I've just stuck watercooling in one of my i7 920 4u folding rigs but the temps appear far too high compared to what I'm expecting. I've got a Laing DDC with xspc res on top 240mm rad Swiftech GT block. Approx 50cm of tubing (had to put a big loop of tubing in to go round an angle) approx 500ml liquid 2x high flow 120mm fans (can't remember the brand) Just the cpu is cooled & the fans/rad are attached to the front of the case sucking cold air in. I'm getting 55 degrees at stock @ 100% load & 75 degrees @ 3.8GHZ (1.325v) @ 100% load. It's a C0/C1 920 on a P6t-SE with 6GB XMS3 RAM. I was wondering if the pumps flowing too fast so the liquid isn't spending enough time in the rad being cooled or if I'm just being a muppet.
Yup - check cpu mounting... The length of the loop makes pretty much no difference (barring it being too long for the pump to force water round) to temps. If the water is moving faster, it passes through more times...
CPU block mounted correctly, still getting up to 75 degrees at 100% load (although the room is quite warm). I'm going to add a couple more fans to the rad & possibly fit a new block as I'm not 100% sure about the backplate on the GT (it has to bend to go round the cpu back bit on the mobo as it's a 775 fitting)
I'd say it was your CPU waterblock's cooling ability rather than anything else. I recall the GT being an old block, so I would suggest picking up anything like a GTZ or a Heatkiller 3.0 LT and it should knock about 12c off your temperatures. As an example, when I ran my 920 D0 at 3.8GHz with 1.3v, I was hitting around 60-61c on core 0 under Prime95. With a new block, i'd imagine you would see a similar result considering yours is the C0 while I have two 5870s adding heat to the loop.
i'd try reversing the airflow and blowing out of the case, this way you are pretty much dumping all the heat the rad gives off back inside the chassis.
That temp is pretty high. It should be hitting 55 max at load with a hefty overclock. What pump are you using? D3 or D5? Im guessing your rad is single thickness swiftech? Did you bleed your system of bubbles because the pump might not be circulating water very well with lots of bubbles in it. Make sure the water is moving well. Did you apply thermal paste? (stupid question I know but it would explain the temps) What liquid are you using? Where is the radiator? If it is mounted to the top of your case, is it pulling air in or pushing air out? More fans would not help airflow just like a car with more wheels wouldn't be faster. Now faster spinning fans or wheels would make a bigger improvement. What speed are your fans running anyway? What is your room temperature. If it is sweaty hot, then you might want to open a window.
MPC350 pump with a single thickness RAD. The fans are pumping out approx 50cfm each. THe loop is bleed & water is moving rather fast. THe liquid is feser one (all I had sitting around). AS5 on the cpu. THe rad is mounted at the front of a 4u short rack case, nowhere else to fit it in that case, there are 2 80mm fans on the back panel sucking heat out (+ a meshed back on the case) The rad is warm & the room is a stable 23 degrees (AC is now on). The temps have dropped to 70-71 degrees now the room has settled down. I'm going to recase it & check the air cooled temps first then consider a new block if I decide to continue watercooling it (& maybe another rad if the new case is bigger)
Like i said, try reversing the airflow. The rad is dumping all it's heat back inside the case, so basically those two 80mm fans are tasked with removing ALL the heat from the enitire system. With 23 degrees in the room the temp inside your case may well be 40 degrees at this point, making a temp of 70 under full load OCed much more probable. Please take the advice from the only guy giving a response to your specific situation
Doesn't matter how fast the liquid is moving. Yes it will spend less time in the rad but because it's moving faster it will be going round the loop faster too, so will reach the rad again in a shorter time. Over long periods, the amount of heat energy transferred is basically the same. The difference is tiny to the point you won't see a difference in temps. If your flow rate is too low for the CPU block you're using though, temps may be a bit higher as some can make good use of the extra pressure and flow to create turbulence which increases heat transfer. I suspect the rad isn't up to the job. A warm rad is a rad that's overloaded although as Sh0cKeR says, the GT wasn't really designed for i7's. The XT is the best block we've used for cooling modern CPUs.
Blowing hot air from inside the case through the rad is only going to increase the temps. Cool air from outside the case blowing through is better. Just means the motherboard will get a little bit hotter.
