Grinding and Polishing a Meniscus Mirror

In many respects, the grinding and polishing stages of fabricating a thin meniscus telescope mirror are not different from any telescope mirror. There are lots of resources, online and in print, describing that. This article will focus on the specific issues introduced by working on such a mirror.

If you have been following along, you now have a slumped blank that is of the correct curvature for your telescope design. Hopefully, it is f/3 (or faster) because you may as well push the envelope a bit. Regardless, the curve is reasonably correct and the surface does not have any major defects. If there are small defects, hopefully they are small bumps and not pits, because it is a whole lot easier to grind down bumps than it is to grind the entire surface to meet the depth of a pit. Sadly, I have had to do both, at times cursing why I bothered with the meniscus when I had to do all this coarse grinding anyhow. But I still ended up with a thin mirror I was really proud of, and so I persevered.

While tools for grinding telescope mirrors is covered extensively elsewhere, I will say that the challenge posed by a fast mirror is that the sharp curvature requires your grinding surface to be made of small “bits”, whether those bits are tiles, coins, nuts, or whatever. In my case, I cast the base of my tool from Hydrostone, using my curve as a mold, and once that was cured I then epoxied 1” ceramic tiles to the top using J-B Weld epoxy. The curvature dictated that my tiles were 1” or smaller, otherwise they would wear very unevenly. Since you will be using this same tool for the entire grinding process, you want to make sure it is solid, and the tiles are very well attached because once the tool and glass have been “mated”, you won’t be able to replace a tile that falls off. Inevitably though, a few tiles will still come loose and fall off, which is another reason why small tiles are better: there are more of them.

When you are working with a tool on top (the preferred approach at this stage), you will need to support your glass appropriately and, as you can imagine, the curved back presents some challenges. Some ATMs use a bunch of towels that conform to the curved back, but in my experience I found the glass moved around too much. Another option is to actually cast another tool, this time using the back of the mirror as a form. The benefit of this is that you can use the tool to grind the back of the mirror into a regular shape, more regular than what simply comes out of the kiln. Some argue this is essential but, frankly, I skipped this step and haven’t seen ill effects. The bonus of doing this, however, is that you end up with a curved support and so by simply putting some non-slip material in between the glass and tool, you will end up with a rock solid support for grinding.

In my case, my mirror support was a styrofoam “donut” used to make decorative wreaths. The styrofoam supported the mirror at about 80% radius, and attaching some non-slip material to the donut with reversed duct tape kept my base nice and stable.

If your surface out of the kiln is pretty smooth, you ought to be able to start grinding with 120 grit; otherwise, you may need 80 or 60. If you do go fairly coarse, then you will need to keep an eye on “wedge”: making sure the thickness of the glass doesn’t start to vary too much from middle to edge. This is tough to measure without a decent set of calipers, but is doable with a depth gauge that spans the width of your glass. Also keep in mind that when you grind with tool on top, you flatten your curve, and mirror on top deepens it. The goal at this point, though, is to first get your tool to have excellent contact with the mirror surface, and then to get rid of any surface anomalies introduced by the slumping.

The progression of grit sizes that worked best for me was 120, 220, and 400 of carborundum. Then switching to aluminum oxide, it was 25µ, 15µ, and then 9µ. A very common question is: how much time on each grit? The answer is, frustratingly, that it depends. The way to know you are done at one grit is when you inspect the surface with a magnifying glass or loupe and see an even surface “roughness”. Each stage grinds away large pits (and…yikes…scratches), but leaves smaller pits, and the pits should all be the same size before moving on. In my case, I spent between 1-2 hours at each grit size.

You see, there are two goals for the fine grinding of the surface. The first is to get it smooth. And the first is to get it regular. Yes, they are equally important, because a smooth but astigmatic surface is useless. So while you are inspecting the surface for pits of increasing fineness, you are also measuring your surface to check two things: the radius of curvature of the centre of the mirror, to ensure it is where you want it for your target f/ratio, and the radius of curvature across the entire surface to ensure that it is the same.

EXCEPT FOR THIS.

The tool to measure the central radius of curvature is very simple: a rigid bar of wood or metal that spans the diameter of your mirror, and a radial or digital depth gauge positioned at the midpoint of the bar. When properly calibrated, this allows you to measure the depth of the curve at the centre of your mirror, the sagitta. For any circle (the 2D projection of a sphere), you can easily compute the precise sagitta for a chord that cuts the circle, and that represents the depth your curve must be if you position the bar holding the gauge precisely across the diameter. All you need is to know your diameter precisely (within 1mm), and ensure that the gauge is precisely at the centre of your mirror. Oh, and the gauge must be able to precisely measure the depth to better than 0.001”. As you may have gathered, precision is critical.

I stress precision because at one point, I was lost chasing the sagitta. I would grind, and grind, and yet my “sag” never reached where I needed it to be. It turns out that the reason for that was because my mirror was not precisely circular. The diameter varied 2mm depending on where I measured it. And when you are measuring the sag to a few thousandths of an inch (my target sag was 0.342”), that turned out to be significant. When I used the correct diameter, suddenly my sag was exact.

If you have the correct sagitta at the centre of the mirror, you then need to measure the sag around the entire surface. This lets you determine how regular the curve is because, for a sphere, the sag will be precisely the same at every point. The tool for this is a spherometer and, again, you can find many designs online. Mine was a very simple 4” triangle of MDF, with three marbles epoxied to three accurately positioned anchor points (holes) such that the marbles were precisely (there’s that word again) equidistant. A hole in the middle of the triangle allowed me to attach my depth gauge. Because the triangle is going to be much smaller than your mirror, the sag it measures will also be smaller. But the key is that regardless of what measurement you get, the value should not change as you slide the spherometer around the surface. You have a very good sphere when you don’t see any variation, to the precision of your gauge.

As I switch over to aluminum oxide for the final stages of grinding, I switched to having the mirror on top. There were two reasons for this. First, by this point, the mirror was nice and transparent, and I could keep a closer eye on contact and surface wetness. But the other, more important reason, was because I found that when I separated the tool and mirror between wets, I was more apt to scratch the mirror with the tool on top than the mirror. I seemed to have better control at taking the glass off the tool than vice versa. It took a few times being forced to go back a grit before I learned that lesson.

Finally, some ATMs would advocate going past 9µ to 5µ for the final stage. I found this to not be necessary as the risk of marring the surface far outweighed the slightly long polish time that stopping at 9µ requires.

And speaking of polishing…

Polishing a fast meniscus

As with grinding, the hand polishing of a fast meniscus mirror is not significantly different than traditional telescope mirrors. So again I will focus on the aspects that make these mirrors unique.

First, the clear goal of this stage is to have a completely smooth and regular surface that is ready for figuring. Note that it is not vital that the surface be spherically smooth, just regular. Small aspheric deviations are totally fine as the final figuring will take care of that.

Second, while the making of pitch laps has a vast amount of coverage online, the key points that you need to focus on for a fast meniscus are:

1. The thickness of the lap should be 1/4″ to 1/2″ to ensure that it is not too stiff, to allow it to easily adapt and flow as the figure changes and to keep contact consistent.
2. The base of the lap should be curved, ideally cast from the mirror, to ensure that the pitch thickness is consistent across the surface; variable stiffness leads to irregular polishing.
3. The channels need to be well cut, kept open, and not symmetrically centred on the face, to avoid polishing in zones.
4. Making a somewhat oversized lap tends to help avoid a “turned down edge” while polishing. If you make a base from wood (as opposed to casting from the mirror) you can make it an inch or more larger in diameter.

The general process for polishing follows the usual ATM descriptions, but here are some key points that I found helpful and/or painful to learn:

1. Polishing with mirror on top allows you to keep a very close eye on contact because, without full and consistent contact, polishing will be very irregular. It also allows you to watch for channels closing so that you know when you need to rework them.
2. Cleanliness is an absolute must. You will know very quickly when a stray bit of grit got onto your lap, and hopefully you have strong enough lap hygiene to avoid that. But for me, a contaminant that was unexpected was bits off a wire brush I was using to rough up the lap. While some ATMs use them successfully, in my opinion, wire brushes have no business being near a polished mirror.
3. As the mirror polishes out, the figure will change and so, to keep excellent contact, you should “press” your lap often. I pressed after every hourly session of polishing, and that seemed to work well. Using a heat gun to quickly and controllably warm the lap makes for efficient pressing.
4. For me, a mixture of 1:8 to 1:10 of cerium oxide to water worked very well as a good dilution that balanced burning through cerox with decent polishing time.

When polishing, it is very normal to wonder how your surface looks from a figure perspective. While you should not get into the habit of regularly checking the figure while your are polishing, it is totally reasonable to check once when you get a bit of polishing done (say after and hour or two) just to confirm that the figure is not horribly astigmatic. So give yourself license to break out the Ronchi tester and have a quick look. Then get back to polishing.

The key to polishing is to not rush it, and just do nice, steady 1/3 centre-over-centre (CoC) strokes. The easiest way to see how your polishing is progressing is to use a laser pointer, aimed at a 45 degree angle, and closely inspecting the reflection. When you are polished out (i.e., done) you will not see the laser on the surface. Any image of the laser on the surface indicates scattering from surface roughness, so get back to it. Typically, you will polish out from the middle of the mirror to the edge, and the last bit can be agonizingly slow. But you cannot rush this and it will be done when it is done. Interestingly, as I went through several cycles of this, I found the time to polish out remarkably consistent: 17 hours. When you are done, reward yourself with another Ronchi image which, if you have been regular with your stroking, should be a fairly decent sphere.