it's true you don't need an inner can container to act as the wall of the foundry when you use fireclay. But he said the walls were filled with crushed fire brick, so you would need something to keep it from caving in. For this type of furnace, I would recommend going the fireclay route.
But personally, I prefer using whole fire brick arranged into walls. Just seems cleaner and like it would last longer (also easier to fix: just take out broken brick and put in a new one).
charcoal with an air blower can be used to melt aluminum. the coals should be below and around the crucible (the more surface area that's in contact with the coals, the better) with air being blown in somewhere near the bottom.
I want to share
My big aluminum furnace runs on coal simply because coal is cheaper for the same weight
and produces more heat for the same weight.
I was able to design mine so there is a cylinder of dense fire brick on sides and floor (yellow in picture), surrounded by perlite (white in picture). Air is blown in through multiple holes in the floor. It even allows for spilled aluminum and ash to fall through the holes onto the gravel road below for easy cleanup. The holes are actually drilled into the dense fire brick, and the brick is held up by a grate made of aluminum I made. The bars of the grate are spaced perfectly so they are exactly between each hole. Air flowing through the holes is enough to cool the bottom brick so the aluminum grate doesn't melt, and being completely hidden protects it from IR heat. The walls are a 25 gallon metal trash can (brown in the picture), the floor that the air tube goes through is cement and perlite (black in picture), the lid is fire clay / cement / sand / perlite slab I made a while ago for another furnace (black in picture).
After some use, there are a couple holes that are plugged by molten iron from a crucible that failed, but there are still like 100 good ones, so it'll be a while before I need to replace those bricks. It would be too hard to melt the iron out, but maybe I could rapidly oxidize or dissolve with chemicals... HCl maybe? would that react with the fire brick?
The outer walls do get hot, but it seems to happen from the top down, meaning it gets hot because the fire touches the top of the can (around the rim) and travels down, not because heat is lost through the walls. Perlite is VERY insulating, but melts at high temps (well above the melting temp of aluminum, but the coal/charcoal can get hot enough), hence the fire brick.
Everything except the cement floor is loose. The perlite keeps the bricks in place, and each brick has a beveled edge that keeps it from falling in. If a brick breaks, I'll probably leave it in place. If several break or perlite starts falling in, I can take everything out and replace the broken parts with spares I made. The floor, being the most delicate and with the most strain on it, can be replaced without taking anything else out.
Overall: The inside diameter of the cylinder is 31 cm (12 inches). The inside height of the cylinder is 42 cm (16 inches). The cylinder is a decagon (10-sided), 2 bricks tall. The whole thing can be picked up and moved by 1 person awkwardly or 2 people easily, maybe 25 kg (55 lb). Crucible is a 4 quart cast iron pot, can hold about 9.5 kg (20 lb) of aluminum. Takes about 60 minutes of preheating before it starts melting.
