Pro/Forums - View Single Post

pHaestus · 11-30-2003, 08:23 PM

Some test results!

I ran a test with the MC5000-A waterblock today in the hopes of getting an estimate of "W" from CPU. By measuring the waterblock inlet and outlet temperature and flow rate, you can solve for W using the equation:

W = (lpm * 4186 * deltaT)/60

(from Bill's thermochill radiator reviews

So one should be able to get a decent average of CPU wattage by running tests over a range of flow rates. Here's the data:

http://www.procooling.com/users/phae...un1numbers.jpg

Oops W isn't all that constant! I know why though. My wb outlet water probe is about 2' from the wb outlet at the moment and I think at low flow rates that this resulted in delta Ts being erroneously low. At very high flow rates the resolution of my Digitecs is not sufficient and I'd really need 0.001C to get accurate W estimates (a difference of 0.01C in deltaT adjusts W by 9W at 2.25GPM). Between 1.25 and 2 GPM though the W estimate seems reasonably constant and reasonably valid. I'll replumb the wb in and out probes tomorrow to see if I can't improve low end W estimates.

Once I had the W estimate of 81.53W at 2200Mhz 1.81V, I could generate a C/W vs flow rate graph for the wb:

http://www.procooling.com/~phaestus/mc5000a-run1.jpg

This also looks pretty reasonable compared to Bill's MC5002 curve; somewhat lower C/W but that's not too unexpected since I have secondary heat losses in my PC that aren't in Bill's tests.

I can also reanalyze the MCX462-V data with the 81.5W value:

http://www.procooling.com/~phaestus/cwmcx462v.jpg

And there you have my day's efforts.

//edit typos

11-30-2003, 08:23 PM	#17
pHaestus Big Player Making Big Money Join Date: Aug 2001 Location: irc.lostgeek.com #procooling.com Posts: 4,782	And now for something completely different... Some test results! I ran a test with the MC5000-A waterblock today in the hopes of getting an estimate of "W" from CPU. By measuring the waterblock inlet and outlet temperature and flow rate, you can solve for W using the equation: W = (lpm * 4186 * deltaT)/60 (from Bill's thermochill radiator reviews So one should be able to get a decent average of CPU wattage by running tests over a range of flow rates. Here's the data: http://www.procooling.com/users/phae...un1numbers.jpg Oops W isn't all that constant! I know why though. My wb outlet water probe is about 2' from the wb outlet at the moment and I think at low flow rates that this resulted in delta Ts being erroneously low. At very high flow rates the resolution of my Digitecs is not sufficient and I'd really need 0.001C to get accurate W estimates (a difference of 0.01C in deltaT adjusts W by 9W at 2.25GPM). Between 1.25 and 2 GPM though the W estimate seems reasonably constant and reasonably valid. I'll replumb the wb in and out probes tomorrow to see if I can't improve low end W estimates. Once I had the W estimate of 81.53W at 2200Mhz 1.81V, I could generate a C/W vs flow rate graph for the wb: http://www.procooling.com/~phaestus/mc5000a-run1.jpg This also looks pretty reasonable compared to Bill's MC5002 curve; somewhat lower C/W but that's not too unexpected since I have secondary heat losses in my PC that aren't in Bill's tests. I can also reanalyze the MCX462-V data with the 81.5W value: http://www.procooling.com/~phaestus/cwmcx462v.jpg And there you have my day's efforts. //edit typos __________________ Getting paid like a biker with the best crank... -MF DOOM Last edited by pHaestus; 11-30-2003 at 10:37 PM.