First post, here I go. I'm attempting to create a high-temperature (<1600C) furnace for melting mild steel. I plan to achieve the necessary from a forced waste oil/diesel burner, driven by an air compressor (propane is expensive in the UK, at least in my area).
I've read a lot about refractories recently. I'm aware commercial refractories are far superior to anything I could plan to make myself, but I propose the question: can I actually melt cast iron/mild steel using materials I can obtain for more or less free?
I have certain plans already: the foundry will be housed in a 20-gal drum, with a 1.5-inch 'hotface' and a 2-inch insulation layer. The general problem seems to be that materials resistant to high temperatures tend not to be very insulating, so I've devised layers to specialise to both purposes.
The hotface will act more like a crucible than insulation, and will protect the insulation layer. Portland cement cannot be used here as it will kind of explode. Silica (play/fine) sand won't hold up very well either, which seems to melt at exactly 1600C -- my planned max. temp. I therefore need to rely on clay and grog to resist the 1600C heat, mixed with polystyrene beads which will burn-off and create air pockets which will allow the clay room to expand when heated. Grog will also help reduce expansion/shrinking of the clay, and therefore reduce cracking over time. I can obtain grog from old porcelain (completely white the way through) tiles for free, and the soil in my area is very high in clay (found barely a foot down), and is a yellow-grey colour. I'm aware due to the yellow it is slightly contaminated with FeO or some other contaminant, which may act as fluxes and reduce its melting point. Will this clay still work as a fireclay, or if not, is there a way to purify it? Another alternative binder is hardwood ash, so I've heard, but I lack any understanding of exactly how it behaves under super-high temperatures. I know it forms a sort of concrete, and could provide an additional dimension of physical strength when used with clay (increases the cold crushing force of the refractory, pretty insulating - more than clay I think, reduces expansion/shrinking of the clay). I really want to add ash to the hotface mix, mostly due to its availability, but I'm just concerned that it might melt (I'm also sure purity matters, and have no understanding of what type of ash I actually need).
The insulation layer will be (I'm hoping) below 1200C, at which temperature perlite can be used. I can also rely on silica sand, which I will obtain from the ground here. In the South of the UK, we have lots of 'flint' or chert, which can be ground down (with effort) to an almost pure quartz-silica structure, once the CaO and CaO3 have been dissolved-off. That's another question - can I use CaO in the hotface? I recognise it's high melting point, but why is it never used? Has anyone ever had any experience with chert or flint as a refractory sand? I doubt so but I'd like to check before I try it. Silica (chert) sand will also form a stronger ceramic bond with the clay when fired than perlite, so less clay is needed in the insulating layer (reducing expansion/shrinking -- cracking). I can also employ fibreglass (just the glass) near the edges of the insulation layer to act as a binder (which used to be done with asbestos fibres, before it became semi-illegal over here).
The layers might be reinforced with steel mesh (think: chicken-wire) around the outside, but not sure how useful that would turn out.
I've heard varying suggestions on the amount of clay I should be using, anything from 5% to 25%. That actually seems like quite a big difference. I understand too much clay can introduce a multitude of problems, so is there a recommended amount, especially considering the home-sourced (yet incredibly sticky, and semi-grey/yellow purity) clay I hope to use.
I hope this post will be of use to newbies also attempting homemade refractory, as I've just dumped most of the knowledge I currently have after hours of research into one post. I hope this project will also help prove something quite important - that completely homemade refractories are possible, and than anyone can melt and pour iron/steel on a budget of pretty much nothing.