Sign in with
Sign up | Sign in
Your question

Core2 E4300 Temps - Need help after reading the guide!

Last response: in Overclocking
Share
February 20, 2007 7:00:59 AM

Hi everyone, I've read the guide for C2D temps and it was really usefull. However, I'm having some issues with my system, and I would appreciate some feedback from the experts.

Specs are:
Core2 E4300
Scythe Infinity - AS5
Asus P5B Deluxe BIOS 1004
Kingston HyperX DDR800 4-4-4-12
Chieftec Case with pretty good airflow.

Tu begin with, my motherboard reports Core/BIOS temps reversely (Core temps are about 12-13ºC lower than BIOS readings).

So here's my story. I bought the E4300 with a Zalman 9500AT. I managed to overclock the CPU to 3150Mhz (9x350) stable after 8hs of 2xPrime95 with Vcore Bios = 1.35V, which whent to 1.26V under load with speedfan 4.32. The thing is, idle temps were around 30/42ºC and I was getting 62/74ºC after running TAT for 20 minutes with a room temp of 22.5ºC, so I knew this was too high (and also too much delta, ~30ºC). I tried reseating the HSF with no luck. After a couple of days temperatures dropped a couple of ºC, I assume due to AS5 break-in period. These are the results I was getting:

Idle
Load - TAT
Load - Prime95

So I went ahead and bought a Sycthe Infinity which is supposed to be better than the Zalman. I installed it without problems, and used AS5 in the way their instructions advise to (one thin vertical line in the middle of the IHS). Now, I'm getting more or less the same temperatures as before, maybe 2 or 3 ºC less considering AS5 hasn't broke-in yet, but the CPU is not stable anymore after 3 hs of prime95. I also tried reseating the Sycthe with no luck.

Basically, I know my temps are off, and I suspect it might be the CPU being concave, because after I reseated the Sycthe, I could clearly see there was more pressure being applied to the sides of the CPU than the center of it (although contact was all along the IHS, the sides were "clearer" of AS5, and the center was more covered up). Is there anything else I should look into? Any help or tip you can give to try and solve this?

Thanks in advance,

Matias
February 20, 2007 1:13:13 PM

OK, some new info. I've reseated the heatsink (again) and took some photos of how the TIM looked. Here they are:

Heatsink
CPU
Both together

How do those look? Is what I should be getting or should there be signs of more pressure being applied to the CPU? When speedfan reports 70ºC on the CPU (CoreTemp reports around 58ºC for each core) the heatsink (both the fins and heatpipes from the base) are not even warm I would say (at least not compared to the NB/SB heatsink which are hot). Please, any help or advice on this would be appreciated as I'm kind of stuck here and don't know how to move on with the system or what else to try.
a b à CPUs
a b K Overclocking
February 20, 2007 9:20:11 PM

MatiasZ, thank you for providing your results and variables. This is exactly the detailed information I need to analyze temperature problems.

Quote:
Chieftec Case with pretty good airflow

Core/BIOS temps reversed
Core temps are about 12-13ºC lower than BIOS readings


idle temps around 30/42ºC
62/74ºC after running TAT 20 minutes
too much delta, ~30ºC.

I know my temps are off, and I suspect it might be the CPU being concave, because after I reseated the Sycthe, I could clearly see there was more pressure being applied to the sides of the CPU than the center of it (although contact was all along the IHS, the sides were "clearer" of AS5, and the center was more covered up)


Tcase (SpeedFan) = 40c Idle, 69c Load (Offset by +15c)
Tjunction (Core Temp) = 28c Idle & 58c Load (Offset by -15c)

Ambient = 22.5
Chipset = 965
C2D = E4300
CPU cooler = Scythe Infinity - AS5
Frequency = 3.15 Ghz
Load = TAT @ 100%
Motherboard = Asus P5B Deluxe BIOS 1004
Vcore = 1.35

Corrected Temps:

Tcase (SpeedFan) = 25c Idle, 54c Load
Tjunction (Core Temp) = 43c Idle & 73c Load


(1) A common 965 scenario of Tcase / Tjunction temps transposed with 15c offsets.

(2) Definitely into the warm region of the temp scale, considering your Vcore, CPU cooler and case.

(3) Since the Delta is extended and out of spec, this underscores your thoughts regarding uneven IHS / HSF contact. The images suggest heavy pressure in the corners, so you may want to focus on the CPU as a possible RMA item.

I think your analysis is correct. You've done a great job with working the problem to this degree, (no pun intended). Hope this helps.

Comp 8)
Related resources
February 21, 2007 6:44:49 AM

Thank you A LOT for your answers. OK, so at least temps now make sense regarding the MB reading it wrong, thank you for clearing that up since I saw it on the guide but wasn't able to interpret it well, with the example it was really clear.

Now, regarding the CPU do you think a local shop would exchange it based on this "problem"? Is there any kind of warranty provided by Intel where I'm eligible for a RMA because the CPU is not flat enough?

I'm pretty sure I won't be able to exchange it, in a best case scenario they'll have the CPU for a couple of days (week?) testing it and conclude it's stable with the stock HSF, and return it to me as it is. What about lapping? Would that solve the problem? Is there any guide to Core2 lapping?

Are there any other choices?

Thank you for the all the help. If I can try anything else (for example test it with the stock HSF) let me know.

Regards,

Matias
a b à CPUs
a b K Overclocking
February 21, 2007 12:17:37 PM

Quote:
Now, regarding the CPU do you think a local shop would exchange it based on this "problem"? Is there any kind of warranty provided by Intel where I'm eligible for a RMA because the CPU is not flat enough?


I wouldn't presume to advise you in this regard, but I am continually seeking relevent information to include in the Guide, and will look into Intel's specifications and warranty policies concerning flatness.

Quote:
I'm pretty sure I won't be able to exchange it, in a best case scenario they'll have the CPU for a couple of days (week?) testing it and conclude it's stable with the stock HSF, and return it to me as it is. What about lapping? Would that solve the problem? Is there any guide to Core2 lapping?


Lapping may be a viable alternative. Check the Cooler and Heatsinks section of the Forum for information on lapping CPU's with Integrated Heat Spreaders.

Hope this helps. Let us know how it works out.

Comp 8)
February 21, 2007 5:25:25 PM

OK, I've lapped the CPU and HSF. The CPU showed to be pretty concave as I was lapping it (picture), so I expected a big temp difference. Sadly, after putting it on again, I see no more than maybe 3-4ºC, which wouldn't be a bad result in a regular situation, but considering I was expecting a 10-15ºC difference this is not so good. I've also tried the stock cooler, and what shocked me the most is that there seems to be little difference (~10ºC) between stock vs Infinity under TAT. Right now, I don't know what to think/do. Maybe the MB is reading everything OFF, maybe I just have a "defective" CPU regarding temps (is that even possible?), or may be I'm missing something subtle somewhere, but I don't think so.

Any more ideas?
!