NBSIR 76- 979

INTERCOMPARISON PROCEDURES
FOR GAGE BLOCKS USING ELECTROMECHANICAL COMPARATORS

J. S. Beers C. D. Tucker

Institute for Basic Standards National Bureau of Standards Washington . D. C. 20234

January 1976
Final

S. DEPARTMENT OF COMMERCE, RogersC. B. Morton, Secretary James A. aker, III, Under, Secretary Dr. Betsy Ancker-Johnson. Assistant Secretary for Scie"ce arid Tech"ology
NATIONAL BUREAU OF STANDARDS, Ernest Ambler,

Acti"g Director

CONTENTS

Page
In troduc tion
The Compara tor

C. Environment
Gage Block Requirements

Temper~ture Effects and Their Control

Intercomparison Procedure

Block Preparation
Comparator Preparation

Block Storage
Thermal Equalization Titne Temp era tureMeasuremen t Handling Techniques

Observations
Temperature and Deformation
Computation and Analysis

Corrections

Summary

References
Appendix:
Experimental Evidence of Thermal Effects in Long

Gage Blocks

INTERCOMPARISON PROCEDURES FOR GAGE BLOCKS

USING ELECTROMECHANICAL COMPARATORS

J. S. Be. ers and C. D. Tucker

Introduct. ion
The techniques described here for intercomparing gage blocks with electro-mechanical gage block comparators are the results of considerable experimental work, statistical design and process analysis. Some laboratories will find these techniques ideally suited to their needs while others may wish to modify them, but in either event the Measurement Assurance Program (MAP) described here, in part, and in other papers (1, 2 , 3, 7)* will promote the analysis of measurement requirements and the application of statistical principles.

Length values are assigned to unknoYlIl gage blocks by a transfer process from blocks of known length. Briefly, the transfer process is a systematized intercomparison of 4 blocks of the same nominal size, called a " 1" series, using an electro-mechanical comparator. Two of the blocks are unknowns and two are standards. Such a process is ideally suited to comparing objects with nearly identical characteristics such as gage blocks and the process provides the redundancy needed for statistical Derived corrections can be applied where such charac as coefficient of thermal expansion and resistance to mechanical deformation differ (4). The method applies to . all sizes but speciaJ, precautions will be described for long gage blocks where temperature effects

analysis.

teristics

predominate.
The Compara tor

There are a number of suitable comparator designs and a typical one is shown schematically in figure 1. An upper stylus and a lower stylus contact the gaging faces of a block supported on an anvil. Each stylus

is attached to a differential transformer core. An integrat.ed signal

from these two transducers is displayed on a meter graduated in length units (usually microinches). The difference in length between two gage blocks is obtained by inserting the blocks, one at a time, between the stylus tips and taking the differe nce between the two readings.
*Figures in brackets indicate the literature references at the end of this

paper.

SECONDARY COIL

TRANSDUCER CORE PRIMARY COIL

MEASURING STYLUS
AMPLIFIER

TEST BLOCK
ANVIL

STYLUS

REED SPRING

ELEMENTS OF A MECHANICAL
. COMPARATOR OF LENGTHS~
FIGURE I.

An important comparator feature is the point t.o point measurement along an axis through the gaging point of the block. Other suitable comparator designs use only one transducer but by various means insure the point to point measurement.

For the most precise measurements, it is important that the defined

gag. e point of the blocks be contacted by the comparator stylus. A metal

or plastic bar about 1/4 inch thick is fastened to the anvil (with laboratory wax if no other means is provided) behind the stylus and is positioned to stop the gage blocks so the stylus will contact the gage point. The bar can be L-shaped to position the gage block both laterally and transversely if desired and special configurations can be devised for comparing blocks of different shapes.

Environment
A temperature controlled laboratory is necessary for intercomparisons of the highest precision. The degree of temperature control needed depends on the length of the blocks being compared, differences in coefficients of thermal expansion among the blocks, and the limiting uncertainty required for the unknown blocks. At NBS, blocks in the size range up to 4 inches are intercompared in a temperature controlled laboratory at 200 + . 25 o Celsius. Blocks longer than 4 inches are intercompared in a laborato ry at 200 050 Relative humidity is held below 50% to prevent corrosion of blocks

and instruments.
Gage Block Requirements Nearly all gage blocks are either square or rectangular in cross section, but a few have a circular cross section. Any of these cross sections are easily handled in intercomparison procedures in sizes up to 4 inches. From 5 to 20 inches, however, the rectangular cross section bec.omes prone to tipping.
Transferring the length unit to a gage block by intercomparison does not require rigorous constraints on the flatness and parallelism of its gaging faces because the transfer is made only along a single line through the block. Gage block use, however, may be more demanding on the geometry of the gaging faces. A separate paper covers the measurement of flatness

Celsius.

and parallelism (6).

Thermal expansion coefficients are generally taken from bulk values and these may vary by as much as 10% from actual value. s for gage blocks. This problem can be circumvented in the intercomparison process by insuring that blocks being intercompared are as close as possible to 200 Celsius.

If non- standard temperatures are used, the coefficients must be knoYlIl unless the attendant systematic errors are tolerable.

Temperature Effects and Their Control

A large uncertainty in the comparison process can be introduced by temperature effects. For example, a temperature difference of . Celsius between two one-inch steel blocks will cause an error of nearly 6 microblocks are noteworthy:

inches in the comparison. Two

causes of temp~rature differences between

room temperature gradients or nearby heat sources such as electronic equipment can cause significant temperature differences between blocks even when they are stored relatiVely close to each other before comparison and

(1)

blocks with different surface finishes on their noncan absorb radiant heat at different rates and reach different gaging faces equilibrium temperatures. The magnitude of these effects is proportional to gage block See the appendix at the end of this paper for further discussion of this subject.

(2)

length.

A number of remedies have been developed. For short blocks (through 4 inches) the remed For example, storing the blocks, both standards and unknoYlIls, on a thermal equalization plate of smooth surface and good heat conductivity close to the comparator but away from heat Also, the use of tweezers or tongs to handle the blocks and a systematic, rythmic block handling technique in the comparison procedure to insure identical thermal environment for each block.

ies are quite simple.

sources.

Precautions for long blocks (over 4

NBS the comparator is enclosed .in

Blocks are stored on the comparator anvil and each block is wrapped in 3 layers of Mylar* reflective film to equalize their thermal characteristics. remain off except for one lamp well removed from the The laboratory lights compar,ator but giving enough illumination to work by. Heat sources are kept as far away from the comparator as possible and the comparator is located away from other activities where laboratory personnel might congregate. As. further precaution during intercomparisons the operator Wears a cape of aMylar reflective film and heavy cotton gloves while handling the blocks with tongs. Finally, as with short blocks, the handling procedure is systematic, rythmic and quick but not rushed.

the front to allow handling of the blocks with tongs.

inches) are more elaborate. an insulated housing partially open on

*Certain commercial equipment, instruments, or materials are identified in this paper in order to adequately specify the experimental procedure. In no case does such identification imply recommendation or endorsement by the National Bureau of Standards, nor does it imply that the material or equipment identified is necessarily the best available for the

purpose.

Temperature problems can be detected in long block intercomparisons by reversing the storage positions of the 4 blocks and repeating the intercomparison after suitable equalization time. Temperature gradients in the storage area will be revealed by a significant change in the differences between blocks. Still another method is to measure the temperature of each block with a thermocouple. A temperature measuring system is described in reference (7) Where gradients are not a problem , a mercury-in- glass thermometer mounted on the comparator will be adequate.
Intercomparison Procedure

Block Preparation
The master and test blocks must be thoroughly cleaned, examined and deburred using procedures in reference (5). The identification numbers are recorded for inclusion in the test r~port or records.
Gage blocks longer than 4 inches are wrapped in three layers of Mylar reflective film (available as BSA Rescue Blanket Stock No. 1090) with the reference edge and identity of the block marked on the outside of the wrapping.

Comparator Preparation

The instrument anvil should be deburred and cleaned. The comparator transducer pressure, magnification and alignment should be checked (5). Periodic cleaning of the instrument anvil is recommended during the work day to help reduce the number of spurious readings that result from minute particles that can contaminate the anvil surface.
It is necessary, when making preparations to intercompare long gage blocks, that after adjusting the comparator for accepting a particular size gage block that all four blocks in that series be inserted between the probes to insure that all readings will be on scale at that magnification when the intercomparison run is made.

Block Storage
Short blocks, up to 4 inches in length, are arranged in groups of 4 (2 standards and 2 unknoYlIls) of the same nominal size thermal equalization plate next to the comparator. From here they can be moved to the comparator anvil by groups at the time of comparison. There is some thermal advantage for sizes from about 0. 3 inch to

ona

4 inches t.o leaving the blocks on the plate at all ,times except when the bl.ock is being inserted in the comparator for measure-

ment. Alternatively, a

group can be moved to the comparator anvil and allowed to equalize there for an appropriate period (see section F4).

The four long gage blocks in a grc.fup are stored .on the compara tor anvil after preparation. Additional long blocks that are to be measured during the day are prepared and placed on a thermal equalization plate next to the instrument. As comparisons of One group of 4 are completed a new size group can be moved from the plate to the anvil and allowed to equalize before comparison. All gage blocks are oriented on the comparator with the top surface

uppermost.
Thermal Equalization Time

Equalization time va.ries with block size, treatment and allowable measurement uncertainty. Blocks prepared in advance and kept in the gaging area are placed on the equalization plate , or in the
case of long blocks, moved from the equalization plat. e
to the

comparator anvil. They may then be intercompared using the following table as a first approximation:

Block Size (inches)

to 0. 250 300 to 1. 000 000 to 20. 000

Equalization Time
(minu tes)

Experiments establishing optimum equalization times should be conducted in youroYlIl laboratory because of the many variables involved and differing measurement uncertainty requirements.
Temperature Measurement

Temperature measurements are made with a calibrated mercury-inglass thermometer. The thermometer is mounted on the block storage plate in the case of short blocks and on the instrument anvil for Apply the test described in the last paragraph of Section E to reveal significant temperature differences among the

the long blocks.

blocks.
In cases where thermocouple measurement of temperature in long

blocks is necessary because of incurable gradients 01;" exceptionally high accuracy requirements , the thermocouple leads should be attached to the instrument column to preclude their interfering with vital parts of the comparator but with enough slack to permit the blocks to be moved on the anvil. The leads are labeled before insertion into the blocks. The thermocouple is inserted halfway down the hole in the block and the hole is plugged at each end with a small piece of polyurethane. A rubber band placed about one- half inch below the top of the block keeps the thermocouple Wire in place. Be sure the thermocouple junctions are insulated Transparent mending tape is sufficient insulation.

electrically from the block.

Handling Techniques The comparison sequence is described in Section F7 but the success of this " 1" intercomparison is largely dependent upon block handling techniques. These techniques include the insertion of all blocks between the styli in a like manner. The operator should develop a rhythm , after he acquires some experience with the process, that will ensure that each pair of blocks is handled for approximately the same length of time as all other pairs in the series. The time interval to make the sixteen observations should be approximately 2 to 4 minutes for an experienced operator with automatic data recording or an assistant doing the recording.

A camel'

s hair brush or an air bulb is useful for sweeping or blowing dust particles from the blocks and the anvil just before

insertion.
The short blocks are moved about by grasping them with rubber tipped 10-inch tweezers. When handling square style blocks, the tips of a pair of tweezers may be bent to accommodate this con-

figuration.
Gloves may be worn in addition to using tweezers when handling blocks from 0. 400 inch through 4 inches. The blocks above the 4-inch size are preferably intercompared on a different instrument situated in a more precisely controlled temperature environment and the operator may wear a Mylar reflective cape as well as

gloves.
A special pair of tongs can be made to facilitate handling (see fig. 2) when sliding long blocks (5 to 20 inches) about the an-

RUBBER
DIP

~ APP~OX

A LUMINUM ALLOY
ALL DIMENSIONS IN INCHES'

HOKE BLOCK TONGS

FIGURE 2

vil and between the measuring probes. With these tongs , long blocks should be grasped about one-half inch above the bottom to effect the surest, safest method of moving the blocks.
The anvil stop plays an important part in the seating of the probe on the block surface as well as the positioning of the gage point between the probes. The block is moved tangentially toward the stop allowing one corner ' to touch the stop and then with an angular motion (which se. ats the probe on the block surface) proceed to abut the entire edge of the block to the stop.
Observa tions

The sequence of observations is from a design series developed to compensate instrument and temperature drifts (3). The series of readings for each size utiliz.es two calibrated master blocks, S and S The order of obserz' and two unknown blocks, X and Y. vations is as follows:
l = Sl - S

== S - Y
= Y
- S

= Y

- S

A=

- Y

A = S -X
= X
- S

4 = S2 - X

Temperature and Deformation Corrections

Deforma tion correc tions for various stylus radii and pressures

(4),

can be applied to the observations when comparing blocks of different materials if the correction is of sufficient magnitude to

be significant.
Temperature correct.ions
are applied to all blocks above 0. 350 inch in size when the blocks are of different materials. This size limit can be raised if larger uncertainties are acceptable.

Computation and Analysis

A computer program and discussion of data processing is in refer-

ence 3. Figures 3 and 4 show the file format, input and output.

== ...

y"

...

PRELMY FILE FORMAT

SAME AS PRELMX FILE FORMAT, BUT NOMINAL SIZES MUST AGREE J~ INE FOR LINE

REPORT FILE FORMAT (OUTPUT FOR 1108)

LINE 0010 DATE OF OBSERVATIONS LINE 0020 TWO INITIALS OF OBSERVER

LINE 0030 INSTRUMENT

LINE 0040 LINE 0050 LINE 0060 LINE 0070 LINE 0080 LINE 0090 LINE 0100

USED, FEDERAL #1 OR FEDERAL #2

TEST NUMBER FOR " SET OF GAGE BLOCKS SUBMITTED BY " X" CUSTOMER ADDRESS OF " ADDRESS OF " SERIAL NUMBER OF SET " MAKER OF SET " GRADE OF GAGE BLOCKS IN SET "

LINE 0110 TEST NUMBER" FOR "y" SET OF GAGE BLOCKS LINE 0120 SUBMITTED BY "y" CUSTOMER LINE 0130 ADDRESS OF " LINE 0140 ADDRESS OF " LINE ' 0150 SERIAL NUMBER OF SET " LINE 0160 MAKER OF SET " LINE 0170 GRADE OF GAGE BLOCKS INSET "
LINE 0180 LINE 0190 LINE 0200 LINE 0210 LINE 0220 LINE 0230
~OM, SER(X) , COR(X) , COEF(X) , PEN (X) , SER(y) , COR(Y) , COEF(Y) , PEN(Y) (REPEATS' THE ABOVE LINE FOR EACH GAGE BLOCK)

COLUMN SUMMATION CHECKS VALUES

NOM. SIZE , TEMP , WITHIN, TOTAL , DIFF , T-TEST, STD. DEV. , F-TEST

(REPEATS THE ABOVE LINE FOR EACH GAGE BLOCK) COLUMN SUMMATION CHECK VALUES Key to abreviations

NOM

... nominal size

SER
COR

... serial no. ... correction (deviation

from nominal size)
coefficient of

TOTAL

DIFF
T-TEST
SID . DEV.

... measured difference

... process standard deviat ion (long term)

between the two masters

COEF

PEN
TEMP

WITHIN

... within- group standard deviation

expansion penetration correction temperature

... standard deviation of

statistical t-test
this run

F-TEST

= statistical f-test.

FIGURE 3

. .

-.

XBAS IC GAGE6 RUN

Computer input data

STARTING WITH NOM. SIZE !0. 130000 ENTER NO. o.F GAGE BLo.CKS 1 1

WRITE IN DATE wlo COMMA !MAY 15 1974 TWo. INITIALS o.F OBSERVER 1 GC INSTRUMENT F1, F2, OR F3 !F2
(l)KEYBOARD OR (2)DATAFILEl1

SIZE . 13 INCH X = 111 153. 51.5 Y - S. 152. 52.

S. - 5...
- Y 150. 51.
50..

Y = GDW

Y-

S.. - X 150.. S.. - Y 150.. S. 150..

S. - X

ALL SYSTEMS GO
THIS CALIBRATION IS NOW COMPLETE

- S.. 149. 50..

151.

50. 51. 49.

(inmicroinches)

Observations

~mER o.F REPEATS =

TURN o.N PUNCH AND TYPE G.
?LIST REPORT

0010 MAY 15 1974

0.020 GC

Print- out of results

0030 F2 0040 XX123

0050 C. E. JOHANNSo.N co. 0060.

0070
0.0.80 515

0.090. CEJ 0.10.0 GRADE A

0110 XY123 0120 Fo.RDMOTOR co.. 0130 DEARBORN MICH.

0.140
0150 132 0160 WEBBER 0170. GRADE AA 130000 111

0180 . -1.315 11.5 0190 . 130000 -1.315 11.5 0200 0210 . 130000 20. 000 340 550. 0220 . 130000 20. 000. 340 550

.1

0.00. GDW 0.00.

335 335

11.5 11.5
527 527

0.00 0.00

150. 150.

167 167

401
2 . 40.1

* See line numbers in Figure 3 for explanation of print-out. Nominal size Corrections (deviations from nominal size) and standard deviations are in micro Coefficients of expansion are x 10-6 per

is in inches.

degree Celsius.

inches.

FIGURE 4

The results of the 4-1 series intercomparison are explained as to statistical handling of data, history of master blocks and predicted lengths in all size ranges in references 1 and 2.

Summary
The electro-mechanical length comparator can be utilized to transfer length values very precisely. It is useful to understand the various parameters that affect the comparative proce$s, namely, instrumentation, temperature, block handling techniques and intercomparison design. This paper should be used in conjunction with the references on page 13.

References

(1)

Pontius, P. E., The Measurement Assurance Program - A Case Study, Part I, Short Gage Blocks (0. 1 to 4 inches), NBS Monograph 149 (in process).

(2)

Pontius, P. E., The Measurement Assurance Program - A Case Study, Part II, Long Gage Blocks (5 to 20 inches), will be NBS Monograph (in process).
Cameron, J. M. & Hailes, G. E., Designs for the Calibration of Small Groups of Standards in the Presence of Experimental Drift, NBS Tech Note 844 (1974).

(3)

(4)
IS)

Bowman, H. A., Beers, J. S., Gage Block Deformation from Electro-Mechanical Comparators, NBSIR (in process).
Tucker, C. D., Preparations for Gage Block Comparison Measurements, NBSIR 74-523, 1974. Beers, J. S. & Tucker, C. D., Gage Block Flatness and Parallelism Measut:"ement, NBSIR 72- 239, August 1973.
Beers, J. S., A Gage Block Measurement Process Using Single Wavelength Interferometry, NBSMonograph 152 , in process.

(6)

(7)

APPENDIX

Experimental Evidence of Thermal Effects in Long Gage Blocks

ments on long gage blocks. o

foam.

Figures 5 through 9 summarize the results . several thermal experi-

In all cases tem~erature was measured with thermocouples inserted halfway down the hole in the square-type blocks and the hole was plugged at each end with a small piece of polyurethane

Figure 5 shows the temperature of two blocks, one in the shadow of the other, as they were exposed to a 40 watt heat source 18 inches away. (Note: The average human generates 100 to l25 watts. As expected , the temperature of the block in the .shadow did not rise as much. Both blocks had nearly identical ground finishes on their sides.
In figure 6, two blocks side by side were exposed to the hea t source. In this case, one block bad a dull finish on its sides and its temperature rose considerably higher than the brighter finished block. The same two blocks were then individually wrapped in reflective Mylar film and exposed to the heat source again. Figure 7 shows the results - almost identical temperature response and a reduced rise.

Further tests were made of the internal temperatures of four l6-inch blocks while they were being intercompared in the normal " 1" procedure. Each point in figure 8 indicates the maximum temperature rise of any single block during the 3- 1/2 minutes it took to. make the intercomparisons. There was no covering on the blocks for the points in the first group, one layer of reflective Mylar was on each block during the runs in the second group, and three layers of Mylar were on each block in the last group of runs. figure 9, the temperature differences among the four blocks are plotted. Maximum temperature difference between any two blocks at ~he moment of the first comparison is indicated by a dot for each run and the maximum temperature difference between any two blocks at the moment of the last comparison is indicated plus sign.

bya

Temperature differences among long blocks being intercompared are by

f~r the largest sources of uncertainty in the process. These experiments

show that temperature stability and equality can be greatly improved by wrapping the blocks in reflective film.

use OMM. NBS-DC

:;)

....-+ ' """"-.'

' - -+-+-

~ 20.100
...I

090 080
40W FLU

ORESCENT
LAMP

lli

' T

+ ..... B

.... A

070 CD .060 20.050 Uj , .040 030

LAM

It .020
010

w-I' '
20 30 40
TIME IN MINUTES

~. 20.000

FIGURE 5.

TEMPERATURE RESPONSE OF TWO 16-INCH GAGE BLOCKS.

(/)

~.

140 130

120
110

.r
L.:::.l

DARK FIN ISH

FIN ISH

f71. r;1..- LIGHT

.J 20.

;2 . 100 U .090
Itl .oao
070
UJ

40 W FLUORESCENT LAMP

:J .040
()30

060 20.060

I- .010
20.000

020

20 30 40
fIGURE 6.

TIME IN MINUTES

TEMPERATURE RESPONSE OF TWO 16-INCH GAGE BLOCKS.

(/)
' , " ~ -- ---....

(j) . (!) . ~
'"

,,~. ---

CARKFINISH
- ::::b..-

0.
ILl

, tL.~HTFINISH

070 en 060 1M 20.060 It: 040 ILl 030 ILl .020

010
L.AMPON

0.

n::

~ 20.000
ILl

a..
ILl

TIME IN MINUTES

;FIGURE 7..

TEMPERATURE RESPONSE OF TWO 16-INCH GAGE BLaCKS WRAPPED IN REFLECTIVE FILM~

~~.
(!)
(/)

~ , ;:.

q;. , .,? '

..
q,

".

..

NO COVERING ON BLOCK5
I REFLECTIVE I I '

11L.AYER OF!

I MYLAR

3 LAYERS OF REFLECTIVE MYLAR

1 .-.f f!,;;

. E) 5 E) x
c( :J 1

1: -

5t 51

UJUJ 9
0::0 :J

..J

I EJ52EJY
. . I

I- UJ

t/)

I BLOCK STORAGE PO51TIONS

521

c(UJ 7
0:: 0:: UJQ')

15

2'

0. UJ

~ C UJ -

I ~'

52 52

I- ::i

~ ~ 4
;;2z

I .I 152 I
I 152'
I. 5

25

X -

10 11 12 8 9NUMBER 13 14 RUN

19 20

FIGURE 8.

TEMPERATURE RESPONSE OF l6~INCH. GAGE BLOCKS DURING MINUTE INTERCOMPARIS~N PERIODS. 3~1/2

+ + (+).

NO COVE~ING ON BLOCKS 11 L.AYER OF I 3 LAYERS OF

I REFLECTIVE I
MYL.AR

MYLAR I

REFLECTIVE

11.1

Of/) 9

. UJ

Z':J
I.

Ci)' 0: ..J 8

+1
I.
.I
I .

UJ 7

IL. O.

E ~ 6

~ ~ 6

UJ 4

(( Q

1+ +

~ ~ 3

. I1234667

~ ~ 2

Q. . ;:E

8 9 10 11

12 13 14 16 16 17 18 19 20

RUN NUM6E~

FIGURE 9. MAXIMUM TEMPERATURE DIFFERENCE BETWEEN 16-INCH GAGE BLOCKS AT THE TIME OF

(FIRST COMPARISON (0) AND AT THE TIME OF LAST COMPARISON