I use the matching Ronchi test for the majority of my testing. Light passes through the bottom half of a precision ruled glass Ronchi grating, reflects from the mirror, and passes through the upper half of the same grating on into the eye. If the mirror is spherical, the bands will appear straight. If the mirror is parabolic, a series of curved bands can be seen projected on the mirror face. The bands, when compared to their ideal shape as seen on a computer monitor or printout, reveal the mirror figure, in addition, overall smoothness is easy to see, and turned edge is easy to detect. It is key when applying the matching Ronchi test, that the bands be compared to their ideal shape at precise distances from the mirror's radius of curvature. Inside of the radius of curvature, the bands bow outward as you center your attention to the middle of the mirror. Outside of the radius of curvature, the bands bow outward as you move your eye towards the edge of the mirror. The Ronchi test is sometimes inappropriately applied by eyeing the curvature of the bands. But this will not work because the curvature can look very similar for a whole series of situations. For instance, a 10" f/5, fully parabolized at 0.3" outside radius of curvature will look very similar to the bands of a 10" f/5, only half parablized, at 0.2" outside radius of curvature. So we must match bands at precise distances from the radius of curvature. I use a series of tests, 0.10" apart, starting from 0.30" inside of radius of curvature, to 0.70" outside of radius of curvature.
Make a stage sliding on teflon that's pushed by a used micrometer head. Make a little four sided box to hold the light, and place the grating so that half of it sticks down in front of the box, and the other half sticks up in front of the box. Look through the upper half starting inside of the radius of curvature until you pick up the reflection of the grating. Now move the entire assembly away from the mirror until bands fill the reflected mirror image. I like to use a music stand for easy adjusting.
How closely must we match the bands? If we calculate the percentage of under correction that yields ¼ wavefront error, and the percentage of over correction that yields the same error, then plot the under corrected and over corrected bands, we can get a feel for the accuracy of judgement needed. In the software, I show the amount of under and over correction for +- ¼ wavefront error. With patience in comparing the bands, one can figure a mirror very nicely - no math, no zonal measurements, just comparing - something the eye does very well. The goal is to get us close enough to switch over to the star test for the final touchup figuring at high power.
Here are the Ronchigrams for a 16" f/5 mirror, using a grating of 100 lines per inch (light source always 'moves' because the Ronchi grating is moved to and fro, the light passing through the bottom half of the grating and returning through the upper half):
Download my Ronchi software here (Windows 2K/XP/Vista with .Net 2.0 Framework required)
Here is a picture of the little Ronchi tester sitting on a Foucault/Caustic stage that I have used for many years:
Berthold
Hamburger's ronchi page
http://www.geocities.com/CapeCanaveral/9601/ronchi.htm
John
Upton's Ronchi for Windows http://members.aol.com/RonWin20/