RF Circuits, Systems, and Wireless Communications Standards

Prof. Jenshan Lin University of Florida

jenshan@ufl.edu or jenshan@ieee.org http://www.rfsoc.ece.ufl.edu/

About the Instructor Joined UF in July 2003. Worked in industry for 9 years.
AT&T/Lucent Bell Labs (1994-2001) Agere Systems (2001-2003)

RF/high-speed circuit/system design for wireless and broadband communications.

antennas, transceivers, standards, system specifications, Using different technologies from silicon to III-V

Professional Activities Senior Member, IEEE Microwave Theory and Techniques Society, Solid State Circuits Society, Communications Society, Engineering in Medicine and Biology Society, Standards Association. MTT Administrative Committee, elected voting member 2006-2008. Members of several conference committees
RFIC TPC Chair 2007 (Hawaii) RFIC General Chair 2008 (Atlanta) RWS Finance Chair 2007-2008 (Long Beach, Orlando) IMS 2007 (Hawaii) Exhibitors Liaison for Taiwan

RWS: Radio and Wireless Symposium. Co-sponsored by MTT and COMSOC. IMS: International Microwave Symposium. Large annual event with exhibition.

Research Projects High-Density 3-D Packaging Technology for RF Devices (Air Force Research Lab) Hydrogen Sensors and System (NASA) Wireless Power Transmission (NASA) Remote Non-contact Cardiopulmonary Monitoring (Vital Signs Sensing) (NSF) Bandwidth-Efficient Modulation for High Speed Interconnect (ITRI) Low Power Phase Shift Modulator for RF Subsystem Research (DARPA)

Workshop Objectives Overall picture of RF/wireless systems The importance of RF design for system integration How to determine RF component-level specifications from wireless communication standards

Workshop Outline Demo of FM Radio Interference RF Overview Antenna and Radio Propagation Wireless Communication Standards RF Specifications from Standards RF Transceiver Architectures Reference Transceiver Design Example Answers to the FM Radio Interference

What is RF?

Radio Frequency Bands - IEEE

IEEE: (A) The frequency in the portion of the electromagnetic spectrum that is between the audiofrequency portion and the infrared portion. (B) A frequency useful for radio transmission. Within this frequency range electromagnetic radiation may be detected and amplified as an electric current at the wave frequency. (loosely defined)

Radio Frequency Bands - FCC

FCC: The radio spectrum is the part of the natural spectrum of electromagnetic radiation lying between the frequency limits of 9 kHz and 300 GHz.

Military Radar Bands (IEEE Standard 521-1984)

Radar Band HF VHF UHF L S C X Ku K Ka V W Frequency 3-30 MHz 30-300 MHz 300-1000 MHz 1-2 GHz 2-4 GHz 4-8 GHz 8-12 GHz 12-18 GHz 18-27 GHz 27-40 GHz 40-75 GHz 75-110 GHz

ITU Bands

Millimeter-wave (mm-wave) ITU: International Telecommunication Union, headquartered in Geneva, Switzerland is an international organization within the United Nations System where governments and the private sector coordinate global telecom networks and services.

Definition of Microwave
American Heritage Dictionary: A high-frequency electromagnetic wave, one millimeter to one meter in wavelength, intermediate between infrared and shortwave radio wavelengths. 300MHz to 300GHz IEEE Standard Dictionary: A term used rather loosely to signify radio waves in the frequency range from about 1GHz upwards. NTIA (National Telecommunications and Information Administration): Loosely, an electromagnetic wave having a wavelength from 300 mm to 10 mm (1 GHz to 30 GHz).

So, what is RF? What is Microwave? What is Millimeter-wave?

Household Electronics

Equipment AM radio Short Wave (SW) radio FM radio TV/Cable Ch. 2-83 GPS Cordless phone Cellular phone Wireless LAN Microwave Oven

Frequency Range 535-1605 kHz 3-30 MHz 88-108 MHz 54-890 MHz L1: 1,575.42MHz, L2: 1,227.60MHz 46, 49, 900MHz, 2.45GHz 800-900MHz, 1.8-2 GHz ~2.45, ~5.2GHz ~2.45GHz

Why use RF?

Baseband transmission v.s. RF transmission Through the wire: no problem, as long as the signal bandwidth does not exceed the wires bandwidth limit. Through the air: How? Need antenna. How big is the antenna? Hint: dipole antennas length is usually . Need to modulate on to a high frequency carrier. Frequency Antenna size

Why use Microwave? Why use Millimeter-wave?

Why use integrated circuits?

RF component evolution Waveguide component Connectorized module RFIC package

Acronyms
RFIC: Radio Frequency Integrated Circuit MMIC: Monolithic Microwave Integrated Circuit MIC: Microwave Integrated Circuit Integrate microwave transistors, passive elements, and transmission lines on a microwave substrate. A departure from old microwave waveguide components.

10

Who control RF spectrum? (Spectrum Regulation)

Federal Communications Commission (FCC)

FCC is an independent United States government agency, directly responsible to Congress. The FCC was established by the Communications Act of 1934 and is charged with regulating interstate and international communications by radio, television, wire, satellite and cable. The FCC's jurisdiction covers the 50 states, the District of Columbia, and U.S. possessions.

http://www.fcc.gov

11

FCC Organizations
Five Commissioners appointed by the President and confirmed by the Senate for 5-year terms, except when filling an unexpired term. The Commission staff is organized by function. There are six operating Bureaus and ten Staff Offices. The Bureaus responsibilities include: processing applications for licenses and other filings; analyzing complaints; conducting investigations; developing and implementing regulatory programs; and taking part in hearings. The Offices provide support services.

12

FCC Wireless Telecommunications Bureau The Wireless Telecommunications Bureau (WTB) handles nearly all FCC domestic wireless telecommunications programs and policies. Wireless communications services include Amateur, Cellular, Paging, Broadband PCS, Public Safety, and more.

Cellular and Broadband PCS Cellular band: 824-849 MHz and 869-894 MHz. PCS band: 1850-1990 MHz.

13

FCC Radio Spectrum

In the United States, regulatory responsibility for the radio spectrum is divided between the FCC and the National Telecommunications and Information Administration (NTIA). FCC, administers spectrum for non-Federal government use and the NTIA, which is an operating unit of the Department of Commerce, administers spectrum for Federal government use. Within the FCC, the Office of Engineering and Technology (OET) provides advice on technical and policy issues pertaining to spectrum allocation and use. OET also maintains the FCC's Table of Frequency Allocations. http://www.fcc.gov/oet/spectrum/

FCC and ITU

The International Telecommunication Union (ITU), headquartered in Geneva, Switzerland is the international organization within which governments coordinate global telecom networks and services. The United States is a member of the ITU. The ITU maintains the [International] Table of Frequency Allocations, which is reproduced in columns 1-3 of the FCC's Table of Frequency Allocations. The FCC and the NTIA assist the Department of State in developing U.S. proposals to revise the ITU's Table of Frequency Allocations.

14

FCC Rules CFR 47

FCC rules are located in Title 47 of the Code of Federal Regulations. Although the Office of Engineering and Technology (OET) is responsible for the maintenance of FCC rules located in Parts 2, 5, 15, and 18 of the Title, the official rules are published and maintained in the Federal Register. http://wireless.fcc.gov/rules.html Part 2: Frequency allocations and radio treaty matters; general rules and regulations Part 5: Experimental radio service (other than broadcast) Part 15: Radio frequency devices Part 18: Industrial, scientific, and medical equipment

ISM Bands
18.107 Definitions. (c) Industrial, scientific, and medical (ISM) equipment. Equipment or appliances designed to generate and use locally RF energy for industrial, scientific, medical, domestic or similar purposes, excluding applications in the field of telecommunication.

15

FCC Safety

FCC is required by the National Environmental Policy Act of 1969 to evaluate the effect of emissions from FCC-regulated transmitters on the quality of the human environment. At the present time there is no federally-mandated radio frequency (RF) exposure standard. However, several non-government organizations, such as the American National Standards Institute (ANSI), the Institute of Electrical and Electronics Engineers, Inc. (IEEE), and the National Council on Radiation Protection and Measurements (NCRP) have issued recommendations for human exposure to RF electromagnetic fields.

RF Safety Guidelines (old)

INIRC: International Non-Ionizing Radiation Committee

16

IEEE RF Safety Guideline (new)

Controlled Environment industry Uncontrolled Environment public

Reference: http://www.arrl.org/news/rfsafety/hbkrf.html

RF Safety some numbers

Sets limit from 3kHz to 300GHz. In general, controlled environment allows higher power than uncontrolled environment (public) except for >12GHz and {<20MHz(H-field) or <1.2MHz(E-field)}. Above 100MHz, safety limit for E-field and H-field radiation are the same. Below 100MHz, E-field radiation has lower power density limit than H-field radiation. Most stringent limit is 0.2mW/cm2 between 30MHz100MHz, in public environment.

17

RF Safety low frequency

At 10MHz, the limit is 2mW/cm2 for E-field and 100mW/cm2 for H-field. At 1MHz, the limit is 100mW/cm2 for E-field and 10W/cm2 for H-field. For inductive coupling, e.g., RFID, H-field is dominant. The RF penetration depth (field intensity drops 50% from boundary) of muscle tissue is ~15cm@10MHz, ~5cm@100MHz.

Questions: How to do the conversion between field intensity and power density? Why the RF safety sets the lowest limit for 30MHz-300MHz?

18

Antenna

What is an antenna? An antenna is a device for radiating or receiving radio waves. A transitional structure between free space and a guiding device. It converts radiated waves into guided waves, or vice versa.

Antenna Circuit

19

Antenna examples Dipole antenna Monopole antenna Patch antenna Horn antenna Reflector antenna (dish, parabolic) Yagi-Uda Antenna (VHF, UHF TV) Phase array antenna

Important notes about antenna Antenna is reciprocal.

An antenna can be used for either transmitting or receiving. Therefore, in general, only one antenna is needed for the wireless transceiver (cellular phone, WLAN), if the transmit and receive frequencies are close enough and within antenna bandwidth.

20

Antenna radiation pattern

A graphical representation of the radiation intensity of the antenna as a function of space coordinates, in most cases, directional coordinates (angle).

3-D radiation pattern

21

Isotropic antenna

A hypothetical antenna having equal radiation in all directions. However, in system design and calculation, isotropic antenna is usually used.

Directional antenna

Having the property of radiating or receiving signal more effectively in some directions than in others. Non-isotropic.

22

Omnidirectional antenna

Having a nondirectional pattern in azimuth and a directional pattern in elevation. Example: Dipole

E-plane and H-plane patterns E-plane pattern=the plane containing the electricfield vector and the direction of maximum radiation. H-plane pattern= the plane containing the magnetic-field vector and the direction of maximum radiation.

23

E-plane and H-plane patterns

z

y z x y

E-plane

H-plane

Polarization Polarization of an antenna=the polarization of the radiated wave, when the antenna is excited.
Linear polarization Circular polarization
RHCP (right-hand circular polarization) LHCP (left-hand circular polarization)

Elliptical polarization

24

Far-field and Near-field Regions

Far-Field Region: Angular field distribution (antenna pattern) is essentially independent of the distance from the antenna. D = largest dimension of the antenna

Far-field region R > 2D2/ R >> D for D >> for D ~

Near-field region

D R

Radiation Intensity
Prad r

U = radiation intensity (W/unit solid angle) Wrad = radiation density (W/m2) Radiation intensity of an isotropic source

U = r 2 Wrad

U0 = r 2
U= EIRP 4

P = rad 4 4 r 2
EIRP = 4 U = 4 r 2 Wrad

Prad

Effective Isotropic Radiated Power

25

Directivity Directivity Gain=the ratio of the radiation intensity in that direction to the radiation intensity of an isotropic antenna
Dg ( , ) = U ( , ) 4 U ( , ) EIRP( , ) = = U0 Prad Prad

Directivity=the maximum directivity gain

U 4 U max D = max = U0 Prad

Antenna gain and efficiency Antenna gain=the ratio of EIRP to the input power EIRP 4 U ( , )
G=

Antenna efficiency=the ratio of total radiated power to the input power

P e = rad Pin G ( , ) = e 4 U ( , ) = e Dg ( , ) Prad

Pin

Pin

26

Dipole Antenna
Transmission line

Directivity = 1.5 for short dipole (1.76 dB) Directivity = 1.64 for dipole (2.15 dB)

Antenna Efficiency

Loss
Reflection due to mismatch between transmission line and antenna I2R loss (conduction and dielectric)

Total overall efficiency:

et = er ec ed = (1 2) ecd
Reflection Conduction dielectric Conduction-Dielectric efficiency

27

Antenna input impedance The impedance measured at the input port of the antenna.
loss RL
ecd

Z A = Rr + RL + j A
XA

Rr

Radiation resistance

Antenna reactance Rr Power delivered to Rr ecd = = Power delivered to Rr and RL Rr + RL

Radiation resistance of dipole

L

L 2 2 L 2 Infinitesimal dipole L</50 Rr = 80 ( ) = 790 ( )

Short dipole L</3 dipole L=/2

L L Rr = 20 2 ( ) 2 = 197 ( ) 2

Rr = 73

28

Antenna aperture Effective aperture. Could be much larger than the actual physical aperture. The ratio of the power delivered to the load to the incident power density.

ZT

I 2R /2 P Ae = T = T T Wi Wi

ZT = RT + jT

Load

Antenna

IT RL RT XT XA VT +

Rr

29

 V 2 RT Ae = T 2Wi ( Rr + RL + RT ) 2 + ( X A + X T ) 2

Under conditions of maximum power transfer (conjugate matching):

Rr + RL = RT X A = XT Maximum effective aperture

VT 2 1 V 2 RT Aem = T = 8Wi ( RL + Rr ) 2 8Wi RL + Rr

Maximum effective aperture of an infinitesimal dipole with RL=0

V 2 1 Aem = T 8Wi Rr

Antenna length E-field of incident wave

dipole is very short Incident power density

Aem = ( EL) 2

VT = E L

Induced voltage
Wi =

E 2
= 32 8

= 377 (120)
L

8( E 2 / 2 )(80 2 L2 / 2 )

30

A1 Dg1 A1m D1
Antenna 1 TX

A2 A2m

Dg 2 D2
RX

Antenna 2

Pt

R
Dg = U ( , ) Wi = W0 U0

Pr

remember

Wi Radiation density (W/m2)

Isotropic

Wi = W0 Dg1 = Pr = Wi Ar =

Pt Dg1

4R 2 Pt Dg1 A2 4R 2

P Dg1 A2 = r (4R 2 ) Pt

A2 A2m
TX

Dg 2 D2

A1 Dg1 A1m D1
Antenna 1 RX

Antenna 2

Pt

Pr

Switch Antenna 1 and Antenna 2

Wt Radiation density (W/m2) Pt Dg 2

Wi = W0 Dg 2 = Pr = Wi Ar =

4R 2 Pt Dg 2 A1 4R 2

Notes: 1. Antenna is reciprocal. 2. Transmission medium is linear and passive.

P Dg 2 A1 = r (4R 2 ) Pt

31

P Dg1 A2 = r (4R 2 ) Pt P Dg 2 A1 = r (4R 2 ) Pt Dg1 Dg 2 = A1 A2 D1 D = 2 A1m A2m

If antenna 1 is isotropic,

D1 = 1

A A1m = 2m D2
A2m = 32 8

If antenna 2 is infinitesimal dipole,

D2 = 1.5

A1m =

2 4

Aem =

2 D for any antenna 4

(Maximum) effective aperture of isotropic antenna

Consider the loss (antenna efficiency)

G = e D

Aem =

2 2 eD = G 4 4

Gt Antenna 1
TX

Gr Antenna 2

Pt

Pr

RX

If maximum directivity aligned:

Wt Radiation density (W/m2)

PG A Pr = Wt Ar = t t rm 4R 2
Gt 2 2 Pr Gt Arm = = ( )Gr = ( ) Gt Gr 4R Pt (4R 2 ) (4R 2 ) 4 Pr 2 =( ) Gt Gr Friis Transmission Equation 4R Pt

Free Space Path Loss Equation

32

A closer look at finite length dipole

Dont forget the IM part!

Dipoles radiation pattern, directivity, and radiation resistance

33

Cellular Base Station Antenna

Why the antennas were designed this way?

Antenna Array Radiation pattern from a single antenna is relatively wide. Notice that directivity increases as antenna size (electrical size) increases.
2L Dipole antenna L

34

Antenna Array Another way to increase antenna size is to repeat single elements array.

Far-field

Far-field

Antenna Array

y z

Radiation Pattern ?

35

Cell Sectors

2 3 120 120 120

Advantages of Antenna Array Higher directivity (higher gain). Control the phase to steer the beam phased array. Control the phase and amplitude beam forming array.

36

RF Propagation

RF Propagation
Free space path loss equation: Relates the power received to the power transmitted between two antennas separated by a distance R>2D2/ (far field)

Sometimes called Line of Sight (LOS) Reflection, Diffraction, Scattering will change the propagation loss complicated models. Either higher or lower loss is possible.

Pr 2 =( ) Gt Gr Pt 4R

37

Free Space Path Loss

Pr =

2 4 4R 2 2 = Pt Gt Gr 4R
Pt Gt Gr
4R

2 (Pr )dBm = (Pt )dBm + (Gt )dB + (Gr )dB + 10 log

Far-field

10 log <0 4R 4R 4R 10 log = 20 log >0

2

Free Space Path Loss

RF attenuation
Notice the peak and dip of O2 curve at 60GHz and 100GHz, respectively.

dB/km

38

dB

94GHz

Attenuation due to rain

Problem at high frequency, e.g., satellite TV.

39

Ground Reflection

d1

d2

= d 2 d1 = (ht + hr ) 2 + d 2 (ht hr ) 2 + d 2 2h h Assuming d>> h t r d

Phase difference due to time delay:

c = c = 2

40

Wi =

E2 Pt Gt = 2 4r 2

E=

2Pt Gt 1 Pt Gt E0 d 0 = 4r 2 r 2 r

E d j (t d1 ) E d j (t dc2 ) Etotal = 0 0 e c c + 0 0 e c d1 d2
E d j (t d1 ) j ( d1 d 2 ) Etotal = 0 0 e c c [1 + e c c c ] d

E d j (t d1 ) j c ( ) c ] Etotal = 0 0 e c c [1 + e d E d j (t d1 ) Etotal = 0 0 e c c [1 + (cos j sin )] d

1 E d Etotal = 0 0 [(1 + cos ) 2 + ( sin ) 2 ] 2 d

1 E d Etotal = 0 0 [(1 + cos ) 2 + ( sin ) 2 ] 2 d

Wi =

Etotal 2 Pt Gt = [(1 + cos ) 2 + ( sin ) 2 ] 2 4d 2 Wi = Etotal 2 Pt Gt = (1 + 2 cos + 2 ) 2 4d 2

=1

PG PG Wi = t t (1 + cos ) = t t 2 cos 2 ( ) 2 2 2 2d 2d PG PG Wi = t t (1 cos ) = t t 2 sin 2 ( ) 2 2d 2 2d 2

= 1

41

 =1
Assuming d>> h

PG Wi = t t cos 2 ( ) 2 d 2 PG Wi t t d 2
P G 2 2 Gr = Pt Gt Gr ( ) Pr = Wi Aem = t t 2d d 2 4

4 times of free space

= 1
Assuming d>> h

PG Wi = t t sin 2 ( ) 2 2 d
2 hh PG PG PG Wi t t ( ) 2 = t t ( 2 t r ) 2 = t t 42 ( ht hr ) 2 4 2 d 2 2 d d d

P G 4 PG G 2 Pr = Wi Aem = t t (h h ) 2 G = t t r (ht hr ) 2 4 2 t r 4 r d d4 40dB/decade Higher path loss than free space disappeared Independent of frequency at large d

Building a radio system

TX

Channel

RX

Very weak signal. To process the signal at baseband, we need large signal while trying to maintain the best signal to noise ratio.

42

Wireless Communication Standards

Wireless v.s. Wire

Media
RF Wireless

Bandwidth
kHz, MHz, GHz

Example
Cellular, WLAN, UWB

Twisted Pair(Wire)

kHz MHz MHz

phone, ISDN DSL, Ethernet analog video, digital

Cable(Wire)

Optical Fiber

GHz

Very high speed data

43

Wireless Communication Standards

Mobile Communication licensed bands Wireless LAN unlicensed bands Bluetooth - unlicensed bands Fixed Wireless Link (MMDS, LMDS) licensed bands UWB (Ultra Wide Band) unlicensed band Others

Wireless Communications Standards

Analog Cellular Telephones
Standard AMPS/NAMPS Narrow Band Advanced Mobile Phone System Rx: 869-894 Tx:: 824-849 TACS Total Access Communication System ETACS: Rx: 916-949 Tx: 871-904 NTACS: Rx: 860-870 Tx: 915-925 FDMA FDD ETACS: 1240 NTACS: 400 ETACS: 25 kHz NTACS: 12.5 kHz FM n/a NMT Nordic Mobile Telephone NMT-450: Rx: 463-468 Tx: 453-458 NMT-900 Rx: 935-960 Tx:: 890-915 FDMA FDD NMT-450: 200 NMT-900: 1999 NMT-450: 25 kHz NMT-900: 12.5 kHz FM n/a

Mobile Frequency Range (MHz)

Downlink/Uplink

Multiple Access Method Duplex Method Number of Channels Channel Spacing Modulation Channel Bit Rate

FDMA FDD AMPS: 832 NAMPS: 2496 AMPS: 30 kHz NAMPS: 10 kHz FM n/a

44

Wireless Communications Standards

Digital Cellular Telephones
TDMA CDMA GSM IS-54/136 Time Division IS-95 Code Division Global System for Multiple Access Multiple Access Mobile Communications Rx: 869-894 Rx: 869-894 Rx: 935-960 Tx:: 824-849 Tx:: 824-849 Tx:: 890-915 DCS 1800 Digital Communication System Rx: 1805-1880 Tx:: 1710-1785 PDC Personal Digital Cellular Rx: 810-826 Tx:: 940-956 Rx: 1429-1453 Tx:: 1477-1501 TDMA/FDM

TDMA/FDM

CDMA/FDM

TDMA/FDM

TDMA/FDM

FDD

FDD

FDD

FDD

FDD

832 (3 users/channel) 30 kHz /4 DQPSK

20 (798 users/channel) 1250 kHz

124 (8 users/channel) 200 kHz

374 (8 users/channel) 200 kHz

1600 (3 users/channel) 25 kHz /4 DQPSK

QPSK/OQPSK

GMSK (0.3 Gaussian Filter) 270.833 kb/s

GMSK (0.3 Gaussian Filter) 270.833 kb/s

48.6 kb/s

1.2288 Mb/s

42 kb/s

Multiple access/Multiplexing

Wireless Communications Standards

Personal Communication Systems

PCS Rx: 1930-1990 Tx: 1850-1910 PCS TDMA (based on IS-136 cellular) PCS CDMA (based on IS-95 cellular) PCS 1900 (based on GSM cellular) Wideband CDMA

45

Downlink and Uplink

Downlink Uplink

Downlink: mobile RX, base station TX Uplink: mobile TX, base station RX RX: receive TX: transmit

Multiple Access v.s. Multiplexing

Multiple Access: Frequency Division Multiple Access (FDMA) Time Division Multiple Access (TDMA) Code Division Multiple Access (CDMA) Space Division Multiple Access (SDMA), Polarization Division Multiple Access (PDMA) Carrier Sense Multiple Access (CSMA) Remote sharing of a communication resource, e.g. satellite, base station. Multiplexing: Frequency Division Multiplexing (FDM) Time Division Multiplexing (TDM) Sharing a communication resource at a local site, e.g. circuit board, antenna.

46

Example: GSM
FDM 935.2 MHz 935.4 MHz 890.2 MHz 890.4 MHz Ch. 1 Ch. 2 Ch. n

TDMA - 8 users/channel
0.577mS

BTS Each transmit/receive unit carries one channel Multiplexing: sharing a fixed or slowly changed communication resource.

4.615mS

Mobile

Multiple Access: sharing a dynamically changed communication resource.

BTS (Basestation Transceiver System)

Duplex
Simplex: Provide only one-way communication. Half Duplex: Provide two-way communication by using the same radio channel for both transmission and reception, but the user can only either transmit or receive at a time. An example is the two-way radio you can buy from Radio Shack. Full Duplex: Simultaneous two-way communication, and the user can transmit and receive at the same time. An example is the cellular phone. Typically on two different channels (FDD) but some new digital systems are using TDD in which transmission and reception share a single radio channel. FDD: Frequency Division Duplexing, TDD: Time Division Duplexing How does a TDD work in Full Duplex? When radio channels data rate is much higher than users data rate, digital radio can store information bursts in-time to appear like full duplex to end user. Sensitive to timing, so only for in-door or small area applications like cordless or wireless LAN.

47

FDD v.s. TDD

Receiver FDD Duplexer Transmitter

Receiver TDD Switch Transmitter

Wireless Communications Standards

Analog Cordless Telephones
CT0 Cordless Telephone 0 2/48 (U.K.), 26/41 (France) 30/39 (Australia) 31/40 (The Netherlands, Spain) 46/49 (China, S. Korea, Taiwan, USA) 48/74, 45/48 (China) FDMA JCT Japanese Cordless Telephone 254/380 CT1/CT1+ Cordless Telephone 1 CT1: 914/960 CT1+: 885/932

FDMA

FDMA

FDD

FDD

FDD

10, 12, 15, 20 or 25

89

CT1: 40 CT1+: 80 25 kHz

1.7, 20, 25 or 40 kHz

12.5 kHz

FM

FM

FM

n/a

n/a

n/a

48

Wireless Communications Standards

Digital Cordless Telephones
CT2/CT2+ Cordless Telephone 2 CT2: 864/868 CT2+: 944/948 TDMA/FDM DECT Digital European Cordless Telephone 1880-1900 PHS Personal Handy Phone System 1895-1918

TDMA/FDM

TDMA/FDM

TDD

TDD

TDD

40

10 (12 users/channel) 1.728 MHz

300 (4 users/channel) 300 kHz /4 DQPSK

100 kHz

GFSK (0.5 Gaussian Filter) 72 kb/s

GFSK (0.5 Gaussian Filter) 1.152 Mb/s

384 kb/s

Mobile Communication Standards

49

3G Standard(s)

North America

Europe

Asia/Pacific

Digital Modulation

50

Digital Modulation
Amplitude Shift Keying (ASK) (analog counterpart: AM)

Frequency Shift Keying (FSK) (analog counterpart : FM)

Phase Shift Keying (PSK) (analog counterpart : PM)

QPSK Modulator

I channel

Q channel

51

How QPSK data is encoded

I channel

Q channel

Reference: Sklar, Digital Communications

Signal Space Diagram (Constellation)

(dI, dQ)
(1,1)

(1,-1)

(-1,-1)

(-1,1)

Reference: Sklar, Digital Communications

52

How OQPSK data is encoded

I channel

Q channel

Also called staggered QPSK (SQPSK)

Reference: Sklar, Digital Communications

QPSK and OQPSK waveforms

Non-constant envelope

QPSK

OQPSK

No transition through zero

Reference: Sklar, Digital Communications

53

How MSK data is encoded I

Reference: Sklar, Digital Communications

How MSK data is encoded Q

Combined I and Q waveform:

No discontinuous phase transition!

Reference: Sklar, Digital Communications

54

Impact on Power amplifier

Linear PA is needed for non-constant envelope modulation. Nonlinear PA can be used for constant envelope modulation. Nonlinear PA has higher efficiency. Higher efficiency lower DC power consumption

TX PA

Pout (dBm)

Saturation Nonlinear

Pin (dBm)

Comparison of QPSK, OQPSK, and MSK spectra

Normalized power spectral density G(f) (dB)

Normalized frequency offset from carrier, (f-fc)/R (Hz/bit/s) MSK is spectrally more efficient than QPSK/OQPSK.

Reference: Sklar, Digital Communications

55

Modulator

Baseband

RF
Reference: Sklar, Digital Communications

DQPSK (Differential QPSK)

DPSK: Differentially coherent detection of differentially encoded PSK. In coherent detection of PSK, signal is compared to a reference synchronized in phase. Phase has to be aligned. DPSK is differentially coherent detection does not require phase synchronization, but compare to previous phase. DPSK requires higher bit energy noise ratio than PSK to keep the same error rate (~3dB higher for 10-1, ~+2dB higher for 10-2 10-3, ~1dB higher for < 10-5). Reference:
Sklar, Digital Communications, p.149, p.160. Couch, Digital and Analog Communication systems, p.344.

56

Reference: Couch, Digital and Analog Communication Systems, 6th ed.

Bit Error Rate (BER) Performance

Reference: Sklar, Digital Communications

57

FSK
FSK can be non-coherently detected, in addition to coherent detection. No phase synchronization in any kind is needed for noncoherent detection. Non-coherent detection FSK receiver is easy to implement lower complexity, lower cost, lower power. Reference:
Sklar, Digital Communications, p. 151. Couch, Digital and Analog Communication systems, p. 352.

Reference: Sklar, Digital Communications

58

QAM (Quadrature Amplitude Modulation) APK (Amplitude Phase Keying) combination of ASK and PSK, QASK (Quadrature Amplitude Shift Keying) ASK in 2-D

cos(0t )
1, 3

y x

1, 3

x
sin(0t )

Pulse Shaping and Baseband filter

Reduce Inter-Symbol Interference (ISI) Where does it come from?

Reference: Couch, Digital and Analog Communication Systems, 6th ed.

59

Raised Cosine Filter

Wireless LAN Standards

60

Wireless Communications Standards

Wireless Data (WAN/LAN)
CDPD Cellular Digital Packet Data (WAN) Rx: 869-894 Tx: 824-849 RAM - Mobitex (WAN) Ardis - RD - LAP (WAN) Rx: 851-869 Tx: 806-824 IEEE 802.11 Wireless LAN (North America/Europe) 2400-2483 (Japan) 2470-2499 CSMA

FDMA

(North America) Rx: 935-941 Tx: 869-902 (European/Asia) 403-470 TDMA/FDM

TDMA/FDM

FDD

FDD

FDD

TDD

832

480

720

FHSS: 79 DSSS: 11 FHSS: 1 MHz DSSS: 11 MHz FHSS: GFSK (0.5 Gaussian Filter) DSSS: DBPSK (1Mb/s DQPSK (2 Mb/s) 1 or 2 Mb/s

30 kHz

12.5 kHz

25 kHz

GMSK (0.5 Gaussian Filter)

GFSK (0.5 Gaussian Filter)

FSK (2 and 4 level)

19.2 kb/s

8 kb/s

19.2 kb/s

Evolved into a, b, g

Wireless LAN Standards

Standard Frequency Range (GHz) IEEE 802.11b 2.4-2.4835 (US, Canada, Europe) 2.471-2.497 (Japan), 2.4465-2.4835 (France), 2.445-2.475 (Spain)* CSMA/CA TDD 5 BPSK, QPSK, CCK Modulation Data Rate(Mb/s) Web site 1, 2, 5.5, 11 IEEE 802.11a US U-NII 5.15-5.35, 5.7255.825 (300MHz) ETSI HIPERLAN/2 5.15-5.35, 5.4705.725 (455MHz) Japan MMAC (100MHz)

Multiple Access Duplex Channel Spacing (MHz)

CSMA/CA TDD 20 OFDM/ BPSK, QPSK, 16QAM, 64QAM 6, 9, 12, 18, 24, 36, 48, 54
http://standards.ieee.org/ getieee802/802.11.html

TDMA (centrally scheduled) TDD 20 OFDM/ BPSK, QPSK, 16QAM, 64QAM 6, 9, 12, 18, 27, 36, 54
http://www.etsi.org/techn icalactiv/hiperlan2.htm http://www.arib.or.jp/ mmac/

http://standards.ieee.org/get ieee802/802.11.html

61

History of Wireless LAN

Research and standardization activities started in late 80s, blossomed in late 90s. IEEE Workshop on Wireless LANs began in 1991.
the oldest IEEE workshop in wireless broadband local and ad-hoc networks.

AT&T WaveLAN and Motorola Altair were the first two commercial wireless LAN products.
AT&T WaveLAN was the first one capable of 2Mb/s

Hollemans and Verschoor, Performance study of WaveLAN and Altair radio-LANs Personal, Indoor and Mobile Radio Communications, 1994. Wireless Networks Catching the Mobile Future. 5th IEEE International Symposium on , Volume: 3 , 18-23 Sep 1994, Page(s): 831 -837 vol.3

62

IEEE 802.11 standards

IEEE:
Institute of Electrical and Electronics Engineers

802.11:
Family of standards by the IEEE to define the specifications for wireless LANs Defines:
Medium Access Control (MAC) Physical Layer (PHY) Specifications

IEEE 802.11 and the ISO Protocol stack

63

IEEE 802.11b b in IEEE 802.11b

High Rate amendment by the IEEE in September 1999. Affects the physical layer, basic architecture is the same
Added two higher speeds
5.5 and 11 Mbps

More robust connectivity

802.11b has been more popular than 802.11a

also known as Wi-Fi (Wireless Fidelity)

IEEE 802.11b Uses Direct Sequence Spread Spectrum (DSSS) or Frequency Hopping Spread Spectrum (FHSS) technology Operates in unlicensed 2.4 GHz unlicensed band
2.4 GHz to 2.4835 GHz for North America

64

IEEE 802.11b Supported data rates and distances

1, 2, 5.5, 11 Mbps at distances of 150-2000 feet without special antenna Greater distances can be achieved by using special antennas Distance (or signal strength) greatly depends on obstructions such as buildings and other objects Maximum data rate obtained depends on signal strength

IEEE 802.11a Date rate of up to 54 Mbps 5 GHz (U-NII band) instead of 2.4 GHz
Unlicensed National Information Infrastructure

OFDM instead of DSSS for encoding

Orthogonal Frequency Division Multiplexing

65

IEEE 802.11a Advantages

higher data rate less RF interference than 2.4 GHz
2.4 GHz used by Bluetooth, cordless/cellular phones, microwave ovens, other unlicensed radios, etc.

Disadvantages
shorter range, need to increase Access Point (AP) density or power to compensate the higher free space path loss (same antenna gain as 2.4 GHz).

IEEE 802.11g Another high data rate standard but still works at 2.4 GHz Moving 802.11a into 2.4 GHz band Data rate of up to 54 Mbps Advantages
RF compatible with 802.11b better range than 802.11a

Eventually, 802.11b/g devices dominate.

66

References
References for WLAN history are posted on the website. Most of them you can download from IEEE website. Except Tuch, Development of WaveLAN, an ISM band wireless LAN, AT&T Technical Journal, July/August 1993. Wireless LAN standards are also posted on class website.

Wireless LAN Physical Layer

67

IEEE 802.11b Physical Layer Spread Spectrum

spreads the transmitted signal over a wide range of spectrum After de-spreading, interferences are suppressed. 2 major approaches to spread spectrum:
FHSS: Frequency Hopping Spread Spectrum DSSS: Direct Sequence Spread Spectrum

IEEE 802.11b Physical Layer FHSS

hop to other frequencies at a fixed time interval using a predetermined sequence hopping allows the system to avoid interferences

DSSS
a different approach: artificially broaden the bandwidth needed to transmit a signal by modulating the data with a spreading code allows for error detection

68

Spread Spectrum (IS-95)

IEEE 802.11b Physical Layer 1Mb/s and 2Mb/s

modules the data with an 11-bit sequence called the Barker code
10110111000

modulated sequence is a series of data objects called chips chips are sent out by the wireless radio
wireless radio modulates a 2.4 GHz carrier modulation techniques: DBPSK, DQPSK 22MHz BW (DSB, null-to-null)

69

IEEE 802.11b Physical Layer

Data
1 Mb/s Information 1 0 11 Mb/s Barker Code Spread Signal 1 0 1 0 1 1 0 1 1 0 1 1 1 0 0 0 0 1 1 1 0 0 0 1 0 1 0 0 1 1 0 1 0 0 0 80211b1.MCD 1 1 1 0 0 0 1 1 1

Chips

70

IEEE 802.11b Physical Layer

5.5Mb/s and 11Mb/s
Use CCK (Complimentary Code Keying) modulation Spreading code length is 8-complex-chip, not 11-bit The 8-complex-chip code word pattern changes as data pattern changes. The encoding is based on QPSK. Essentially a rotating pattern on QPSK constellation just like turning the combination lock on safe. At 5.5Mb/s, each symbol (each 8-complex-chip code word) contains 4 bits of data. At 11Mb/s, each symbol (each 8-complex-chip code word) contains 8 bits of data. The chip rate is always 11Mchips/s, producing 22MHz RF BW null-to-null.

CCK

71

72

IEEE 802.11b

FCC: US IC: Canada ETSI: Europe

73

IEEE 802.11b Channel Agility Option

Frequency Hopping Channels for High Rate 5.5Mb/s and 11Mb/s

Non-overlapping:

Half-overlapping: enable interoperability with 1Mb/s and 2Mb/s FH systems

IEEE 802.11b

74

IEEE 802.11b

The measurements shall be made using a 100 kHz RBW and 100kHz VBW

OFDM Concept
Orthogonal Frequency Division Multiplexing

A special case of multi-carrier transmission. Split a high-rate datastream into a number of lower rate streams that are transmitted simultaneously over a number of subcarriers. Subcarrier frequencies are orthogonal. Can be viewed as either a modulation technique or a multiplexing technique. Robustness against frequency selective fading. Used in IEEE 802.11 wireless LAN a and g versions. Use PSK or QAM on each subcarrier. Reference: Van Nee and Prasad, OFDM For Wireless Multimedia Communications

75

IEEE 802.11a

76

IEEE 802.11a

IEEE 802.11a

77

HIPERLAN/2

78

Standard Mobile Frequency Range

802.11a 5.150 5.250 GHz (USA U-NII Lower band, channels 36,40,44, and 48 25 mW/MHz Max. Tx 5.250 5.350 GHz (USA U-NII Middle band, channels 52, 56, 60,and 64 12.5 mW/MHz, Max. Tx 5.725 5.825 GHz (USA U-NII Upper band, channels 149,153,157,161 50 mW/MHz Max Tx

802.11b 2.41-2.462 GHz (N. America, 11 channels 1000 mW Power Allowance) 2.412-2.472 GHz (Europe, 13 channels 100 mW Power Allowance) Japan: 2.484 GHz (1 channel 10mW/MHz Power Allowance)

802.11g Same band as 802.11b

Multiple Access Method Duplex Method Users per Channel Channel Spacing

CSMA/CA TDD 127 OFDM: 20 MHz

CSMA/CA TDD 127 US 25 MHz non overlapping (3 channels), 10 MHz, overlapping (6 channels) Europe 30 MHz non overlapping (3 channels), 10 MHz, overlapping (6 channels) DSSS DBPSK (1 Mb/s) DSSS DQPSK (2 Mb/s) Complimentary code keying (CCK): QPSK (5.5 and 11 Mb/s)

Same as 802.11a Same as 802.11a Same as 802.11a Same as 802.11a

Modulation

OFDM: QPSK, QAM (0.5 Gaussian filter) OFDM: BPSK (5.5 Mb/s) OFDM: 16QAM, (24, 26 Mb/s) OFDM: 64 QAM (54 Mb/s)

Same as 802.11a

Bluetooth A wireless connectivity at 2.4 GHz

Mobile phone with headset Mobile phone with PDA Mobile phone with computer Keyboard and mouse with computer Computer with printer Automobile with mobile phone

79

Bluetooth Radio
Frequency

Output Power

Bluetooth Radio
Modulation: GFSK (Gaussian Frequency Shift Keying) 1

<20ppm

80

Spurious Emission

Receiver Sensitivity: -70dBm with 0.1% BER (Bit Error Rate)

References: http://www.bluetooth.com/ http://www.bluetooth.org/

The Trend of Wireless Communications

Cellular
Voice Mobility Wireless

Internet
Data Broadband High Speed Connection

+ other functions Bluetooth GPS IrDA Remote controller (TV, VCR)

Broadband Mobile Internet

81

RF System Specifications and Parameters

Channel

Propagating medium or electromagnetic path connecting transmitter and receiver

82

Building a transmitter
TX

Tx output
RX

Other radio
Large signal. Cannot interfere with other radio systems!

For transmitter design, a clean output spectrum is the key.

Of course, need to meet output power requirement and get the efficiency as high as possible.

Building a receiver
TX Very weak signal at RX antenna. To process the signal at baseband, we need large signal while trying to maintain the best signal-to-noise ratio. RX through RF receiver Noise signal

Channel

For receiver design, S/N or S/(N+I) is the key.

Whats wrong here?

83

SNR (Signal-to-noise ratio)

SNR S signal power = N noise power
S

Power ratio SNR will be degraded due to Signal loss Noise increase Interference
Degradation is measured by Noise Figure (NF)

NF

SNR in SNR at input = SNR at output SNR out

An example of RF system
LNA Antenna IF LO (Synthesizer, and/or PLO, DRO) Baseband BPF Mixer BPF AGC

Duplexer

HPA

Driver

BPF

Mixer

BPF

IF AMP

Notes: Duplexer can be two filters back-to-back, a switch, or a circulator Diplexer refers to a diplex filter separating two frequency bands.

84

An ideal receiver
Assuming we have transceivers with perfect components noiseless linear amplifiers (NF=0dB, IIP3=infinity), noiseless mixers which generate the mixing product you want and no image, no spurs. Perfect frequency source without phase noise or jitter. Lossless filters with infinite out-of-band rejection and extremely sharp roll-off. Of course, they drain no power from battery or power supply. Also, cost nothing and so small. Then, designing a system would be so easy! These are the wish list from system engineers. Unfortunately, none of them is true. Thats why we need so many RFIC designers in the world.

Pin

Pout

noisy and nonlinear

Non-Ideal RF Component
1dB Gain compression P1dB fA, fB
op e sl =1

P1dB, out
Pout (dBm)

k: Boltzmanns constant = 1.38x10-23 J/K

T: absolute temperature Output noise floor B: bandwidth NF

Input noise floor = kTB

P1dB, in

Pin (dBm)

85

3rd Order Intermodulation

?
fA fB nonlinear amp 1st order

2nd order 3rd order 3rd order

2nd order

fB - fA

2fA - fB

fA

fB

2fB - fA

fB + fA

Filter Bandwidth

Pin

Pout

noisy and nonlinear OIP3

IP3
IP3 3rd order Intercept Point

Pout (dBm)

fA, fB
op e sl =1

Output noise floor

NF

Input noise floor = kTB

slop e=3

2fA-fB , 2fB-fA IIP3 Pin (dBm)

86

How to measure OIP3

OIP 3 = Plin + Plin PIM 3 2

in dB

OIP3

Plin
Pout (dBm)

fA, fB
op e sl =1

PIM3
receiver noise floor
slop e=3

2fA-fB , 2fB-fA Pin (dBm)

Make sure the slope is 3 before you apply this equation!

System Specifications
Now, you have built a non-ideal radio with non-ideal components, how do you know if it works? (meets the spec?) Where did the spec come from? System specifications came from standards documents. Documents were developed and published by standards organizations. For example: ITU (International Telecommunication Union) ETSI (European Telecommunications Standards Institute) TIA (Telecommunications Industry Association)(US)

87

System Specifications Example: GSM

Receiver sensitivity

* Published by ETSI

System Specifications Example: GSM

Transmitter output spectrum

30kHz RBW 100kHz RBW * Published by ETSI

88

System Specifications Example: CDMA IS-95

Receiver sensitivity

* Published by TIA

System Specifications Example: CDMA IS-95

Transmitter output spectrum

Measured in 30kHz RBW 0dBc -42dBc -54dBc fc Fc+900kHz Fc+1.98MHz

* Published by TIA

89

System Specifications Example: DECT

Specifications: From system to component

System Reference sensitivity level Reference interference level Receiver blocking characteristics Receiver intermodulation characteristics Receiver spurious emissions Transmitter output power Output RF spectrum Spurious emissions Rx NF Rx LO phase noise Rx LO phase noise, Rx spur from mixer Rx 3rd order intermod Rx radiated spur Power amplifier output Modulator, Filter, PA Tx spur from mixer Component

90

Receiver Spurious Emissions Remember the FM radio demo? The FM receiver did cause interference! In GSM Standard (GSM0505), 5.4 defines the receiver spurious emissions.

Specifications: Circuit Design Issues

Circuit Rx LNA Rx Mixer Rx VCO Tx Modulator Tx Mixer Tx AGC Tx PA Duplexer Transceiver Integration Challenges NF(sensitivity), IP3(intermod) IP3(intermod), Spur(blocking) Phase noise(blocking), jitter(timing) Noise floor(output spectrum), Signal balance(I&Q) Spur(spurious emission) Gain control(output power level) Linearity(output spectrum), efficiency Insertion Loss(sensitivity), Rejection(spurious emission) Isolation/cross-talk, DC Power, Optimized Partition.

91

Link Budget

Link Budget is a term used to determine the necessary parameters for a successful transmission of a signal from a transmitter to a receiver through space. Includes Tx PA output, gain and loss throughout the system and link, and the S/N level required at receiver for desired bit error rate(BER) or detection. The most simple, basic system specifications analysis to begin with only gain, loss, noise (linear characteristics).

Link Budget Analysis Example

http://www.ardentech.com/

Download trial version or Lite-version

92

EIPR and ERP

Effective Isotropic Radiated Power, or Equivalent Isotropic Radiated Power EIRP = (power delivered to the antenna) X (antenna gain) in a given direction

Same EIRP

ERP = Effective Radiated Power = (power delivered to the antenna) X (relative antenna gain with respect to maximum directivity of half-wavelength dipole) ERP (dB) = EIRP (dB) 2.15dB

Link Margin
Difference in dB between (Eb/N0) received and (Eb/N0) required Bit energy Noise power spectral density Noise power per Hertz (noise energy)

E E M dB = ( b ) received,dB ( b ) required,dB N0 N0
Varies from one system design to another Depends on modulation, coding schemes.

93

Bit Error Rate

Eb x0 , PE P0 N0

PE
0.5

For

P0
0 Shannon Limit of error-free communication -1.6dB

x0

Eb N0

BER Orthogonal Modulation e.g., FSK

BER M-PSK

BW

M=2k BW M-ary M symbols on signal space

Eb/N0 (dB)

94

Bit Energy to Noise Ratio and Signal to Noise Ratio

E T E R S = b = b N N0 B N0 B
S = signal power
RF

Eb S T S B = = N0 N / B N R
Digital baseband

N = noise power Eb = signal energy per bit N 0 = noise power per Hertz T = time duration per bit R= 1 = bit rate T B = bandwidth in Hertz

Energy

Sensitivity
S Eb R = N N0 B
E S ( b ) required ( ) required Sensitivity N0 N

SNRdB = ( Eb / N 0 ) dB + 10 Log10( R ) 10 Log10( B)

SensitivitydBm = (kTB) dBm + NFdB + SNRdB,reqd = (kT ) dBm / Hz + NFdB + 10 Log10( B) + SNRdB,reqd = 174dBm / Hz + NFdB + 10 Log10( B) + SNRdB,reqd = 174dBm / Hz + NFdB + ( Eb / N 0 ) dB,reqd + 10 Log10( R)
At T = 290K, kT = -174dBm/Hz

95

RF System Specifications
Pin Pout 1dB fA, fB
op e =1

P1dB

nonlinear system

IP3 3rd order Intercept Point

Pout (dBm)

sl

Output noise floor

slop e=3

2fA-fB , 2fB-fA

SNR required NF Pin (dBm)

Input noise floor = kTB

Sensitivity

MDS = Minimum Discernible Signal

Dynamic Range (DR)

There are many definitions for DR:
IP3 Pout (dBm) P1dB

3rd-order IM SFDR Usable DR

Max. IM3 allowed Sensitivity Output noise floor SNR required MDS IM3 = MDS NF Pin (dBm)

Input noise floor

96

Glossary Summary
IMD: Inter-Modulation Distortion IM3: 3rd-order Inter-Modulation product(dBm) OIP3: 3rd-order Output Intercept Point IIP3: 3rd-order Input Intercept Point SFDR: Spurious Free Dynamic Range NF: Noise Figure SNR: Signal Noise Ratio

OIP3(dBm) = IIP3(dBm) + Gain(dB) OIP3(dBm) = Pout + (Pout - IM3)/2

Noise

97

Major Noises
Thermal Noise: most basic type of noise. Caused by thermal vibration of charges. Also called Johnson or Nyquist noise. Present in every resistor or resistance in circuits. Shot Noise: happens when DC current flows through a potential barrier, e.g., a PN diode junction. a.k.a. Schottky noise. Flicker Noise: a.k.a. 1/f noise. Caused by charge trapped due to surface defects and impurities. BJT is better than FET. Important to VCO and direct conversion receiver.

Noise Figure / Noise Factor

GA NA Sin Nin Sout Nout noisy amplifier

NF =

Sin / N in S N GA + N A NA = in in = 1+ Sout / N out N in Sin G A N in G A

NF = N out N in G A

NF

Total Output Noise Power Output Noise due to Input Source

T
If not specifically mentioned, T=290K

NF>1, NF(dB)>0dB

98

Thermal Noise and Temperature

Random voltage generated by a noisy resistor at T>0K. v(t) has a zero average value, but a nonzero RMS value 4hfBR f : center frequency (Hz) vn = v(t ) 2 = B : Bandwidth (Hz) hf / kT 1 e R: Resistance value ()
h=6.546x10-34 J-sec (Plancks constant) k=1.38x10-23 J/K (Boltzmanns constant) Black body radiation law

vn = v(t ) 2 =

4hfBR e hf / kT 1

If hf << kT, e.g., f=300GHz and T=290K, hf=1.96x10-22 J << kT=4.0x10-21 J

hf e hf / kT 1 kT

vn = 4kTBR

Not valid at very high frequency or very low temperature noise power independent of frequency (white noise source) noise power proportional to the bandwidth These white noise sources are independent Gaussian distributed random variables, they are additive. AWGN =Additive White Gaussian Noise

99

Pn

Available noise power = maximum noise power delivered to load resistor

vn 2 vn 2 Pn = ( ) R = = kTB 2R 4R
Smaller bandwidth collects less noise power Cooler devices and components generate less noise power Impedance has to be matched

Equivalent Noise Temperature

An alternative way of describing noise contribution of a noisy network (a source or an amplifier) Ps: noise power from source
Assume a matched R at input with temperature Te

Ps = Pn = kTe B

P Te = s kB

Thermal noise floor at 290K = kTB, Te=290K

100

Contribution of noise from amplifier is modeled as an equivalent noise source at input

Pi = 0 Po = Pi G + N A = N A

Pi = kTe B

NA = 0

Po = Pi G + N A = GkTe B

N A = GkTe B

T0 = 290K

NF = 1 +

GkTe B T NA = 1+ = 1+ e N in G GkT0 B T0

Total output noise power:

N out = GkT0 B + GkTe B = Gk (T0 + Te ) B

Te

The temperature increase at source resistance to produce same amount of noise power at output when modeling the amplifier as noiseless.

101

Some textbooks use F for Noise Factor (linear) and NF for Noise Figure (in dB), but some others do it the other way. Some treat Noise Factor and Noise Figure the same, and can be in either linear scale or dB. As long as you know whether its in linear scale or dB, doesnt matter what symbol and term you use.

Example from Lee, The Design of CMOS RFIC

Cascaded Gain and NF

G1 NF1 Sin Nin G2 NF2 Gn NFn Sout Nout

G cas = G 1 G 2 G 3 L G n G cas = G 1 + G 2 + G 3 + L + G n
NFcas = NF1 +

not in dB in dB not in dB

NFn 1 NF2 1 NF3 1 + +L+ G1 G1G2 G1G2 LGn1

en Tcas = Te1 + e 2 + e3 + L + G1 G1G2 G1G2LGn 1

not in dB

102

NF of a passive lossy network

Terminate with a resistor R at same temperature T

L=1/G: Loss
N out = kTB = GkTB + N added
1 GkTe B = N added = (1 )kTB L Te = ( L 1)T

T T NF = 1 + e = 1 + ( L 1) T0 T0
If T= T0

Reference to T0

NF = L

sky noise temperature

Antenna collects noise from galactic, solar, and terrestrial sources.

103

Antenna temperature from sky noise

Brightness Temperature of objects

Typical numbers: ground ~300K, sky at zenith ~5K, sky at horizon ~100-150K
Why 5K? 3K background radiation + noise due to atmospheric absorption

Why use T0=290K? K = 1.38x10-23 J/K KT0 = 1.38x10-23 J/K x 290K = 4x10-21 J = 4x10-21 W/Hz = -174dBm/Hz

104

Linearity

P P = out OIP3 P + out IM3 = P + G + in 2 2

IIP3= P + in

in dB!
Output

OIP3

Pout

fA, fB
op e =1

PIM3
receiver noise floor
slop e=3

sl

2fA-fB , 2fB-fA Pin

Input

105

Cascaded IP3
G1 IP31 G2 IP32 Gn IP3n Gtotal IP3total

Output:

Not in dB!

1 1 1 1 = 1 + + + L+ OIP3total OIP3n GnOIP3n 1 GnGn 1OIP3n 2 Gn LG2OIP31

Input:
G G LG G GG 1 = 1 + 1 + 1 2 + L + 1 2 n1 IIP3total IIP31 IIP32 IIP33 IIP3n

IPn
n 1 OIPn = Pout + n 1 IIPn = Pin +

nth-order intercept point in dB!

Cascaded OIPn
OIPn total (1 n ) / 2 = OIPn M (1 n ) / 2 + ( G M OIPn M 1 ) (1 n ) / 2 + ( G M G M 1OIPn M 2 ) (1 n ) / 2 + L + ( OIPn 1G 2 L G M ) (1 n ) / 2

106

Summary of Noise and IP3 in Cascaded System

G1 NF1 IP31 G2 NF2 IP32 Gn NFn IP3n

Noise dominated by first stage. (noise floor amplified by first stage gain and NFn all refer to 290K) IP3 dominated by last stage. (large signal easily gets into saturation)

Nonlinearity
Pin Pout nonlinear system Pout (dBm) P1dB 1dB fA, fB
op e =1

IP3 3rd order Intercept Point

2fA-fB , 2fB-fA Pin (dBm)

Where does gain compression come from? IP3 is usually ~10dB higher than P1dB, why?

sl

107

Gain Compression: Two Scenarios

Scenario 1 Large signal input No blocker or weak blocker G=?

fS

nonlinear amp

Scenario 2 Small signal input But, large in-band blocker G=?

fS fB

nonlinear amp

Gain Compression: Mathematical Model

Nonlinear behavior of a component

Signal at

1, Blocker at 2.

Consider up to 3rd order term:

108

Scenario 1, large signal input at

1, no blocker or very weak blocker

V2 = 0

3 Vo = (a1 + a3V12 )V1 cos 1t + L 4 = a1'V1 cos 1t + L

To have gain compression, 1dB compression point

a3 <0 a1
20 log(1 + 3 a3 2 V ) = 1 4 a1 1

Scenario 2, small signal input at

1, large blocker at .
2

V2 >> V1

3 2 Vo = (a1 + a3V2 )V1 cos 1t + L 2 = a1'V1 cos 1t + L

109

Signal at

1
1dB

Gain (dB)

P1dB Blocker Power (dBm)

Comparison of Two Scenarios

scenario: 1dB compression point:

fS

nonlinear amp

fS fB

nonlinear amp

V 0.381 20 log( 1 ) = 20 log( ) = 3(dB ) V2 0.269

1dB compression point due to blocker is 3dB lower than due to signal itself.

110

1dB Compression Point in dBm

20 log(1 +

3 a3 2 V ) = 1 4 a1 1

IP3: Mathematical Model

At IP3,

a1 V =

3 a3 3 V 4

V2 =

4 a1 3 a3

111

Summary of Noise and IP3 in Cascaded System

G1 NF1 IP31 G2 NF2 IP32 Gn NFn IP3n

Noise dominated by first stage. (noise floor amplified by first stage gain and NFn all refer to 290K) IP3 dominated by last stage. (large signal easily gets into saturation)

Transceiver Architectures

112

Receiver Heterodyne
LNA BPF Mixer IF Amp SAW

DEMOD Duplexer To Antenna LO2 LO1 90 From Transmitter LPF LPF I ADC Q ADC

Single conversion or dual conversion Channel selection by SAW filter IF frequency planning Need at least two LO sources Integration level is low due to filters

Receiver - 2nd IF Sampling

LNA BPF 1st Mixer IF Amp IF1 SAW

Duplexer To Antenna

LO2 LO1

2nd Mixer IF2 Filter

From Transmitter ADC Faster ADC Digital IQ demodulation Need IF2 filter (LO2IF1). Why not make IF2=0? Still need two LO sources Integration level not much higher

113

Receiver - Direct IF Sampling

LNA BPF Mixer IF Amp SAW ADC

Duplexer To Antenna

LO

From Transmitter

Very fast ADC Single LO source Digital IQ demodulation Channel selection by SAW

Receiver - Direct-Conversion Zero-IF

DEMOD LNA LPF ADC Duplexer To Antenna 90 LO LPF ADC

From Transmitter

Very high level integration No image frequency (single carrier) No SAW. Channel selection by baseband LPF Adjustable channel BW by baseband LPF Problem: DC offset and IM2 (IP2) Problem: 1/f noise (CMOS) LO pulling by in-band interferer (injection locking) AC coupling can be used in broadband system

114

Receiver - Direct-Conversion Low-IF

DEMOD LNA ADC

ADC Duplexer To Antenna 90 LO Image Rejection and BPF near DC

From Transmitter

Channel selection by BPF Image rejection by image-reject architecture DC offset handled by AC coupling 1/f noise is not so critical as Zero-IF

Transmitter - Indirect Up-Conversion

PA VGA BPF Mixer IF Amp SAW

MOD Duplexer To Antenna LO2 LO1 90 I To Receiver DAC Q DAC

SAW filter cleans up modulator noise Need two LO sources

115

Transmitter - Direct Up-Conversion

PA BPF VGA

MOD Duplexer To Antenna LO 90 I To Receiver DAC Q DAC

No SAW filter High level integration - only one LO needed Modulator noise floor must be low enough LO pulling

Details on Receiver Architectures

116

Typical duplexer frequency response

To Receiver
e.g., PCS 1850-1910 MHz 1930-1990 MHz

S21
Duplexer Antenna

From Transmitter

Frequency

PA leakage affects LNA

?W ? dBw ? dBm Affects the sensitivity desensitization. Problem in FDD with high power transmitter. Gain Compression G1 NFtotal G1

117

Rejecting out-of-band interferers (band selection)

Band: the entire spectrum in which users of a standard can use. Examples: GSM downlink 935-960 MHz, FM 88-108 MHz Channel: a portion in the band that one user occupies. Examples: GSM/FM channel bandwidth=200kHz

Rejecting in-band interferers (channel selection)

A 900MHz phone receives 30kHz channel while rejecting interfering channels 60kHz away. If the BPF needs to provide 60dB rejection at 45kHz away, the Q would be very high! (~107 for 2nd order LC filter) Hard to do channel selection at RF.

118

Interference in wireless communications

Transmitter

Receiver

Where do those interferers come from?

Co-existence of wireless Systems

Sprint PCS CDMA CDMA phone
RX 1930-1990 MHz

GSM phone T-Mobile GSM1900

RX 1930-1990 MHz

Which phone suffers the worst interference problem?

119

How does IMD3 affect the signal?

3rd-order intermodulation products

Image problem in heterodyne system

mirror

1 cos 1t cos LO t = [cos(1 LO )t + cos(1 + LO )t ] 2 1 cos imt cos LO t = [cos(im LO )t + cos(im + LO )t ] 2

LO - 1 = IF = im - LO

120

Image rejection by filter

IRF

IRF also rejects large output noise of LNA. An alternative is to use Image Reject Receiver to be discussed later.

Channel selection by another filter

Important: IP3 is not an issue after channel select filter.

121

Channel Select Filter SAW

Channel Select Filter Crystal

122

Channel Select Filter Ceramic

Problem of Half-IF Interferer

In-band interferer/blocker

To solve the problem: 1. Minimize the 2nd order distortion. 2. IF frequency > 2 x (RX Bandwidth)
RX Bandwidth for different systems: EGSM: 35MHz, DCS1800: 75MHz, PCS1900: 60MHz AMPS/TDMA/CDMA(800MHz band) and GSM: 25MHz

123

Tradeoff between image rejection and channel selection

High IF

Image Reject Good Channel Select Bad

Low IF

Image Reject Bad Channel Select Good

Dual-IF Architecture
IP3

Relaxed Q

124

Quadrature Downconversion
The last stage before ADC in most digital communication systems

1 [ a (t ) cos + b(t ) sin ] 2

1 [a (t ) sin b(t ) cos ] 2

Direct-conversion Zero-IF (Homodyne)

fRF = fLO No image!

AM, ASK OK, but FM, FSK or QPSK not good

No IRF monolithic integration No need to have 50 ohm

125

IP3

Direct-conversion Zero-IF Receiver Channel Selection

LPF: linear & low noise Amp: can be nonlinear

LPF: noise not critical Amp: linear & low noise

Filtering in digital domain ADC: high linearity

DC Offset

Due to self-mixing

f t Self-mixing due to LO

Self-mixing due to Interferer

126

DC Offset Cancellation
High pass filtering.
Affect BER performance. Need large capacitors.

DC-free coding encode the signal to reduce DC energy. Calibration when not transmitting signal. DC offset may not be constant. Could vary in time.
Makes the DC offset cancellation more difficult.

Quadrature Generation

Quadrature in RF Difficult for broadband signals

Quadrature in LO

Very narrowband (Single-tone)

127

Quadrature Generation
RC-CR network

I/Q Mismatch

128

I/Q Mismatch

Gain error

Phase error

I/Q Mismatch
In time domain:

Gain error

Phase error

I/Q mismatch in heterodyne is less a problem: 1. I/Q balance at low frequency is easier. Less sensitive to mismatches in parasitics Large devices can be used to improve matching 2. Homodyne has I/Q separation before filtering (linearity). Heterodyne has gain and filtering before I/Q separation. 3. Heterodyne can have digital I/Q separation. Homodyne can only do analog I/Q separation.

129

Even-order distortion
Example: IP2 (second-order)

And, nonlinearity of mixer also contribute!

Reference Design Example RF Specifications Case Study

130

GSM Receiver
Task: Define RF specifications for a GSM handset receiver front end and design an RFIC architecture that meets the specifications. Documents: 1. GSM 05.05 version 5.5.1 (ETS 300 910) January 1998. Radio transmission and reception. 2. GSM 11.10-1 version 5.6.1 (EN 300 607-1) December 1998 Mobile station conformance specification. 3. GSM 01.04 version 5.0.0 March 1996 Abbreviations and acronyms. 4. GSM 05.10 version 5.1.1 (ETS 300 912) May 1997 Radio subsystem synchronization. Tools: 1. SysCalc. 2. Lecture notes. 3. Hand calculation or calculator.

Why GSM? The most popular cellular phone standard worldwide. Its standard documents are the most well organized and the requirements are the most well defined. Its famous blocker requirement gives the most stringent phase noise specifications for VCO.

131

GSM 05.05 version 5.5.1

Scope

132

133

Mobile Station Power Class

134

Mobile Station Power Control

Base Station Power Class

58 dBm 55 dBm 52 dBm 49 dBm 46 dBm 43 dBm

135

Output RF Spectrum Mask

Comparison of QPSK, OQPSK, and MSK spectra

Normalized power spectral density G(f) (dB)

Normalized frequency offset from carrier, (f-fc)/R (Hz/bit/s) MSK is spectrally more efficient than QPSK/OQPSK. But, has wider main lobe.

136

Equations for Power Spectral Density

First Null For QPSK and OQPSK:

G ( f ) = 2 PT (

sin 2fT 2 ) 2ft

fT =

f = 0.5 R

For MSK:

cos 2fT 2 G ( f ) = 16 PT ( ) 2 2 2 1 16 f T

fT =

f 3 = R 4

P: Average power in modulated waveform T: Bit period = 1/R

3 270.833 = 203.12 200 4

Baseband!

137

In RX band

138

Mixer Spur/Intermod Table

Base Station RFIC

139

Figure

270.833 S Eb R = = 8.398dB + 10 log = (8.398 + 1.354)dB = 9.752dB 200 N N0 B

Eb = 8.398dB N0

Note: B=Bandwidth, either in baseband or RF

140

Receiver sensitivity

S/N ratio

141

TU50: Typical Urban 50 km/h Control Channel

RA250: Rural Area 250 km/h

FH: Frequency Hopping

HT100: Hilly Terrain 100 km/h

Eb = 9.588dB N0
TCH: Traffic Channel

S = (9.588 + 1.354)dB = 10.942dB 11dB N

142

SID: Silence Descriptor.

143

A Receiver Front End for GSM Mobile Station

LNA

BPF

Mixer IF Amp

SAW AD6459

Duplexer To Antenna

VCO

Synthesizer

From Transmitter

http://www.analog.com/product/selection_guides.html

144

Find a Duplexer

http://www.murata.com/develop/index.htm

145

Dielectric Duplexer For GSM 900

Find a Dielectric Filter for Image Rejection

http://www.murata.com/develop/index.htm

146

Find a SAW Filter

> 50 MHz (2x Receiver BW)

http://www.sawtek.com/products/catalog/gsm.html

Receiver Specifications
Pin Rx Pout IP3 P1dB

Pout (dBm)

Sensitivity = MDS + 9 Output noise floor Gain=Pout-Pin

slop e=3

sl

op e

=1

MDS = KTB + NF

SNRmin= 9 dB NF Pin (dBm)

Input noise floor = kTB = -174dBm/Hz x 10 log(200000)= -121dBm

147

SNR Required
In digital radios like GSM, sensitivity level is actually related to BER, FER or RBER under different channels and propagation conditions, not a simple SNR. We use the most stringent BER requirement of 10-5 in Data Traffic Channel to estimate SNR ~ 11dB. Frequency Hopping reduces the SNR requirement from 11dB to 9dB, (An introduction to GSM, Artech House, 1995, p. 142.) 9dB SNR also ensures meeting Co-Channel Interference requirement C/I=9dB. Signal is at 20dB above sensitivity level. Advanced proprietary coding may further reduce SNR requirement. For now, lets use 9dB for SNR requirement.

Reference Sensitivity NF

Sensitivity = -121dBm + NF + 9 < -102dBm NF < 121dBm - 9 -102dBm = 10 dB

148

Intermodulation Characteristics IP3

Assume noise power is much lower than IMD3 power.
(Refer to input)

-49dBm

-99dBm
-99dBm-9dB

= -108dBm

Intermodulation Characteristics IP3

IIP 3 = 49 + 49 ( 108 ) = 19 .5 2

IP3

-49dBm

-108dBm

In general, IIP 3 = 49 +

49 ( 99 SNR ) SNR = 24 + 2 2

149

SysCalc Calculation
GSM MS Receiver GSM MS Receiver
Duplexer LNA Image Rejection Filter Mixer IF Amp 71 MHz SAW AD6459

Total NF (dB) Gain (dB) IIP3 (dBm) Po (dBm) NF+ (dB) IP3+ (dBm) Input Pwr (dBm) 3.20 -3.20 100.00 -105.20 1.09 0.00 -102.00 2.50 15.00 -10.00 -90.20 1.74 0.56 2.60 -2.60 100.00 -92.80 0.00 10.00 8.00 0.00 -84.80 1.08 1.03 290.00 -114.05 9.00 N/A 0.00 -105.05 -21.54 Input IP3 (dBm) -15.96 Output IP3 (dBm) 88.74 OIM3 (dBm) -169.39 ORR3 (dB) 172.09 IRR3 (dB) 57.36 SFDR3 (dB) 65.39 5.00 15.00 3.00 -69.80 0.04 4.78 6.50 -6.50 100.00 -76.30 0.00 10.00 79.00 100.00 2.70 0.02 0.00 6.92 104.70 -15.96

System Temp (K) MDS (dBm) Eb/No (dB, Req'd) Es/Eb (dB) Sens. Losses (dB) Sensitivity (dBm) G/T (dB/K)

Modulation: MSK, Coherent System BW (MHz) 0.20 BER (Req'd) 33.63e-6 Eb/No (dB, Actual) 12.05 Srce Temp (K) 290.00 Te Eff. (K) 1136.53

SysCalc Calculation
GSM MS Receiver GSM MS Receiver
Duplexer LNA Image Rejection Filter Mixer IF Amp 71 MHz SAW AD6459

Total NF (dB) Gain (dB) IIP3 (dBm) Po (dBm) NF+ (dB) IP3+ (dBm) Input Pwr (dBm) 3.20 -3.20 100.00 -43.20 1.09 0.00 -40.00 2.50 15.00 -10.00 -28.20 1.74 0.56 2.60 -2.60 100.00 -30.80 0.00 10.00 8.00 0.00 -22.80 1.08 1.03 290.00 -114.05 9.00 N/A 0.00 -105.05 -21.54 Input IP3 (dBm) Output IP3 (dBm) OIM3 (dBm) ORR3 (dB) IRR3 (dB) SFDR3 (dB) -15.96 10.74 -61.39 48.09 16.03 65.39 5.00 15.00 3.00 -7.80 0.04 4.78 6.50 -6.50 100.00 -14.30 0.00 10.00 1.00 100.00 -13.30 0.02 0.00 6.92 26.70 -15.96

System Temp (K) MDS (dBm) Eb/No (dB, Req'd) Es/Eb (dB) Sens. Losses (dB) Sensitivity (dBm) G/T (dB/K)

Modulation: MSK, Coherent System BW (MHz) 0.20 BER (Req'd) 33.63e-6 Eb/No (dB, Actual) 74.05 Srce Temp (K) 290.00 Te Eff. (K) 1136.53

150

SysCalc Calculation
GSM MS Receiver GSM MS Receiver
Duplexer LNA Image Rejection Filter Mixer IF Amp 71 MHz SAW AD6459

Total NF (dB) Gain (dB) IIP3 (dBm) Po (dBm) NF+ (dB) IP3+ (dBm) Input Pwr (dBm) 3.20 -3.20 100.00 -52.20 1.09 0.00 -49.00 2.50 15.00 -10.00 -37.20 1.74 0.56 2.60 -2.60 100.00 -39.80 0.00 10.00 8.00 0.00 -31.80 1.08 1.03 290.00 -114.05 9.00 N/A 0.00 -105.05 -21.54 Input IP3 (dBm) Output IP3 (dBm) OIM3 (dBm) ORR3 (dB) IRR3 (dB) SFDR3 (dB) -15.96 34.74 -64.39 66.09 22.03 65.39 5.00 15.00 3.00 -16.80 0.04 4.78 6.50 -6.50 100.00 -23.30 0.00 10.00 25.00 100.00 1.70 0.02 0.00 6.92 50.70 -15.96

System Temp (K) MDS (dBm) Eb/No (dB, Req'd) Es/Eb (dB) Sens. Losses (dB) Sensitivity (dBm) G/T (dB/K)

Modulation: MSK, Coherent System BW (MHz) 0.20 BER (Req'd) 33.63e-6 Eb/No (dB, Actual) 65.05 Srce Temp (K) 290.00 Te Eff. (K) 1136.53

IIM3 (input IMD3)=-64.39-50.70=-115.09<-108

Blocking Characteristics VCO Phase Noise

LO - Blocker = -43 dBm + Gain Blocker IF = -99dBm + Gain RF

LO Blocker = -43dBm x dBc/Hz RF = -99dBm

600KHz 600KHz

-43 dBm + Gain + x dBc/Hz + 10 log(200kHz) < -108dBm + Gain x dBc/Hz < -108 + 43 - 53 = -118 dBc/Hz

151

Blocking Characteristics VCO Phase Noise

Offset 600 kHz 800 kHz 1.6 MHz 3 MHz

Phase Noise (dBc/Hz) -118 -118 -128 -138

Blocker level (dBm) -43 -43 -33 -23

Reference Interference Level VCO Phase Noise

In GSM 05.05 6.3, adjacent channel interference also affect VCO phase noise. -33 dBm + Gain Interference IF = -82dBm + Gain RF

LO Interference = -33dBm x dBc/Hz RF = -82dBm

600KHz 600KHz

-33 dBm + Gain + x dBc/Hz + 10 log(200kHz) < -82dBm + Gain - 9dB x dBc/Hz < -91 + 33 - 53 = -111 dBc/Hz

152

Reference Interference Level VCO Phase Noise

Offset 200 kHz 400 kHz 600 kHz

Phase Noise (dBc/Hz) -71 -103 -111

Interference C/I (dB) -9 -41 -49

Combined VCO Phase Noise Requirement

Offset 200 kHz 400 kHz 600 kHz 800 kHz 1.6 MHz 3 MHz

Phase Noise (dBc/Hz) -71 -103 -118 -118 -128 -138

153

GSM Receivers Phase Noise Requirements

A Comparison
Phase Noise (dBc/Hz)
-100 -110 -120 -130 -140 -150 -160 100 1000 10000

1 2 3 4 5 6

Offset Frequency (kHz)

Offset 1. DCS 1800 2. GSM 900 3. DCS 1800 4. DCS 1800 5. GSM 900 6. GSM 900 @kHz MS small MS(<=2W) Micro BTS Normal BTS Micro BTS Normal BTS 600 -116 -118 -121 -128 -130 -137 800 -116 -118 -131 -138 -140 -147 1600 -126 -128 -131 -138 -140 -147 3000 -133 -138 -131 -138 -140 -150

Assume C/I=9dB

Which offset frequency is most critical? MS: 600kHz and 3MHz, BTS: 800kHz

DCS 1800 MS

-20dB/decade or -30dB/decade

GSM 900 small MS

DCS 1800 Micro BTS

154

RFIC Design Issues

Impedance Matching
A typical transistor (FET) on Smith Chart MAG

NFmin

Match to 50

S22 S11
Transistors in/out usually capacitive

155

Why 50 ?
A compromise of 77(minimum attenuation for air dielectric) and 30(maximum power capability) CATV uses 75.

Note: 50 was not optimized for Integrated Circuit(IC) environment.

Reference: T. S. Laverghetta, Microwave Measurements and Techniques

Impedance ()

S parameters
S21 (gain or insertion loss) a1 S11 (input return loss) b1 S12 (isolation) S11 = b1 a1 a = 0 2 S21 = b2 a1 a2 = 0 b S12 = a 1 2 a =0 1 S22 = b2 a2 a1 = 0 DUT a2 S22 b2 (output return loss)

S-matrix =

156

VSWR v.s. Return Loss v.s. Mismatch Loss

VSWR 2.5:1 2:1 1.5:1 1.22:1 1:1

Return Loss 7.4 dB 9.6 dB 14 dB 20 dB ?

Mismatch Loss 19%(0.89dB) 11%(0.52dB) 4%(0.18dB) 1%(0.045dB) ?

VSWR and Return Loss

Reflection Coefficient

b1 a1

Voltage Standing Wave Ratio

VSWR =

1 + | | 1 - | |

Return Loss

RL = - 10 log

b1 = - 20 log | | = - S11 a1

157

Power Gain
PAVS
Source

PIN
DUT

PAVN
S11 S 21 S12 S 22

PL
Load

IN

OUT

Transducer Power Gain

GT =

1 S PL power delivered to load = = PAVS power available from source 1 IN S

2 2

S 21

1 L

2 2

1 S 22 L

Available Power Gain 2 1 S power available from DUT 1 P 2 G A = AVN = = S 21 PAVS power available from source 1 S11S 2 1 OUT Operating Power Gain P power delivered to load 1 GP = L = = PIN power input to DUT 1 IN
2

S 21

1 L

2 2

1 S 22 L

Power Gain
For 50 source and load:

GT =

PL 2 = S 21 PAVS

Most commonly used

GA =

1 PAVN 2 = S 21 PAVS 1 OUT 1 PL = PIN 1 IN

2

GP =

S 21

Note:

PAVS = PIN when IN = S

* *

IN = S11 +

PAVN = PL when L = OUT

S12 S 21L 1 S 22 L S12 S 21S 1 S11S

OUT = S 22 +

158

Power Gain
Now, if the DUT is matched to non-50 , but equipment source and load are still 50 , how do we get real DUT gain in non- 50 ? Hint: 11 12 is measured. S 21 S 22 50
S S

GT =
S = Z S 50 Z S + 50 L =

1 S

2 2

1 IN S

S 21 50

1 L

2 2

1 S 22 (50 ) L

Z L 50 Z L + 50

If S12 (50 ) = 0 and S11(ZS ) = S 22 (ZL ) = 0

(Input/Output matched to Z S / Z L )

IN = S11(50 ) =
1 S 1 S
2

Z S 50 = S Z S + 50
2

OUT = S 22 (50 ) =

Z L 50 = L Z L + 50

GT =

2 2

S 21 50

1 L 1 L

2 2

1 1 ( Z + 50) 2 ( Z + 50) 2 2 2 S 21 50 = S S 21 50 L 2 2 4 Z S 50 4 Z L 50 1 S 1 L

Power Gain
Example 1 : You have an amplifier designed for 75 system but can only measure S21 in 50 system. What is the S21 in 75 system? Also assume that you did a perfect job to match input and output to 75.

GT = S 21 75 =
2

( Z S + 50) 2 ( Z + 50) 2 2 2 S 21 50 L = (1.042) 2 S 21 50 dB( S 21(50 ) ) + 0.36 dB 4 Z S 50 4 Z L 50

( Z S + 50) 2 ( Z L + 50) 2 (75 + 50) 2 (125) 2 4 Z 50 = 4 Z 50 = 4 75 50 = 15000 = 1.042 = 0.18 dB S L

Example 2 : You have an amplifier designed for 500 in/out matching. How do you determine its gain from 50 measurement? Again assume that you did a perfect job in 500 matching.

( Z S + 50) 2 ( Z L + 50) 2 (500 + 50) 2 (550) 2 = = = = 3.025 = 4.81 dB 4 Z S 50 4 Z L 50 4 500 50 100000

dB ( S 21( 500 ) ) = dB ( S 21(50 ) ) + 9.62 dB

159

Interference due to Cross-Talk

Major causes of cross-talk: Capacitive coupling could be reduced by ground plane. Inductive coupling could be reduced by proper design Common ground coupling due to finite ground impedance (conductive substrate or ground inductance) Power supply line coupling due to finite impedance looked into power supply could be reduced by de-coupling capacitor Radiation could be reduced by proper shielding

Single-ended v.s. Differential

Single-ended circuit without good ground is liable to interference due to all links of corss-talks. Differential circuit has common mode rejection. It reduces common ground coupling and power supply line coupling, but still liable to capacitive coupling, inductive coupling, and radiation coupling. Single-ended circuit with good ground and good de-coupled power lines may be better than differential circuit since ground plane reduces capacitive coupling. Inductive coupling seems to be the most difficult one. Suggestion: use high impedance in your circuit to reduce current flow on interconnections. Watch the voltage swing on active devices.

160