Page 71 - Instrumentation Reference Book 3E

56 Measurement of length

relative position can be obtained accurately even the new definition should be, “The meter is the
if the base triangle moves in space. The major length of the path travelled by light in vacuum
practical problem in adopting this method is that during the fraction (11299, 792,458) of a second.”
the three length measurements each require track- This new definition takes reproducibility of
ing arrangements to keep them following the length definition from parts in lo8 to parts in 10”.
point. The accuracy of pointing, however, is only In common industrial practice, few people require
subject to easily tolerated cosine forms of error the full capability of the standard but adequate
which allow relatively poor following ability to margin is needed to provide for loss of uncertainty
give quite reasonable values. each time the standards are transferred to a more
The three alternatives can also be combined to suitable apparatus.
provide other arrangements but in each case there To establish the base standard takes many
will always be the need to measure three variables months of careful experimental work using very
(as combinations of at least one length with length expensive apparatus. This method is not applicable
andlor angle) to define point position in a general to the industrial workplace due to reasons of cost,
manner. Where a translational freedom is con- time, and apparatus complexity. The next level of
strained the need to measure reduces to a simpler uncertainty down is, however, relatively easily pro-
arrangement needing less sensing channels. vided in the form of the industrial laser interfero-
meter that has been on the market for several years.
Figure 3.5 is such an equipment in an industrial
3.4 Standards and calibration application. The nature of the laser system shown
of length is such that it has been given approval by the
National Institute for Standards and Testing (NIST)
With very little exception length measurements for use without traceable calibration, for, provided
are now standardized according to SI measure- the lengths concerned are not too small, it can give
ment unit definitions, length being one of the an uncertainty of around 1 part in 10’ which is
seven base units. It is defined in terms of the unit adequate for most industrial measurements.
called the meter. Optical interferometer systems operate with
Until early 1982 the meter was defined in terms wavelengths of the order of 600 nm. Subdivision
of a given number of wavelengths of krypton-86 of the wavelength becomes difficult at around
radiation. Over the 1970 decade, however, it 1/1000 of the wavelength making calibration of
was becoming clear that there were improved submicrometer displacements very much less cer-
methods available that would enable definition tain than parts in 10’. In practice, lengths of a
with reduced uncertainty. micrometer cannot be calibrated to much better
The first was to make use, in the manner than 1 percent of the range.
already adopted, of the radiation available from Laser interferometers are easy to use and very
a suitable wavelength-stabilized laser source, for precise, but they, too, are often not applicable due
this is easier to produce and is more reproducible to high cost and unsuitability of equipment. A less
than krypton radiation. At first sight this might expensive calibration method that sits below the
be an obvious choice to adopt, but a quite differ- interferometer in the traceable chain uses the
ent approach was also available, that which was mechanical slip and gauge block family. These are
recommended in 1982. specially treated and finished pieces of steel whose
The numerical value of the speed of light lengths between defined surfaces are measured
c(c = 299 792 458 m-l) is the result of numerical using a more certain method, such as an interfero-
standards chosen for the standards of time and of meter. The length values are recorded on calibration
length. Thus the speed of light, as a numerical certificates. In turn the blocks are used as standards
value, is not a fundamental constant. to calibrate micrometers, golno-go gauges, and elec-
Time standards (parts in 1014 uncertainty) are tronic transducers. Mechanical gauges can provide
more reproducible in terms of uncertainty than of the order of 1 in lo6 uncertainties.
length (parts in 10’ uncertainty) so if the speed of For lengths over a few meters, solid mechanical
light is defined as a fixed number then, in principle, bars are less suitable as standard lengths due to
the time standard will serve as the length standard handling reasons. Flexible tapes are used which are
provided suitable apparatus exists to make the calibrated against the laser interferometer in stand-
conversion from time to length via the constant c. ards facilities such as that shown in Figure 3.6.
Suitable equipment and experimental proced- Tapes are relatively cheap and easy to use in the
ures have now been proven as workable. By field compared with the laser interferometer. They
choosing a convenient value for c that suited can be calibrated to the order of a part in lo6.
measurement needs (that given above) it was, in For industrial use little difficulty will be experi-
1982, agreed by the signatories of the committee enced in obtaining calibration of a length-meas-
responsible for standardization of the meter that uring device. Probably the most serious problem