ice point with a triple point cell, checking the bridge against a calibrated standard resistor and checking the thermocouples for equality when they are all at the same temperature in a well of an insulated copper cube.

4.2. Atmospheric Pressure

The interferometer housing is not airtight and pressure inside it is the same as laboratory pressure. An aneroid barometer is located adjacent to and on the same elevation as the interferometer. Aneroid barometers can be read much faster than mercury barometers and this is an important characteristic in interferometric length measurements. The aneroid used for gage block measurements has good stability, and frequent comparisons with NBS reference barometers insure its reliability.

4.3. Water Vapor

A commercially available Dunmore type hygrometer is used. This type has a sensing element whose electrical resistance changes with moisture content of the air. Periodically calibrated, it has an uncertainty of about 1.5% relative humidity [8, 9]. The sensing element was originally mounted adjacent to the interferometer, but was later moved inside the interferometer housing (see section on process evaluation).

5. GAGE BLOCK MEASUREMENT PROCEDURE

In preparation for measurement, the gaging faces of the NBS standards are cleaned with trichloroethylene and ethyl alcohol. Wiping the alcoholrinsed gaging faces is done with paper tissue or a clean cotton towel. Lint or dust is removed with a camel's hair brush.

The gaging faces are examined for burrs and, if present, these are removed with a deburring stone. The bottom end is tested for wringability with a quartz optical flat. The transparency of the flat makes possible a judgment of the wring. A uniform grey color at the interface indicates a good wring. Any colored or light areas indicate a gap of several microinches between the faces in that area, and such a gap may result in an erroneous length measurement. Deburring and cleaning is continued until a good wring is achieved. The quartz flat is left wrung to the gaging face to keep it clean until preparations are completed for wringing the block to the steel optical flat (platen).

Trichloroethylene and

Preparation of the steel platen is similar. ethyl alcohol rinses are used, then deburring if necessary, followed by more rinsing and drying. After sweeping with a camel's hair brush, a thin uniform film of light grease is applied to the wringing surface with a bit of tissue. This film is rubbed and finally polished with tissue or towel until the film appears to be gone. After carefully sweeping the platen with the brush, it is ready for wringing.

The block is removed from the quartz flat and the exposed gaging face

platen, maintaining perpendicularity by feel. Sliding is continued until wringing resistance is felt, and then with slight downward pressure the block is slowly worked into its final position. Square style blocks, such as the NBS reference standards, are positioned so that the gage point* is at the right in the viewing field of the interferometer. One to four blocks can be wrung to the platen for measurement.

The platen with its wrung blocks is placed in the interferometer and two thermocouples are fed down the center hole of each block, one about three quarters down and one about one quarter down the length of the block. A small wad of polyurethane is pushed into the hole to seal and hold the wires.

A preliminary alignment by autocollimation with the Gaussian eyepiece as the support table is adjusted, will produce fringes. Rapid changes in the fringe pattern occur as the block cools from handling.

It is convenient to wring the blocks in late afternoon and allow overnight temperature stabilization. No effort has been made to determine minimum stabilization time. Two length observations are made several hours apart, but the first measurement is not taken until the laser is in equilibrium. Leaving the laser on continuously when observations are being made over several days eliminates delays. The laser beam is blocked from entering the interferometer except when measuring.

Final fringe pattern adjustment is made so that the platen fringes are parallel to the top and bottom edges of the block and one fringe on the block goes through its defined gage point. The direction of increasing fringe order is verified by pressing down on the eyepiece and observing fringe flow. A photograph is taken of the pattern. Block temperature, air temperature, barometric pressure, and humidity are measured immediately before and after the photograph and then averaged.

The photograph can be either a positive or a negative and image size is determined by the distance between the film plane of the lensless camera and the viewing aperture of the interferometer. A compromise is made between image size and exposure time, but short exposure is desirable because the fringe pattern may shift. Further image magnification takes place in the fringe measuring instrument. A compromise is also made between magnification in the photograph and magnification in the measuring instrument. Too much total magnification degrades image sharpness.

*NBS reference standards in the 5 to 20 inch size range are of the square type with an axial hole, and the gage point is on the measuring face midway between the edge of the hole and the edge of the block nearest the size marking. In figure 2 the right hand edge is the reference edge and the gage point is midway between it and the right hand edge of the hole. In sizes below 5 inches the reference standards are of the rectangular type with the gage point at the center of the measuring face as in figure 5.

parator which has a fixed stage in the focal plane of a microscope movable on X-Y slides by precision screws having drum readouts. Four measurements of distances "a" and "b" are recorded and averaged to obtain fringe fraction f, figure 5. Settings on the block fringe to obtain "a" are made at the point where the fringe intersects the gage point. Settings on the platen fringes to obtain "a" and "b" are made as close as practical to the edges of the block because it is here that the platen best represents an extension of the bottom face of the block. Parallelism between gage block and platen is also read from the photo because this information is useful in detecting a defective wring. Poor wringing contact is indicated if the parallelism is different from its usual value and this is sufficient cause to discard a particular length measurement.

Photographing fringe patterns has several advantages. The photograph is a permanent record which can be reinterpreted at any time with the same or other techniques. Changes in block geometry are readily seen by comparing photographs in time sequence. The coordinate comparator is superior to the common practice of fringe fraction estimation by eye because it is more objective, lends itself to multiple measurements and averaging, and is more precise. Finally, photography is fast and thus permits readings of the ambient conditions to closely bracket the fringe recording. This is especially important because atmospheric pressure is usually in a state of flux.

Calculating gage block length from the data is done in 3 steps:

(1) Calculation of wavelength, A

tpf'

at observed ambient conditions.

(2) Calculation of the whole number of fringes in the gage block length from its predicted length and the laser wavelength in ambient air.

(3) Calculation of the gage block length from the whole number of fringes, the observed fringe fraction, wavelength in ambient air, gage block temperature, and interferometric correction factors.

6.1. Calculation of Wavelength

Edlen's 1966 formula [10] for the refractive index of air is used to calculate the wavelength from the relationship

nt,p,f

P[8342.13 + 2406030 (130-02)1+ 15997 (38.9-02)−1] 10-8

.720.775

This method for determining the whole number of fringes is valid only if the block length is known to better than +1/2 fringe (1 fringe ≈ 12uin. for helium neon laser light) from either its history or from an independent measurement process. The calculation is as follows:

where F is the number of interference fringes in the length of the block at temperature t', and in air of temperature t, pressure p, and vapor pressure f.

L is the predicted block length at 20°C taken from an independent measurement or from the block's history.

e is the linear thermal expansion coefficient per degree Celsius of the block.

t' is the block temperature at the time of measurement.

λtpf

is the wavelength in the ambient air at the time of measurement. The fractional part of F is retained temporarily for the reason explained below.

6.3. Calculation of Block Length from the Observed Data

Generally, the observed fringe fraction is simply substituted for the fractional part of F, but there are cases where the last digit in whole number F must be raised or lowered by one. For example if the retained fractional part of F is .98 and the observed fraction is .03, obviously the last whole number in F must be raised by one before substituting the observed fraction.

The new measured block value, in fringes, is

F' = F + $

Finally, the interferometer aperture correction, 62, is added, the block is normalized to 20°C, and a conversion to length units is made to arrive at the final value at 20°C as follows: