Molding Refractory Forms From Multi-Part Templates

I'm building a stained glass and bronze desk lamp, made with glass sheets that have been shaped over refractory forms in a kiln. One of the big challenges has been making the forms, which are molded from a castable refractory material. The molds must have a sufficiently smooth surface to prevent any marks from becoming part of the forms, and then showing up on the glass during kiln firing. In this case, the molds were pressure formed from sheet plastic over 3D printed templates.

In the past, the molds were CNC milled from large blocks of aluminum. Any glitch or error in the milling process would likely result in an expensive chunk of scrap metal, and a perfectly machined mold still required significant labor to smooth the milling marks out of the cavity. This led to very long moldmaking cycle times and a “do it once” mindset that opposed flexibility and experimentation. 

The molds are large enough that the templates had to be printed in two pieces each, and then joined together. One difficulty with attaching the template parts together was preventing any witness line from showing in the pressure formed mold. The temperature and pressure of the pressure forming process can cause bulging in an epoxy seam, no matter how well sanded. And re-sanding after an initial pressure forming doesn't prevent a repeat bulge, but does risk distortion of the template.

I developed a hybrid gluing and filling process that sufficiently reduces or completely eliminates the witness line from pressure forming templates assembled from multiple pieces. This process uses an engineering epoxy to form the primary bond between the parts of the template, and a hand application of UV curing resin to fill the exposed part of the seam. 

The main tools used were a Form 3 resin printer, a Form Wash station, a Form Cure UV oven, a Mayku Multiplier pressure former, and a Foredom Flex Shaft with abrasive disks. I also used a home-made UV curing box. Hand tools included files, sanding blocks, and small paintbrushes. Materials and consumables included Formlabs High Temp resin, epoxy, a syringe with a narrow needle tip, sandpaper, aluminum strip, plywood, wood glue, masking tape, packing tape, and denatured alcohol for resin and epoxy cleanup.

The templates are printed as hollow shells with a reinforcing structure underneath, generally a grid with walls ranging from 4mm to 6mm in thickness. These have consistently been strong enough for the pressure forming process. Cleanup in the recesses of the grid can be minimal since it has no effect on the rest of the process.

Ideally the template pieces are as symmetrical as possible and have keying features between them for precision alignment.

A recessed pocket about 1/2mm deep should be provided between the mating surfaces for the epoxy, both to allow some thickness of epoxy and to form a barrier against the epoxy leaking up to the critical surfaces. This pocket is ideally split between the two mating surfaces.

Prior to using the epoxy, only enough cleanup is done to ensure precise mating of the template pieces. The epoxy is applied, and the pieces are clamped together. Using a dark epoxy makes it easier to see where it has gone.

After the epoxy has cured the bottom of the template is sanded flat, generally with 80 grit sandpaper. All the sandpaper mentioned here is preferably silicon carbide wet or dry paper. Of course, wear an appropriate particulates respirator when doing any sanding of cured resin! It's also important to have a sufficiently flat surface to work on.

Depending on the condition of the critical seam, there are two ways to proceed. In the first method, the seam can be directly filled with High Temp resin using a syringe and narrow needle and/or a small paintbrush. The resin needs to end up higher than the seam so that it can be sanded down after curing, otherwise multiple cycles of resin and curing will be required. Getting resin a couple of millimeters to either side of the seam is pretty unavoidable, but more resin also means more cleanup needed. (Note that the brushed filling photo shows a different template being worked on than the other photos in this article.)

The second method starts with intentionally opening up the seam to allow deeper application of resin. This is best done with a Dremel or Flex Shaft type rotary tool and a small abrasive disk. This method gives even better results in terms of zero witness line, but adds the risk of scarring up the critical surfaces and requiring more resin fill and cleanup. If this method is followed, all abrasive grit must be cleaned away before adding the resin per the first method.

Since the resulting assembled template is too big to fit in the Form Cure oven, a UV curing chamber can be made with an inexpensive UV curing lamp, a corrugated cardboard box, and a lining of aluminum foil. The template with wet resin is placed in the box and cured for 4-6 hours.

Assuming that the resin seam is completely above the surface of the template, the seam can be sanded down and the template can be finished. Initial lowering of the filled seam is done with curved files and 150 grit sandpaper wrapped around a cylinder. Sand this close to flush, but it doesn’t have to be perfect.

This is followed by sanding of the critical surfaces with 220 grit sandpaper, which I do with a sanding block faced with a thin layer of high density neoprene foam. Use more 150 grit at this stage only if 220 grit isn’t getting anywhere. Keep the sanding block moving uniformly over the template and don’t sand any more than absolutely necessary to achieve a smooth and even surface. Use cylinders wrapped with sandpaper where necessary. If there are any imperfections too deep to sand out, clean the form thoroughly and fill with more UV resin and repeat the curing process. Non-critical surfaces can be sanded smooth with 150 grit paper, followed by 220 grit. Be careful not to sand near the edges in any way that would interfere with the release draft on the template. Also, try to keep the marks from the final sanding passes parallel to the direction of pull for releasing the template from the pressure formed sheet. 

Once the entire template (except for the bottom) is sanded smooth to 220 grit, wash it off, dry it, and inspect it carefully. If it all looks good, do a final sanding with 400 grit. After washing and drying from that sanding, you may find some areas that need more sanding. Repeat as needed. If your application needs an even finer surface finish, proceed as necessary. 

The template is now ready for use in pressure forming the basic mold. Here, a 1mm PETG sheet was used to form the mold.

After pressure forming, I made a plywood frame and trimmed the pressure formed mold sheet to fit. I scuffed the underside of the mold for better epoxy adhesion.

Masking tape is all that’s needed to hold the mold sheet to the frame. The underside of the mold sheet and the adjacent portions of the frame are then coated with thickened epoxy. I used a bulk marine grade epoxy thickened with silica or sawdust; whatever is at hand that can make it thick enough to not sag. A layer about 4-6mm thick will suffice. This stiffens the mold and joins it to the frame. Ideally, wet the surfaces with a thin layer of non-thickened epoxy first in order to enhance adhesion. 

As shown here, the completed mold was used with an additional sheet aluminum dike (held together and secured with packing tape) to provide a riser for the refractory form. In most cases this isn’t needed.

Compressed air was used around the edges of the refractory form to release it from the mold. Patience is required. The finished refractory form surface is smooth and free of marks. (The refractory material used should have instructions for fully drying the form and coating it with a release agent.)

The refractory form is shown in the kiln with a sheet of formed glass.

The title image at the top of the article shows the goal of all this; a sheet of formed glass that has been cut and precision ground to final shape and then wrapped with copper foil, test fitted into the metal frame.