zman

Hi unclewebb! I'm new here, but as an owner of an E2180 I have been following the temperature debacle for quite a while. Up till today I used the Tjmax 100C when looking at the DTS readings of my processor. However, I began wondering why the temperature of my CPU cooler base (measured with a thermocouple) was 20-30 degrees lower than the one calculated from DTS using the 100C figure. This did not seem to be in line with my previous experiences on the AMD64 platform: If the CPU is "70C" the cooler should be HOT too, not luke-warm. (Coolers tested: Xigmated HDT-S1283, Zalman CNPS7700Cu and Intel stock)

Finally I decided to do a similar IR-gun test as you did. I clocked my processor at 1.2GHz (6x200) using Vcore of 1.164V (which was 1.18V as reported by SpeedFan). Then I loaded up Windows XP with the cooler still in place, started CoreTemp (configured to use 85C Tjmax) and put the computer in stand by state. At this point I removed the cooler, cleaned the CPU and put some masking tape on it.

I held the IR gun on the processor while waking the computer from standby with shaking hands . The temperature started to rise surprisingly slowly, and when the Windows desktop with the CoreTemp software appeared I was at 31C as per CoreTemp and about 32C on the IR gun.

When idling on the desktop, the temperature kept rising and reached about 65C. For the whole range from 31 to 65C the IR gun was showing temps that were 1 - 1.5C lower than the ones displayed by CoreTemp.

I then started to stress the CPU a bit (starting some programs etc.) until I reached 89C on the IR gun and noticed that the CoreTemp reading stopped at 85C. All the way to the 85C point the 1-1.5C difference between the IR gun reading and CoreTemp reading remained the same. At this point I shut down the mains power, preventing possible CPU damage.

Unfortunately I did not have an opportunity to take pictures of the setup, so you do not have to believe me. Just wanted to tell you that people are repeating your experiments with similar results.

At least for me this ended the whole Tjmax debate since now I know that with the Tjmax set at 85C the DTS temp = Tcase temp + 1-1.5C. And since Intel specifies the thermal allowances using exactly the Tcase parameter I really couldn't care less about what the _actual_ core temperatures are. They can be 200C for all I care. As long as I keep the DTS temps within a reasonable range and take care not to overshoot the Tcmax of 73.xC, I will be within safe limits.

BTW. There is simply NO way I'm going to believe that there exists a constant delta (eg. 10C) between the core temps and the Tcase temps. Simple physics dictate that the delta must grow as the core temperatures rise. Otherwise the thermal conductivity of the materials between the cores and the top of the IHS should be variable, which makes no sense.

Sorry for the rant-style post. It is by no means directed at you. This temp uncertainty is just getting on my nerves.

TheSubtleKnife

Bump for relevance/wondering if there are any updates

__________________