I was pleased with this, and moved on to insulating the back of the door. I spent a long time moulding a nice plug for the door entry out of my insulating mix (vermiculite, fireclay, cement and sodium silicate), attaching it to the back of the oak with metal ties. The insulating mix dried, cracked a bit, dried some more, then fell apart when I lifted the door up.
Lesson 1: my insulating mix is good for filling cracks, and coating chimney insides. It's not good for making doors.
Door number 2 was made by using the wooden front from door number 1, then carving two thermalite blocks to act as a plug for the oven. They were fixed to the oak door using screws - the carving was very easy thanks to the softness of the blocks. This looked promising.
Pleased with this, I had a nice big fire and stuck the door in place after the coals had cooled off for 10 minutes. By watching the temperature (graphs in C and F here, C top, F below)...
... I could see that this door worked well. Notice that this graph's over 28h, and the dome surface now takes 6 hours to cool from 300C to 200C - twice as long as it did without the door. I thought I'd cracked it, but unfortunately:
- the sodium silicate round the edges of the door had erupted into a white fuzzy mass in the heat, and looked horrible
- the wood around the door had charred badly, causing the aluminium frame to come loose, and in one place, fall off
- the wood on the back of the door had warped with the heat and cracked, causing the thermalite block to move and crack as the wood bent. You can see one of the big cracks in the block below. The bit at the bottom is a loose chunk that fell off when I lifted the door.
The thermalite block was still solid, if a bit more brittle than before, but the wood was badly burned around the edges. This wasn't a great example of how to build a door - perhaps more of a warning to others! Don't use wood for the door - I'd underestimated just how hot it would get round the edges. Now I need to make door number 3 - an all metal version.