Re: ATM Are you parallel?

["Frank Ward" <thewards@mindspring.com>]

Bill's analysis is perfect.

I have found a way to do it a little faster and you
might consider it also.

With the tester on a tripod at the measured position
for the ROC turn on the light and step back four feet
or so.  The light will appear on the mirror. Twist the
mirror left or right until its right edge is on the edge
of the knife.  Up and down will be put in by cranking
the tripod.

O.K. That took 45 seconds.

What I did a little differently is adapt the tester with a
fulcrum handle that pushes the slide against a depth
gage accurate to .001".  By pushing or pulling the fulcrum
handle I can run thru the whole shadow spectrum in a second
or two.  When ever it starts leaning one way or the other I
turn a screw on the tripod that centers the light.

An easier way is to adapt the Ronchi Test. If the bands leave
the mirror quickly on one side or the other you are not parallel.
Swivel your tripod so they go off the mirror equally on both sides.

The above unfortunately took two minutes more, but once done
you need only replace the mirror exactly, look for the reflection
and click the mirror left or right to get it on the blade. Then you
are ready to see what trauma you have done to the surface
"this time".

In friendship,

Frank Ward
Atlanta, GA
----- Original Message -----
From: Wm. D. Hanagan, Jr. <hanaganw@attglobal.net>
To: James Lerch <jlerch1@tampabay.rr.com>
Cc: <atm@shore.net>
Sent: Saturday, April 07, 2001 6:13 PM
Subject: Re: ATM Are you parallel?


>
> James Lerch wrote:
> >
> > Greetings all,
> >
> > Question:  How do you determine if your Foucault tester moves parallel
to
> > the optical axis And the knife edge is perpendicular to the optical
axis?
> > This is important, yes?
> >
> > Details:
> > Slitless Foucault tester has 2 axis of measured movement
> >     X axis moves the Knife edge across the mirror diameter
> >     Y axis moves the Knife edge towards/away from mirror
> >
> > The tester sits atop a camera tripod with which I can adjust the Tilt,
Pan,
> > elevation
> >
> > More often than not, in my limited experience, I have found that as I
move
> > the Y axis, I often 'lose' the mirror return beam (Or find the
top/bottom
> > Left/right side of the returned beam).  I have determined that the Y
axis is
> > not parallel with the optical axis.
> >
> > Asides from trial and error, is there a more accurate way to ensure Y
axis
> > is parallel to the optical axis, and how close is "close enough"?
> >
> > So far its not a problem, as I have not needed to take any real
> > measurements, but I presume that the time is nearing.
> >
> > Thanks,
> > James Lerch
> > St. Petersburg, FL
> > http://lerch.yi.org/atm for my latest ATM adventures
>
> James,
>
> Here's an easy way to align your tester's y axis to the optic axis of your
> mirror.  It takes me a while to explain it clearly, but once you
understand it
> you will find it easy to do.
>
> First of all, make sure that the center of your mirror and your knife edge
are
> at the same height to within an inch or better.  Since your test stage is
on a
> tripod, you need to take some precautions to make sure your test stage is
> reasonably level and nearly impossible to move accidentally.
>
> Start aligning the y axis by simply AIMING your tester at the mirror.  You
> should have a straight line on your test stage that you know to be
parallel to
> the direction of travel produced by your y axis drive screw.  If you don't
have
> such a line on your test stage, draw one on it.  This sight line could be
as
> simple as the corner edge of a board.  Using that line like a gun sight,
aim the
> tester at a point on the mirror.  You want to aim for a point on the
mirror that
> is offset from the center of the mirror by the same distance and direction
that
> your sight line is offset from your knife edge.  Ordinarily, you only need
to be
> approximate with this--there is no need to mark a point on the mirror or
make
> precise measurements, just don't ignore several inches of offset,
especially
> with a small mirror close to the tester.
>
> To determine if your Foucault tester y axis is reasonably parallel to the
optic
> axis of the mirror, move the knife edge far enough inside the radius of
> curvature (ROC) of your mirror that you get a sharp shadow on the mirror
when
> you cut the knife edge into the return beam.  The shadow of the knife edge
> should be on the same side of the mirror as the knife edge itself.  Adjust
the
> x-axis position to place the edge of the shadow as close as you can get it
to
> the
> center of the mirror.  Now, using the y axis screw, move the knife edge
far
> outside of the ROC.  Here, the shadow of your knife edge should be sharp
again,
> but should fall on the other side of the mirror.  If your tester is well
> aligned, the edge of the shadow will be close to the center of the mirror.
> Viewing the mirror from above your tester, hold your little finger up at
arms
> length in front of the mirror.  If the edge of the shadow moved away from
the
> center of the mirror to either side by less than the distance blocked out
by
> twice the thickness of your little finger, which is roughly 2 degrees,
then your
> alignment is more than good enough for the quantitative Foucault test.
>
> If your y axis alignment is off, your null readings will be compressed by
a
> factor equal to the cosine of the angle by which you are misaligned.  For
> example,  let's say that your y axis alignment is off by 2 degrees
entirely in
> the plane defined by your x and y axes--the plane in which your tester can
> move.  Two degrees is roughly twice the width of your little finger held
at arms
> length.  The cosine of 2 degrees is 0.9994, so your readings are all going
to
> shrink by 0.06%.  This miniscule error is certain to be dwarfed by other
errors
> in the zonal readings, in particular your ability to judge the null
position for
> each zone.  Even if you are using a good quality 1" micrometer that reads
to
> .001" your reading error is likely to be at least 0.1%.
>
> A more serious concern related to misalignment of the y axis is that the
knife
> edge will move in and out of the beam as you move the knife edge along the
y
> axis.  It's easy to get the impression from reading books that the
Foucault test
> requires you to immobilize the knife edge in the x axis while you make
your
> zonal readings by moving the knife edge only along the y axis.  That's not
true,
> and
> if you try to do the test that way, there is a chance that the error in
the y
> axis alignment will cause the knife edge to become completely withdrawn
from the
> light path and fool you into believing you have achieved a null on your
last
> zone when what you have actually done is let the knife edge drift out of
the
> beam.  More likely, the knife edge will be shifted in the x axis to a
point
> where the sensitivity for detecting the null is reduced.
>
> You should get in the habit of adjusting the extent of penetration of the
knife
> edge into the beam using the x axis screw as you close in on each null
> position.  Unless the light level you judge your null at is clearly less
than
> the maximum, you can't be sure that you have found the null position for
that
> zone.  Also, as you move
> the knife edge rapidly through the null position, you should see that
first one
> side of the zone is brighter, then the two sides of the zone are equal,
and
> finally the opposite side is brighter.  If that's not happening, you
aren't
> actually moving through the null position for that zone.
>
> According to both Texereau and Howard, who also describe the use of slit
> testers, you want the knife edge to be at a point in the beam where it
produces
> only a small reduction of the intensity over the zone to be nulled.  I
find that
> my ability to see the change in the null position with my slit based
tester is
> better when the knife edge has cut the intensity to a comfortable viewing
level,
> usually about 1/2 of the beams unblocked intensity.  I like to drive the
knife
> edge in and out at the null position to see if I can find a midpoint where
I see
> an imbalance in the light level, then move the y-axis again to more
closely
> determine the null.
>
> I should mention that, if you have a tester similar to Texereau's in which
the x
> axis rotates the knife edge into the beam instead of moving it along a
line, you
> will have a little more reason than most to get the y axis well aligned
before
> making measurements.  This is because the same adjustment affects the
> parallelism of the slit and the knife edge.  It's probably best in this
instance
> to refine the alignment of the y axis as much as you can and to only use
the x
> axis to 1) make the knife edge parallel to the slit and 2) to insure that
the
> knife edge is at least part way in the beam before you read a null
position.
>
>
>
>
> Clear Skies,
>
> Bill Hanagan
>