CPU/Case temp differential
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
CPU/Case temp differential
I've been keeping an eye on my temp readings and I'm noticing that there seems to be about a 20 degree C difference between my case temp and my cpu temp readings. I know it's normal for CPU temp to be higher, but how much higher do you guys get? I'm worried that heat may not be transferring to the heat sink efficiently.
When I got my AX-7, it had a sticker on the bottom that had to be removed and it left a lot of sticky residue on it. I used some Goo Gone to get rid of most of it, but I think it might've affected its cooling ability.
When I got my AX-7, it had a sticker on the bottom that had to be removed and it left a lot of sticky residue on it. I used some Goo Gone to get rid of most of it, but I think it might've affected its cooling ability.
-
- Patron of SPCR
- Posts: 543
- Joined: Mon Oct 21, 2002 1:01 am
- Location: Vancouver, BC, Canada
- Contact:
A 20C difference seems about right, expecially if it's a higher wattage CPU. If you look here at MikeC's review of the SLK900 you'll see a similar temp. rise (depending on the speed of your fan).
Just to be on the safe side, I'd clean up the googone with some 99% isopropyl alcohol that you can get from the pharmacy. Add a fresh ultra-thin coat of paste to the CPU and just tint the HS with a smudging of grease. Reseat...
Just to be on the safe side, I'd clean up the googone with some 99% isopropyl alcohol that you can get from the pharmacy. Add a fresh ultra-thin coat of paste to the CPU and just tint the HS with a smudging of grease. Reseat...
My TbredA XP1700 @ XP2000 is 33oC idle, with case temp of 39oC.
my P4 1.6 @ 2.133 is 35oC idle w/ case temp of 32oC.
I don't get why the Tbred has a cooler CPU than case temp. The BIOS reports similar too! Doesn't make sense. Maybe something is wrong with my CPU sensor (Though I have observed this on other GA7VAX boards with this CPU).
Both systems run panaflo L1A @ 7V on an AX7/478
my P4 1.6 @ 2.133 is 35oC idle w/ case temp of 32oC.
I don't get why the Tbred has a cooler CPU than case temp. The BIOS reports similar too! Doesn't make sense. Maybe something is wrong with my CPU sensor (Though I have observed this on other GA7VAX boards with this CPU).
Both systems run panaflo L1A @ 7V on an AX7/478
-
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
Somethin's wrong....
Hello:
Geez -- there's gotta' be something wrong there! What is the HS, and the fan? Is there a flaw/grooves/no thermal grease on the bottom of the HS? At the very least, the life of the poor thing will get shortened...Tigr wrote:Well, I wish I could be that lucky My diff is about 35-40 degrees C Right now, it's running full load with CPU at 70 C and case temp is 28. And that's a Duron...
Re: Somethin's wrong....
If I knew what's wrong it would not be wrong for long, you knowNeilBlanchard wrote:Hello:
Geez -- there's gotta' be something wrong there! What is the HS, and the fan? Is there a flaw/grooves/no thermal grease on the bottom of the HS? At the very least, the life of the poor thing will get shortened...Tigr wrote:Well, I wish I could be that lucky My diff is about 35-40 degrees C Right now, it's running full load with CPU at 70 C and case temp is 28. And that's a Duron...
Anyway, I always have exactly the same problem. I installed and reinstalled the CPU HSF so many times I can do it with my eyes closed. And I have exactly the same problem always - it just keeps running hot.
Ok, I used Arctic Super Silent or whatever it is called HSF which is aluminium with 8 cm fan. The Arctic Silver is applied, tried thicker or thinner layers - no noticeable difference. On another machine, running Athlon XP 1500+ - the same problem, using GlobalWin full copper HSF with Arctic Silver etc etc - no luck, running hot.
I do not know, maybe I just screw it up somehow although I cannot imagine where. Or, maybe I should use a Volcano
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Arctic Super Silent -- that is the CPU HSF you are using? If so, it may be that the thermistor-controlled fan is miscalibrated. When I tested this HSF, I found that the thermistor sped the fan up only when the CPU get pretty dang hot. In my sample, the speed increase still occurred at a low enough temperature to keep the CPU just cool enough.
BUt I could see that it would only take a slight variance in manufacturing tolerance to make it run too slow to providfe effective cooling. Maybe check to see how deeply that termistor is buried in the fins. Maybe it is riding too close to the top where it doesn't really see the heat.
One option, if you are soldering-iron savvy, is to extend the themistor on a longer lead so you can position it at a hotter spot to make the fan kick up sooner / faster -- like in the gap between CPU and HS base. Can't remember if the thermistor is small enough...
BUt I could see that it would only take a slight variance in manufacturing tolerance to make it run too slow to providfe effective cooling. Maybe check to see how deeply that termistor is buried in the fins. Maybe it is riding too close to the top where it doesn't really see the heat.
One option, if you are soldering-iron savvy, is to extend the themistor on a longer lead so you can position it at a hotter spot to make the fan kick up sooner / faster -- like in the gap between CPU and HS base. Can't remember if the thermistor is small enough...
LOL! I did not know Arctic HSF was termistor controlled. I am running it with a Zalman fan controller turned down to 2000 RPM, otherwise it is too loud. If I leave it alone, it gives me circa 5-10 degrees temperature drop but it is somewhat loud at 3500 RPM. I am ready to push the CPU to 70 degrees just to get rid of the noise.
On the subject of termistors, the GlobalWin in the other machine was termistor controlled but the fan was no fun, awfully loud. So I cut away the termistor and running a Papst NGML on the heatsink instead. Makes things much quieter.
Anyway, I think if getting a bit more noise meant a significant drop in temperature I would probably do it. As things are now, even using a Volcano drops the temperature only to 55 degrees. That's not good enough for the noise it makes. Others report temperatures of 45 degrees under load but I have never seen this sort of figures on AMD CPUs.
On the subject of termistors, the GlobalWin in the other machine was termistor controlled but the fan was no fun, awfully loud. So I cut away the termistor and running a Papst NGML on the heatsink instead. Makes things much quieter.
Anyway, I think if getting a bit more noise meant a significant drop in temperature I would probably do it. As things are now, even using a Volcano drops the temperature only to 55 degrees. That's not good enough for the noise it makes. Others report temperatures of 45 degrees under load but I have never seen this sort of figures on AMD CPUs.