Casting a Large, Light-Weight Telescope Mirror From Recycled Glass

The first photo below shows a 12.5 inch diameter, 1.5 inch thick, telescope mirror blank, made from recycled glass, that has been lightened by casting the mirror in a mold that made lots of hexagon-shaped pockets in the back of the mirror. This is the story of how I did it.

Why light-weight mirror blanks? Well, telescope mirrors are made from thick glass. Thick glass is heavy, no, really heavy. Most people don't realize how heavy thick glass actually is because they usually only deal with small pieces of thin plate glass. Handing a large telescope mirror to someone who has never seen one before, is a revelation to them about just how heavy glass really is. I like to tell them that it is as heavy as granite. So imagine the piece of granite that was cut out of a granite countertop to make a sink opening. That piece of granite is about the same size and weight as a large telescope mirror. It's heavy stuff.

In a large telescope, the primary mirror can constitute 1/3 or more of the weight of the entire instrument. Supporting the massive weight of the mirror requires that the rest of the telescope be sturdily built. The result is that large telescopes can be shockingly heavy. I know this first hand because I have thrown out my back several times moving my relatively "easily portable" 17.5 inch Dobsonian Telescope. So I decided to try making some-light-weight mirror blanks, just to see if I could do it.

Let's start this story at the beginning. About two years ago I bought a small kiln at a garage sale and began experimenting with making my own telescope mirror blanks by melting together pieces of thin scrap glass in my kiln to make the thick glass needed for telescope mirrors. The process works great. I started out small, producing 6, 8 and 10 inch diameter blanks up to 1.5 inches thick in the small kiln (Photo #2). Then I saw a much larger kiln for sale cheap on Craig's List. I bought it (Photo #3) and scaled up my mirror making process. The kiln was in rough shape. Well, to be honest, it was near the end of its useful lifetime and about ready for the landfill. But I fixed it up and gave it new life. So not only am I using recycled glass, I am also using a recycled kiln. Fixing up the kiln though is a story for another Instructable. I have now produced wonderful solid mirror blanks in the bigger kiln, up to 14.5 inches in diameter and 1 1/2 inches thick (Photo #4). Once I had the process of making solid blanks pretty much perfected, I wanted to try making light-weight mirror blanks.

Two years of trial and much error went into developing this process. This Instructable just presents the finished product. If you would like to see the whole history of my glass casting (mis)adventures, please visit my web site at http://www.mdpub.com.

I played around with some layout ideas in Google Sketchup. I wanted to make a mold that is the negative shape of the finished mirror. I had settled on a mirror 12.5 inches in diameter and 1.5 inches thick for my initial experiments.  So I tried arranging hexagonal prism shapes inside a circular wall or dam 12.5 inches inside diameter that will retain the glass. I had the idea of assembling the mold using one of my kiln shelves as a base. So I built up all the parts and played with sizes and spacing of the hexagons. The hexagonal prisms would form pockets in the back of the glass mirror that would lighten it. I settled on hexagons 1 7/8 inches in diameter (as measured across the flats) and 1 inch tall. That would leave 1/2 inch of glass on the top face of the mirror for grinding in the curve. The spacing between the hexagons would be 3/8 inch.

All the parts of the mold would need to be built of materials that could withstand glass fusing temperatures. The hexagons would be made from plaster and silica mixture. The outer wall would be 2.5 inches tall and cut from soft firebrick with a band saw and the pieces glued together with furnace cement. The kiln shelf base will easily withstand the temperatures involved.

With an initial design in hand, I set out to make it happen. I wish I could just CNC mill the mold out of a big block of refractory material. Or even use a 3D printer to print a plastic negative of the mold that I could pour castable refractory into. Oh well, maybe someday. Till then, I have to piece it all together the hard way.

How to make the hexagonal inserts was a vexing problem at first. I would need them in large quantities. I tried casting them a few at a time using various methods. It took a long time to make the necessary quantity of hexagons for a mirror. Then I got the idea of making large, hexagonal log molds. I could cast long hexagonal logs and cut individual hexagons off the logs with my diamond wet saw.

Photo #1 shows the beginnings of a hexagonal log mold. I cut two hexagons out of wood with my band saw. I arranged 6 pieces of aluminum bar stock the same width as a side of the hexagon around the wooden pieces and held everything together temporarily with rubber bands. Then I used aluminum duct tape to tape all the seams together (not shown in this photo).

Photo #2 shows me filling one of the log forms with a plaster/silica mixture. Here I have removed the wood hexagon from the top so I could fill the mold, but left the other one in as a plug at the bottom. I first sprayed in a little teflon mold release agent. The mold leaks a little, but it looks worse than it is. The plaster mix is equal parts Hydrocal Plaster, 200 Mesh Silica Flour, and water. All measurements are by weight. This is a well known recipe to anyone who has kiln-cast glass. This plaster mix is used by glass artists to make milds because molten glass will not stick to it, and molten glass sticks to almost everything. A respirator should be used when mixing up the plaster. The silica dust is very bad for you if inhaled, and the fine plaster dust can be irritating.

Photo #3 shows two hexagonal logs after curing and removing them from the mold. To get them out of the mold, I simply slit the tape along one of the seams in the mold, and unroll the mold off the log. The teflon mold-release agent allows everything to come apart cleanly.

Photo #4 shows one of the logs on the table of my diamond wet saw. I have the fence set for cutting 1 inch thick slices off the logs. I find that the cutting actually goes better if I don't use any water spray, and cut the logs dry. The saw slices the logs like a hot knife through butter. The process makes a lot of dust though, so wear the respirator again.

Photo #5 shows some freshly cut hexagons. I needed to make 19 hexagons for all the pockets in the mirror. I cut all those and a few extra, just in case. Later I decided to add a few partial hexagons around the edge of the mirror to further lighten it. Those extra hexagons came in handy then.

The outer wall of the mold acts as a dam to contain the molten glass and prevent it running all over the kiln. I made the wall out of soft firebrick. A half inch wide template was cut from cardboard. The cardboard template was used to mark out arcs on a K-23 soft firebrick. The arcs were then cut out with a band saw. The saw blade is pretty much ruined in the process, but will last long enough to cut out enough pieces to make several dams.  New blades for my band saw cost less than $10 at Harbor Freight, so it is no big deal to ruin one now and then cutting up firebricks. With care, one brick will yield enough pieces to make one 12.5 inch ID dam with a wall thickness of about 1/2 inch.

I cut a circular disk 12.5 inches in diameter out of thick Styrofoam using a hot-wire foam cutter. This disk serves as a template for assembling the arcs to make the dam, and keeps the finished product reasonable round. The pieces are assembled as tightly as possible around the Styrofoam disk. High-temperature furnace cement is used to glue the pieces together.

Molten glass is dense, and exerts considerable hydrostatic pressure on the inside of the dam. A failure will cause molten glass to run out all over the inside of the kiln. So after the cemented ring sets up, I usually reinforce the joints by cementing on short scrap pieces of firebrick left over from cutting out the arcs. Several wraps of thin stainless steel wire wound tightly around the mold and held in place with dollops of cement will help prevent dam from opening up and spewing out molten glass if it should crack or if one of the joints fails. These features are shown in the last photo below, taken at a later stage of mold construction.

The inside surface of the mold will get several coats of kiln wash before firing. Kiln wash prevents the molten glass from adhering to the mold.

Photo #1 shows a mock-up of the pattern of hexagons I wanted to layout. I wanted to actually see the pattern and make sure there were no problems before continuing. So I mocked up the pattern of the hexagons on a paper circle 12.5 inches in diameter. I made the late addition partial hexagons at the edges by cutting up  some of the left-over whole hexagons with a diamond tile saw and shaped them with a belt sander to make them fit better around the edge. I used square wooden dowels 3/8 inch in size as spacers to get the spacing between the hexagons the way I wanted it and keep it uniform. Everything looked good, so I moved on to building the mold for real.

Photo #2 shows the start of building the mold with a kiln shelf that has been well coated in kiln wash. I drew a 12.5 inch diameter circle on the kiln shelf. This is the footprint of the finished mirror blank. The hexagons and partial hexagons will need to be arranged within this circle. I also drew on some light layout lines to help me get the hexagons arranged correctly. Again I am using the wood dowels to get the spacing right. To keep the hexagons from shifting around, I glue them down with a slurry of kiln wash.

Photo #3 shows the process of continuing to glue down the hexagons. To glue them down, I mix up a slurry of soupy kiln wash and use a spray bottle to wet down both the hexagon, and the position on the shelf where it will go. I then dip the bottom of the hexagon in the soupy kiln wash and quickly put it in place. The plaster hexagon and the porous kiln shelf wick the water out of the kiln wash slurry in a matter of seconds, and the hexagon locks in place. If I don't wet the pieces first, the kiln wash slurry hardens instantly. I have only a very short period of time to maneuver the hexagon into place after it touches town on the kiln shelf before the kiln wash seizes up. If it isn't in the right place, I have to break it loose, scrape off the hardened kiln wash, and try again.

Photo #4 shows all the hexagons and partial hexagons in place. Then I put the outer wall ring of the mold in place. This particular ring has been used several times before. It is beat up and has multiple repairs. Once the ring is in place, I use a small artist's paint brush and more kiln wash slurry to fill any gaps between the kiln shelf and the bottom of the ring. This prevents glass from leaking out under the dam. Once mold is completely finished, I usually let it air-dry for a few days to reduce the moisture content before firing. If I am in a hurry, I will put in my kitchen oven overnight at 180F to dry it out. The initial heatup of the kiln will be very slow to prevent steam explosions and cracking.

Where to get the glass to make a mirror blank? The preferred glass for telescope mirrors is low-expansion borosilicate glass. This glass is very expensive though, and hard to find. My previous experiments with casting solid mirror blanks showed that ordinary soda lime glass makes excellent mirrors, and scrap soda lime glass is all around us. It is very easy to come by. Scrap glass can be had for a song, and recycled into telescope mirrors. See my web site at http://www.mdpub.com to see my early glass casting experiments.

So, where to get the glass? Bottles immediately come to mind. It isn't hard to collect a lot of used bottles and recycle them into telescope mirrors. Just make sure that you use all the same kind of bottles for each blank, so that you don't have to worry about glass incompatibilities. You should also choose clear or light-colored bottles so that the finished blank isn't too opaque to check for internal stresses. More on that later. Bottles require label removal and extensive cleaning before they can be used though. That is tedious.

Photo #1 shows my favorite source of glass. Thick glass tabletops. Tabletops 1/2 to 3/4 inches thick are an ideal source of glass to recycle into telescope mirrors. I pick these up at yard sales, thrift stores, and furniture consignment stores. They can be had amazingly cheap, especially if they are chipped or badly scratched. Ask the store owner if they have any tabletops in the back room that are damaged or don't have matching bases. They will usually sell you any glass they have for just a couple of bucks just to get rid of it. Sometimes I even get them free. One thrift store now sets their unusable glass tabletops aside for me rather than tossing them into the dumpster.  A buddy of mine also just gave me a large, thick glass tabletop that his neighbor was throwing away. He salvaged it from the trash and gave it to me. All the glass for a mirror should came from the same sheet of glass to avoid glass incompatibilities.

The glass needs to be broken up into pieces between roughly 1/2 and 2 inches in diameter. This seems to be about the ideal size range. Smaller pieces entrain too many air bubbles in the melt. Larger pieces are hard to arrange in the mold.

Photo #2 shows the amount of glass needed for a mirror being weighed out into a plastic colander. All this glass came from the same large sheet of 3/4 inch thick plate glass. Some of it has been melted before into blanks that failed for one reason or another and got broken up again for re-use. Glass can be recycled over and over again.

In Photo #3 I am cleaning the chunks of broken plate glass in my dishwasher. I just put the colander full of glass in the dishwasher and run it through a cycle. The glass should be free of dust, oil and grease.

The mold needs to dry out for a couple of days before firing. The process could be sped up by drying the mold in the kiln with the temperature set to about 180F for about 12 hrs.

Photo #1 shows the mold loaded into the kiln. The kiln needs to be carefully leveled so that the molten glass will form a uniform thickness blank. If the kiln is out of level, the glass will run to the low side of the mold and be thicker there, and thinner on the high side of the mold.

Photo #2 shows the glass loaded into the mold. Making the outer wall of the mold an inch taller then the finished blank allows it to hold all the needed glass. The thin tube sticking into the kiln from lower left holds a thermocouple. I get it as close to the surface of the glass, without touching it, as I can for best accuracy.

The firing process starts with a slow initial heat-up, to drive off any remaining moisture in the mold. Later it ramps up to 1800F. Here is a peek inside the kiln with the glass at over 1700 F. The chunks have melted and slumped down filling the mold. By the time the glass gets to 1800F, it is fully liquid and fills all the voids in the mold. It is still quite viscous though, like cold honey, so the glass is soaked at this temperature for two hours to allow it to level out and let bubbles rise to the top and pop. Then comes a very rapid cooldown to the annealing temperature of 1050F This is accomplished by opening the kiln lid. The cooldown must be rapid because the temperature region around 1350F is a danger zone where devitrivication of the glass can occur (crystal growth). That temperature region must be passed through as quickly as possible. Next comes annealing. Annealing removes internal stresses from the glass. It is a vitally important step. There is an initial annealing soak at 1050F, then a very slow cooldown until the temperature of the glass is below the strain point, then a faster cooldown in two stages to ambient temperature. The entire firing schedule takes  almost four days to execute. I use a Ramp/Soak temperature controller of my own design to control the kiln. See my web site at http://www.mdpub.com for details on that.

Here is the firing schedule that was used to make the blank.

Firing Schedule
Ramp Time	300	60	120	AFAP	2000	1000	1400
Temperature	950	1250	1800	1050	950	850	90
Soak Time	0	60	120	400	0	0	0

Times are in minutes. Temperatures are in degrees F. AFAP = As Fast As Possible (I open the kiln lid to get the temperature down quickly so as not to linger in the devitrification zone). The above photo was taken during the rapid cool-down to the annealing temperature.

I am always tinkering with the firing schedule to try to improve the finished product. I think that on my next blank I would soak the glass at 1800F even longer in order to get more of the small bubbles in the melt to rise up to the surface and pop.

After the blank had finally cooled down to ambient, I took it out of the kiln, and removed the outer mold. The outer mold usually comes off quite easily, and can often be reused. Minor repairs to it may be needed before reuse.

Then it was time to start cleaning the plaster out of the hexagons and partial hexagons. A very messy and delicate job. Fortunately the plaster is rendered quite soft and powdery by the temperatures involved in casting glass. So it comes out fairly easily. The plaster has the consistency of a very dense dirt clod at this stage. The glass and plaster do not stick together. However, the plaster is tightly compacted into the pockets by the shrinkage of the glass as it cooled. So it takes a little persuasion to get it out. I use an electric drill with an old spade bit to drill several holes in the plaster of each pocket. Then I use a screwdriver to break out chunks of plaster. Care must be taken not to exert too much force or the glass can be chipped or cracked. It usually takes me about an hour to clean out a mirror this size.  Use a respirator to avoid breathing any of the dust created.

People viewing this process on my web site at http://www.mdpub.com have suggested using a pressure washer to clean out the plaster. I haven't tried that yet. It might work. I don't have a pressure washer, but I guess I could take a blank down to the car wash and use the high-pressure wand there.

Here is the finished blank after giving it a thorough rinse to wash away dust and stuck-on bits of plaster. It is a thing of beauty, if I do say so myself.

It is not quite done yet. It is slightly oversize and the edge is a little irregular because of how I made the outer wall of the mold. This will be fixed by grinding the edge true in the next step.

Photo #1 shows the setup for edging the mirror blank. The blank itself is on a lazy-suzan table mounted on the table of the drill press. My home-made grinding wheel is chucked up in the drill press. The blank is pressed against the grinding wheel with one hand while applying braking pressure with the other hand to prevent the blank from spinning out of control. The differential speed provides good grinding action. In short order the blank was trued up, perfectly round, and cut down to the proper diameter. This particular photo was taken when I was edging an older blank made when I was using a different design that used round holes instead of hexagons. I still use this same edging process though. You can see more about the different types of mirror blanks I have made on my web site at http://www.mdpub.com

Water must be continuously applied during the grinding operation to keep down the heat, or large chips will begin breaking out of the edge of the mirror blank as the edge overheats. The entire blank may crack if allowed to overheat.

During the grinding operation I wear a face shield, a good quality dust mask, and leather gloves. Chips of glass tend to go flying in all directions, especially at the start of the roughing operation when the edge is still ragged. The dust produced is very bad for your lungs. Of course there are always sharp edges when dealing with glass, so I wear leather gloves. I also swaddle my drill press in plastic sheeting to prevent the water, mud, dust and glass chips from getting all over and into it.

Photo #2 shows my home-made edge grinding wheel. The proper tool for edging a mirror blank is a thick grinding wheel. Well, thick grinding wheels are very expensive, and somewhat hard to find. So I decided to make my own grinding wheel. While in Harbor Freight I noticed they sell thin masonry grinding disks in packs of 10 at a very inexpensive price. Looking at them, I got the idea of stacking a bunch of these thin disks to make a thick wheel. I had a machinist friend of mine make a mandrel for me that could hold a stack of thirty disks. Shazam! Just like that I had myself a thick grinding wheel on the cheap. When I wear it down, the disks are easily, and cheaply replaceable.

Here is a photo of the finished 12.5 inch diameter honeycomb-back mirror blank after final edge trimming. It is side-by-side with a solid 12.5 inch blank of the same 1.5 inch thickness that I also cast in my kiln. The honeycomb blank weighs 10 lbs 3.6 oz. The solid blank weighs 17 lbs 10.4 oz. The honeycomb mirror is 42% lighter. Not bad at all.

This Instructable is all about the process of actually casting the light-weight mirror blanks. The process of grinding, polishing and figuring them into finished mirrors is really fodder for another Instructable, because it is another long, complex, highly technical process with many steps. I will add a little bonus material in the last two steps that gives you a peek at the process of finishing a mirror.

You can see all the steps from raw glass to finished mirror on my web site at http://www.mdpub.com

Photo #1 shows the rough grinding process. The mirror blank is ground against a ceramic tile tool with #60 carbide grit in between. Using the correct grinding stroke results in the mirror becoming concave as the tool becomes convex. This process is continued until the mirror reaches the desired curve. The process was all done by hand, and took several weeks of Saturdays, (and raised a few blisters).

Photo #2 shows the mirror at the end of rough grinding. I was aiming for an f/4.5 mirror, so the required depth of the curve is 0.175 inch. Once the required depth is reached, the rough grinding is complete. Amazingly, the rough grinding has removed over a pound of glass. The mirror now weighs in at only only 8 pounds 15 ounces. Now the fine grinding can begin to clean up the rough surface left by the rough grinding.

Photo #3 shows the sequence of abrasive grits that must be used to get from rough grinding, through fine grinding, to polishing.  Successively finer grits are used to clean up the pits left by the previous stage. Eventually the mirror reaches the stage where polishing with Cerium Oxide can begin.

Photo #4 shows the mirror at the end of fine grinding. The surface is not quite transparent. It still has a slight frosted look to it. Rough and fine grinding are done outside to prevent contaminating the polishing room with large grits that could cause scratches on a mirror. Once it is time to polish, the mirror gets thoroughly cleaned and I go into the polishing room. See the next step for polishing.

Once polishing begins, the ceramic tile tool that has been used from the beginning of rough grinding is set aside. A pitch polishing lap is constructed for polishing and figuring the mirror.

Photo #1 shows the pitch polishing lap. It was created by pouring hot pitch on the protected surface of the mirror, over a channeling template, then placing a plaster or wood disk on top of the still hot pitch. Once the pitch cools, the polishing lap is separated from the mirror and trimmed to size. The lap now has a lot of square facets on it. Pressing the soft pitch lap against the mirror with nylon mesh between them makes lots of micro-facets. The soft pitch lap conforms exactly to the shape of the mirror. The edges of all the facets catch and hold polish particles and rub them against the surface of the glass.

Photo #2 shows the surface of the mirror with polishing nearly complete. Note that it no longer looks frosted. It looks perfectly transparent. It actually still needs more polishing.

Photo #3 shows the mirror on a Foucault test stand. The mirror is still at an early stage, but I wanted to see if the pattern of hexagons on the back of the mirror was "printing through" to the front surface. So far, as long as I use light polishing pressure, print-through doesn't seem to be a problem.

After polishing is complete, it will be time to figure the mirror. The shape of the mirror we aim for at the end of polishing is a sphere. Figuring means turning the sphere into a parabola by using special polishing strokes with the pitch lap. This mirror hasn't made it to that stage yet. I'll post updates as polishing and figuring progress.

Meanwhile, I continue to turn out new mirror blanks from my kiln. See my web site at http://www.mdpub.com for more information.