Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: Tenchusatsu who wrote (55684)	9/17/2001 7:38:55 PM
From: andreas_wonisch	Respond to of 275872

Tench, Re: I don't know if T-bird can survive a CPU fan failure, however. I would bet that it can't. It can, I've already tried this myself (with my Swiftech MC462 heatsink). The temperature is slowly rising after the fan is deactivated (about 1°C every ten seconds or so) until it locks the PC. Then you can just turn it off without any harm to the CPU. Most motherboards have also a user specified shutdown temperature (however, it doesn't work with my MSI K7T Turbo, don't know why, I tried all settings). It works over the onboard thermal diode which is somewhat inaccurate of course. Andreas

To: Tenchusatsu who wrote (55684)	9/18/2001 12:51:24 PM
From: pgerassi	Respond to of 275872

Dear Tench: The problem is the slow sampling rate of most thermal diode monitoring circuits. If MB manufacturers were more worried about heatsinks falling off (something that the new methods of screwing in heatsinks ala Swiftech 462A, Zalman and Alpha 6045 would not do) they would use AM music speed ADCs, 8 bits at 8KHz sampling rate, more than fast enough to power off CPUs with heatsinks falling off. VIA seems to like those that sample every 5 seconds (0.2Hz). I do not know how fast the other chipsets sample the thermal diode. I once used an IR camera using a thermal diode. It could sample at 2MHz by about 10 bits, so the thermal diode itself on an Athlon MP is not the problem as to sample rate or slew rate (an obvious error in the analysis on Tom's article). The two major methods to check temperature on a thermal diode are to keep a constant current running through the diode checking the changes in voltage to get the temperature and keeping a constant voltage and using the changes in current to get the temperature. In both cases, an operational amp is used to amplify the small changes into a voltage suitable for an ADC to get a value which is converted to temperature in a table lookup. All this means is that the latency of the thermal monitor of the Siemens MB used is too long for a heatsink off condition. It is relatively simple to increase the sample rate to forestall the condition. 1ms would be short enough (1KHz sample rate) tied to something that shuts the CPU off when a certain value is reached. Possibly something that MB manufacturers need to do, but not AMD (I believe that a HS falling off is very rare). Pete