re:Re: Interferometer designs for DRO or servo control?

Jon Elson <elson@...> replied:
Subject: Re: Interferometer designs for DRO or servo control?



michael pettengill wrote:

>Out of curiosity, has/is anyone making use of interferometry for sensing

position?

>
>I was browsing Lindsay Pub's reprint of "Precision lead screws, gears, and

Pantographs - Machinery's Industrial Secretes" and see pictures of setup to
measure lead screw quality using interferometer setups that remind me of the
kinds of physics lab experiments we were doing in high school and college in
the 60s.

>
>I'm sure that there are widgets that use interferometry internally, but I'm

thinking more in terms of DIY.

>
>

Interferometers are used in precision measuring gear and in machine
calibration and testing
systems. But, they are extremely sensitive to vibration, and so can't
be used on live cutting
machines. Many of them are so sensitive that someone walking near the
machine will
cause the digits to just be a blur. When the person stops walking,
everything settles down
in a few seconds, but the readout doesn't settle completely for a minute
or so. I've never
heard of any machine tools using interferometers. The ruling engines
used to make
diffraction gratings do use interferometers, but these machines are
wiping a tiny diamond
across an aluminized glass plate, making scratches a few microns deep,
at most. And, they
have to be built to eliminate all vibrations, anyway.

Jon

Jon and michael,
While interferometers are quite sensitive, they are no more prone to
problems from vibration than other phase measurement techniques, such as
grating type DROs. The period of the signal is small, on the order of 316
nm (12 µinch). What is required to measure in the presense of vibration is
adequate bandwidth.

Consider a vibration of 50 micron (0.002 inch) peak to peak amplitude of
frequency 100 Hz (caused by 6000 rpm spindle).
The equation of motion is .025 sin(2 pi 100 t) (mm), the velocity is 5 pi
cos(2 pi 100 t) mm/s. Peak velocity is 5 pi (about 15) mm/s.
With fringe spacing of .000316 mm, there will be 49708 fringes/s. A
bandwidth of twice that, 100 kHz, will be able to measure this signal
without loss of phase information.

100 kHz bandwidth is not much for interferometers, the Teletrac ones I used
20 years ago has bandwidth on the order of a few MHz. The ones that use
Zeeman splitting, like the HP design, may be limited to less than 1 MHz,
single the Zemax splitting is only a few MHz itself (unless they have
increased the magnet size over the original design).

The limiting factors in interometer use in the shop are two fold:
1. the index of refraction of the air the laser goes through changes with
pressure and humidity. This must be corrected by measurement of the local
pressure and humidity and use of an appropriate algorithm. Along a long
path it may be necessary to measure at several points. If this is not done
you will see all sorts of interesting signals.

2. the laser wavelength is not completely stable unless you have a (duh)
stabilized laser. There are two sorts of common lasers used for
interferometry, the HeNe and the diode. The HeNe has the remarkable
property that its wavelength is known to 1 ppm, so if care is taken (see 1.
above) distance measurements of this precision can be made. The HeNe can
mode hope, which can cause the fringe count to jump even then there is no
motion, but there are ways to prevent this I won't get into here.

The second common laser is the diode type. A cheap 650 nm laser pointer
will make a nifty source for your interferometer, but its wavelength is not
exactly known (can be calibrated) and is unstable with temperature and drive
current. Either the wavelength must be measured continuously or stabilized.
Neither of these things is particularly easy. One cool thing about diode
lasers is that the wavelength can be modulated by changing the current.
This makes heterodyne interferometry easy. Known length reference cavities
can be used to calibrate the laser wavelength and pressure humidity, if you
want to make a multichannel system.

Interferometers are used in machine tools and measuring instruments that
require their accuracy. For example, a diamond turning lathe uses them. It
turns surfaces to a precision of a few nm as large as 1 m in diameter. I've
used them on lots of measurement equipment with great success.

As Jon wrote, they are sensitive enough to detect very small motions while
the machine is settling. I consider this a great benefit, as the dynamics
of the machine are then made apparent and can be quickly tuned up.

One great advantage of the interferometer over gratings DROs is that Abbe'
offset error can be reduced or eliminated. At a high level of measuring
precision this can dominate other errors.

Is the Abbe' error well understood by this group? This getting to be a long
post, that could be the subject of a second post.

Elliot