View Full Version : Tbird 1.2Ghz, Cu HSF, silver paste = 60C ???
Yeah, I got my "better" cooling today and I'm getting 60C. Rather dissapointing taken that I bought an expencive Cu heatsink (unknown) and Titans silver heat paste. The heat paste claims to have C/W of 0.06 and the heatsink is a ~5mm Cu base with 5-6cm folded Cu plating soldered to the base. On top is a, what I would describe as, medium sized (5000rpm) fan.
The 60C was measured from the CPU backside with the MB thermistor. On my MSI K7T Turbo the thermistor comes into perfect contact with the CPU back (even added some heat goo). Generally they seem to give 5C more than actual heat, and even if it gave 10C more id still be at 50C. I was expecting somethin like 45C.
The CPU "backside" could be a sealed cavity, and thus collecting heat. So I might be best off with a actual external heat sensor. But still, 60C from backside seems alot.
So whats wrong?
And I fogot to mention, I also moded my PSU for larger fan, and the air stream from back to side. Now the PSU sucks air directly above the CPU (I have a medium tower).
My first thought is unless your ambient temps are real high, the HS is not flush with the CPU die.
I checked it. It is straight and the ambient was some 26C (on MB).
Mayby I should "finger" test the sink if it gets warm... Also Sisoft Sandra reports the temp all wrong. Says its 4C for the CPU. Mayby the thermistor is busted?
The Doors
08-16-01, 05:14 AM
My suggestion is to Monitor your temps with MBM, but something is wrong, 4 degree C cannot be real.
My higher temps with a TB1200@1400Mhz 1.775v and a cheaper Fop32 are around 57-58C @full load with Folding@home 24/7, but with an ambient temp of 30-35C.
I hope it helps you :-)
Maybe you used to much of that thermal paste, you should only have a very thin layer. This is about the only thing I can think of that hasn't been mentioned yet. What cooler do you have? It doesn't sounds to me as a cheapo generic, but I would like to see some tests before I can give you any advice.
The Doors
08-16-01, 08:20 AM
Or the Hsf wrongly seated, take a look to this picture (posted by Savi)
When I fiddled with the HSF, I could feel it "tilt" (before tightening) on top of the rubber pads of the CPU freely from all corners. So I dont think its an issue with the HSF connection.
I also noticed that the MSI software reports Vcore to be 1.85V. Then I set it from the BIOS to 1.7V, and the software showed 1.8V. Which is wrong? Should I set it to 1.65V? This ways the software would show 1.75V.
Mr.Lansing
08-20-01, 05:43 AM
well this mite be ur problem
i first used MBM5 to check my temp and it was so freaking high 55C
lata i used Sandra 2001 to check it and it appeared to be 42C
so try checking ur temp with Sandra 2001
u can download it from download.com
The BIOS adjuster for Vcore is not accurate at all. At default the CPU asks for 1.75V and gets 1.85V. There is a 0.1V difference. The voltage probe reads clear 1.85V from both, BIOS and the MSI win app. I then forced the voltage to 1.7V and the reading went to 1.8V. Then I lowered to 1.65V when the reading went to 1.75V. Now my CPU tems are some 52C under load and some 45C idle.
So my MSI K7T Turbo board does not set the Vcore accurately. This MB particularly has +0.1V inaccuracy. So howbout yer MBs? Id like to hear if other MBs have similar problems.
The Overclocker
08-21-01, 08:26 AM
i hope you are happy now
The Doors
08-21-01, 08:56 AM
It's the same with my Abit KT7A-Raid, more voltage then the value setted into Bios :D
One thing that needs to be mentioned is that Titan silver heat paste has only 10% silver. Unfortunately this means it will perfrom very close to the standard white zinc oxide based compounds.
Originally posted by Colin
One thing that needs to be mentioned is that Titan silver heat paste has only 10% silver. Unfortunately this means it will perfrom very close to the standard white zinc oxide based compounds.
Yeah, I kinda suspected that it aint gonna be that good, but here where I live, I should be gracious that the local shop sported this Cu cooler and the paste. There is one mailorder shop (here in Finland) selling cooling goods, but then Id had to spend $10-$15 more on postal charges...
I'm idling along at 60C as well according to my Asus a7a probe. Tried watercooling, cheap HSF, now my SK-6 (all copper one) with an 80mm fan on it. Still getting 60C at idle. I have the same problem with my voltage. It's set to 1.75 in the bios (which is the lowest it will go) but the monitoring software is reading around 1.824 all the time. I believe that has something to do with the higher temps. But also I believe the probe is just faulty or just not very accurate. I have no stability problems at all. I never get errors or system crash's. So I don't really even worry about it anymore.
The Doors
08-22-01, 04:05 PM
Probably the only way to go is an external thermal sensor in between the Cpu & Hsf to check the true temp.
Originally posted by The Doors
Probably the only way to go is an external thermal sensor in between the Cpu & Hsf to check the true temp.
I've been meaning to do that, Just havent gotten around to ordering one yet.
Hmm, the new BIOS for my MSI K7T Turbo seems to have a fix to the overgrown Vcore as it can "Adjust Vcore Voltage Sensitivity on Hardware Monitor". I gues this means a read back to see if the Vcore went right...
Gotta DL it today...
OK, I checked the BIOS update. It's a hoax! After I installed it, the "hardware monitor" tab in BIOS showed the Vcore to be what I had set it to be. But after I got to windoze, the MSI win app showed that the Vcore was once again 1.85V and the heat was back up at 60C.
So the BIOS update seemed to just "fix" the BIOS hardware monitor tab to show incorrect voltage. The voltage U had set for it. Instead of actually calibrating the Vcore accordingly to the probe.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.