From: "James Lerch" 
To: "ATM List" 
Reply-To: "James Lerch" 


----- Original Message -----
From: "Michael Peck" 

> I'll try this one more time. The closest thing you'll find to an objective
> statistical analysis of the Foucault test (Robo or otherwise) is my
article
> at http://home.netcom.com/~mpeck1/astro/modal/modal.pdf>. It's mostly
> theoretical, but there's some real data in there too.

Hi Mike,

Just wanted to ensure everyone knew that its your work that got me started
down this path! :)

>I've looked at some
> more real data, both that I've taken and one data set from James Lerch,
but
> I don't feel motivated to publish an analysis of it right now. I will say
> that I think James is getting adequate repeatability in his results. I
also
> think he is throwing away good data for no very good reason, and I think
he
> could get better results if he changed his data collection protocol.

Thank you, and I concur, I do discard a lot of possibly useful data!  There
is one reason of course, that being I wanted to keep the ability to
directly compare zonal readings taken by RTAFT, with those taken by a
classic Couder Mask / Foucault test.  With the current code this is easily
done.

In addition, I could be testing mirrors with up to 12 zonal measurements,
in practice we rarely do over 5 on mirrors 8 - 10".  The reason is
simple, if the surface is smooth, increasing the zones tested over doesn't
really make a whole lot of difference in the results.  For our 4hr Class
room environment the difference between theory and practice comes into
play.  In Theory I could nail a Strehl down to 2 or more significant
digits.  In practice knowing it to one significant digit suffices. :)

I guess it all depends on your goals.  When I wrote RTAFT the goal was to
maintain the accuracy of the classic Foucault test in skilled hands, while
significantly reducing the test time.  I'm pretty confident I met these
goals.  In addition there is nothing preventing the code from gather more
data if the need arises, its just that in our class environment the need
isn't there (Yet!)

> Well, ATMJ is defunct and Suiter hasn't had a visible presence on the
> internet in years that I'm aware of. What he did was a computer-aided
> Foucault test that wasn't fully automated and that still relied to some
> extent on subjective judgement.
>
> One thing Suiter did that nobody else has done (to my knowledge) is
compare
> his Foucault test results with an interferometric test of the same mirror.
> Until someone else does that, preferably on a statistically meaningful
> sample of mirrors, the jury is going to remain out...

What can we do to eliminate this?  Anyone know of someone with an
interferometer that might be interested in donating some test time?  I can
scrounge a significant number of test mirrors from our class members.  I'd
even volunteer to transport the mirrors and donate a week of vacation time
if it could help!

Speaking of mirror testing, What is the status on the Mirror Round Robin????

James

--- BBBS/NT v4.00 MP