Re: ATM Reading foucalt software

["Bill T." <twentiethwave@hotmail.com>]

> Well I finally made it to a sphere, no TDE or other defects.  Now how do I
> interpret the software to remove material to get to a parabola?  I am
using
> Richard Dunnas's windoze software couder (easier to see the progress),
when
> showing the wave front error the graph shows the error to be below the
> center line, is this the area that needs worked on. Or do I need to do
> something different. I started doing this and the ronchi looks like Mel's
> software pics but the software numbers got worse.  Thanks in advance to
all
> who helped me getting to a sphere.  Time and patients will overcome a
> mirror.
>
> Ron DePack

Ron,

It sounds like Richard Dunnas' graphs are similar to those of Tex30 and
Admir, which I use.  A parabola is represented by a straight line.  A
perfect sphere is represented as a "U" shaped curve with the low point at
the 70% zone.

In general, you can parabolize a mirror by "polishing down" those parts of
the graph that are "high."  But it is *very important* NOT to polish
directly on the edge of the mirror!  Concentrated polishing directly over
the outer zones can roll the edge very quickly!

But that edge can be lowered indirectly...

If the curve of your graph shows a high outer edge, and a high or flat
center, concentrate your polishing on the inner part of the mirror.  Be sure
to always bring your lap out slightly over the edge on each stroke, but do
not concentrate work at the edge.  As you push down the center of the
mirror, the edge will "magically" flatten out almost by itself.  That's
because polishing on the center will shorten the focal length there.  By
comparison the edge, which is "falling behind" by not get as much polishing,
will start to seem longer.  That's how a parabola differs from the nearest
sphere -- the center is "shorter" and the edge is "longer."  This technique
will lead to a parabola of slightly shorter focal length.  Subsequent runs
of your software will start to find this new, shorter parabola, complete
with flatter edge.

Remember that it's still important to keep the edge blended into the overall
curve, that's why all (or almost all) strokes should include the edge.  If
you consistently stop short of the edge, you'll develop an obnoxious zone at
the "edge of  avoidance."

These comments apply mainly to figuring with a sub-diameter lap, and amount
to a strategy of digging a paraboloidal hole in the center of the mirror in
a way that minimizes messing with the edge.  For me, concentrating on the
center of the mirror would mean W strokes with a sub-diamter lap that go
almost the full width of the mirror, always overhang at least slightly, and
which strongly favor the center of the mirror.  Something like this:

http://www.kupercontrols.com/misc/lowright.jpg

You can try that stroke if you like, modifying it as you obtain feedback
from Foucault and Ronchii observations.  At first you might want to try a
little less overhang than is shown until your software starts to report
wavefront errors better than about 1/2 wavelengths.

But as always, all figuring advice must be taken only in the broadest of
terms.  One man's parabola is another man's mystery curve.  There are many
other paths to the parabola.


Bill Tondreau