Captain obvious here. There's a problem in that picture.. water blocks should NOT be connected in series...only one asic will get useful cooling, the other will overheat quickly and probably fry itself or the board.
you dont understand water loops. take a look at SLI/xfire water loops some time.
I lol at the low hash rate and high error rate though. wonder how many watts that thing is drawing..
sooo glad I bailed the BFL ship after getting singed on my 65 nm gear.
Q=-kA(Ts-T) - Hot water absorbs energy slower than cold water!
True, but that's only if the water sits around long enough to get hot. Assuming the pump and radiator are working correctly and efficiently, there should a steady stream of cold water to quickly whisk away heat from several chips in a series. If it works for double/triple SLI and Crossfire setups, it should work for some BFL chips. Then again this is BFL we're talking about, so I doubt they've planned that far ahead.
I suppose it depends on how much energy these chips will consume (and give off as heat). If they meet their power specs (they won't), then it would be reasonable. CoinTerra, for example, has their water blocks set up in a parallel configuration (last I looked), which may be the result of each chip consuming ~500W.