Warped motherboard = bad temps???

cornelious0_0

[H]F Junkie
Joined
Apr 6, 2003
Messages
12,783
This isn't a ridiculously serious issue, nor is it a motherboard defect. What I'm referring to is actually that after the stock Intel heatsink is installed, there is a very heavy bend in the motherboard under the processor. At first I didn't think anything of it, but I have now re-seated the cooler with better paste and made sure that all 4 push-pins were securely pressed down. At the moment my E4300 is clocked at 3.2GHz and set in BIOS to 1.350v (runs at 1.312v load) and hits load temperatures of 70C+. I'm using TAT to monitor temps, so there's no way it's being reported incorrectly, and I know that my temps should be a fair bit lower, compared to any HSF reviews I've seen where an X6800 or other Conroe is used.

I'm planning on getting a Thermalright Ultra-120 Extreme and either lapping or removing the IHS to really "fix" my temps...but until then I'd really like to know what's going on.

I know that the Ultra-120 comes with a back plate which should eliminate any flexing, but do you guys think it's actually possible that the board bending enough under the CPU could effect HSF contact?

If not, what do you think? I've tried everything I could think of and still the temps remain far too high for my current clock speed/voltage as far as I've seen.
 

Vengance_01

Supreme [H]ardness
Joined
Dec 23, 2001
Messages
6,308
Those 4300's had an issue with reading there temps. Your real temp may be 15C lower than being reported.
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
As the previous poster said, your temperatures might not be correct and you can't trust TAT just because it was programmed by Intel. It was designed for heat soaking the laptop Core processors and came out long before the E4300 ever did and hasn't been updated.

The problem is that the E4300 came in two different types. One with a Tjunction or TjMax temperature of 85C and the other one has a TjMax of 100C. Intel has no documented way to tell the difference via software so ALL temperature monitoring programs are left guessing.

If your E4300 has an actual TjMax of 85C but the software guesses at 100C then the reported temperatures displayed to you will be exactly 15C higher than what they actually are.

I found that processors with a date code of L643 or L644 were 85C processors. The early ones with a date code of L640 had a TjMax of 100C. Have a look at your Intel box and see if you can get a date code off of it.

Use CoreTemp 0.94 for the 85C processors and CoreTemp 0.95 for the 100C processors. You can also use SpeedFan 4.32 which I think uses 85C by default but lets you set an offset of +/- 15C to correct for this lack of info from Intel.

A temp of 70C while running TAT with the Intel cooler may not be out of whack. It depends on your room temperature and your case air flow. The Intel cooler is pretty inadequate for running TAT. Try Orthos instead.
 

cornelious0_0

[H]F Junkie
Joined
Apr 6, 2003
Messages
12,783
Well...the week is Q702. I'm really dying to know if I should "trust" .95 or not...sure it reads the same as TAT, but it's right that it was released before the E4300...and if only certain E4300's actually had TjMax of 100C and mine's not one of them then that would make sense.

Any more input? Should I go by 0.94 based on the week of the CPU?

Sorry to sound so newbish, but this whole thing is really weird and it's my first C2D. :)
 

abcdelight

[H]ard|Gawd
Joined
Nov 23, 2006
Messages
1,275
I had a mobo with a bend in it once and suspected the heat sink but it turned out a standoff wasn't seated properly. Took me 3 check overs to notice it. Those Lian Li standoffs can be finicky.
 

cornelious0_0

[H]F Junkie
Joined
Apr 6, 2003
Messages
12,783
No problems here related to the standoff's at all, and I'm not so much worried about the bend in the board right now as I am trying to find out whether I should be relying on Core Temp 0.94 or 0.95. The newer 0.95 reports the same as TAT, but it also shows that E4300's typically run MUCH hotter then E6xxx CPU's...which I find very hard to believe. Part of me just wanted to believe that my CPU isn't actually running that hot, but it doesn't make sense how I could idle just shy of 50C at stock voltage and a mild OC on my E4300 compared to what I've seen from E6xxx results.

One more way I can test it I suppose...I have the shutdown temperature set to 85C in my BIOS and when I bumped up my voltage the system would mysteriously shut down after about 10 minutes without warning. I'll try and reproduce that issue, but I guess the shutdown temperature is based on the BIOS readings, which doesn't accurately report information from the digital sensors embedded within the cores of E4xxx chips.

Argh...who would have ever thought it could be so difficult to just find out what my REAL cpu temps is?!?!?
 
Joined
May 24, 2007
Messages
739
No problems here related to the standoff's at all, and I'm not so much worried about the bend in the board right now as I am trying to find out whether I should be relying on Core Temp 0.94 or 0.95. The newer 0.95 reports the same as TAT, but it also shows that E4300's typically run MUCH hotter then E6xxx CPU's...which I find very hard to believe. Part of me just wanted to believe that my CPU isn't actually running that hot, but it doesn't make sense how I could idle just shy of 50C at stock voltage and a mild OC on my E4300 compared to what I've seen from E6xxx results.

One more way I can test it I suppose...I have the shutdown temperature set to 85C in my BIOS and when I bumped up my voltage the system would mysteriously shut down after about 10 minutes without warning. I'll try and reproduce that issue, but I guess the shutdown temperature is based on the BIOS readings, which doesn't accurately report information from the digital sensors embedded within the cores of E4xxx chips.

Argh...who would have ever thought it could be so difficult to just find out what my REAL cpu temps is?!?!?
Touch the side of the heatsink, if it is over bearably hot then get a new cooler.
 

unclewebb

[H]ard|Gawd
Joined
Jun 21, 2006
Messages
1,807
cornelious0_0: Here's the test I came up with to try and determine which version of CoreTemp you should believe.

Go into the bios and lock your multiplier at 6 and drop your FSB back to the default of 200 MHz. Next drop your core voltage as low as you can get it. My P5B lets me set it to 1.150 which equals approximately 1.104 volts at idle. My computer won't boot up if I try to set it any lower than that.

Dropping the MHz and especially the voltage will reduce the amount of heat your cpu is putting out down to next to nothing. Open the case and set your cpu fan to its maximum speed. The point of all this is to run your cpu as cool as possible.

When I do this with my E6400 my core temps start to approach my room temperature and actually are equal to my room temperature. Depending on how low you can get your core voltage and how well your cpu heatsink is working you should be able to get your temps down very close to room temperature.

If a program shows that you are running at more than 15C above ambient temperature then it is likely wrong. If a program shows a temperature less than ambient then it is also wrong. This is a very simple test and you don't have to bring your cpu up to crazy temps like 85C or 100C. If you can prove a program is wrong then you'll know what program is right.

Most people who have used this test are pretty convinced afterwards what their Tjunction of TjMax really is.

Edit: I've done a lot of research on this topic so if anything in this post works for you then give it a bump before it gets lost.
http://www.hardforum.com/showthread.php?t=1174781

This was the first time I learned about 85C E4300 processors. It gets interesting around page 4 if you don't have time to read it all.
 

cornelious0_0

[H]F Junkie
Joined
Apr 6, 2003
Messages
12,783
Thanks for the reply man, I really appreciate all the time that you have put into researching this.

I actually just finished going through all 6 pages and I'm quite intrigued. I haven't had time or a chance to try the low clock/voltage testing, but I did notice it said that later then Q644 were 85C chips...and mine is Q702.

I'll try the testing later tonight or tomorrow and let you know...hopefully things work out and it turns out that I can rely on Core Temp 0.94 and relax a bit. :)
 

5aboy

Weaksauce
Joined
Sep 17, 2005
Messages
92
i heard that purchasing the thermalright s775 backplate is known to reduce cpu temps, and so is a vital part of a new c2d setup. i know that when i finally bite the bullet i shant be doing it without it.
 
Top