Cooler test on Tom's Hardware Guide at . . .

[adamslaw]

Two models are recommended: the Swiftech MC370-0A, which had the best cooling performance but also led in the noise test; and the Global Win FOP32-I. Interesting read, especially for the conclusions about the Hedgehog and other highly regarded coolers.

Adam

Not sure how to post the line, but it's at:

Tom's Hardware Guide

 

[Insomnium]

I think these cooler reviews are crazy. Every site that does one seems to get different results. In fact, if you compare cooler reviews from about 10 different sites, you'll see a ton of inconsistency. The reviews are helpful, but most of the stats, such as how much cooler Combo A will run than Combo B should be taken with a grain of salt. Those are just my thoughts on these reviews and i know a lot of other people will agree with me.

 

[aznmist]

It's inconsistent becase they use different ways to measure the temp. Some use the mobo and some use a thermister...and they place it different places too..

 

[Renob]

Well I just ordered the swiftech. I hope tom was right in his testing. I hope to get it friday or saterday . I post my temps ASAP.

 

[tonyou]

The Alpha heatsinks in his test may have done better if they had the same high speed fans as the GlobalWin or Swiftech. I think the Sanyo Denki fans on the two Alphas are in the low 20CFM range.

Tony

 

[Mikewarrior2]

My response to this pretty much crap-filled review is In This Thread. THis review clearly has hte same problems, if not more problems, as the Anandtech Heatsink "comparison".


Mike

 

[Renob]

Mike why dont you do a review?????????????

 

[Liquid_Lenny]

Yea, Mike, why don't you do your own heatsink review??

 

[Mikewarrior2]

Because I neither have the time nor the equipment to do it. ANd with recent developments at overclockers.com, I hvae reason to believe that their future heatsink reviews will be much better than the run-of-the-mill reviews.


Mike

 

[Mustanggt]

well I dont know if Toms review is worth taking to heart. I do know that I got a brand new TaiSol CGK742092 and this thing kicks but on my Alpha 6035!! I nave have my Duron 600 @1050 and temps never go higher than 41 C at full load. my alpha would be running temps around the 45C t0 47 C.

 

[Brian48]

Yeah, you gotta take all these reviews with a grain of salt. I'm using a cheap, "generic" heatsink with a fairly quiet (sub 5000rpm) fan. The only thing special I did was reverse the fan (ala Alpha) and made my own heatsink shroud out of a piece of aluminum foil. With a little bit of silver grease from CircuitWorks applied, here are my results.

Max temps never go above 45C under heavy gaming conditions.

 

[Mikewarrior2]

you could potentially be running warmer than that depending on what heatsink you are using. The majority of low end socket A heatsinks "temps' are read lower than what they really are(ala C-orb, the cheaper coolermasters).


Mike

 

[Brian48]

I should also mentioned that the heatsink that originally came with the CPU had a faster, but louder fan. Although I could tolerate the noise, the heatsink didn't seem to cool as well. Idle temps hovered at around 40-41C with max at 54-55C.

 

[Brian48]

<< The majority of low end socket A heatsinks &quot;temps' are read lower than what they really are(ala C-orb, the cheaper coolermasters). >>



Yeah, I thought about that too, but I'm pretty sure the results are accurate. The thermal sensor, in both instances, was properly contacting the back of the CPU, not the heatsink. Any reading would have resulted from what it could gauge off the back of the chip, not any other heat source. Whatever was happening, a lower temp was the end result.

I'm a firm believer that you can get decent cooling with just about any heatsink so long as it's made from decent materials, mounted with good clips, a proper layer of thermal grease applied, and is carefully mounted. I really took my time and spent nearly an hour getting it installed &quot;just perfect&quot;. Taking the time to do it correctly is the key in opinion. Then again, not many folks are as anal as I am when it comes to these things.

 

[Mikewarrior2]

Contacting the backside of a cpu, because it is a secondary heat pathway, does not show true/full cpu core temp changes nor is it accurate.

I will opst a link to a long discussion about this later.


Mike

 

[Renob]

Mike what resent developments at overclockers.com are you talking about??????????

 

[Mikewarrior2]

Heatsink Testing Methodology

and

Thermocouple Tests.


They ran some tests with an isolated thermocouple directly behind the cpu core. This temp was very significantly higher than the kt7(which touches the edge of the core, on the backside of the chip).

At a loss of around 25% at compared to an isolated back-of-core reading, i'm guessing the thermistor reading location loses only reads 70% or so of the primary heat pathway. that number goes down when the thermistor is not isolated.

Their Swiftech Review, Very Nicely done

Mike