----------------------------------- Mircea Pteancu 14 Feb 2009 11:38 ----------------------------------- In corespondenta de mai jos de pe ATM List,dintre Peter si Mike Lockwood,veti gasi mentionat: - testul cu stea,reala sau artificiala, ca metoda de verificare a rezultatelor obtinute prin alte metode -ca astfel testul Foucault poate fi folosit la oglinzi mult mai luminoase decat se crede -Mike recomanda grupul Yahoo al lui Carl Zambuto celor care vor sa studieze in profunzime calitatea optica a obiectivelor -Mike isi doreste sa adauge testul interferometric printre metodele sale de testare,lucru deja realizat de ceva timp de catre Tavi Stanescu -aflam ca Mike lucreaza la o oglinda de 36 cm diametru F/2,6 Sublinierile imi apartin.Mircea Peter, Peter wrote: > Wonderful work Mike!! > I've tried making a few fast mirrors but I found it difficult to get a > decent figure when using the Foucault test. The shadows are tricky to detect > and the tolerance in the 1/4 wave front tornado diagram is unbelievably > small. As a result I have not had very good success parabolizing mirrors in > the f/4 or faster range so I stay away from them now. I wanted to ask if > you'd shed some light how you've managed to do this? Thanks. The first thing I'd say about testing is don't believe everything you read about making masks for Foucault testing or where its limits lie. What theoretically gives the best test results does not always do it in practice. You need to make masks with zone widths (and heights!) that suit testing the particular focal ratio, your eye, and how you view the shadows (I use a small telescope), and you need to validate it. Experimentation and practice are the ways to do this - find the combination that gives you the most repeatable results, then star test the mirror to verify that the figure you measure is close to what you measured. This telescope will give me good feedback at F/3. At F/3.6-3.7 I have gotten excellent agreement from my own observations and from Rick Singmaster's feedback, so I know that my measurements are accurate. It is very important that the mirror be a figure of revolution. I check this before, during, and after figuring with an artificial star test. When the mirror passes this test, the only astigmatism I have seen in the scope is due to thermal effects during cooldown. If the mirror is not a figure of revolution, then it may test differently across different axes, and that leads to inconsistent measurements. You must support the mirror properly on the test stand. The artificial star test will tell you if it's potato chipping or being distorted. A smooth mirror, lack of air currents in the test cavity, and a sturdy mount for the mirror and tester are also critical for accurate Foucault testing. A mirror with bad zones or nasty roughness cannot be tested accurately and repeatably with Foucault. I test in a basement on a concrete floor with the tester on a steel tripod that would normally carry a large telescope. Finally, I'd like to say that is is OK to look at your previous test results while you take readings. This allows the tester to compare previous readings immediately and one can catch a lot of measurement issues this way. (Testing blind is for those who can't take measurements without bending the truth, in my opinion.) A one-minute figuring session will not change the figure that much, so if the readings are markedly different, or if a new feature has appeared since the last session, or if your second set of readings during a test session is quite different from the first, you will immediately know that something is up. Maybe an air current has corrupted a reading, maybe you bumped the tester, or maybe the reading was written down wrong. You also learn a lot about how to look at data - features that appear consistently are probably real, while those that come and go probably are not, and one can then look for the sources of error that "created" them. As for figuring, I'd start by going through Carl Zambuto's figuring examples on his Yahoo group. Beyond that there's not much I can tell you other than take careful notes and don't use hard pitch. Everyone's techniques and machines are a little different. > Someone once told me that on fast mirrors Kennedy can only go so far with > Foucault testing then he switches to star testing all fast mirrors. I don't > know if this is true but seems like one way to get the performance. I don't know what Steve does exactly, but I do know that he star tests, which is good for the primary and secondary. I believe, based on my own experience and conversations with extremely experienced individuals, that Foucault can be used for much faster mirrors than is commonly believed. As always, a second test to double-check results is a good idea, like star testing or autocollimation with a good flat. I'd like to add interferometric testing with autocollimation to my list of available tests (as a final test), especially as I make faster optics. We just generated a 14.5" F/2.6 mirror, so that will be my next experiment. Mike Lockwood Lockwood Custom Optics