Making a 150mm f4 mirror.
It starts end of januari 2005 with cutting the mirror blank out of the 19mm glass slab. At first I experimented with a speed reduction construction made of wooden pulleys and a V-belt. As can be seen on the image the lowest speed of the drillpress (600 rpm) suited just as well for the 150mm blank. It cuts through the glass quite rapidly when you change grit often, in less than two hours or so.
The first phase is to dig a hole in the middle, of the proper depth of course: for an 150mm F/4 this is approximately 2.3mm. I used a pyrex dessert bowl as a tool, which was the largest flat piece of cheap glass I could find in the shop (90 cents or so). Throw on the nr. 80, some water, and off you go. After about an hour and a half the hole was deep enough to make a tool and move to nr. 180.
The next image shows what is a usual result after hogging:
Due to the excessive overhang the shape has become very hyperbolic. Also, the edge has hardly been touched. To get to a shpere the ridge at the edge needs to be ground away, something I achieved by hogging to the full depth and then course grind with tile-tool on top to wear down the edges. Start with nr.80 to get approximate conformance.
In the following images you can see the making of the tool. Take the blank, put an alufoil wall around it, smear vaseline on the blank, stick the tiles on it. Take care of not spilling any vaseline on the tiles, or they will get loose (as you can see further below). When ready, prepare the plaster (a high hardness plaster) and pour.
Break the tool in with nr.80 grit, until it conforms with the mirror (sharpie test). Then go to nr.180 and get rid of the pits. Meanwhile check the focal length of the mirror (wet mirror test) and tune it by intermittent MOT and TOT.
The image above shows the partly wet mirror and tool during nr.180. Note that one tile has come loose already!
The next grit sizes I used were nr.320 and nr.800. Window (or float-) glass is fairly soft and grinds away quickly. In fact the whole grinding
process could easily be done in an afternoon or so. The next images show the mirror after these respective grits.
Note that another tile got loose! This was in fact caused by an insufficiently bevelled mirror edge. The mirror then needs to tilt only slightly to bump into a tile and loosen it. A side effect is edge chipping...
The mirror was finished with nr.1200 grit (9 micron) and then a final extended wet. Red-out test shows a light bulb reflected at approximately 30°, which is about as far as you get with this grit.
We're almost there you think .... well, not really! The polishing takes more than half of the total time to make a mirror from a blank. The making of the polishing tool was a bit of a hassle, and there are no photos made. Below you can see the result as seen through the mirror.
The contact is not very great, and I had to repair the low (reddish) parts of the patches lateron to get uniform contact and a smooth motion.
Here you can see the polished-out and nicely spherical mirror, ready for figuring. Unfortunately the pitch appears to be very brittle, and chips off along the edges. Here you can see that a complete pitch square came loose! This has to be repaired and conformed to the mirror surface before figuring can start. I do this by placing some lumps of pitchin the hole and play the butane torch over it. Do not melt the whole tool, but only fuse the lump into the tool!
The above ronchi images reveil that working the repaired tool into conformation has resulted in an overcorrected sphere, with a turned down edge. You can feel the TDE happen, when the mirror is on top and strokes tend to draw a vacuum. This means that either the mirror is not in thermal equilibrium and that the edge is expanded, or that the tool itself isn't good.
The TDE can in principle be removed with 30 min or so of short strokes TOT, slightly left and right off center.You need to bring down the zone just inside of the edge.
These images above show the result of 2x20 min of de-TDE treatment. There is still a small edge visible, but not as bad as it was: since it is going to be used with low magnification I define it ready for parabolization...
Figuring proceeds quickly, only two evenings give a 1/15λ RMS wavefront error, and a Strehl ratio of more than 80%. See ronchi images below:
The analysis with Foucault Test Analysis gives the following results:
Conclusion: the 50% zone and the edge could be brought down a little more. However, the inner 20mm will be in the shadow of the secondary and the rest of the mirror gives a wavefront error of P-V 1/7λ. Strehl and RMS won't differ so much.
Because of the intended rich-field use at low magnifications, I call the mirror done!