FAILURE: Glass kiln from scratch

Note: this is a failed project. I'm just putting this here in case anyone else wants to try.

For a while I have been trying to encourage my wife to try her hand at making something. She has always liked fused glass and has told me she wants to try making it, but there is always that significant obstacle of needing a kiln. Buying one is too expensive and there are no classes being taught around here, so the obvious solution is to just make one... right? It can't be that hard. It's just a really hot box, right? Well, I guess it's time to find out the hard fun way.

First I did some research on glass kilns, particularly of the DIY type, and found that it would indeed be a challenge. In fact, I have yet to see a finished, functional DIY fused glass kiln. That is discouraging. Here are some of the challenging points:

• Needs to reach and maintain 800C+ temperatures.
• Needs to be able to heat and cool at specified rates.
• Needs to function unsupervised for several hours.

Can I do it for under $100?? Most of it will be salvaged junk, which is free. The parts I have to buy include:

• insulating bricks - 178yen each x 29? =  5162yen
• thermocouple - 400yen
• instrumentation amp - 400yen
• ATmega328P microcontroller - 250yen
• plastic case - 100yen
• various resistors, protoboard, etc.- 100yen

Now for the making! There are two main parts to the thing, the hot box and the controller, so I'll discuss them separately.

The hot part
The box of heating is based on a 1200watt toaster oven from a garbage pile. This provides a metal box with top and bottom heating elements, which are coils of resistance wire inside quartz tubes. It also has a convenient spring loaded pull-down door that pulls a metal rack out when opened.
Of course, I couldn't just use it as is since it is only designed for around 200C temps. I started by gutting it and removing anything that was not steel or ceramic. Note that even aluminum is no good! Aluminum melts around 660C and I'm aiming for over 800C. Fortunately all the important wires were copper which is good up to about 1080C.
Next I took an angle grinder with cutoff wheel and chopped off the chunk with the toaster controls. There was a glass window on the front that I replaced with some steel sheet metal, which used to be the case of a DVD player.

Here are some pictures of it:

 Inside and outside of the bare metal box with heating elements.

 The complete box with outer shell, rack and door. Note that the right side was chopped off. This is where the toaster controls used to be. Also, the glass window was replaced with sheet metal.

The metal box was then encased in a layer of fire bricks. They are actually high density bricks used for ovens rather than the low density bricks for kilns. This is certainly not an optimal material, but after much searching, this was the only economical and readily available type. They are rated for 1400C, so no problems there. Fortunately, the bricks are sized just about perfectly for the toaster, so there is a very small gap between toaster and brick on all sides. Yay! No brick cutting needed!

The controller
This part I am a little more confident about making. Essentially it is just a microcontroller switching some mains rated relays and monitoring the temperature.
I used a good old ATmega328P. It is connected to the following bits:

• Type K thermocouple via LT1167CN8 instrumentation amp. Gain is set at 80 so that 1000C = 3.3V, which is Vcc. This gives me about 1 degree resolution with the 10-bit analog to digital converter in the atmega.
• Two optoisolators switching mosfet transistors in turn driving 250V 15A relays for top and bottom heating elements. They also switch some LEDs to show when they are on.
• Thermistor for measuring ambient temperature for thermocouple compensation. I pulled this out of some junk and had to calibrate it, but it seems to work pretty well.
• SD card where I can put a firing schedule and keep a temperature log.
• Red/green status LED.
• Another microcontroller and LED display for showing current temp just for convenience.

Here is a rough schematic:

And here is a picture of the controller:

The upper pcb is a power supply from a DVD player that conveniently has +5V, +12V and -12V rails. This was necessary for the inst. amp.

The Result
As you likely know from the title, this was a failure.
The controller works great, no problems there.
The problem is that I could not reach 800C. With the kiln as shown above, the best I got was around 500 or so. I tried all sorts of configurations and a few different heating elements. At long last I achieved over 800 with a 1700watt element in a very small chamber made of only bricks, several layers of bricks. I rejoiced! But then as I was disassembling it, the heating element just crumbled in my hand. I tried again with a larger number of lower wattage elements and I could get close to the desired temp, but at this point it was all looking pretty useless as a glass fusing device.
In frustration I simply gave up and decided that without proper materials, tools and a substantial budget, a kiln was just out of my reach.
On to the next project.