Switch airflow direction?

[Boris Bilinsky]

Hello, I'm using the thermaltake element q case (there was a review here on silentpcreview on a similar case but another brand but I can't find it now) with an i3 540 cpu, and I have problems both with heat and noise.

I did some tests with undervolting the PSU fan and it got about as silent as I want, but I don't really know how much airflow it need to keep a decent temperature in the power supply. I also added a 120 mm case fan on the side of the case sucking air out from the cpu.

My idea now is to switch direction of the PSU fan so it becomes an inlet instead. I think the PSU and cpu fan "competes" about the air as they sit almost right on each other and blow the air in opposite direction. What do you think about sucking air in via the PSU to the cpu and out with the case fan? I think there will be plenty of air coming in from the front and the other side of the case as well, where I have my Palit nvidia GT 220 (with the fan removed... (~40 degrees C idle, ~80 degrees during load) ).

... and what is a safe temperatures for an i3 processor? I've read 72 degrees, but then I also see 105 degrees maximum junction temperature.

cheers

[ces]

That is a nice looking case, but I think SPCR had some cooling issues with it. What about copying their solution?

Personally I get uncomfortable when my cpu temps exceed 50C. But I am a bit on the squeemish side.

[fumino]

seems like a decent idea really. lets us know how it turns out.

[cordis]

Yeah, max running temp before the motherboard shuts a processor is usually 101 to 105. I have systems where the cpu has been running mid 80s for weeks, doesn't seem to hurt them. So if your cpu cooler fan doesn't get too fast, I'd say that temps in the 80s are probably ok, as long as they are stable there. You usually need to run with your highest load for a while to see where the temp ends up. I only really get worried about temps if they spike into the 90s regularly. But then again, I also air conditioning at my place, your headroom may vary if you expect to have a hot home in the summer.

[Boris Bilinsky]

Thank you for the feedback, I will post here when I've tried. Found some rubber grommets today that I can mount the PSU fan with, I think a lot of the noise is vibration related.

[ces]

Thank you for the feedback, I will post here when I've tried. Found some rubber grommets today that I can mount the PSU fan with, I think a lot of the noise is vibration related.

Did it work?

[Boris Bilinsky]

Thank you for the feedback, I will post here when I've tried. Found some rubber grommets today that I can mount the PSU fan with, I think a lot of the noise is vibration related.

Did it work?

Hello, I managed to test it this weekend, and I must say it works really good. The PSU fan is mounted in reverse airflow direction with rubber grommets blowing "fresh air" down to the cpu and out from the case with the 120 mm fan mounted similar to silentpcreviews article. In this configuration I could undervolt the PSU as well, and still keep it cool. The idle temperature of the i540 dropped from around 47-49 to 39-41 degrees C, during load the difference also is about the same.

Considering noise, the result is essentially that the CPU fan can run at a lower speed for the same cooling, thus keeping the noise almost load independent. The noise consist of a slight hum from the 120 mm fan together with the stock intel cooler, but it is a lot better than running the intel cooler alone at a higher rpm. The startup PWM level is set to 20% and it seems that the fan runs at this level most of the time.

cheers

[nutball]

One of the quietest builds I've managed used this technique - flipped the fan in the PSU to draw air inward and blow down on to the CPU. The CPU had a large tower heatsink (Ninja if I recall) and I build a duct enclose it and ensure that all the air from the PSU passed through the heatsink and out of the usual outlet hole in the back of the case (which I'd cut the grill out of an there wasn't a fan).