YOU said it was "wrong" in your eariler post...You said that Foucault readings are in error, and in effect systematically "wrong" -- MIke Lockwood
I never used the word "wrong". That is too broad of a brush as far as I am concerned. I was very specific. The readings are in error because the crossover positions of the zonal Foucault test are not those of the COC of the zones (as assumed), but actually inside the COC. Irvin Schroader calls it "the knife-edge error" (ATM, III, p. 443). If it isn't, what it is? He even suggests how to test for it:
"Now remove the mask and it will be easy to see that the shadows on the mirror move in the same direction as the knife-edge across both sides of the zone previously exposed by the two holes; that is, the knife-edge is inside the center of curvature of these two areas of the mirror surface"..." (p. 432)
In other words, if the zonal test with the mask cutouts showed both zones to darken evenly and simultaneously, it suggests that the knife-edge was at the COC of the zone, but when the mask is removed it's obvious that it's inside the zone. If a test suggests one thing and it turns out it's another thing, what do we call that if not an error?
The problem with the zonal testing using masks with cutouts seems to have a strong subjective element as well, which would tend to vary form person to person. Again, Schroader (my emphasis in bold):
"Although each side a and b of a masked zone is very nearly spherical, the centers of curvature A and B (Figure 2 and 3) of these spheres are
separate from each other and do not lie on the optical axis. In performing the Foucault test, then, we placed the knife-edge at the
crossover point, (that is, inside the actual centers of curvature) and then struggled to force the shadows to behave as though it were at the
center of curvature of both a and b simultaneously; a little like trying to force a nut when the thread is crossed." (p. 433)
E.H. Linfoot, states that the accuracy of the Foucault zonal test is (again, my emphasis in bold) "limited only by the observer's limited power of brightness-discriminaction." (Monthly Notices of the Royal Astronomical Society, Vol. 108, p.428). On the other hand, I think it is generally agreed that the subjective bias is largely removed in the caustic test.
Perhaps, as Nils Olof suggested, I should have perhaps euphemistically called it a "compromise" rather than an error. It's important also to stress that no matter what we call it, it is not always significant, especially for smaller and slower mirrors. The main point is that the caustic test is more accurate then the Foucault test, which is why it was introduced to begin with, and which makes it "better", especially for bigger and faster mirrors.
Edited by MKV, 16 October 2014 - 11:06 PM.