Pump RPM speed and reservoir bubbles

Ronnie30

Distinguished
Jun 15, 2014
71
0
18,540
Hi,

I just built my first watercooling rig. I had my pump plugged into the CPU header on the motherboard, with auto/default settings and had my radiator fans plugged into a Sys_Fan header. I read elsewhere that the pump should be run full speed on a different header, and the radiator fans on CPU header. I did this and moved my pump to the CPU_OPT header, and set it to max speed in the BIOS.

Previously my pump was idling at about 770rpm and went up to 2445 during a stress test. I didn't run this stress test long enough for the water to reach max temp, as I read about the pump needing to be full speed, within a few minutes of running the stress test. Now, with the changes made, the pump runs permanently at 4750 rpm. Speeding it up definitely revealed more air bubbles in my loop. I keep reading that pumps usually have variable/adjustable speeds, so I'm just wondering if running it at Max 24x7 is necessary?
Also, the loop had been running for 24 hours prior to this. Do I need to run another overnight leak test at the higher rpm?

I also feel like my reservoir water/coolant (EK-Cryofuel) got darker, to the point where I can barely see the anti-cyclone device inside. Is that normal?

Lastly, after first setting it up, I left it running for about 8 hours. Upon returning, the reservoir looks to have a lot of bubbles in it, but tapping or tilting the case doens't move them. It almost looks like they are stuck or they are white specs of dirt inside it. This doesn't appear in any of the tubing though, so I doubt it's dirt. Any idea what this could be?

Before pump was set to max speed - bad pic but the anti-cyclone was definitely more visible:
before_pump.jpg


After pump was set to max speed:
after_pump.jpg
 
Solution
Cavitation can damage some surfaces mostly on impeller itself. Baubles are actually made out of steam produced by fast movement of impeller (propellers on boats and ships) and when they collapse they can pull material off whatever they touch.
Solution is same, don't run it any faster than necessary for good cooling. In addition. water is slow to transfer heat to and from so running too fast may produce opposite effect and actually cool less.


Thanks. Is it or will it cause any issues long term? If so, would dropping back to a custom pump speed resolve this?
 
Cavitation can damage some surfaces mostly on impeller itself. Baubles are actually made out of steam produced by fast movement of impeller (propellers on boats and ships) and when they collapse they can pull material off whatever they touch.
Solution is same, don't run it any faster than necessary for good cooling. In addition. water is slow to transfer heat to and from so running too fast may produce opposite effect and actually cool less.
 
Solution
Thanks. I'm not sure how to figure out the optimal speed, but I guess I'll drop it and test how it fares.

Is using the CPU_Opt header on an auto setting OK? As above, it ran at 770 rpm idle and 2445 during stress test when it was set to auto.
 
From within BIOS, I can. My options are Auto/Silent/Full Speed/Manual. With Manual, I can set a percentage with options ranging from 0.75 PWM value /oC ~ 2.50 PWM value /oC, according to the User Manual. I haven't had a play around with it yet, but when I had it on Auto, it only peaked at 50% max rpm.
 
And I definitely have more of these bubbles now, 48 hours after initial build (and about 12 hours after I set pump to max speed), than I did 24 hours ago.

I also have some more air at the top of the res, and some condensation. Pictures added.

pump_flash.jpg


pump_no_flash.jpg
 
There's fewer bubbles now, but they still aren't completely gone. It's been 4 days since I went back to auto speed. (Manual curve, on the CPU_OPT header doesn't seem to do anything. It runs at 100% speed regardless of how I set the curve). Should they have dissipated by now? Is there any harm in leaving them, assuming new ones aren't forming?