Duron 700@700 at 130F (55C) - too hot or not?

[deadlock]

Hi guys.

Someone said that for a Duron 700 running without overclocking of any sort, 130 degrees (55 Celsius) is very hot. Is this so? Does this mean that I can forget overclocking with my current cooling?

My temprature never goes above 130 degrees, but under intensive use it tops out to about 125 degrees. My Vcore voltage is 1.63v, and regular voltage is 0.3v. I have an OEM Duron 700 on a MSI K7T Pro 2 with a standard issue HSF, with the fan running at about 5600 rpm. No thermal grease though, here in Malaysia it's pretty hard to find so most people don't bother.

Do I have a problem, and if so, how do I fix it?

 

[Marsupial]

If the ambient air temp is high and/or you don't have good cooling (circulation) in your case, I wouldn't be surprised. 55 is a tad on the hot side, esp. for a 700 Duron (I hope that isn't the idle temp though).
Are you using any sort of thermal interface? Keep in mind that AMD doesn't recommend thermal grease of any kind, so a stock thermal pad (esp. on a non-overclocked OEM part), should perform adequately. IE. it should give lower temp. than what you're getting.

To fix this problem:
-Lower ambient temp via AC (I do it all the time)
-Check if it's a case circ. problem (run open case)
-Try a good hsf (you should have no problems finding recommendations)

 

[Mikewarrior2]

What is your room temp, and ambient case temp?

Chances are, since you're using a MB without temp compensation, in addition to not using grease, that your CHIP is runnign in the 60Cs. I would try to find some heatsink grease, even if it is generic grease.


Mike

 

[DaddyG]

Hmmm, the "Standard Issue Fan", makes me wonder. Is it AMD approved ??
If it is it should have come with a thermal pad. If not, look for a AMD approved hasf. (Globalwin, Taisol, Coolermaster etc.)

 

[ericb]

How are you measuring the temp?

 

[Mikewarrior2]

The k7tPro2 Uses an in-socket thermistor, one that is likely a bit more inaccurate than the kt7s for referencing.


Mike

 

[buildingacomputer]

I read the thread you posted but it does not say much about the way temp is measured. Would you explain how this is done in Socket A systems?

 

[Mikewarrior2]

Socket A mb's use a thermistor inside the socket to "approximate" cpu temp. Some mb's have a sensor that contacts the backside of the cpu, while other's don't come close.

The problem is, measuring from a secondary heat pathway is both highly inaccurate, and is only a percentage of the true-cpu core temp change.

Hence why you get "compressed" results, a la the anandtech review. The c-orb only reports to be higher than the alpha pal6035 by 2C.

If you actually compare the the two heatsinks based on C/W, the difference is at least 10C, or more.

Accuracy is also another issue, with un-compensated MB readings giving temps that are too-low. This is why the newer Asus and Abit bios' compensate the temperature +10C or so. It is an attempt to "approximate" cpu core temp. And for hte most part, the updated bios' can be fairly close in many instances. There are times they are still significantly wrong, though.

Anandtech Heatsink Comparison Discussion

and
Socket A temps, and why they're inaccurate

Some extra reading on this issue.

Mike