mb side: (notice the y splitter with shutoff valves i installed for gpu wb upgrades)
TT mozart is the most practical case:
core i7 lapped: check!
nb wb installed: check!
cpu wb installed: check!
and a shot from behind:
So i just bought the i7 920 and asus p6t deluxe as planned. No waterblocks are available atm so i bought a koolance 340 cpu block, fabricated a backplate and sligtly modifyed the retension bracket so the WB would sit firmly. I also bought a nb wb from coolance since the asus board has this feature- hybrid nb heatsink.
I tried running the cpu at stock (2.66) speed and afer a few minutes of priming the water in the cold side loop would stabilise around 11C (but would still drop slowly since i still do not have a temp controller yet), and the cpu temps were a bit high, in the 40s. I am still testing so i will update with 4g results really soon ( 920 overclocks easily to 4G but the problem is i cant have the pc running for more then a few minutes cause i fear condensation.
I still have to do more testing but i am puzzled by the cpu temps, which are high even at idle (the wb is cold to touch and water is at abouut 10C), i have lapped it and resitted the wb a few times to make sure it makes good contact... maybe it is just the temp sensor, cause with stock cooling and at stock speed i would get load temps in the 80s!!!
EDIT: after more testing i got the following results: after half hour prime cold side water would heat up to 17C and the cpu would stay at around 55C on all 4 cores (cpu Vcore 1.4v both in bios i=and during prime. asus p6t has loadline callibration to eliminate vdroop) but as soon as i stop prime it would drop to 25-30 in a sec. Actually i am shocked how hot the chip runs, after reading all the amazing stories here at xs i thought 4g temps were easily manageble but i was in for a surprise. my qx9650 at 4g and 1.45v runs at 65C load after half hour priming on a aluminum 120x240 rad with 2 low cfm fans (same wb)!
also i noticed something wiered, realtemp would show 55C and asus pcprobe 51, but when i tested at 1.5V i got 70C in pcprobe and 65 in realtemp...
in any case what really matters to me is the cold loop temp, and even though it is a bit high it does rise only a few degrees compared to idle temps.
I am planning to add just one gpu to the loop, most likely 280gtx 55nm as soon as such a product is released, and i expect to get 20C in cold side loop at load. i think that should be attainable cause in real life it is really hard to have the cpu running 100% on all 4 cores. if i find that the cold side loop is running over 20C i'll add another 240 or 480 rad, and that should do the trick as the tecs are capable to handle the extra load with better cooling.Edited by teatchme - 11/29/08 at 11:17am