Core Temp WRONG!?

[commandercup]

Well I've been a firm believer of Core Temp for quite some time now... but after installing my Xigmatek cooler... the temperatures seem to be reporting wrong?

I'm sure its not a problem with the seating since I've spent over 2 hours securing it... and it wont move and when pulled causes the motherboard to pull with it, no its definately on tight

Heres a screenshot of the temperatures.

As you can see, Everest and SpeedFan report relatively similar temperatures ~1-2 C while Core Temp reports the overall CPU temperature as the two cores... whats going on?

I'm running an E2200 which has been lapped...

 

[ZenEffect]

Well I've been a firm believer of Core Temp for quite some time now... but after installing my Xigmatek cooler... the temperatures seem to be reporting wrong?

I'm sure its not a problem with the seating since I've spent over 2 hours securing it... and it wont move and when pulled causes the motherboard to pull with it, no its definately on tight

Heres a screenshot of the temperatures.



As you can see, Everest and SpeedFan report relatively similar temperatures ~1-2 C while Core Temp reports the overall CPU temperature as the two cores... whats going on?

I'm running an E2200 which has been lapped...

is there a back plate with it? it could be the tension of the mounting system is causing the motherboard to bow slightly providing inadequate contact. if that is not the case then i dont know what to tell you other than mabye the base of the hsf may be bowed which on a flat lapped cpu will hurt performance.

 

[commandercup]

nope, its a push-pin system, but there is no visible warping

also, its definately not bowed because its not a normal heatsink base, it just has exposed aluminum/copper (3 heatpipes are directly exposed on the base) unlike the normal flat copper base

its the HDT-1283... so it should have amazing temperatures, which is why I'm inclined to beleive the lower of the two groups (30's rather than 40's)

 

[sneekypeet]

other than speed fan and everest read a sensor on the mobo...where IIRC coretemp actuall reads from the core itself, not some sensor on the mobo!

Edit...my E2200 runs hot too!...I figured its because its a glorified Pentium CPU....or as just explained by Jr to me in another thread....these are E4XXX CPU's that didnt make the cut!

 

[KBD]

I heard that Coretemp reads wrong temps on nvidia chipsets. It does on mine anyway. Everest is also having the same problem on my rig. I found temp probes from my fan controller to be most reliable.

 

[commandercup]

well my point is that the temperatures were better with the intel heatsink... after I switched, the temperatures went from ~35 to ~45...

and I'm 100% sure its not a mounting problem

the E2200 I have is fine, its a great CPU and went to 3.7 on liquid

2.8/9 on stock volts

 

[sneekypeet]

With my DFI boards Ive always trusted Smart Guardian as being really close....does EVGA have such a utility on the Driver CD?

was this set up on another mobo with the stock cooler...or water?

 

[commandercup]

nope, I just upgraded the cooler, it was reporting temperatures fine on the stock HSF

and yeah, I did get 3.7 on a DFI motherboard, but thats irrelevant

btw, BIOs reported the temperature at ~35

 

[sneekypeet]

If all you did was switch the coolers and the Xig is reading higher temps....read previously with coretemp I assume?

I have to say there is something wrong in the application or seating of the new HS!

 

[erocker]

What method did you use to apply thermal paste to the Xigmatek?

After much testing the method in the pics below give the best results. First fill in the gaps with thermal paste, then two lines across the aluminum part.

 

[Jarman]

Alot of the 45nm E series have faulty thermal sensors. Try Real Temp instead of core temp for them.

 

[commandercup]

real temp agrees with everest and speed fan... so should I just go with that?

and erocker, I just spread it out on the base... but I'll reapply it like that when I take it off in the future

 

[theonetruewill]

My chip is in the low 40's for idle. Yeah it's high, but it's a £35 chip that clocks to a £100 one so I don't give a twiddle either way. As long as it doesn't exceed 75*C when under load I'd say to hell with it. I wouldn't normally but these things are throwaways for the price. I know cash is cash but c'mon they are NOT expensive. I also have different tempe for real and coretemp (differ by about 12*C at load, but I have the L2 stepping and I go for coretemp which shows the higher just to be safe.)

 

[commandercup]

well but my point is that, the temperatures were reported fine before I changed the cooler... so it must be improper mounting?

 

[sneekypeet]

that was where I was leading earlier...if all other things havent changed...what its the logical issue...the cooler is!

 

[theonetruewill]

What were the original temps? And are the load temps also very different?

 

[sneekypeet]

What were the original temps? And are the load temps also very different?

"well my point is that the temperatures were better with the intel heatsink... after I switched, the temperatures went from ~35 to ~45... "

10*C is even too much for me to say AS5 hasnt set in....most likely used something better anyways!

 

[theonetruewill]

"well my point is that the temperatures were better with the intel heatsink... after I switched, the temperatures went from ~35 to ~45... "

10*C is even too much for me to say AS5 hasnt set in....most likely used something better anyways!

Apologies I simply did not see that - I rescanned the thread and still didn't so I used Ctrl-F and hey presto.

Try replacing the cooler with the Intel one - check the temps. Then reseat the Xigma one again and check the temps again. I know it's tedious but give it a go.

 

[commandercup]

eh, I'll try it later... lol, it took me so long to get the xigmatek on...

 

[theonetruewill]

eh, I'll try it later... lol, it took me so long to get the xigmatek on...

I can imagine you're pretty damn bored of putting that cooler on right now.. hard luck mate.

 

[echo75]

Actually commandercup, i have been looking at your screenshot carefully and it seems coretemp is the only one reading the temps that are likely correct and both everst and speedfan are likely wrong , i will explain why.

Both speedfan and Everest are reporting that the core temps (temps in the die centers) are the same with CPU temp (surface cpu temps) . The core temps are the temps in the CPU die and ALWAYS quite a few degrees higher than the surface CPU temp .
Everest claims they are the same and speedfan even reports one core temp cooler than the surface CPU temp.
Also notice that Speedfan is reporting what seems to be 3 cores? core , core 1 and core 2 . you have just 2 cores mate, all points to misreading.

so before u suffer the stress of repeatedly uninstalling your Xigmatek , i actually dont think it has any effect on the way software reads temps. just leave it on if you installed it properly in the first place.

what does your EVGA motherboard temp monitoring software say?

 

[commandercup]

the board didnt' come with any temperature monitoring software since its the T1 revision, the stripped version which doesn't have nearly as much box content as the A1 did

however... the bios is reading the temperature as 36C... which is close to everest/speed fan

and the extra core readings in those are the cpu temperature itself

but yeah, this is my second time reseating my Xigmatek... but the problem with just leaving it is that it must mean that the Xigmatek for some reason is worse then stock intel cooling!? now that makes no sense at all

 

[echo75]

and the extra core readings in those are the cpu temperature itself

the cpu temp itself is labeled as "CPU temp" not Core and u can see that value further up (reported as 34 C in ur screenie).

like i said the core temp and CPU temps CANT almost the same values as what Speedfan reports.

 

[dadi_oh]

What method did you use to apply thermal paste to the Xigmatek?

After much testing the method in the pics below give the best results. First fill in the gaps with thermal paste, then two lines across the aluminum part.

I agree with this assessment. The thermal paste application is critical with these exposed heatpipe designs. I like the example given above. I used a slighty different method on mine. Filled the gaps with compound but instead of applying the compound on the aluminum part I actually spread a very thin layer with a razor blade down the entire length of all 3 heat pipes (basically covering the copper surfaces completely.

I have not tried the other method so I can't say for sure which would work better. However, what I can say is if you just put a blob on the CPU heat spreader and expect it to spread out nicely as it does with a flat heatsink bottom you are going to lose a bunch of compound up the crevices and it will not travel to the extremities of the CPU. That would leave areas of the CPU not contacting the heatsink.

Did you examine how well the thermal paste spread when you removed it the first time? If you do decide to reseat using one of the above methods it would be good if you took a photo of the surface of the CPU right after you remove it so we can see how well it spread.

 

[ghost101]

Thats such an old version of core temp. M0 chips ahve been corrected since then.