[ehume]

I've never bothered to test my fans in pull-only mode, but tonight I accidentally attached my San Ace 9S1212H401 fan backwards: it was facing me rather than the Megahalems.

It was a few minutes before I noticed that the fan was sucking air rather than pushing air through the stack. So I let OCCT finish the run. An hour later, I learned that the fan, pulling air, had achieved a difference of 48.5c over ambient.

The rig is my sig rig, running at Vcore 1.28v, 4004MHz with a Vtt=1.19v, PCH=1.2V, PLL=1.9V and DRAM=1.616v. The 48.4c diff is not bad. The best I've gotten so far is with twin Ultra Kaze 3000's, at 44.5c difference, and with twin San Ace PWM fans running nearly 4000 rpm, also with a diff of 44.5c.

So of course I flipped the fan around and tried again. The ambient went up to 18c from 17.5c. Trivial, right? But the diff with the San Ace Silent in push mode was 50c.

Of course, I will run this comparison again tomorrow. But please, will a bunch of you test it out? Sometimes when theory meets fact, theory must die.

------------------------------------------------
2010-03-28 Did some testing today. I used OCCT set to Linpack, 1h5m, which gives a 60-minute test. I tested three fans, first pushing, then pulling. Ambient was 17.5c throughout the day. The system was shut down to reposition or replace the fan each time.

Scythe Gentle Typhoon 1850 rpm, rated at 58 cfm.
Push: Temp over ambient = 52c
Pull: Temp over ambient = 52.5c

San Ace 9S1212H401, rated at 86.5 cfm.
Push: Temp over ambient = 49.5c
Pull: Temp over ambient = 50.5c

San Ace 9G1212H101, rated at 99 cfm.
Push: Temp over ambient = 48c
Pull: Temp over ambient = 48.5c

In essence, given the fact that the cpu/mb reports temps in whole numbers, there is little difference between push and pull - at most 1c favoring push.

Again, people should compare for themselves and post the results.

 

[makecoldplayhistory]

My CPU's about 3c cooler at load with pulled air.

 

[ehume]

I've started my runs. I set OCCT to 1h5m. That gives 60 minutes of stress testing (1min pretest idle, 4min posttest cooldown). Today I'm doing first push, then pull. I'm starting with the 9S1212H401, then I'll move on to the 9G1212H101, then a Gentle Typhoon 1850. I suppose I should try a 9G1212P4G03, which is a PWM fan spec'd at 4100 rpm (mine run just under 4000) and 130 cfm. A 5v, this fan runs a bit over 2000 rpm and blows a lot of air. That will give us a good bracket of output to test.

 

[BlackHoleSon]

I Get a 4-5C decrease with an 80-ish CFM fan pulling. The stock Cooler Master Hyper 212+ is pushing, by the way.

 

[TFB]

Very interesting. Sub'd

 

[letsgetiton]

Quote:

Originally Posted by BlackHoleSon I Get a 4-5C decrease with an 80-ish CFM fan pulling. The stock Cooler Master Hyper 212+ is pushing, by the way.

Yes, but he's talking about 1 fan. (either on the exhaust or the intake side. Pushing OR pulling, not both) We all know a pull fan added to a push fan lowers temps. What we're learning here is a pull fan out performs a push fan in a single fan setup. I never would've thought that and I pondered that thought the first time I installed a CPU cooler and came to the conclusion that the single fan should (difenatly) be blowing on the cooler. Seems against logic but as they say, you learn something new every day.

Thanks for enlightening my day.

 

[ehume]

Quote:

Originally Posted by letsgetiton Yes, but he's talking about 1 fan. (either on the exhaust or the intake side. Pushing OR pulling, not both) We all know a pull fan added to a push fan lowers temps. What we're learning here is a pull fan out performs a push fan in a single fan setup. I never would've thought that and I pondered that thought the first time I installed a CPU cooler and came to the conclusion that the single fan should (difenatly) be blowing on the cooler. Seems against logic but as they say, you learn something new every day. Thanks for enlightening my day.

Well, I posted this as a provocation. The result might have been a fluke. I'll be done with my last test in about an hour, and I'll post the results.

I was hoping other would try out their own fans and see. Some people use only a single 38mm fan one the Megahalems, since at that level a second fan drops temps only another degree. So, for a single fan user it's important to know what really works better.

 

[ehume]

Finished testing on three fans, push and pull. Results posted in OP.