From: "Jerald F. Wright" 
To: tkrajci{at}san.osd.mil
CC: atm{at}shore.net
Reply-To: "Jerald F. Wright" 



    I agree with most of what you say at least as far as it would go toward
mathematically describing to someone something they will have to learn by
rote. By practice they will learn about their repeatability in readings and
hopefully what causes them not to be the same all the time. Control all of
the variables that are controllable to the extent able and willing. The big
things are stable temperature and still air. And the glass being as
thermally stabilized as possible.  If you look at the spread of your
readings within sets when the conditions and mirror are ideally stable that
would be the spread left due to test equipment and operator.

    The best advise on Focault and Ronchi testing, or any other test for that
matter, would be to test a bunch of mirrors a bunch of times.  It is one of those
things that practice will make you much better at it.

    Leave those mirrors on the test stand a couple of days and take sets of
readings under varying temperature conditions.  I know how much my mirrors change
shape with the small daily cycle of rising and falling temperature. I know
how repeatable my numbers are when the glass and air in the room is stable.
If my numbers are falling outside the box for my repeatability for a set of
knife readings I am watching the air move. If I am inside my repeatability
box for a set of readings but it is out of line with the numbers from other
sets of readings with good repeatability in the sets I am watching the
glass move. All my
mirrors have been pyrex so I don't know what plate glass will do yet.

    Everyone will not be able or willing to make their conditions ideal.  The
accuracy of their testing will limited.  Maybe not useless.  If you are trying to
get to 1/8 wave and conditions limit test accuracy to 1/20 wave you are ok.

    One must assume that to make accurate measurements one must have a well
made
test device capable to the required accuracy.

    The various Focault programs make it easy for some enterprising ATM to run
the programs entering various quantities of knife errors and see just how
much difference you can have in those readings and still have a good curve.
 It can give you a good feel for how close you have to be able to measure
the numbers and
how close to the desired numbers you need them be to have what you will
find acceptable for your mirror.

    I suppose one could calculate Q if they wished but it will only give them a
number to hang on a property that they should be able to understand better
through experience.

    So my advise would be practice and control the conditions and test in the
best conditions you can make.  If you aren't already doing that your testing will
improve without regard for whether you know your Q or not.


Jerry

--- BBBS/NT v4.00 MP