Corsair H100

[TheDarkWraith]

Why is it that PC hardware manufacturers are now releasing bug ridden/unfinished/untested hardware just like PC software manufacturers have been doing for some time now

Case in point: Corsair's H100 CPU water cooler. Installed it last night into a first gen Antec 1200 case (had to do a little surgery to the case to get it to fit correctly on the back fans). Fired up computer and everything was great for the first 10-15 mins then an audible noise starts emanating from the pump. The noise comes and goes now and it's REALLY annoying as it can be heard over everything else.

Being the engineer that I am I pulled out my mechanic's stethoscope to take a listen to see if I could identify the source of the problem. It's clearly evident that the noise is impeller cavitation - the impeller is spinning way too fast.

Did a Google search to verify my findings and it appears that many others are experiencing the same problem. Why is a manufacturer allowed to sell these products when they know they have a design problem?

Took a standard 1N4007 diode and placed it in series with the +12VDC supply to the pump to drop the voltage to the pump by 0.65VDC thus decreasing it's speed. This was the only way to lower the speed of the pump as the unit does not come with any way to control it's speed. In theory this should decrease the impeller cavitation since it would now be spinning slower. Fired up computer and voila - noise problem eliminated

Why is it that an ex-Nuclear/Electrical engineer can diagnose/find a fix to this problem but the manufacturer cannot

My only worry now is how long this band-aid fix will last

EDIT:

I should clarify that I had to install the radiator with tubes pointing down as that's the only way I could get it to install on the back fans (with some case surgery). That makes me wonder if that is somehow decreasing head pressure to the inlet of the impeller? The tubes of the radiator are lower (height wise) to the pump. Two-thirds of the total radiator sit higher than the pump so I would think that the volume of water contained in that 2/3rds would be ample head pressure to the pump's inlet.

[longam]

You could always test that buy trying a U-tube at different heights for head pressure.

[TheDarkWraith]

Can't really move the radiator anywhere else in the case. Where it's at is where it has to be.

I removed the 1N4007 diode and the pump has been making the noise more often now (after only one day of use!) so I decided to try a little test to prove my cavitation theory.

Fired up Intel BurnTest and put in on maximum and ran the test 5 times. The radiator is setup in a push-pull configuration using Artic Cooling fans. The test didn't even complete the first test when I saw the CPU temp hit 85C then computer shutdown (CPU overtemp). Now that's very strange in a water-cooled setup with a push-pull configuration. The pump was making the noise the whole time during the test. Knowing that cavitation greatly decreases efficiency my theory that the pump is cavitating is holding true.

Reinstalled the 1N4007 diode and the pump quit making the noise, totally quiet. Ran the Intel BurnTest again for 5 iterations at Maximum. The test completed all 5 iterations and max CPU temp was 68C.

For those of you using any of Corsair's water-cooled CPU coolers if you're pump is making noise it's because it's cavitating (impeller spinning way too fast for the head available). Installing a 1N4007 diode in the +12VDC line to the pump will drop the voltage to the pump just enough to slow it down and prevent cavitation.

For those of you who don't know cavitation will eventually erode the impeller blades to the point they are totally inefficient. Do your pump a favor and install a diode in it's +12VDC line or install a fan controller and hook the pump's molex connector to it to control the pump's speed.

Corsair's engineers should be ashamed of themselves. Well probably like every other company they don't care as long as they get your money