TSK's Crazy Theory on TJmax and the DTS - Page 4 - Overclock.net

unclewebb · 03-11-08

In terms of construction, I don't think there is a hell of a lot of difference between the mobile and the desktop CPUs.

**grunion** · 03-11-08

Quote:

Originally Posted by unclewebb

This is what Intel says about the Mobile Core2Duo chips:

Section 5.1.3
http://download.intel.com/design/mob...s/31674505.pdf

"Regardless of enabling the automatic or on-demand modes, in the event of a catastrophic cooling failure, the processor will automatically shut down when the silicon has reached a temperature of approximately 125°C."

I think the last screen shot must have been very close to thermal shut down.

For the Desktop Core processors there is nothing written in stone in the Intel docs. I might have to go buy one of these cheapy processors tomorrow to do some proper testing with my IR thermometer.

Question about your IR, is it mounted in the same fixed position for all of your tests?

TheSubtleKnife · 03-11-08

Quote:

In terms of construction, I don't think there is a hell of a lot of difference between the mobile and the desktop CPUs.

higher quality silicon, i remember the old mobile bartons...OC'd like beasts compared to their desktop counterparts, and the mobile ones probably have THERMTRIP# set higher

GigaByte · 03-11-08

Quote:

Originally Posted by TheSubtleKnife

I found this info from intel

which means that is should shut down at 93.2C (73.2+20) actual temps. Take the average of the last readings of 111C and 123C that i got before it shut down, and you get 118C. subtract 118 by 21.8(Realtemp Tjunc 95 - Tcasemax73.2)= 96.2C, which is pretty darn close to the temp that its supposed to shut down at.

Where did you find that? Iam not seeing it in the link unclewebb posted, although there is some interesting stuff in there.

unclewebb · 03-11-08

Quote:

Originally Posted by grunion

Question about your IR, is it mounted in the same fixed position for all of your tests?

The IR thermometer was not fixed which lets you position it to get a maximum temperature and not just the Intel TCase temperature which is measured at the geometric center of the IHS.

Hopefully I can find a cheap E2160 for some testing tomorrow. Be back then.

**grunion** · 03-11-08

Quote:

Originally Posted by unclewebb

The IR thermometer was not fixed which lets you position it to get a maximum temperature and not just the Intel TCase temperature which is measured at the geometric center of the IHS.

Understood
Just curious because with my IR gun if I vary the distance to the measured surface, temps are thrown off as much as +/-10°c

BTW Welcome to OCN

TheSubtleKnife · 03-11-08

this is where i got mine
look for thermtrip in Table 25
http://download.intel.com/design/pro...hts/316981.pdf

did you guys see my new hypothesis?

Quote:

I can say that the TJmax for my E2180 is NOT 85C (Coretemp) or 95C (realtemp), but closer to 75C. Note that the DTS is calibrated for higher temperatures, so this TJmax is not accurate at idle temps

opt33 · 03-11-08

Quote:

Originally Posted by TheSubtleKnife

The tests were done on an E2180, M0 stepping.

For the temperatures, I rigged a thermometer with a probe so the probe is in contact with fins of the stock intel heatsink. No fan on the heatsink, so we can better assume that the core temperatures are close to the temperatures of the heatsink.

While the gradient across a couple millimeters of core, thermal bond adhesive, and IHS, that intel has perfected over years, is very likely less than 1C, by leaving the heatsink in place and measuring at the coolest point...fins, you have set up a passive cooling system with unknown gradient through a large block (even without a fan). If using measuring probe, better to measure at center of back of IHS, with ~known offset. Might be ok to leave heatsink on if reading calibrated cpu temp from onboard diode.

Quote:

Just fired up orthos, and you can see that the difference between cores is close to 12C, which is not too surprising, because the core heats up faster than the heatsink.

Using orthos increases the gradient. Better to remove the heatsink, and let idle increase temps slowly.

Quote:

which means that is should shut down at 93.2C (73.2+20) actual temps. Take the average of the last readings of 111C and 123C that i got before it shut down, and you get 118C. subtract 118 by 21.8(Realtemp Tjunc - Tcasemax)= 96.2C, which is pretty darn close to the temp that its supposed to shut down at.

With the above as evidence, I can say that the TJmax for my E2180 is NOT 85C (Coretemp) or 95C (realtemp), but closer to 75C. Note that the DTS is calibrated for higher temperatures, so this TJmax is not accurate at idle temps

Though even with the error you incurred, from unknown gradient to fins of heatsink, and shutdown temp for Exxx series is also quoted as 20-25C hence another possible 5c, and using orthos load possibly inducing gradient (though granted this effect minimized if did reach steady state), it is nevertheless interesting (and you got some serious kohonies testing shutdown temp). Just wish you had done this with heatsink off, orthos off, and probe to center cpu, to lessen the "unknowns". But I think you may be right, in that tjmax may be in the lower range.

Not suggesting you do this, but if others try, better to remove heat sink, and definitely no orthos then, just record tcase temp at delta to Tjmax of 40, 35, 30 (less risk and accurate) with temp probe on center cpu directly. Then simply add cpu temp to 40, etc and get 3 measurement for tjmax. Would eliminate error of unknown gradient of temp on fins, tim2, and 5c variance in shutdown temp.

Also, Exxx series has tjunction diode, you could calibrate that at ambient, and use that even with heatsink on as absolute temp test.

TheSubtleKnife · 03-11-08

yeah, but i just thought that it was interesting that it was 12C even after 2 hours of orthos, and the temps maxed out.

**Retrospekt** · 03-11-08

Quote:

Originally Posted by unclewebb

The IR thermometer was not fixed which lets you position it to get a maximum temperature and not just the Intel TCase temperature which is measured at the geometric center of the IHS.

Hopefully I can find a cheap E2160 for some testing tomorrow. Be back then.

In case some of you don't know, unclewebb is the developer of RealTemp. I posted a link to this thread on XS and he seems to have found his way here.

Maybe he can straighten some things out for us once and for all.

		Overclock.net - Overclocking.net > Intel > Intel CPUs
TSK's Crazy Theory on TJmax and the DTS

03-11-08	#31 (permalink)
unclewebb New to Overclock.net Join Date: Mar 2008 Posts: 29 Rep: 8 Unique Rep: 8 Trader Rating: 0	In terms of construction, I don't think there is a hell of a lot of difference between the mobile and the desktop CPUs.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)