Greg : if the tin cells aren't supposed to be produced by uum, how are they supposed to be produced?
If you think a player should mine for a perpetual supply of tin, then
a. This is lame, because tin isn't even consumed in the reaction
b. It makes the fusion reactor uncompetitive with renewable energy sources
Anyways, it sounds pretty simple to me : just add another output slot to the side of the fusion reactor that the products go, and put the empties in that slot. I think this is the best solution, as it makes the most sense. It doesn't sound like a whole lot of coding, mostly cut and pasting.
Second, you should total up the total amount of energy invested to make a fusion reactor, and make it pay for itself within some defined time period of continuous operation. The EU/second output can be defined by this. I'll do the math here in a minute.
Third, another nasty problem with your fusion reactor : it requires absurd amounts of water to be processed. (the base ingredient). This is laggy, especially in SMP. (pumping water in minecraft sends a block update to all clients, and you have to do it a LOT). For this reason, I would multiply the fuel durations for fusion fuel by a factor of 10 or so.
Fusion reactor iridium cost : (0.5+ 1+ 0.25)*27 + 8+1+6+ 1+2 = ~186 iridium plates. That's 747 iridium ore. Rounding to 750 because I may have miscounted, these recipes are complicated. Each iridum ore is 7 uum. Each uum is 16.6 million energy. Thus, with default settings, you need to make 87,150 million energy.
Fusion reactor makes 4096 EU/tick. Net result, a fusion reactor needs 300 hours to pay for the energy that had to be used to make it.
Ok this really is pretty absurd. If a player needs energy, they aren't going to have 87,150 million energy to invest in a way to get more. Also, at the current absurd tin burn rate : over the 300 hours, it'll take 41,556 tin ingots. Or 20,778 pieces of tin ore. That's 6 doublechestfuls of pure tin ore.