Well, spring and summer brought many chores arround the house, a constant scramble after enough business to keep marginally afloat financially, and a bazillion other distractions. Time available for atm-related activity was scarce and pretty much entirely taken up with Stellafane Convention. For the first time I pitched my tent a few days early and helped with setup and preparation for the big event. That was a lot of fun. It really is remarkable how the STM crew works. People just see something that needs doing, and do it. Somehow, without much overt organization, the whole immense project comes together.
Finally, last week, I got back to the mirror. When I left off after the last post below, I had been advised continue to do a narrow W to both deepen the correction in the central zone and spread it toward the edge, so, after several hours of pressing to get the lap back into contact, I did that for fifteen minutes or so. Then on Saturday I took the mirror to Stellafane to work on it where I could get guidance from experts. A standard Foucault test showed about 75% of the total needed correction (in terms of knife-edge motion to move a null from the center zone to the edge zone), so once again we did a short session of narrow W strokes (I say "we" because Dave Kelley, one of Stellafane's mirror-making gurus, did some of this work both to show me exactly what he meant and because, lacking a cleat setup to hold it, we were having trouble getting the tool to stay put on the table—so I held the damn thing with extended fingers while he worked the strokes. This brought the total correction right to the correct value, but when we looked at the mirror using autocollimation, we could see that the edge zone was still somewhat discontinuous from the desired paraboloid.
Autocollimation testing involves a reference flat, and as I've noted below, Dave has a spectacularly good one. The light reflects twice from the tested mirror—originating behind a central hole in the flat, traveling to the tested mirror (at focal length, not radius), reflecting back to the flat, and then from the flat back to the mirror, and finally back through the central hole in the flat to the knife edge. The result is that with a true paraboloid you see an evenly illuminated field, which "winks out" evenly and abruptly when the knife edge intervenes—a null test. Any zonal shadows represent departures from the paraboloid. It's a qualitative test, with no easy way to measure errors, but very sensitive because any error affects the rays twice.
On my mirror we could see a brighter zone on the right edge and a darker one on the left—an outer zone on the mirror that was focusing longer than ideal. Dave prescribed a short session of chordal strokes, with varying overhang around 1" or less and a little thumb pressure on the overhanging side. I did three rotations of the mirror like this, maybe ten minutes, and then cleaned it off and put it back on the tester. Dave took a look and said “you bagged it” (it took me a second to process that—at first I thought he meant something had gone terribly wrong). But what I saw when I looked at it was beautiful. The knife edge cut off the light almost like flipping a switch, and the surface looked velvety smooth. It seems I have a very good mirror, considerably better than the first one. Can't wait to get it into a scope!