[Faded_Mantis]

This is regarding an overclocked E2180, M0 @ 2.67ghz 1.3125v in bios (1.336v in CPU-Z)

What I was doing, was using core temp 0.99.5 and ensuring my temperatures don't go over 73C (this CPU's max temp I believe). However I've been doing some reading on tjmax recently and I have some context specific questions about my situation.

coretemp gives me the same readings as my bios, ie when I set my bios to idle at 40C, the coretemp reading is also at 40C +- ~1-3C. tjmax is set to 85C. When I run prime95 the temp can get as high as 69C, with an AC freezer 7 pro.

I've seen that the Tjmax for the M0 stepping is 100C, so I believe that means I need to add +15C in the compensation settings in coretemp. Does this mean that even though the sensor is telling the computer it is idling at 40C, it is actually at 55C? and when Prime95 is running the CPU is getting up to 84C? 11C above the max for the CPU.

I'm hoping that I'm just confused about the way tjmax works, and I appreciate any advice on this.

[EDIT]
Just incase someone mentions the high temp with a non-stock cooler. I believe it is attached properly. It doesn't rock or wobble, but when I push the top of the cooler it does move slightly.

 

[guyladouche]

Is there a source from Intel for the TJmax for the M0 E1/2xxx being 100C? The only source I found for the TJmax for the 65nm dual core E series is for the celerons, which is 85C--info on the E2xxx seems to be lacking... I assume all M0's should have the same TJmax give or take--which would be 85C as the only value Intel has quoted for an M0 stepping. I've read on other forums people say that for the E2xxx the TJmax is 100C, but haven't seen anything from Intel stating so.

If it is the case that the actual TJmax is 100C (I'm still skeptical), then yes, you'd be adding 15C to the temps read by coretemp/etc. if the TJmax is reported as 85C--this makes me skeptical of the TJmax. But just realize at idle temperatures, TJmax is a very inaccurate value, so an offset based on TJmax for idle temps will likely not be giving accurate temperatures. For load values, as you approach the TJmax temperature, the reported temperatures become more accurate--though who knows just how inaccurate it might be.

I'm pretty sure, regardless of TJmax issues/questions, your CPU is well within thermal specifications. But I'm curious--what's your case airflow like? 40C with a TJmax of 85 (or conversely 55C for TJmax of 100C) does seem to be a little warm. With 20C ambients I get about 25C core temp idle--don't recall at other temperatures. This is with stock vcore and speedstep enabled--I'm about 5C higher with 1.45v vcore.

 

[justarealguy]

Quote:

Originally Posted by Faded_Mantis [EDIT] Just incase someone mentions the high temp with a non-stock cooler. I believe it is attached properly. It doesn't rock or wobble, but when I push the top of the cooler it does move slightly.

This little bit here causes a bit of concern. Put the chip under load, monitor temps, then apply pressure to the heatsink while keeping an eye on the temps.

If it decreases then you'll know where the problem is. The AC7 is a great cooler for a dual-core.

 

[Faded_Mantis]

I popped back to point out that I'm aware CPU temp and core temp are different values, and that the 73C max is the cpu temp (I'm naturally cautious, so I try to apply the core temp to this).

Guyladouche (mXc?)

I re checked where I saw the M0 tjmax 100C, and I miss read what was being written, they were talking about a few different CPU's and I skipped a bit to another one withour realising *Embarassed*. Also it was just a forum post that I got it from, not a credible source.

My case airflow isn't too good (crap cable management), I've added a low rpm 120mm intake fan which has improved things, it has a 90mm exhaust fan. The 40C I quoted at idle is the setting I've told bios to keep it to, to make my fan quieter when I'm just browing the internet etc. I should have made this clear

justarealguy
That is a good tip, I don't have time to do it now (about to go to bed) but I'll do it tomorrow and post the results here. Like you, I was surprised at how high it was. Room temp is around 21C.

Also to clarify, by "push the top of the cooler" I meant with sideways pressure

 

[justarealguy]

Quote:

Originally Posted by Faded_Mantis Also to clarify, by "push the top of the cooler" I meant with sideways pressure

A little side "slide" is fine, but it's worth a shot anyways.

 

[Faded_Mantis]

Ok, I tried pushing down on it while running a stress test. It didn't seem to make any difference to the temperature, infact at one point I pushed it on slightly the wrong angle and the temps quickly went up.

I tried to do it with the case still upright, to simulate how it would normally sit, and because I can't be bothered moving my desk to get it on its side while still plugged in.

When gaming the CPU stays in the 50-59C range.

 

[Faded_Mantis]

I've decided that I'm going to get some MX-2 paste, and re seat the heat sink. I think I pushed it off balance when I was pushing on the top earlier, now it can get up to the high 50's just loading the web browser.

 

[justarealguy]

Quote:

Originally Posted by Faded_Mantis I've decided that I'm going to get some MX-2 paste, and re seat the heat sink. I think I pushed it off balance when I was pushing on the top earlier, now it can get up to the high 50's just loading the web browser.

Hmm...Well, I don't know what to tell you. If pressure is fine then the other parts of your cases' airflow may be at fault. A picture perhaps?

 

[Faded_Mantis]

The origional problem was my misunderstanding of tjmax, but this thread opened up another problem, which was confirmed by people here, that my temperatures were strangly high.

They seem to be fixed now, with load temperatures generally being about 10C lower at the same point in the prime95 stress test, ie no higher than 62C after 30 mins (short I know, but I was just comparing to the old situation).

Now I can't be sure what the solution was. I used cellotape (lol) to stick all the lose cables to the roof of the case, and some are tapped to the CD drive/HDD racks to keep them out of the way. This would have made it easier for air to travel from the 120mm low rpm intake fan into the rest of the case.

I also removed the heatsink, and cleaned the old paste off it and the CPU. Admitedly I used meths, but I didn't have any isopropyl alcohol around. I then applied the MX-2 paste. I noticed that when I tried to smooth it with my credit card, when stroking in one direction it touched the sides of the heat spreader, but not the middle, so I put a dab of extra thermal paste there (still keeping it quite thin). I also paid extra attention to seating the heatsink this time, last time I just pushed them in and turned to tighten them, this time I use a torch to look at where the black part of the clip was in relation to the ridges on the white part of the clip they are meant to go under. On 1, maybe 2, they black part of the clip hadn't properly gone over the white part, and needed extra force to get it fully secure.

So the temperature problem was solved with either:

1) Better case airflow
2) Better thermal paste
3) Better seating of the heat sink
4) Combination of the above.

Thanks for the help.