Temps w/ new mobo are nuts!

[tno]

So, I used an Asrock 775Dual for a while and that was nice, though CPU temp was highish (for a Zalman 9500) at 35-38 and my core temps were equally warmish at 44-49, all at idle. Got a DS3 and installed last night. During the install I made a point to provide a goodly amount of AS5 because I noticed that during my last application I hadn't really provided enough, I expected my temps to come down is the point. So color me surprised when during the first boot the bios reads my cpu temp at 40, core temps are 42. It gets odder. I try to trouble shoot with speedfan and the such, to no avail. I figure, well let's burn her in tonight and see what happens. So a full night of orthos and some good things to report, my core temps never crested 50. But my CPU temp was solidly 51 the whole night. Weird.

Here are my theories:

1) A delta of ~10 should exist between the CPU temp and the Core temps, so since the temps move pretty consistently then this problem is one of calibration, the CPU temp was actually 41 at load but some how the calibration is off by ~10. Does that happen? Can it be fixed?

2) I did something wrong when seating the HSF, possibly scraped the paste of the center leaving a hot spot over the CPU temp sensor. I'll reseat the HSF today to check but I'm doubtful this is it.

3) Something is broken. This scares the bejeezus out of me, as I really can't afford broken right now.

Any one have any thoughts or advice? Did I miss something?

tno

 

[ArchAngel777]

Delta from Tcase to Tjunction is 15c for C2D and 10c for C2Q.

 

[tno]

Yeah that's what I thought, so you know what I'm talking about. The delta shouldn't be the 0-2C that I'm looking at right now. And really the core 0 to core 1 delta shouldn't be two either so I think it's time for a quick tear down, wipe down and reapply. Thanks AA.

tno

 

[Emission]

Remember, with AS5, less is more. Try the razor application method, smoothing it out across the CPU surface with a razor or similarly thin object, so as to get a nice even thin coat on the CPU. I had noticeable temperature drops after I did this, almost 10C worth!

Like I said, less is more.

 

[tno]

Yeah I tried the thin stripe down the middle approach last time and struck out, when I pulled the HSF off there was almost no residue it was as if it had been so thin it cooked off, but this time I probably went too nuts. I'll dial it back and yeah try the razor this time. Thanks.

 

[ArchAngel777]

Some motherboards in BIOS report Tcase and some report the Tjunction of the hottest core. So it is like comparing apples with oranges when switching motherboards. My opinion is that the DS3 reports Tjunction just like my IP35-E does as of BIOS version 11+. Not something I would be worried about.

Tcase is just a guestimate anyway, as there is no thermal diode there. It requires one to machine a nice slit into the IHS and place a probe there per the Intel thermal specification guide.

My advice is to simply use CoreTemp and monitor the individual cores.

 

[tno]

AA,

That's what I had planned, I did reapply the arctic silver in a nice single smooth layer and that has done nice things to my core temps (up to 2.5 and idle is 45) but my delta to Tcase is still almost nill, but since I can verify Tjunction in all of the most popular temp softwares I'm going to call this issue settled. I'll monitor and be concerned with my Tjunction temps and generally pay little mind to my TCase temp. Thanks all, thanks AA. Now, to see what a rev 1.0 DS3 and a B2 E6300 can really do! (My rig's so poor!)

tno