Hi Ironman,
Phil has it right, although an oxidising flame is useful for ally - the fluxes help remove dissolved hydrogen (which ally has a major affinity for in the molten state) which will otherwise outgas during solidification, making the castings porous... An oxidising flame will help reduce this (umm, incorrect term, reducing is not enough oxygen, but you know what i mean! - minimise this?), I'm told most of the hydrogen in molten drinks cans comes from the paint and plastic coatings, similar for oily scrap - it seems a bit much degreasing scrap before melting it, but it helps! SOME of the hydrogen comes from the fuel though, unless you're using charcoal/forced draught.
Some professional foundries use both flux and an inert gas bubbled through the melt t remove the hydrogen - never tried gas myself but I' told it can be very effective, argon's used for top-quality castings for aerospace... Spendy!
If you're after best quality castings, it's worth starting with an alloy that's intended for casting - call at an engine rebuilders and offer them over scrap price (about 25 cents a pound) for old pistons, cracked cylinder heads etc., they're a really good, strong, easily-cast alloy (high silicon content - be sure to unscrew the bolts and studs though, and remove piston gudgeon/wrist pins) and if you have to dilute it go for extrusions (window frames etc.) up to 20 or 30% - extruding alloys (the 6000 and 7000 series) are relatively very low in silicon, have all kinds of oddities in 'em!
For flux, lo-salt wrapped in ally foil works well and is available in most grocery aisles

The big problem with DIY crucibles (reduced with Phil's suggesed refractory wash) is that molten ally will dissolve the iron from a steel crucible, ruining it and buggering up the ally's properties (it gets brittle, doesn't flow and cast well, you name it) - this is why tin-can crucibles normally leak on the first melt and pee in the bottom of your furnace! Your poor result in the experiment with adding the angle brackets may have been down to dissolved iron in the melt?
The wash doesn't need to be anything fancy, fireclay watered down to a creamy consistency and fired works well enough, specially if you go to the bother of checking it and renewing if needed betweeen melts.