re:Re: how straight is straight[rails]?

Dave writes:

>Hi Carl,

>Straight line is much easier than posistional.

>I thought about a simple laser pointer and a couple photosensors on
>the table. The beam passing between them so they both had some
>reading. They would be on a micrometer table and the output of them
>would be fed to a simple comparator. as the needle drifted one way
>or the other, one would move the micrometer table. Tested every foot
>or so would reveal the accuracy of either axis.

There is a detector built for this purpose in every CD player. It considers
of 4 silicon photodetectors on a single die, very closely spaced. Its
purpose is to generate a servo signal for tracking and focus of the CD.

This is ususally instrumented with a difference/sum circuit, but if you are
looking for is a null the difference may be good enough.

BTW, don't be sure that such a simple scheme will accurately determine
straightness.
There will be some mechanical drift and noise of the laser beam. As the
diode heats up, things expand, the heat path out of the laser is not always
symmetrical, which can cause the direction to change until it reaches
thermal equilibrium.
Further, if the beam has some assymetric component because of lens or laser
imperfections, the center power point of the beam may not propagate in a
(perfectly) straight line. For some highly unlikely cases this could amount
to a large fraction of the beam diameter as on progresses from the lens out
towards infinity.
Good optics will reduce this effect, I suppose you could rotate the laser
and see if the straightness measured the same.

Another effect to keep in mind is local variations in the air temperature.
A gradient of air temperature or humidity or pressure will bend the beam.
Usually the gradients have zero time averages so a long enough integration
will average them out, but sometimes there can be a somewhat stable vertical
gradient.
I have made large tents filled with helium to minimize this effect, a real
PITA. Fans work well in many cases.

All these effects are larger with long path lengths.

Consider a very good viewing night for star gazing. Atmospheric turbulance
might be on the order of 1 arc second ( 5 microradians). Thats traveling
through the entire atmosphere, the path in your shop will be much shorter,
but probably not as still. Most of the effect is in the lowest few hundred
meters of the atmosphere.

There are some interferometers that measure straightness that are a bit more
robust than the split detectors, but are still subject to atmospheric
turbulance.

Its kind of fun to set up the split detector/laser setup you mention, output
it to an oscilloscope, then whack the machine. Many arguments have been
made about stiffness of machines, this is an easy way to measure it. For
example, a lathe can make a facing cut good enough to get a specular
reflection. The variation of the angle of the reflection during rotation
tell one a lot about the spindle and axial movement of the tool.

Elliot