Polishing of a 6" f/8 Mirror

After grinding, it's time to make a pitch lap and polish the glass to a clear shine. This is done using a pitch lap and a very fine abrasive called Cerium Oxide.

I used Gulgoz 87 pitch, for a somewhat hard lap. Here's the 6" Pitch lap, after channeling with a razer. The pitch chips are lying scattered about, but mostly on the newspaper. These are thrown away. Channeling with a rzor is a slow process, and getting the right feel for it takes some practice. It is often helpful to cut the channels with water running over the pitch.

After making the pitch lap this way, I warm it to about 100 deg F and then press the mirror on it until I see good contact. Be sure to use polishing compound or the mirror may stick to the tool. Whn I have pressed for 10 minutes or so, with about 25 lbs weight on the mirror, I use my body weight (through my hands) to make sure it is seated well. The squares of the lap should all show contact with the mirror.

Now take the razor blade used to cut the channels and micro-score the squares. This is done by lightly running the razor along at 45 degrees to the channels. The blade is held directly on to the lap, and a light stroke is used- just enought to scatch the pitch.

When this is done so that each square has half a dozen 'micro-facets' I then use a soft brass wire brush to roughen the entire lap. This really makes the friction increase when using the lap and makes polishing very fast. It is also necessary to repeat this procedure after every half hour of polishing, as the lap hardens and does not respond well after a time. If your lap pushes easily or has a 'glazed' appearance, it may need to be roughened up.


I use a thin mixture of cerium, lightly colored, about the shade of this page. At times you may hear a slight screetch. This is OK when polishing, but may cause micro-ripples if continued into figuring. When figuring use a little thicker application of cerium and use less hand pressure. I add a drop of dishwashing soap to the cerium mixture. To figure this mirror took well under 1 oz of cerium.




Time (minutes)

80  20  10-12 wets, 1/2 tsp each, MOT, COC, sag reached.
 120  20  10-12 wets, 1/2 tsp each, MOT, COC
 220  20  10 wets, 1/2 tsp each, TOT, COC
 320  20  10 wets, 1/2 tsp each, TOT, COC
 30 micron  12  3 wets, 1/2 tsp each, MOT
 9 micron  12  3 wets, 1/2 tsp each, MOT
 Polishing-Cerium Oxide 
(micro 90 liquid polish)
 30 3 wets, MOT. After 20 min, looked polished out, figure tested was hyperbolic.
2 wets TOT. Ready to figure after 10 min more, close to sphere. 
Foucault shows some biscuit (see photo below)

10, 10,


W stroke, some biscuit still-1/20 wave ripple, prob.

Change stroke: 3 trips around barrel 1/6 chordal, 4 trips 1/3 chordal, pressure on part of mirror over tool.


And the Results:

6 Inch Mirror

101.5 Raduis of Curvature, FL 50.75

Moving light source, 4 zone mask .5 in inner zone radius

 Zone  Radii  KE Reading  M-L Tolerance  Wavefront Err
 0  .5




 1  1.56 .295  .0441  -.0173
 2  2.15  .303  .0245  .0017
 3  2.61  .317  .0191  -.0090
 4  3.0  .325  .0162  .0082

1/20 wave, with about 1/20 wave ripple - 1/10th overall.


When testing a mirror, one should alwyas roate the mirror 90 degrees and make new readings again to ensure there is no astigamtism. On this mirror, interferrometer testing revealed the ATMers most feared problem, next to turned down edge. This was noticeable when I tested, but I attributed the slight irregularity to uneven cooling, as the mirror was tested after only 1/2 hour after a figuring spell. The astigmatism is about 1/2 wave, too much to allow, so I will continue figuring, with mirror on top in hopes of reducing the astigmatism.

When I got the mirror back from the Interferrogram Tests, the results showed the mirror to be 1/8 wave at best. What happened? It appears that when I placed the mirror on the stand I got lucky and tested the area of the mirror that had the least astigmatism. In the plane that the mask was on, the mirror test quite good. Turn the mirror 90 degrees and retest, and the results aren't so good.... star testing would also quickly show this kind of defect.

I hope to make and show Foucaultgrams and Ronchigrams before doing the final figuring.