Selecting and Inspecting Glass

Glass Thickness: The thinnest mirror that I have made was 8" by 3/8". I had some trouble with flexing while working this piece, but I taught myself how to correct the astigmatism. I presume that if I can do it, you can do it. Remember that for a given stiffness, the thickness of the glass goes up as the square of the diameter. So you have the following possibilities: 13"x1" (if you can find glass that thick), 11"x3/4" (3/4" glass is commonly available as broken book shelves, table tops, etc), 9"x1/2", 8"x3/8, 5.6"x1/4". (note that 1/4" glass is 0.20" thick, while 3/8" glass is about 0.40" thick). If you use glass that is thicker than these values, you will have an easier time in grinding and polishing. 3/8" glass is so commonly available, there is really no reason to use anything thinner. If you are forming your curve by grinding, you should add the saggita of the curve to the above required thicknesses. It is best to mold the curve in a high temperature kiln.

Strain Your glass must be strain-free. First of all, your glass must be untempered so that you can cut and grind it. Tempered glass has so much strain that it will shatter into a bozillion tiny pieces if you cut, scratch, or grind it deeper than the outer skin that is under compression. Conventional orthodox wisdom in telescope making urges that glass be fine annealed: there must be NO visible strain... however, I am not convinced of this. You may can easily de-temper glass by heating it above 575 degC in a conventional potter's kiln. Cool it with the kiln lid closed. Tempered glass is manufactured with a special high speed jet of cooling air; tempering does not happen by accident. Detempered glass may be cut and ground, but may still have considerable internal strain. For slumping and fine annealing, see the article on ribbed mirror blanks.

"Satanic cross" indicates strained glass

You should examine your glass for any etchings; sometimes tempered glass has a small etching "Armor-All" (but this junk is NOT bullet proof). Also, you can sometimes identify tempered glass by little dimples where it was supported by tongs. If your glass has etching or tong dimples, DO NOT TRY TO CUT IT! You can see strain in your glass if you examine it between crossed polaroid filters. Use LINEARLY polarized filters, not the circularly polarized filters that the camera store salesman will try to sell you. If you cannot get polaroid filters, you can simulate one by passing light through several layers of window glass that is tilted about 45 degrees with respect to the light path. The light from a laptop computer is polarized, so you can use your laptop for a light source and make do with only one small polaroid filter. If the polarizing filters are crossed, no light should get through. Strain in the glass causes a rotation in the optical vector and allows light to come through. The result is a "satanic cross" as shown in the photo. If you have such a strain pattern in your glass, it can be easily fixed by annealing.

Good annealing is no guarantee that the glass is stable. Here is a disturbing quote from G.D.Dew, "The Measurement of Optical Flatness", Journal of Scientific Instruments v.43 p.411 (1966).

"Contrary to expectations, the results showed that the combined figures of the two flats had undergone a change from the first to the last [of a series of photographs in the course of a few minutes], the total change being 0.1 fringe over the full 30 cm diameter... All customary precautions had been taken: the flats had been undisturbed for at least 24 hours before the photographs were taken; the room temperature was controlled, nominally to 0.1 degC; and the flats themselves were screened from draughts and other thermal disturbances.

"Although the blank was fine annealed and showed no detectable strain on a conventional strain-viewer, changes of figure of approximately 0.2 fringes occurred in the first few months immediately following the completion of the glass polishing."

Rough Cutting a Glass Circle

Score and Break: In this method of glass cutting, a scratch is made on the surface of the glass. A crack is initiated, and the scratch guides the propagation of the crack. This operation requires a certain amount of skill, but you will be doing it with cheap and available 3/8" or 1/4" salvaged glass, so you can get a lot of practice. With this glass you can make mirrors up to 7"x3/8" or 4"x1/4" with confidence, and if you are skilled in polishing and figuring you can go somewhat larger with the thin glass.

When the crack in the glass is propagating, it does not do so perpendicular to the glass surface. When cutting a circle, it is best to leave a margin equal to the glass thickness. This will be cleaned up in the final edging of the glass circle. If the crack comes too close to the edge of the glass, it will leave its intended path and break out to the edge. When scoring the glass, make ture that the scored circle does not come closer to the edge than six times the glass thickness. It may seem wasteful, but you can do this because you are working with cheap, commonly available glass.

Be sure your circle is not too large for your piece of glass

Score and break sequence for glass circle cutter

The preferred tool for scoring circles in glass is similar to a beam compass with a suction cup at the center and a glass-scoring wheel on an adjustable carrier. When scoring glass, you want it to be clean and well supported on a flat carpet surface. Scoring seems to work better if you lubricate the scoring wheel with light oil or turpentine. Take the compass around once to be sure it will not bump into anything, and then apply firm, but not heavy pressure as you score the circle.

After you score the circle, you must initiate and propagate the crack. There are several ways of doing this, all of which apply a concentrated force to the glass behind the score. One way to start the crack is to turn the glass over and wack the glass over the score with the ball end of a glass cutter. I have also used a large steel ball bearing for this. Once the crack is started on the circle, it must be propagated. There are several ways to do it. One way is to turn the glass over again so that the score is on top. Position the glass so that one end of the crack is over a steel ball bearing, and gently press. As the crack propagates, move the glass so that the pressure point is always on the end of the crack. If you have both luck and skill, the crack will follow the score around, and when it completes the circle and meets its other end, you will hear a satisfying little "tick". You will soon learn that it is easier to propagate cracks around a large circle than a small one. That is one reason why I don't like to make mirror blanks smaller than 7".

Once the crack is propagated around the entire circle, use a hand glass cutter to score breakout lines as shown in the picture. These do not need to be perfectly straight, just do them by hand. However, take care to assure that the breakout lines are colinear. Turn the glass over and initiate and propagate the cracks on the breakout lines. When you complete this, you should be able to pull a piece of glass away from the circle. Repeat this several times, and you will soon have a nice glass circle.

Broken glass has very sharp edges. You can use a carbo whetstone or wet or dry carbo sandpaper to dull the edge, or you can simply cover it with masking tape. If you do not do this, you will bleed!

Having taught you all of this, I must now say that I have bought a Gemini ring saw. It does an excellent job of cutting straignt edges and circles, and I do scoring of the surface only as a guide mark. (Ink markings are not durable in the wet glass grit environment of a diamond saw.) Wet Sawing: For glass thicker than 3/8" (10 mm), I use a Skil saw with a carbide masonry blade to cut out a square. Adjust the saw blade so that it protrudes about 1 mm beyond the bottom of the glass. Support the glass on a flat carpeted surface. (Old salvage carpet is cheap, so you don't care if the carbide blade cuts it.) Put on your gas mask, and saw out a square that is about 1/2" larger than your intended circle. Do not be alarmed if the corner area breaks off in an irregular shape as in the picture; you will soon cut the corners off.

Clean up your glass square with Windex and use the circle cutter beam compass to inscribe the desired circle on the glass. I have found that the best marking pen will not stand up to the environment of a wet saw, so I use the circle cutter to scribe a mark that will not fade.

Use a diamond wet saw to trim the glass as close as you dare to the circle. Start by cutting off the corners, and then work your way down. Take care not to cross the scribed circle! The when you are finished, you will have something like the glass shown on the scale. Clean up your work area, your gas mask, and your self. You don't want youself or your environment to be contaminated by glass dust.

Trepanning: another way to do it

This trepanning machine is easy to build and cuts nice circles

Edging a Perfect Circle

Mirror blank stuck to chuck with pitch

The process of grinding an edge is pretty simple: chuck up the glass centered on a rotating platform, and rotate it against a water or turpentine cooled diamond bit.

Chuck center hole solidly rests on a conical point with this simple jig

This chuck is adjustable for centering the workpiece

You get a lot of grinding for $16

Edge chips are shallow and easily ground out

I made two chucks: one for smaller work is simply a metal plate with a shaft; for larger work, I use a thicker plate with a small hole in the center. Glass is attached to both with pitch. It is easy to center the glass on the chuck by heating with a blowtorch. (My shafted chuck is adjustable to center the glass, but I have never needed this feature; save yourself the needless complication and keep it simple!) Note that the shafted chuck idea is easily adaptable to edging blanks for elliptical secondaries: just tilt the plate 45 degrees.

Once you have your glass chucked up, you need to rotate it against a diamond tool, with enough water or turpentine to keep the diamonds happy. If you use water, be sure to clean, dry, and oil your drill press when you are done, or you will turn it into a pile of rust. If you use turpentine, be sure it is REAL turpentine distilled from wood. The fake petroleum distillate sold in some stores is OK for painting and pitch cleanup, but does not provide the required cooling and lubrication for cutting glass with diamonds.

The diamond tool costs about $50 in an art supply store, but only about $16 from a ceramic industry supplier. Diamonds last a long time if you run them fast and don't let them dry out. I simply run a round-head bolt through the grinder head and put it in the drill press chuck. The hole in the chuck plate rides on a conical point in the drill press jig shown. Be patient and take light cuts, and you will end up with a nice perfect circle. Chips that extend into the surface are very shallow; they will most likely be ground away before you establish perfect contact with your tool.

Now that you have read all this, I confess that it is a pain in the neck to edge a perfect circle this way. Attaching the glass to the chuck with pitch is time consuming, smelly, messy, and exposes the glass to thermal stress. I am now experimenting with a promising chuckless method. If you want to learn more, Google "centerless grinding". There are other possibilities... for example, suppose you have the edge of the glass resting between two rollers 120 degrees apart, and a grinder head cutting on the opposite side... Now replace each of those rollers with two on a pivoting lever... If you know how to do computer simulations of such things, please contact me!

Slumping and Annealing the Thin Blank

On the other hand, there are reasons not to slump. A suitable furnace with a pyrometric programmable controller may not be available. You may simply believe in orthodox mirror blankology, which does not allow for innovation. Support of a meniscus, especially a deeply curved one, is uncharted territory in mirror cell technology. The best mirror cell software available does not model the potato chip deformation that results when a load is applied to the edge of a meniscus. It is up to you.

Glass: what you need to know. Glass manufacturers define four temperatures for glass in terms of the viscosity of the material. These temperatures depend on the type of glass. From hottest to coldest, they are, for Corning 7740 pyrex:

• Working Point: 10E4 poises - Glass is a thick fluid, like honey. Only the large bubbles will rise to the surface; but small bubbles are trapped in the glass. 1252 degC.
• Softening Point: 10E7.6 poises - Glass flows slowly; any structures collapse into a horizontal pile. This is hotter than you need for slumping and fusion welding. 821 degC.
• Anneal Point: 10E13 poises - Glass is solid, but internal strain is relieved in 15 minutes. This is not hot enough for slumping and fusion welding in a reasonable time. 560 degC
• Strain Point: 10E14 poises - Internal strain is relieved in 4 hours. 510 degC

The temperatures for LOF soda-lime float glass (typical plate glass) are: Softening Point 715 degC, Anneal Point 548 degC, and Strain Point 511 degC, not very different from pyrex for annealing.

After a great deal of experimentation, I have found that the following cooking schedule works well:

• Heat the glass to 620 degC for slumping, to 640 degC for fusion welding of strucures. The heating rate is not at all critical.
• Maintain the glass at the above temperature for about a half hour.
• Cool the glass to about 575 degC immediately by opening the furnace door. Don't worry, glass at or above this temperature is soft and will not crack from thermal shock.
• Cool the glass to 375 degC at a constant rate. The cooldown from 575 to 375 degC should take 60 hours per inch of thickness. At 375 degC, the glass is hard and will crack from thermal shock, so don't open the oven door below 500!
• Cool the glass from 375 degC to room temperature at a constant rate. I have found that 24 hours per inch of thickness is safe for this final cooling.
• Open the furnace and put your hand on the glass; if it feels warm, leave it in place with the furnace door open until it is at room temperature. Warm glass that is comfortable to handle can crack when removed from the furnace. Haste makes waste!
• All of the above times assume that heat is entering the glass from both sides. If the glass is on an insulating platform such as firebrick, double these times.

I make my slumping platform out of ceramic kiln shelves. I form the curved surfaces by grinding two such shelves together with carbo grit, as if making a mirror. The ceramic material is much harder than glass, so grinding by hand requires a little patience. However, once you make one of these platforms, you can use it to mold the curve into hundreds of mirror blanks. To prevent the glass from sticking during cooking, I coat the platform with a slurry of "kiln wash" (my best guess is that this is aluminum oxide powder). After the slurry is dry, I wipe the surface to remove the brush marks, leaving kiln wash only in the pores of the ceramic kiln shelf. When I tried to make a slumping platform out of dental stone, it worked OK, but fractured badly enough to prevent a second firing. I have tried a castable refractory cement; it has a very rough surface that is not amenable to grinding, but it works fine if coated with a creamy slurry of dental stone. The superficial cracks in the dental stone do not compromise the structural integrity of such a sluming form. CAUTION It is a lot of work to make a slumping platform. You must never, ever, let molten metal contact it. I ruined one platform when copper melted in a ceramic crucible and soaked right through the crucible and into my platform like water on a paper towel.