
Time Error (ms)

Temp Stability = 2 x 10-8
20 Resync Interval = 4 days
15
10
0 10 20 30 40 50 60 70 80 90 100
Days Since Calibration
TCXO OCXO MCXO RbXO

Avg. Temp. Stab. 1 x 10-6 2 x 10-8 2 x 10-8 2 x 10-8
Aging/Day 1 x 10-8 1 x 10-10 5 x 10-11 5 x 10-13

Resynch Interval*
(A/J & security) 10 min 4 hrs 6 hours 4 days 6 hours 4 days 6 hours 4 days
Recalibr. Interval * None

(Maintenance cost) 10 yrs 80 days 50 years 1.5 yrs 94 yrs 3 yrs needed 300 yrs
* Calculated for an accuracy requirement of 25 milliseconds. Many modern systems need much better.
8-10
In the graph above is an example that illustrates how the time error of a clock grows with days
since synchronization and syntonization (i.e., calibration). The clock is assumed to be in a system that
requires a 25 millisecond accuracy, the clock is assumed to have an average frequency offset due to
temperature of 2 x 10-8, an aging rate of 5 x 10-10 per day, and negligible other error sources. Upon
calibration, the clock is assumed to have zero frequency and time error.
The clock is assumed to be resynchronized every four days. During the initial period, the time
error is caused almost entirely by the frequency error due to temperature, i.e., 2 x 10-8 in the example.
Aging, however, adds 5 x 10-10 to the frequency error every day, so, after 40 days, the frequency error
due to aging equals that due to temperature. After 40 days, aging is the dominant cause of the
frequency error. The time error increases more and more in each four day resynchronization interval.
Eventually, after about 100 days, the time error at the end of the the four days reaches the 25 ms limit.
At that point, either the resynchronization interval must be made shorter, or the clock must be
recalibrated, i.e., its oscillator’s frequency must be readjusted to the correct value.
The table shows other examples - for a TCXO, OCXO, MCXO and RbXO. For each, two
resynchronization and recalibration scenarios are shown that result in a 25 ms time error at the end of
the recalibration and last resynchronization intervals. Whereas a TCXO in this application needs to be
resynchronized/recalibrated frequently, a Rb frequency standard or RbXO can maintain the required
accuracy indefinitely.

Time Error vs. Elapsed Time
10 s
-4
Aging rates are per day 10 -6 -8
1
X -5
10 -7 10
G 10 1
X 10 1
X
1s
IN X G X G
1
AG 1 IN G IN -9
-5
IN
G AG IN AG 10
10 AG
1X AG X
Accumulated Time Error
100 ms 1
ET G
FS IN
-
OF -6 10
X1
0 AG X
1 1
10 ms ET G
FS IN 10 -
OF 10
0
-7
AG X
1 X1 1
1 ms ET G 11
-
FS IN 10
OF AG X
0
-8 1
1 X1 IN
G 12
100 µs
ET AG
FS
OF -9 - 10 -11
10 X 10 X 10
1X T1 T1
10 µs ET FS
E FS
E
FS OF OF
OF
1 µs
10 30 2 4 8 16 2 3 4 5 61 2 3 1 2 4
minutes 1 1 1
hour day week MONTH YEAR
Elapsed Time
8-11
To Estimate The Accumulated Time Error*:
1. Estimate the initial frequency offset plus the average expected offsets due to temperature
and other environmental effects.
2. Find the time error caused by the sum of the offsets.
3. Find the time error caused by the oscillator’s specified aging rate
4. Add the results of 2 and 3 to estimate the total time error
* In the nomograph, the contribution due to aging assumes no resynchronization. If there is

resynchronization, as in the examples on the previous page, then the offset due to aging must first be
calculated and added to the other offsets. For example, in the TCXO example on the previous page,
in the four hour resynchronization interval, the offset due to temperature causes a time error of ~14
ms. The aging after 80 days results in a frequency offset of 8 x 10-7 which results in another ~11 ms
error, so, after 80 days of aging, the combined offsets due to aging and temperature result in ~25 ms
time error four hours after resynchronization.

Synchronization, Syntonization
Clocks are synchronized* when
- They are in agreement as to the time, or
- Output signals or data streams agree in phase, or
- Sync patterns are in alignment
Clocks/oscillators are syntonized** when

- Oscillators have the “same” frequency (the output
signals need not be in phase)
- Clocks run at the same rate (the internal oscillators
need not be of the same frequency)
* Chron Î time
** Tone Î frequency
8-12
Wristwatch examples: a wristwatch is compared to a time standard, e.g., to the time at

www.time.gov, and is found to be 10 seconds off. When the time shown by the wristwatch is adjusted
to that of the standard, the watch is synchronized. When the wristwatch is compared again a month
later, it is found to be off by 30 seconds. The wristwatch is taken to a watchmaker who finds that the
frequency of the internal oscillator is incorrect. He adjusts the frequency to the correct value, i.e., to
32,768.0000 Hz, using a frequency standard (e.g., a GPS disciplined oscillator). At that point, the
watch is syntonized to the watchmaker’s frequency standard. After the watch is disconnected, the
syntonization will last for a finite amount of time (that time depends on the allowed frequency offset.).
When two oscillators are frequency locked, they will remain syntonized as long as they remain locked.
In the slide, “in agreement,” “same”… mean that they are within an acceptable or specified range of
each other.

On Using Time for Clock Rate Calibration
It takes time to measure the clock rate (i.e., frequency) difference between
two clocks. The smaller the rate difference between a clock to be calibrated
and a reference clock, the longer it takes to measure the difference (∆t/t ≈ ∆f/f).
For example, assume that a reference timing source (e.g., Loran or GPS)
with a random time uncertainty of 100 ns is used to calibrate the rate of a clock
to 1 x 10-11 accuracy. A frequency offset of 1 x 10-11 will produce1 x 10-11 x
3600 s/hour = 36 ns time error per hour. Then, to have a high certainty that
the measured time difference is due to the frequency offset rather than the
reference clock uncertainty, one must accumulate a sufficient amount (≥100
ns) of time error. It will take hours to perform the calibration. (See the next
page for a different example.) If one wishes to know the frequency offset to a
±1 x 10-12 precision, then the calibration will take more than a day.
Of course, if one has a cesium standard for frequency reference, then, for
example, with a high resolution frequency counter, one can make frequency
comparisons of the same precision much faster.
8-13

Calibration With a 1 pps Reference
Let A = desired clock rate accuracy after calibration
A' = actual clock rate accuracy
∆τ = jitter in the 1 pps of the reference clock, rms
∆τ' = jitter in the 1 pps of the clock being calibrated, rms
t = calibration duration
∆t = accumulated time error during calibration
Then, what should be the t for a given set of A, ∆t, and ∆t'?
Example: The crystal oscillator in a clock is to be calibrated by

comparing the 1 pps output from the clock with the 1 pps output from a standard. If
A = 1 x 10-9; ∆τ = 0.1 µs, and ∆τ' = 1.2 µs, then, [(∆τ)2 + (∆τ')2]1/2 ≈ 1.2 µs, and
when A = A', ∆t = (1 x 10-9)t ≡ (1.2 µs)N, and t = (1200N) s. The value of N to be
chosen depends on the statistics of the noise processes, on the confidence level
desired for A' to be ≤ A, and on whether one makes measurements every second
or only at the end points. If one measures at the end points only, and the noise is
white phase noise, and the measurement errors are normally distributed, then,
with N = 1, 68% of the calibrations will be within A; with N = 2, and 3, 95% and
99.7%, respectively, will be within A. One can reduce t by about a factor 2/N3/2 by
making measurements every second; e.g., from 1200 s to 2 x (1200)2/3 = 226 s.
8-14
The above results are based on an analysis by Samuel R. Stein, Ball Communication Systems Div.,
private communication, November 1989.

Time Transfer Methods
Method Accuracy ~ Cost (‘95)
Portable Cs clock 10 - 100 ns $45K - 70K
GPS time dissemination 20 - 100 ns $100 - 5K

GPS common view 5 - 20 ns
Two-way via satellite ~1 ns $60k
Loran-C 100 ns $1K - 5K
HF (WWV) 2 ms $100 - 5K
Portable quartz & Rb Calibration interval $200 - 2K

clocks dependent
8-15
Time-transfer techniques provide a method of maintaining synchronization among remote

locations which complements the use of independent clocks. In fact, most systems derive time using
both external time references and internal clocks. The former provide long-term accuracy and
interoperability; the latter provide autonomous capability in the absence of the external references. A
variety of time-transfer techniques are in use today. They vary in capability from a few milliseconds to
a few nanoseconds. GPS, and the GPS common view technique, are discussed in the following
pages.
Radio broadcast services, such as WWV and WWVB in the United States, disseminate time with
modest accuracy. The high-frequency broadcasts between 2.5 and 20 MHz are usually received after
reflection from the ionosphere. As a result, variability in the path delay limits the accuracy to a few
milliseconds for most users.
The highest accuracy time-synchronization is obtained via two-way satellite techniques. Both
the propagation errors and the delays through the receiver are calibrated by transmitting time in both
directions between two sites. Each site measures the difference between the time of arrival of the
pulse from the other site and the time of the local clock. The difference in the measurements made at
the two ends provides the relative time of the two local clocks. The effects of the transmitter and
receiver delays, the uplink and downlink propagation delays, and the delays through the satellite are
substantially canceled. As a result, time-synchronization accuracy of a few nanoseconds has been
obtained using commercial communication satellites and very small aperture terminals (VSAT). A
custom spread-spectrum time-transfer modem is necessary.
W. Lewandowski & C. Thomas, “GPS Time Transfer,” Proc. IEEE, vol. 79, pp. 991-1000, 1991.
D. Kirchner, “Two-way Time Transfer Via Communcations Satellites,” Proc. IEEE, vol. 79, pp.
983-990, 1991.
S. Leschiutta, “Time Synchronization Using Laser Techniques,” Proc. IEEE, vol. 79, pp. 1001-
1008, 1991.

Global Positioning System (GPS)
GPS Nominal Constellation:

24 satelites in 6 orbital planes,
4 satelites in each plane,
20,200 km altitude, 55 degree inclinations
8-16
The Global Positioning System (GPS) is the most precise worldwide navigation system
available. It is also capable of providing nanosecond-level timing accuracies, so, it is also one of the
most accurate time sources.
GPS is a satellite-based radio navigation and positioning system that is designed to provide
global, all-weather, 24-hour, accurate navigation to an unlimited number of users. Each of the
satellites contain atomic clocks. The satellites transmit a navigation message that provides satellite
position, time, and atmospheric propagation correction data. The GPS receiver, which contains a
quartz crystal clock, measures the transit time of the satellite signal and multiplies that time by the
speed of light to compute range to the satellite. The satellite clocks are more accurate than the
receiver clocks. Therefore, although three satellites can provide latitude, longitude and altitude, the
signal from a fourth satellite is used to correct for the navigational error caused by the receiver clock's
inaccuracy, i.e., the receivers calculate their x, y, z, and t from receiving each of four satellites’ x, y, z,
and t. Velocity is determined from the Doppler shifts of the the transmitted carrier frequencies.
A. J. Van Dierendonck and M. Birnbaum, "Time Requirements in the NAVSTAR Global Positioning
System (GPS)," Proc. 30th Annual Symposium on Frequency Control, pp. 375-383, 1976, AD046089.
F. E. Butterfield, "Frequency Control and Time Information in the NAVSTAR/Global Positioning

System," Proc. 30th Annual Symposium on Frequency Control, pp. 371-374, 1976, AD046089
Much information is available on the Internet, e.g., see “Navstar GPS Internet Connections” at
http://gauss.gge.unb.ca/GPS.INTERNET.SERVICES.HTML, and
“Global Positioning System Overview” by Peter H. Dana (from which the above illustration was
“borrowed,” with permission from Peter H. Dana, The University of Texas at Austin) at
http://www.utexas.edu/depts/grg/gcraft/notes/gps/gps.html

GPS
GPS can provide global, all-weather, 24-hour, real-time, accurate navigation
and time reference to an unlimited number of users.
• GPS Accuracies (2σ)

Position: 120 m for Standard Positioning Service, SPS
40 m for Precise Positioning Service, PPS
1 cm + 1ppm for differential, static land survey
Velocity: 0.3 m/s (SPS), 0.1 m/s (PPS).
Time: 350 ns to < 10 ns
• 24 satellites in 6 orbital planes; 6 to 10 visible at all times; ~12 h period
20,200 km orbits.
• Pseudorandom noise (PRN) navigation signals are broadcast at L1 =
1.575 GHz (19 cm) and L2 = 1.228 GHz (24 cm); two codes, C/A and
P are sent; messages provide satellite position, time, and atmospheric
propagation data; receivers select the optimum 4 (or more) satellites to
track. PPS (for DoD users) uses L1 and L2, SPS uses L1 only.
8-17
GPS may also be used in a differential mode, often called common view, to provide improved
synchronization capability. For sites located within several thousand kilometers (km) of one another,
timing errors due to errors in the ephemeris and the propagation delay are approximately equal.
Thus, when the absolute GPS times of arrival of simultaneously observed satellite signals are
subtracted from one another, the differential accuracy can be improved to ~ten nanoseconds.
W. Lewandowski and C. Thomas, "GPS Time Transfer," Proc. IEEE, Vol. 79, pp. 991-1000, July 1991.
C. Fox and S. R. Stein, "GPS Time Determination and Dissemination," Tutorials from the Twenty-third
Ann. Precise Time and Time Interval (PTTI) Applications and Planning Meeting, 1991, NTIS
Accession No. AD-A254745

Oscillator’s Impact on GPS
• Satellite oscillator’s (clock’s) inaccuracy & noise are major sources of
navigational inaccuracy.
• Receiver oscillator affects GPS performance, as follows:
Oscillator Parameter GPS Performance Parameter

Warmup time Time to first fix
Power Mission duration, logistics costs (batteries)
Size and weight Manpack size and weight
Short term stability ∆ range measurement accuracy, acceleration
(0.1 s to 100 s) performance, jamming resistance
Short term stability Time to subsequent fix
(~15 minute)
Phase noise Jamming margin, data demodulation, tracking
Acceleration sensitivity See short term stability and phase noise effects
8-18
D. L. Hessick and W. C. Euler, "GPS User Receivers and Oscillators," Proc. 38th Annual Symposium
N. Ashby & M. Weiss, “Global Positioning System Receivers and Relativity,” NIST Technical Note
1385, March 1999.

Time Scales
• A "time scale" is a system of assigning dates, i.e., a "time," to events; e.g., 6
January 1989, 13 h, 32 m, 46.382912 s, UTC, is a date.
• A "time interval" is a "length" of time between two events; e.g., five seconds.
• Universal time scales, UT0, UT1, and UT2, are based on the earth's spin on its
axis, with corrections.
• Celestial navigation: clock (UT1) + sextant position.
• International Atomic Time (TAI) is maintained by the International Bureau of

Weights and Measures (BIPM; in France), and is derived from an ensemble of
more than 200 atomic clocks, from more than 60 laboratories around the world.
• Coordinated Universal Time (UTC) is the time scale today, by international

agreement. The rate of UTC is determined by TAI, but, in order to not let the
time vs. the earth's position change indefinitely, UTC is adjusted by means of
leap seconds so as to keep UTC within 0.9 s of UT1.
8-19
UTC is generated (~30 days) after the fact, by The Bureau International des Poids et Mesures
near Paris, France, by taking the times of about 230 (as of 1999) of the best (atomic) clocks in the
timing laboratories of the world and combining their readings in an optimum way. Each of the 230
clocks receives a weighting factor, depending on its performance. More than 30 countries’ laboratories
contribute time data to UTC.
The length of the second is determined by evaluations of 11 laboratory Cs primary frequency
standards (located in six countries). A weighted combination is used to make the best estimate of the
second.
The time scale generated from combining the 230 clocks and 11 primary standards is called the
Temps Atomique International, TAI, i.e., the International Atomic Time. The source of the “second” is
the 11 primary standards, and the 230 clocks are the flywheel that maintain the calibration provided by
the primary standards.
To obtain real-time estimates of UTC, timing centers around the world generate their own
current best estimate of UTC. These are called UTC(k), where k is usually the name of the timing
center, e.g., UTC(NIST), UTC(USNO), etc. The BIPM distributes a monthly bulletin which reports all
the UTC - UTC(k) for the previous month.
GPS time is not adjusted for leap seconds. It is steered to agree with UTC, except for an integer
number of seconds. Except for the leap seconds, GPS time is within 40 ns of UTC. UTC, as
estimated by the USNO, is available from the GPS signal.
T. J. Quinn, “The BIPM and the Accurate Measurement of Time,” Proc. IEEE, vol. 79, pp. 894-905,
1991.
D. B. Sullivan, “Time Generation and Distribution,” Proc. IEEE, vol. 79, pp. 906-914, 1991.
D. D. McCarthy, “Astronomical Time,” Proc. IEEE, vol. 79, pp. 915-920, 1991.
The Science of Timekeeping, Hewlett-Packard Co. Application Note 1289

For Frequency Control and<http://www.tmo.hp.com/tmo/Notes/English/>
Timing Applications - A TUTORIAL”
Clock Ensembles and Time Scales
• An ensemble of clocks is a group of clocks in which the time outputs of
individual clocks are combined, via a “time-scale algorithm,” to form a
time scale.
• Ensembles are often used in mission critical applications where a

clock’s failure (or degraded performance) presents an unacceptable
risk.
• Advantages of an ensemble:
- system time & frequency are maintained even if a clock in the
ensemble fails
- ensemble average can be used to estimate the characteristics of
each clock; outliers can be detected
- performance of ensemble can (sometimes) be better than any of
the contributors
- a proper algorithm can combine clocks of different characteristics,
and different duty cycles, in an optimum way
8-20
P. Tavella & C. Thomas, “Comparative Study of Time Scale Algorithms,” Metrologia, vol. 28, pp. 57-
63, 1991.
C. Thomas, “Impact of New Clock Technologies on the Stability and Accuracy of the International
Atomic Time TAI,” Proc. 1996 IEEE Int’l Frequency Control Symposium, pp. 1123-1130, 1996.
Kamas, George and Michael A. Lombardi, Time & Frequency Users' Manual, NIST Special Publication
559, revised 1990.

Relativistic Time
• Time is not absolute. The "time" at which a distant event takes
place depends on the observer. For example, if two events, A and
B, are so close in time or so widely separated in space that no
signal traveling at the speed of light can get from one to the other
before the latter takes place, then, even after correcting for
propagation delays, it is possible for one observer to find that A
took place before B, for a second to find that B took place before
A, and for a third to find that A and B occurred simultaneously.
Although it seems bizarre, all three can be right.
• Rapidly moving objects exhibit a "time dilation" effect. (“Twin
paradox”: Twin on a spaceship moving at 0.87c will age 6 months
while twin on earth ages 1 year. There is no "paradox" because
spaceship twin must accelerate; i.e., there is no symmetry to the
problem.)
• A clock's rate also depends on its position in a gravitational field. A
high clock runs faster than a low clock.
8-21
1385, March 1999.
C. Alley, "Relativity and Clocks," Proc. 33rd Annual Symposium on Frequency Control, pp. 4-39, 1979.
G. M. R. Winkler, “Synchronization and Relativity,” Proc. IEEE, vol. 79, pp. 1029-1039, 1991.
The Science of Timekeeping, Hewlett-Packard Co. Application Note 1289

<http://www.tmo.hp.com/tmo/Notes/English/>

Relativistic Time Effects
• Transporting "perfect" clocks slowly around the surface of the earth along
the equator yields ∆t = -207 ns eastward and ∆t = +207 ns westward
(portable clock is late eastward). The effect is due to the earth's rotation.
• At latitude 40o, for example, the rate of a clock will change by 1.091 x 10-13
per kilometer above sea level. Moving a clock from sea level to 1km
elevation makes it gain 9.4 nsec/day at that latitude.
• In 1971, atomic clocks flown eastward then westward around the world in
airlines demonstrated relativistic time effects; eastward ∆t = -59 ns,
westward ∆t = +273 ns; both values agreed with prediction to within the
experimental uncertainties.
• Spacecraft Examples:
• For a space shuttle in a 325 km orbit, ∆t = tspace - tground = -25 µsec/day
• For GPS satellites (12 hr period circular orbits), ∆t = +38.5 µsec/day
• In precise time and frequency comparisons, relativistic effects must be

included in the comparison procedures.
8-22
1385, March 1999.
G. M. R. Winkler, “Synchronization and Relativity,” Proc. IEEE, vol. 79, pp. 1029-1039, 1991

Relativistic Time Corrections
The following expression accounts for relativistic effects, provides for clock rate
accuracies of better than 1 part in 1014, and allows for global-scale clock
comparisons of nanosecond accuracy, via satellites:
T
∫ ⎢⎣ 2 (v )
− v 2g − (Φ S − Φ g )⎤ dt + 2 A E
1 ⎡1 2ω
∆t = − 2
c 2 s ⎥⎦ c
0
Where ∆t = time difference between spacecraft clock and ground clock, tS-Tg
VS = spacecraft velocity (<<c), Vg = velocity of ground station
ΦS = gravitational potential at the spacecraft
Φg = gravitational potential at the ground station
ω = angular velocity of rotation of the earth
AE = the projected area on the earth’s equatorial plane swept out by the vector
whose tail is at the center of the earth and whose head is at the position
of the portable clock or the electromagnetic signal pulse. The AE is taken
positive if the head of the vector moves in the eastward direction.
Within 24 km of sea level, Φ = gh is accurate to 1 x 10-14 where g = (9.780 + 0.052

sin2Ψ )m/s2, Ψ = the latitude, h = the distance above sea level, and where the sin2Ψ
term accounts for the centrifugal potential due to the earth's rotation. The "Sagnac
effect," (2ω/c2)AE = (1.6227 x 10-21s/m2)AE, accounts for the earth-fixed coordinate
system being a rotating, noninertial reference frame.
8-23
N. Ashby and D. W. Allan, "Practical Implications of Relativity for a Global Coordinate Time Scale,"
Radio Science, Vol. 14, No. 4, pp. 649-669, July-August 1979.
G. M. R. Winkler, “Synchronization and Relativity,” Proc. IEEE, vol. 79, pp. 1029-1039, 1991.

Some Useful Relationships
• Propagation delay = 1 ns/30 cm = 1 ns/ft = 3.3 µs/km ≈ 5 µs/mile
• 1 day = 86,400 seconds; 1 year = 31.5 million seconds
• Clock accuracy: 1 ms/day ≈ 1 x 10-8
• At 10 MHz: period = 100 ns; phase deviation of 1° = 0.3 ns of time

deviation
*
• Doppler shift = ∆f/f = 2v/c
--------------------------------------
* Doppler shift example: if v = 4 km/h and f = 10 GHz (e.g., a slow-
moving vehicle approaching an X-band radar), then ∆f = 74 Hz, i.e.,
an oscillator with low phase noise at 74Hz from the carrier is
necessary in order to "see" the vehicle.
8-24

One Pulse-Per-Second Timing Signal
(MIL-STD-188-115)
Rise Time Fall Time

10 Volts < 20 Nanoseconds < 1 Microseconds
(± 10%)
0 Volts
(± 1 Volt
20 µsec ± 5%
"The leading edge of the BCD code (negative going transitions after extended high
level) shall coincide with the on-time (positive going transition) edge of the one
pulse-per-second signal to within ±1 millisecond." See next page for the MIL-STD
BCD code.
8-25
MIL-STD-188-115, Military Standard, "Interoperability and Performance Standards for

Communications Timing and Synchronization Subsystems," 31 March 1986. Copies are available
from Naval Publications and Forms Center, 5801 Tabor Ave., Philadelphia, PA 19120

BCD Time Code
(MIL-STD-188-115)
Rate: 50 Bits per Second

Example: Selected Time is 12:34:56
Bit Pulse Width: 20 msec
H = +6V dc ± 1V
L = -6V dc ± 1V
LLLH LLHL LLHH LHLL LHLH LHHL
20 8421 8421 8421 8421 8421 8421
msec
1 2 3 4 5 6
Level
Level Hours Minutes Seconds Held
Held Hi
Hi Until
Until 8 4 2 1 8 4 2 1 8 4 2 1 8 4 2 1 8 4 2 1 8 4 2 1 Start
Start of
of Next
Code- Code-
Word Word
24 Bit BCD Time Code*

* May be followed by 12 bits for day-of-year and/or 4 bits for figure-of-merit
(FOM). The FOM ranges from better than 1 ns (BCD character 1) to greater
than 10 ms (BCD character 9).
8-26
Time codes are used primarily to record time-of-day information with data in order to provide a
time reference during data reduction (also called “time tagging”). Time codes originated with
requirements during early missile and space programs to correlate test data with precise time. Today,
time codes are used widely in applications such as communications systems, medical (patient
monitoring) applications, and industrial and commercial data acquisition systems.
MIL-STD-188-115, Military Standard, "Interoperability and Performance Standards for

Communications Timing and Synchronization Subsystems," 31 March 1986. Copies are available
from Naval Publications and Forms Center, 5801 Tabor Ave., Philadelphia, PA 19120
Parallel time codes are defined in IRIG Standard 205-87
Serial time codes are defined in IRIG Standard 200-98,

http://tecnet0.jcte.jcs.mil/RCC/manuals/200/index.html

Time and Frequency Subsystem
Oscillator
Oscillator and
and Power
Power
Clock
Clock Driver
Driver Source
Source
Frequency
Frequency Time
Time Code
Code
Distribution
Distribution Generator
Generator
f1 f2 f3 TOD 1 pps
8-27
The time and frequency subsystem, also sometimes called a “time and frequency unit
(TFU) provides frequencies and time to a system.

The MIFTTI Subsystem
MIFTTI = Modular Intelligent Frequency, Time and Time
Interval
Oscillator/
Oscillator/ Battery
Battery
Clock-Driver
Clock-Driver
Microcomputer
Microcomputer V1
External
External Compensation
Compensationand
and V2
Reference
Reference Control
Control
Clock
Clockand
and
Frequency
Frequency DC-DC
DC-DC
Time
Timecode
code
Distribution
Distribution Converter
Converter
Generator
Generator
>
f1 f2 f3
TOD 1pps Vin
* The microcomputer compensates for systematic effects (after filtering

random effects), and performs: automatic synchronization and calibration
when an external reference is available, and built-in-testing.
8-28
Time and frequency subsystems can include built-in intelligence to improve performance, as was
shown in the MIFFTI development program. Intelligence was used to achieve maximum time and
frequency performance from the available frequency sources by providing automatic fault detection,
synchronization and syntonization, and by providing compensation for such systematic errors such as
short-term aging and temperature sensitivity.
Clocks and timing receivers can be combined in a timing system to provide a broader range of
timing capabilities than either one can provide alone. Such a system uses the received timing signal
to calibrate the local clock, and learn its time, frequency, and frequency aging. When the timing signal
is unavailable, the local clock acts as a "flywheel." Its free-running operation starts using the time and
frequency provided by calibration versus the external source. The frequency may subsequently be
updated periodically for the predicted frequency aging. This procedure produces the minimum
possible free-running timing errors. Commercial "disciplined oscillators" now provide these functions
in an integrated package.
K.D. Lyon, W.J. Riley, and J.R. Vaccaro, “The Modular Intelligent Frequency, Time And Time Interval
(MIFTTI) Subsystem Program,” Proc. 43rd Ann. Frequency Control Symp., pp. 225-231, 1989.
S. R. Stein & R. L. Filler, “Kalman Filter Analysis for Real Time Applications of Clocks and Oscillators,
Proc. 42nd Ann. Frequency Control Symp., pp. 447-452, 1988.


"Time" Quotations
z 3 o'clock is always too late or too early for anything you want to do..................Jean-Paul Sartre
z Time ripens all things. No man's born wise............Cervantes.
z Time is the rider that breaks youth............George Herbert z Time wounds all heels..................Jane Ace
z Time heals all wounds..................Proverb z The hardest time to tell: when to stop.....Malcolm Forbes
z Time is on our side..................William E. Gladstone z It takes time to save time.............Joe Taylor
z Time, whose tooth gnaws away everything else, is powerless against truth..................Thomas H. Huxley
z Time has a wonderful way of weeding out the trivial..................Richard Ben Sapir
z Time is a file that wears and makes no noise...........English proverb
z The trouble with life is that there are so many beautiful women - and so little time..................John Barrymore
z Life is too short, and the time we waste yawning can never be regained..................Stendahl
z Time goes by: reputation increases, ability declines..................Dag Hammarskjöld
z Remember that time is money...............Benjamin Franklin
z Time is money - says the vulgarest saw known to any age or people. Turn it around, and you get a precious truth -
Money is time..................George (Robert) Gissing
z The butterfly counts not months but moments, and has time enough..................Rabindranath Tagore
z Everywhere is walking distance if you have the time..................Steven Wright
z The only true time which a man can properly call his own, is that which he has all to himself; the rest, though in
some sense he may be said to live it, is other people's time, not his..................Charles Lamb
z It is familiarity with life that makes time speed quickly. When every day is a step in the unknown, as for children, the
days are long with gathering of experience..................George Gissing
z Time is a great teacher, but unfortunately it kills all its pupils..................Hector Berlioz
z To everything there is a season, and a time to every purpose under the heaven..................Ecclesiastes 3:1
z Time goes, you say? Ah no! Time stays, we go..................Henry Austin Dobson
8-29
Man must sit in chair with mouth open for very long time before roast duck fly in. - Chinese
Proverb
The time to repair the roof is when the sun is shining. - John F. Kennedy
There may be times when we are powerless to prevent injustice, but there must never be a
time when we fail to protest. - Elie Wiesel
The time to win a fight is before it starts. - Frederick W. Lewis
Nothing is a waste of time if you use the experience wisely. - Auguste Rodin
There is one thing stronger than all the armies in the world, and that is an idea whose time
has come. - Victor Hugo
We didn't lose the game; we just ran out of time. - Vince Lombardi
I have made this letter longer because I lack the time to make it shorter. - Blaise Pascal
Begin to make the kind of investment of personal time which will assure that those who
come after us will live as well. - Charles W. Bray III
Most of us spend too much time on the last twenty-four hours and too little on the last six
thousand years. - Will Durant
Don't spend your precious time asking 'Why isn't the world a better place?” It will only be
time wasted. The question to ask is 'How can I make it better?' To that there is an answer. -
Leo F. Buscaglia
In Germany they came first for the Communists, and I didn't speak up because I wasn't a
Communist.
“QUARTZ CRYSTAL RESONATORS AND Then they came for the Jews, and I didn't speak up because I wasn't a Jew.
OSCILLATORS
Then they came forTUTORIAL”
For Frequency Control and Timing Applications - A the trade unionists, and I didn't speak up because I wasn't a trade
Units of Measurement Having Special Names
in the International System of Units (SI)
SI Base Units
Electric Temperature Luminous Amount of
Mass Length Time
Current kelvin Intensity Substance
kilogram meter second
ampere candela mole
kg m s A K cd mol
S K cd sr
Coordinated Time Celsius Luminous Flux
international atomic time Temperature lumen
SI Derived Units TAI
0
Celsius
0 lm
s-1 C
kg m s-2 Frequency
Force hertz
newton Hz
N m-2cd sr
kg m2s-2 kg m-1s-2 SA Illuminance
Pressure Electric charge lux
Energy pascal lx
joule coulomb
J Pa C
kg m s-3 2 -3
kg m s A -2
Non-SI units Power kg m2s-3 A-1 Resistance sr: the steradian is the supplementary
watt Electric Potential ohm
recognized volt SI unit of solid angle (dimensionless)
W Ω
for use with SI s-1 V rad: the radian is the supplementary
Activity kg-1 m2s4 A2 kg m2s-2A-1 SI unit of plane angle (dimensionless)
day: 1 d = 86400 s becquerel Capacitance Magnetic Flux
hour: 1 h = 3600 s Bq farad weber
minute: 1 min = 60 s m2s-1 F Wb
Absorbed Dose kg-1 m2s3 A2 kg m2s-2A-2
liter: 1 l = 10-3 m3 gray Conductance Inductance
ton: 1 t = 103 kg Gy siemens henry
m2s-2 S H
degree: 10 = (π/180) rad Dose Equivalent kg s-2 A-1
minute: 1’ = (π/10800)rad sievert Conductance
second: 1” = (π/648000)rad Sv siemens Electromagnetic
Health related S measurement units
electronvolt: 1 eV ≈ 1.602177 x 10-19 J measurement units
unified atomic mass unit: 1 u ≈ 1.660540 x 10-27 kg
8-30
Time interval (frequency) is the quantity that can be determined with the highest accuracy. It
can be measured with an accuracy greater than 1 part in 1013. With the help of satellites, it is possible
to compare the time scales kept by the national laboratories, worldwide, to an accuracy of ~1 ns.
Time, therefore, plays a central role in metrology and in the definitions of SI units.
The SI consists of seven base units and a number of derived units, as shown above. Shown on
the next page are the units that do NOT depend on the unit of time.
-------------------------------
R. J. Douglas, et. al, "Frequency Standards, Timekeeping, and Traceable Services at the National
Research Council of Canada," Proc. 28th Ann. Precise Time & Time Interval (PTTI) Applications &
Planning Meeting, pp. 65-80, 1996.
The chart above, and the one on the next page, were provided by R.J. Douglas, National Research
Council Canada, 1997.
E. R. Cohen & B. N. Taylor, “The Fundamental Physical Constants,” Physics Today, pp. BG7-BG14,
August 1997.
B. W. Petley, "Time and Frequency Fundamental Metrology," Proceedings of the IEEE, vol.
79, pp. 1070-1076, 1991.

Units of Measurement Having Special Names in the SI Units,
NOT Needing Standard Uncertainty in SI Average Frequency
SI Base Units
Temperature Amount of
Mass Substance
kilogram kelvin
mole
kg K mol
K
SI Derived Units Celsius
Temperature
0
Celsius
0
C
Non-SI units
recognized
for use with SI
ton: 1 t = 103 kg
degree: 10 = (π/180) rad
minute: 1’ = (π/10800)rad
second: 1” = (π/648000)rad
unified atomic mass unit: 1 u ≈ 1.660540 x 10-27 kg
8-31

CHAPTER 9
Related Devices and Application

Discrete-Resonator Crystal Filter
A Typical Six-pole Narrow-band Filter
Layout
Circuit
9-1
Most crystal filters are bandpass filters. There are two types, one which consists of discreet
resonators, as illustrated above, and the other, monolithic crystal filters which consist of acoustically
coupled resonators, as illustrated on the next page.
R. C. Smythe and R. S. Wagers, "Piezoelectric and Electromechanical Filters," in E. A. Gerber and A.

R. G. Kinsman, Crystal Filters, John Wiley & Sons, 1987.

Monolithic Crystal Filters
Two-pole filter and its response
25.0
20.0 Four-pole filter

electrode arrangement
Attenuation (dB)
15.0
10.0
5.0
0
Frequency
9-2
R. C. Smythe and R. S. Wagers, "Piezoelectric and Electromechanical Filters," in E. A. Gerber and A.

R. G. Kinsman, Crystal Filters, John Wiley & Sons, 1987.

Surface Acoustic Wave (SAW) Devices
λ 2 λ 2
BAW SAW, One-port SAW, Two-port
Simplified Equivalent Circuits

C0 L1 C1 R1
C1 L1 R1 C0 C0
BAW and One-port SAW Two-port SAW
9-3
The primary application of SAW devices is in filters. Applications in precision frequency control
and timing are limited because the long term stability and temperature stability of the best bulk-
acoustic-wave (BAW) devices are better than those of the best SAW devices.
For BAW resonators, the plate thickness determines then fundamental-mode frequency. For
SAW resonators, the interdigital transducers' (IDT) spacings determine the frequency. For quartz, a
300 MHz BAW resonator plate is 6 µm thick. A 2.6 GHz SAW resonator has 0.3 µm IDT spacings,
and can be produced by e-beam lithography.
In SAW resonators, wave motion is concentrated at the surface of the crystal; motion decays
exponentially with distance from surface; 90 to 95% of the energy is within one acoustic wavelength of
the surface.
In one-port SAW and BAW resonators, the static capacitance, C0, provides a low-impedance
path that can mask out the desired resonance at high f's. An external inductor is usually placed in
parallel with C0 to "resonate out" C0. In two-port SAW resonators, C0 does not shunt the motional arm
of the equivalent circuit, therefore, two-port SAW resonators are preferred in many applications. See
chapter 4 for phase noise comparisons with BAW devices.
D. P. Morgan, Surface-Wave Devices for Signal Processing, Elsevier, 1985.
D. P. Morgan, “History of SAW Devices,” Proc. 1998 IEEE Int’l Frequency Control Symp., pp. 439-
460, 1998.
F. S. Hickernell, ”Surface Acoustic Wave Technology - Macrosuccess through Microseisms," in

Ultrasonic Instruments & Devices II, edited by R.N. Thurston, A. D. Pierce, vol. ed. E. Papadakis,
Physical Acoustics Vol. XXIV, Academic Press, pp. 135-207, 1999.
W. R. Shreve and P. S. Cross, "Surface Acoustic Waves and Resonators," in E. A. Gerber and A.
R. S. Wagers, “Surface-Acoustic-Wave Filters,” and “SAW Bandpass and Bandstop Filters,” in E. A.

Gerber and A. Ballato, Precision Frequency Control, Vol. 1, pp. 230-249, and pp. 257-266, Academic
Press, 1985.
Quartz Bulk-Wave Resonator Sensors
In frequency control and timekeeping applications, resonators are
designed to have minimum sensitivity to environmental parameters.
In sensor applications, the resonator is designed to have a high
sensitivity to an environmental parameter, such as temperature,
adsorbed mass, force, pressure and acceleration.
Quartz resonators' advantages over other sensor technologies are:
• High resolution and wide dynamic range (due to excellent short-

term stability); e.g., one part in 107 (10-6 g out of 20 g)
accelerometers are available, and quartz sorption detectors are
capable of sensing 10-12 grams.
• High long-term accuracy and stability, and
• Frequency counting is inherently digital.
9-4
D. S. Ballantine, Jr., et. al, Acoustic Wave Sensors - Theory, Design, and Physico-Chemical
Applications, Academic Press, 1997.
Special Issue on Sensors and Actuators, IEEE Trans. on Ultrasonics, Ferroelectrics, and Frequency
Control, vol. 45, September 1998.
E. P. EerNisse, R. W. Ward and R. B. Wiggins, "Survey of Quartz Bulk Resonator Sensor

Technologies," IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 35, No.
3, pp. 323-330, May 1988.
G. G. Guilbault and J. M. Jordan, "Analytical Uses of Piezoelectric Crystals: A Review," CRC Critical
Reviews in Analytical Chemistry, Vol. 19, Issue 1, pp. 1-28, 1988.
J. J. McCallum, "Piezoelectric Devices for Mass and Chemical Measurements," Analyst, Vol. 114, pp.
1173-1189, Oct. 1989.

Tuning Fork Resonator Sensors
e
rc
Fo
Beam
Motion
Tine
e
Motion
rc
Fo
Photolithographically produced tuning forks, single- and double-ended (flexural-mode
or torsional-mode), can provide low-cost, high-resolution sensors for measuring
temperature, pressure, force, and acceleration. Shown are flexural-mode tuning forks.
9-5
The resonant frequency and Q of a vibrating tuning fork or beam resonator, whether
single, double or triple-beam, are sensitive to forces, temperature, and the fluid surrounding the
resonator. Such resonators have been used as accelerometers, pressure sensors, force sensors, gas
density sensors, vacuum gauges, and thermometers.
--------------------------
E. P. EerNisse, R. W. Ward and R. B. Wiggins, "Survey of Quartz Bulk Resonator Sensor
Technologies," IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, Vol. 35, No.
3, pp. 323-330, May 1988.

Dual Mode SC-cut Sensors
•Advantages
- Self temperature sensing by dual mode operation allows
separation/compensation of temp. effects
- Thermal transient compensated
- Isotropic stress compensated
- Fewer activity dips than AT-cut
- Less sensitive to circuit reactance changes
- Less sensitive to drive level changes
•Disadvantage
- Severe attenuation in a liquid

- Fewer SC-cut suppliers than AT-cut suppliers
9-6
See discussion of the dual-mode technique in Chapters 2 and 3.
------------------------------
J. R. Vig, “Dual-mode Oscillators for Clocks and Sensors,” Proc. 1999 IEEE Int’l Ultrasonics
Symposium

Separation of Mass and Temperature
Effects
• Frequency changes
∆f (m, T, x ) ∆f (m) ∆f (T ) ∆f ( x )
= + +
f f f f
o o o o
total mass temperature other effects
• Mass: adsorption and desorption

∆f ( m ) ∆m
≅−
fo mo
• Temperature/beat frequency
∑ ci ⋅ ∆fβ
i
∆f (T )
= i fβ ≡ 3fc1(T ) − fc 3 (T )
fo fo
9-7
When, for example, a mass is deposited onto a resonator (e.g., a quartz

crystal microbalance), there is always a temperature change accompanying the mass
change. The temperature change is caused by the heat of adsorption, and by the heat
emitted by the evaporation source. The frequency change due to mass deposition is due to
the combined effects of the mass change and the temperature change. The dual mode
technique yields two equations with two unknowns which allows the separation of the mass
change induced frequency change from the temperature caused frequency change.
The mass change alone can be determined without having to calibrate the
beat frequency, fβ, as a function of temperature. However, by calibrating the fβ ασ α
φυνχτιον οφ τεµπερατυρε, ονε µαψ, φορ εξαµπλε, περφορµ συρφαχε στυδιεσ ασ α φυνχτιον
οφ τεµπερατυρε.
--------------------------
D. E. Pierce, Y. Kim, and J. R. Vig, "A temperature insensitive quartz microbalance,“ Proc.
1997 IEEE Int’l Frequency Control Symp., pp. 41-48.
J. R. Vig, “Dual-mode Oscillators for Clocks and Sensors,” Proc. 1999 IEEE Int’l Ultrasonics
Symposium
US Pat. No. 5869763, J. R. Vig and R. L. Filler, "Method of measuring mass change using a
quartz crystal microbalance," Feb. 1999.
Y. Kim, “Sensing of organic vapor adsorption on gold using a temperature insensitive

microbalance,” Proc. 1999 IEEE Int’l Frequency Control Symp.

Dual-Mode Pressure Sensor
End Caps
Electrode
Resonator
9-8
High precision pressure sensors employ quartz thickness shear resonators. The
frequencies of these vary not only with pressure, but also with temperature. Temperature control, or
compensation is used to minimize the errors due to temperature effects. Some pressure sensors have
used a temperature sensor that is external, but in close proximity, to the pressure sensing resonator,
and shielded from the applied pressure. However, when a temperature gradient exists between the
pressure and temperature sensors, such as may occur under pressure and temperature transients, the
accuracy of pressure measurement is degraded.
Dual-mode pressure sensors have been proposed in which the b-mode is stress
compensated and the c-mode is temperature compensated. In such a sensor, the b-mode indicates
primarily the temperature of the sensing resonator, and the c-mode indicates the applied pressure. As
the b-mode’s frequency depends on the temperature of the resonator’s vibrating volume, the effects
of temperature gradients are greatly reduced. Under pressure transients, especially, dual-mode
pressure sensors allow for superior temperature sensing and compensation accuracy, and superior
pressure sensing accuracy. The above diagram shows such a dual mode pressure sensor.
----------------------------
R. J. Besson, J. J. Boy, B. Glotin, Y. Jinzaki, B. Sinha, and M. Valdois, “A dual-mode
thickness-shear quartz pressure sensor,” IEEE Trans. Ultrason. Ferroelect. Freq. Contr.,
Vol. 40, pp. 584-591, 1993.
N. Matsumoto, Y. Sudo, B. Sinha and M. Niwa, “Long-term stability and performance

characteristics of crystal gauge at high pressures and temperatures,” Proc. 1999 IEEE Int’l Frequency
Control Symp.

CHAPTER 10
Proceedings Ordering Information,
Standards, Website, and Index
10

IEEE International Frequency Control Symposium
PROCEEDINGS ORDERING INFORMATION
NO. YEAR DOCUMENT NO. SOURCE* NO. YEAR DOCUMENT NO. SOURCE*
10 1956 AD-298322 NTIS 33 1979 AD-A213544 NTIS

11 1957 AD-298323 NTIS 34 1980 AD-A213670 NTIS
12 1958 AD-298324 NTIS 35 1981 AD-A110870 NTIS
13 1959 AD-298325 NTIS 36 1982 AD-A130811 NTIS
14 1960 AD-246500 NTIS 37 1983 AD-A136673 NTIS
15 1961 AD-265455 NTIS 38 1984 AD-A217381 NTIS
16 1962 AD-285086 NTIS 39 1985 AD-A217404 NTIS
17 1963 AD-423381 NTIS 40 1986 AD-A235435 NTIS
18 1964 AD-450341 NTIS 41 1987 AD-A216858 NTIS
19 1965 AD-471229 NTIS 42 1988 AD-A217275 NTIS
20 1966 AD-800523 NTIS 43 1989 AD-A235629 NTIS
21 1967 AD-659792 NTIS 44 1990 AD-A272017 NTIS
22 1968 AD-844911 NTIS 45 1991 AD-A272274 NTIS
23 1969 AD-746209 NTIS 46 1992 92CH3083-3 IEEE
24 1970 AD-746210 NTIS 47 1993 93CH3244-1 IEEE
25 1971 AD-746211 NTIS 48 1994 94CH3446-2 IEEE
26 1972 AD-771043 NTIS 49 1995 95CH3575-2 IEEE
27 1973 AD-771042 NTIS 50 1996 96CH35935 IEEE
28 1974 AD-A011113 NTIS 51 1997 97CH36016 IEEE
29 1975 AD-A017466 NTIS 52 1998 98CH36165 IEEE
30 1976 AD-A046089 NTIS 53 1999 99CH36313 IEEE
31 1977 AD-A088221 NTIS 54 2000 00CH37052 IEEE
32 1978 AD-A955718 NTIS 55 2001 01CH37218 IEEE
Please check with NTIS or IEEE for current pricing. IEEE members may order IEEE proceedings at half-price.
*NTIS - National Technical Information Service *IEEE - Inst. of Electrical & Electronics Engineers
5285 Port Royal Road, Sills Building 445 Hoes Lane
Springfield, VA 22161, U.S.A. Piscataway, NJ 08854, U.S.A.
Tel: 703-487-4650
Fax: 703-321-8547 Tel: 800-678-4333 or 908-981-0060
E-mail: info@NTIS.fedworld.gov E-mail: customer.services@ieee.org
http://www.fedworld.gov/ntis/search.htm http://www.ieee.org/ieeestore/ordinfo.html
______________________________________________________________________________
Prior to 1992, the Symposium’s name was the “Annual Symposium on Frequency Control,” and in 1992, the name was IEEE Frequency
Control Symposium (i.e., without the “International”).
10-1

Specifications And Standards Relating To Frequency Control - 1
Institute Of Electrical & Electronic Eng
gineers (IEEE) General Specs for:
Order from: IEEE Service Center MIL-C-3098 Crystal Unit, Quartz

445 Hoes Lane
Piscataway, NJ 08854 MIL-C-24523 (SHIPS) Chronometer, Quartz Crystal
Telephone: (732) 981-0060
http://standards.ieee.org/catalog/contents.html MIL-F-15733 Filters & Capacitors, Radio Interference
176-1987 (ANSI/IEEE) Standard on Piezoelectricity MIL-F-18327 Filters, High Pass, Band Pass Suppression and Dual Processing
177-1966 Standard Definitions & Methods of Measurements of Piezoelectric MIL-F-28861 Filters and Capacitors, Radio Frequency Electro-magnetic
Vibrators Interference Suppression
180-1986 (ANSI/IEEE) Definitions of Primary Ferroelectric Crystal Terms MIL-F-28811 Frequency Standard, Cesium Beam Tube
(SH10553)
MIL-H-10056 Holders (Encl), Crystal
319-1971 (Reaff 1978) Piezomagnetic Nomenclature (SH02360)
MIL-O-55310 Oscillators, Crystal
1139-1988 Standard Definitions of Physical Quantities for Fundamental MIL-O-39021 Oven
Frequency & Time Metrology (SH12526)
MIL-S-4933(ER) Surface Acoustic Wave Devices
IEEE Std 1193-1994 (ANSI) IEEE Guide for Measurement of Environmental
Sensitivities of Standard Frequency Generators MIL-STD-683 Crystal Units, Quartz/Holders, Crystal
Department of Defense (DOD) MIL-STD-188-115 Interoperability & Performance Standards for Communications,
Timing & Synchron-ization Subsystems
Order from: Naval Pubs & Form Center
5801 Tabor Avenue MIL-STD-1395 Filters & Networks, Selection & Use
Philadelphia, PA 19120
Telephone: (215) 697-2000 MIL-T-28816(EC) Time Frequency Standard, Disciplined AN/URQ-23
http://www.dscc.dla.mil/Programs/MilSpec/default.asp
http://stinet.dtic.mil/str/dodiss4_fields.html MIL-W-46374D Watch, wrist: General purpose
MIL-W-87967 Watch, wrist: Digital
10-2

US Government Standards EIA-417, Crystal Outlines (standard dimensions and pin connections for current
quartz crystal units, 1974)
FED-STD-1002 Time & Frequency Reference Information in (b) Production Tests

Telecommunication Systems
EIA-186-E, (All Sections) Standard Test Methods for Electronic Component Parts
EIA-512, Standard Methods for Measurement of Equivalent Electrical Parameters
Federal Standard 1037C: Glossary of Telecommunications Terms
http://www.its.bldrdoc.gov/fs- of Quartz Crystal Units, 1 kHz to 1 GHz, 1985
http://www.its.bldrdoc.gov/fs-1037/
IRIG Stdrd 200- EIA-IS-17-A, Assessment of Outgoing Non-conforming Levels in Parts per
200-98 - IRIG Serial Time Code Formats
http://tecnet0.jcte.jcs.mil/RCC/manuals/200/index.htm Million (PPM)
l
EIA-IS-18, Lot Acceptance Procedure for Verifying Compliance with Specified
A source of many standards: Quality Level in PPM
(c) Application Information

American National Standards Institute (ANSI)
1819 L Street, NW
EIA Components Bulletin No. CB6-A, Guide for the
Suite 600
Use of Quartz Crystal Units for Frequency Control,
Washington, DC 20036
Oct. 1987
http://webstore.ansi.org/ansidocstore/default.asp
(d) EIA-477, Cultured Quartz (Apr. 81)
Electronic Industries Association (EIA)
EIA-477-1, Quartz Crystal Test Methods (May 1985)
Order from: Electronic Industries Assoc.
2001 Eye Street, NW International Electro-Technical Commission
Washington, DC 20006 (IEC)
Telephone: (202) 457-4900
Order from: American Nat'l. Standard Inst.
(a) Holders and Sockets (ANSI) 1430 Broadway
New York NY 1001
EIA-192-A, Holder Outlines and Pin Connections for Quartz Crystal Units Telephone: (212) 354-3300
http://webstore.ansi.org/ansidocstore/default.asp
(standard dimensions for holder
types)
EIA-367, Dimensional & Electrical Characteristics Defining Receiver Type IEC Publications Prepared by TC 49:
Sockets (including crystal sockets)
10-3

122: Quartz crystal units for frequency control and
selection 444-1 (1986) Part 1: Basic method for the measurement of resonance frequency
and resonance resistance of quartz crystal units by zero phase technique in a π -
122-2 (1983) Part 2: Guide to the use of quartz crystal units for frequency network with compensation of the parallel capacitance Co
control and selection
444-
444-4 (1988) Part 4: Method for the measurement of the load resonance
resonance
122-2-1 (1991) Section One: Quartz crystal units for microprocessor clock frequency fL, load resonance RL and the calculation of other derived
supply (Amendment 1 - 1993) values of quartz crystal units up 30 MHz
122-3 (1977) Part 3: Standard outlines and pin connection (Amendment 2 - 483 (1976) Guide to dynamic measurements of piezoelectric ceramics with high
1991, Amendment 3 - 1992, Amendment 4 - 1993) electromechanical coupling
283 (1986) Methods for the measurement of frequency and equivalent 642 (1979) Piezoelectric ceramic resonators and resonator units for frequency
resistance of unwanted resonances of filter crystal units control and selection.
Chapter I: Standard Values and Conditions
302 (1969) Standard definitions and methods of measurement for Chapter II: Measuring and test conditions
piezoelectric vibrators operating over the frequency range up to 30 MHz
642-2 (1994) Part 2: Guide to the use of piezoelectric ceramic resonator units
314 (1970) Temperature control devices for quartz crystal units
(Amendment 1 - 1979) 642-3 (1992) Part 3: Standard outlines
314A (1971) First supplement 679: Quartz Crystal Controlled Oscillators
368: Piezoelectric Filters 679-1 (1980) Part 1: General information, test conditions and methods
(Amendment 1 - 1985)
368-l (1992) Part 1: General information, standard values and test
conditions 679-2 (1981) Part 2: Guide to the use of quartz crystal controlled oscillators
368-2 (1973) Part 2: Guide to the use of piezoelectric filters 679-3 (1989) Part 3: Standard outlines and lead connections (First supplement -
1991) (Amendment 1 - 1994)
368-2-1 (1988) Section One - Quartz crystal filters
689 (1980) Measurements and test methods for 32 kHz quartz crystal units for
368B (1975) Second supplement wrist watches and standard values
368-3 (1991) Part 3: Standard Outlines 758 (1993) Synthetic quartz crystal; specifications
and guide for use
444: Measurement of quartz crystal unit parameters
10-4

862: Surface Acoustic Wave (SAW) Filters: 1253: Piezoelectric ceramic resonators - a specification in the IEC quality assessment
system for electronic components (IECQ)
862-1 (1989) Part 1: General Information, standard values and test conditions,
Chapter I: General information and standard values, Chapter II: Test conditions 1253-1 (1993) Part 1: Generic specification - qualification approval
862-2 (1991) Part 2: Guide to the use of surface acoustic wave filters (Chapter 1253-2 (1993) Part 2: Sectional specification - qualification approval
III)
1253-2-1 (1993) Section 1 - Blank detail specification - Assessment Level E
862-3 (1986) Part 3: Standard outlines (Chapter IV)
1261: Piezoelectric Ceramic Filters for use in Electronic Equipment, a specification in
1019: Surface Acoustic Wave (SAW) Resonators the IEC quality assessment system for electronic components (IECQ)
1019-1-1 (1990) Part 1: General information, standard values and test 1261-1 (1994) Part 1: General specifications, qualification approval
conditions, Section 1 - General information and standard values
1261-2 (1994) Part 2: Sectional specifications, qualification approval
1019-1-2 (1993) Section 2: Test conditions
1261-2-1 (1994) Part 2: Section 1, Blank detail specification, Assessment Level E
1019-1-3 (1991) Part 3: Standout outlines and lead connections
1080 (1991) Guide to the measurement of equivalent electrical parameters of
quartz crystal units International Telecommunication Union
1178-1 (1993) Quartz crystal units - a specification in the IEC Quality emissions, List of ITU-
Time signals and frequency standards emissions, ITU-R Recommendations
Assessment System for Electronic Components (IECQ) Part 1: General http://www.itu.int/rec/recommendation.asp?type=products&parent=R-REC-tf
Specification
1178-2 (1993) Part 2: Sectional specification - Capability approval
1178-2-1 (1993) Part 2: Sectional specification - Capability approval, Section 1:

Blank detail specification
1178-3 (1993) Part 3: Sectional specification - Qualification approval
1178-3-1 (1993) Part 3: Sectional specification - Qualification approval, Section

1: Blank detail specification
1240 (1994) Piezoelectric devices - preparation of outline drawings of surface-

mounted devices (MSD) for frequency control and selection, general rules
10-5

Frequency Control Website
A huge amount of frequency control information can be

found at
http://www.ieee-uffc.org/fc
Available at this site are >100K pages of information,

including the full text of all the papers ever published in
the Proceedings of the Frequency Control Symposium,
i.e., since 1956, reference and tutorial information, nine
complete books, historical information, and links to
other web sites, including a directory of company web
sites.
10-6
Some of the information at this website is available only to IEEE UFFC Society members.


Fei Fe-5440a Td1251u

Uploaded by

Copyright:

Available Formats

Fei Fe-5440a Td1251u

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fei Fe-5440a Td1251u

Uploaded by

Copyright:

Available Formats

What are the main components of a Cesium Beam Primary Frequency Standard?

What are the main components of a Cesium Beam Primary Frequency Standard?

What are the output frequencies and accuracy specifications of the FE-5440A Cesium Beam Frequency Standard?

What are the output frequencies and accuracy specifications of the FE-5440A Cesium Beam Frequency Standard?

9/13/2019 FEI Model FE-5440A Cesium Beam Primary Frequency Standard

https://freqelec.com/cesium_fe5440a.html Go MAY SEP OCT 👤 ⍰❎

Vendors & Function Description Technical Characteristics

Meters and Displays

Three Power Sources

Front panel indicators advise of input power availability,

Technical POC: JERRY LONG

NAICS Code: 334419

Contract Specialist Name: Detra A Armstrong

Contract Specialist Phone Number: Information Unavailable

Contract Specialist E-mail Address: Information Unavailable

Current Adjusted Server Time: 3/05/2020 16:28:30 Eastern

Created Date: 07/23/2017

Last Modified By: System Auto Update

Last Modified Date: 07/27/2017

Work sponsored under ESA contract 18040/04/NL/DS

1-4244-0647-1/07/$20.00 ©2007 IEEE 1127

4.00E-009 Toven=130°C, D2:44'σ

3.40E-009 Fig. 4: Measured frequency stability of OSCAR operated

In Fig. 5, we plot the different noise current densities as a

function of the Ramsey peak-valley photocurrent. The

Fig. 3: Current noise density (measured at Ramsey fringe

Bid Protests &

B-204483, APR 5, 1982

B-204483, APR 5, 1982

FREQUENCY ELECTRONICS, INC.:

Quartz Crystal Resonators and

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

The findings in this report are not to be construed as an

The citation of trade names and names of manufacturers

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

IEEE Xplore, <http://www.ieee.org/ieeexplore>, provides full-text access to IEEE transactions,

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

Technology Units Unit price, Worldwide

Atomic Frequency Standards

Hydrogen maser ~ 10 $200,000 $2M

Cesium beam ~ 500 $50,000 $25M

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

Dava Sobel, Longitude, Walker & Co., New York, 1995

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

Analog signal Digitized signal ∆t

J. A. Wepman, “Analog-to-Digital Converters and Their Applications in Radio Receivers,” IEEE

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS

• In AT&T's network, for example, timing is distributed down a hierarchy of

1 1 x 10-11 N.A. GPS W/Two Rb 16

2 1.6 x 10-8 1 x 10-10 Rb Or OCXO ~200

3 4.6 x 10-6 3.7 x 10-7 OCXO Or TCXO 1000’s

4 3.2 x 10-5 N.A. XO ~1 million

J. E. Abate, E. W. Butterline, R. A. Carley, P. Greendyk, A. M. Montenegro, C. D. Near, S. H.

“QUARTZ CRYSTAL RESONATORS AND OSCILLATORS