So I've started to experiment with putting a solid refractory layer on the inside of at least some parts of the system. The material I have started working with is mizzou.
I selected mizzou on the basis of it being on offer from High-Temp Inc. and because of something I read on a forge thread somewhere about protecting blankets "with a mizzou slurry worked into the blanket." This was second-hand info, i.e., the person commenting had seen and had worked with a forge based on this construction, but it wasn't his work, so he didn't have any details. So anyway, I got a 55# bag.
When I opened the bag, I was surprised to discover that it had fairly coarse aggregate in it, up to at least 3/16". Unrealistic to work pebbles into the blanket, even little pebbles. Hmmm.... nevertheless, I weighed out 5#, added enough water to get a somewhat spreadable consistency, put it on a 1'x2' piece of blanket and rolled that onto a letterbox form. It broke immediately when I took it off the form, but the blanket held the pieces together well enough to stand it up, dry it and do an initial firing in the oven, then use it as a burn tunnel for one test run. This was actually back between BOOH V and VI. I just wanted to see if the added mass did anything drastic to the way the system performed with bare blanket. Basically it seemed to somewhat slow and smooth out the ramping up of the system temp.
Recently the stars have again aligned so that I had time to work on this project again. The aggregate is too coarse for what I want to do, so I sifted it out. Using a collander with roughly 2 mm metal mesh, sifting pulled out 40% by weight, pebbles along with a certain amount of refractory fiber lint. The sifting mats this stuff up, so it probably isn't worth the trouble to try mixing it back into the 60% cement/fine aggregate.
Cut to the chase. I molded another letterbox piece, using 4# of the sifted material:
This piece is 8" wide and tall, and 12" long. Based on a density of 140 pcf, the hard layer is about a sixth of an inch thick on average. Obviously it is weak against wishboning, and by the time I got it stabilized in the build, it had two longitudinal fractures. However, the blanket keeps it together and once in place with the sides buttressed, it wasn't in imminent danger of collapse.
I stuck another flat test piece in as a fuel landing pad and otherwise more or less rebuilt BOOH VI around this new burn channel. The heat riser is unchanged.
This build is still pretty much of a hack, so the details aren't interesting. I did manage to get the connections comparatively tight.
So far we've run this system twice. Of course it draws better with the roof not collapsed. It heats up at about 50˚/min, the way we have been running it, and it is possible drive the temp up to 900˚. However it "wants" to run more in the 650-750˚range.
The air/fuel inlet is now the biggest problem. It is just way too shallow and it's somewhat awkward to choke. Plus I have to reduce all the fuel to 10" or less to minimize the smokeback problems.