QuietIce
07-01-06, 09:50 AM
I decided to take some measurements before installing my WC system. Not only did it provide a baseline it also gave all the parts a good rinse. :) Flow rates were measured at P5 with 0' head ±0.5' using two 5 gallon buckets and pumping 11 quarts of distilled water. Results shown are an average of three timed runs per test:
MCP: 3.9 gpm (= 240 gph) (100%)
MCP/Rad (custom (http://www.ocforums.com/showpost.php?p=4791444&postcount=31)*): 3.4 gpm (87%)
MCP/Storm/Rad: 1.6 gpm (40%)
MCP/Rad/MCP/Storm: 2.1 gpm (52%)
Setting v RPM data found on a German OC forum:
P1 = 1800 RPM
P2 = 2550 RPM
P3 = 3300 RPM
P4 = 4050 RPM
P5 = 4800 RPM
Since I have 2 pumps I've been doing some in-use testing. Routing is MCP655#1/Storm/GPU(MCW60)/T-line/MCP655#2/Rad. There are no noticeable differences in performance running 2 pumps at P5 or just the #1 pump at P4. Running only the #2 pump at P4 rises the CPU temps a hair (< 1°C). Room temp varies from 21-23°C and water temps in front of the Storm are +3°C. The CPU idles at +5°C with a load temp of +17°C (38-40°C) per on-board sensors.
(That's what I get for mounting the Storm with the MB installed and the case in an upright position :bang head ).
Currently I'm running both pumps at P3 and have only a very slight loss of cooling. Under sustained use (2x Prime95) the water and CPU temps will rise 1°C in an hour compared to running both pumps at P5. With only one pump active at P3 the temps will rise 2°C over 1.5 hours before stabilizing. Neither configuration has any effect on my OC settings - the Opty still runs at 2.8GHz just fine. Even with two pumps the noise levels were never bad but at 2xP3 the pumps are virtually silent.
BTW - I run two MCP655's for the same reason I have two HDD's on RAID 1 ...
__________________
*The "custom" radiator has since lost it's copper plumbing and is now connected with 1/2" tubing and 1/2" copper T's but still hooked up in parallel. Because the rad is in a fixed position I need 10' of tubing (5' out/5' back) between it and the computer plus about 6' of tubing in and around the computer because the pumps are on external mounts. This has worked out great for a test rig but the next system will be mostly internal with an external (but case mounted) heater core.
MCP: 3.9 gpm (= 240 gph) (100%)
MCP/Rad (custom (http://www.ocforums.com/showpost.php?p=4791444&postcount=31)*): 3.4 gpm (87%)
MCP/Storm/Rad: 1.6 gpm (40%)
MCP/Rad/MCP/Storm: 2.1 gpm (52%)
Setting v RPM data found on a German OC forum:
P1 = 1800 RPM
P2 = 2550 RPM
P3 = 3300 RPM
P4 = 4050 RPM
P5 = 4800 RPM
Since I have 2 pumps I've been doing some in-use testing. Routing is MCP655#1/Storm/GPU(MCW60)/T-line/MCP655#2/Rad. There are no noticeable differences in performance running 2 pumps at P5 or just the #1 pump at P4. Running only the #2 pump at P4 rises the CPU temps a hair (< 1°C). Room temp varies from 21-23°C and water temps in front of the Storm are +3°C. The CPU idles at +5°C with a load temp of +17°C (38-40°C) per on-board sensors.
(That's what I get for mounting the Storm with the MB installed and the case in an upright position :bang head ).
Currently I'm running both pumps at P3 and have only a very slight loss of cooling. Under sustained use (2x Prime95) the water and CPU temps will rise 1°C in an hour compared to running both pumps at P5. With only one pump active at P3 the temps will rise 2°C over 1.5 hours before stabilizing. Neither configuration has any effect on my OC settings - the Opty still runs at 2.8GHz just fine. Even with two pumps the noise levels were never bad but at 2xP3 the pumps are virtually silent.
BTW - I run two MCP655's for the same reason I have two HDD's on RAID 1 ...
__________________
*The "custom" radiator has since lost it's copper plumbing and is now connected with 1/2" tubing and 1/2" copper T's but still hooked up in parallel. Because the rad is in a fixed position I need 10' of tubing (5' out/5' back) between it and the computer plus about 6' of tubing in and around the computer because the pumps are on external mounts. This has worked out great for a test rig but the next system will be mostly internal with an external (but case mounted) heater core.