Why don't people mention room temperature when someone asks for comparisons

[dennilfloss]

e.g. for a Tbird.

We all know that socket A thermistor readings are somewhat inaccurate and that placement of the thermistor and compensation varies amongst boards. However, even identical configurations can only be compared reliably for temperatures if the room temperature is the same (or at least is known). If one person has a room temperature of 30C and another 20C, that is a difference of 10C in the cooling air right from the start and without knowing this variable, the person who reads a 10C higher cpu temp may thing that his/her cooling is way less efficient that the other person.

The Bios rev. is another variable that might be useful in interpreting temperature comparisons.

 

[WarCon]

Actually for an air-cooled processor, the only real pertinent air temp is the one inside the case. If you have an ambient of 20C but a case temp of 30C, then someone can let you know that you need better air circulation in your case. But strictly for whether a hsf works well or not is to compare case temp to full load to get your delta so you can figure out how well it cools. Its just too bad AMD doesn't have an on-board thermistor, because the temps that we see are so board dependent.

P.S. Denni, I am on the first page for TACube. 44 and climbing....oh and if you wanna update your processor page, I am currently running a 1ghz T-bird at 1.33.

 

[dennilfloss]

I get a mobo temp of 34C with an ambient temp of 28C. With rc5 full time, cpu temp is generally 19-20C above mobo (A7V133) temp for my 1.1@1133GHz AJFA Tbird.

The first page of the Borg pproxy is getting more competitive by the minute. What were you using before?

 

[WarCon]

Duron 600 @ 950, I had to tune it down shortly after I told you to like 907 or something cuz I chipped the core. But that only lasted for a couple weeks til I got my T-bird.

Yeah and aircooled my full load temps averaged 22-25C higher (than system)with the T-bird and 19-23C higher with my duron (after I bent up thermistor on KT7A). Before bending up the thermistor it was 14-18C higher.

Now with my water cooler it is 10-11C higher.

 

[Mikewarrior2]

Even when including room temp and ambient case temp, socket-a temp comparisons are still highly suspect...

The only thing you can realyl do is to check if the temps are "reasonable" ocmpared to other temps of your same MB type.


Mike

 

[WarCon]

Yeah, I know Mikewarrior. I really don't even like giving out temps because not everyone who's using the KT7A has made sure the thermistor is touching, and the temps will vary so much depending on where the thermistor really is. And as you have pointed out to everyone, these aren't even die temps, just approximations made by measuring heat behind the chip in the socket area and then applying a compression filter formula to the reading. I have a pretty hard and fast way to tell. If my overclock goes unstable and my hsf is getting warmer then I have a problem.... either my hsf(cooling solution sucks) or my case cooling sucks. Wish there was better way, but until on die thermistor is implimented in palamino(if) we are all stuck guessing and doing the best we can.

Hey Mikewarrior or anyone, do you happen to know the dimensions on the AMD mobo holes around the socket or know a place I could find that info? I am making a new water block to try and want to make sure it slides over my bolts with no problem. No chance of being held up on them.