I got in there this morning to see what was up -- and to document what happened. (you should be able to click on the pictures to see larger versions).
I detatched the circuit board to photograph the inside. As you can see, quite a bit of carbon was splashed about, painting one section of the cylinders black, and coating the back side of the circuit board.
Last night Ryan and I saw what we thought was a drop of water where the limiter screws in to the cylinder body. Wasn't sure if that was it or not. Well, closer inspection this morning reveals that it very likely was (I feel like I should go on CSI or something). You can see the drip of water up by the limiter, and another spot below it when it dripped down. The arcing occurred right vertical with this. This, combined with the fact that the heater had not yet been attached securely to the wall mount yet (until I was sure that it worked) so the unit wasn't completely vertical front-to-back, allowed a drop of this water to fall on to the circuit board and short out 240 volts and ... well, at least 40 amps.
I was trying to be sure I didn't tighten the thing too much like the instructions said. Guess I should have made it a little tighter. A little drip like that. It's a good argument for epoxy-coating the back of that PCB, no?
Anyway, need to call Seisco and see about a new one.