Incorrect Core Temp Readings

[CycoSomat1c]

Ok fellas, I wasn't sure which thread to post this under, but I have a really strange issue since converting to watercooling on my Rampage Extreme. I'm getting ready to do some overclocking and wanted to compare my actual individual core temps to the overall CPU temp readings. PC Probe, Everest, and CoreTemp shows my overall CPU idle temps at 25-27C degrees, however, I noticed in Everest that the Core#1 and Core#2 temps are constantly reading 40C. I know this is incorrect because no matter what kind of load I put on the CPU using Prime95 the individual cores NEVER move from 40c, while the overall CPU temp will go from 25c and max at about 33c under load. I then tried CoreTemp and RealTemp, but all programs show the individual cores frozen (pardon the pun) at 40c. As most of you know, Everest has in its settings a way to change your particular processor's Tj.Max setting. Intel's spec on the E8400 for Tj.Max is 100c. Using that setting yields the same core readings of 40c, but changing the Tj.Max to say 80c causes both cores to then show 20c, setting Tj.Max to 90c causes them to read 30c, Tj.Max of 120c gives both cores reading 60c, etc, etc. Hope I haven't confused anyone? I'm wondering if this is a BIOS issue with the RE in the way it obtains its sensor information. I'm not 100% sure, but it seems like this issue wasn't present when running the system on air-cooling, but I can't really remember for sure. Again, no matter what kind of load I put on the processor, the core temps never fluctuate, but the overall CPU temp does, so I'm sure that the core temps are wrong. Has anyone else seen this issue or know what could be causing this? I really want an accurate reading before o/c'ing. Thanks in advance!

 

[Arctucas]

@CycoSomat1c,

It sounds like your core temperature sensors are stuck.

Use the Sensor Test in RealTemp to see if there is any movement.

 

[newtekie1]

Yep, sounds like stuck temperature sensors. Pretty common in the 45nm processors.

The internal temp sensors won't go below 40C when they are stuck. And since you water cooling is keeping the processor under 40C, it won't move. I be your air cooling wasn't keeping the processor below 40c.

 

[r9]

Stuck sensors. I have one showing 54C when real temp is 0-54C when the temp goes above 54C it is changing. And the other one is working just fine. One not working it is not that big of a problem but two stuck sensors that is bad thing if you are pushing hard your cpu.

 

[pajama]

Stuck senors

I noticed the same thing on my Q9450. Core 0 is 28 degreec C at idle and the other three cores are at 40 degrees C at idle. As soon as the first core goes above 40 during stress testing all of the other cores temps work fine. So I have 3 temp senors stuck at 40 on the low end but function perfect above 40.

 

[CycoSomat1c]

After searching the web for similar problems I believe you guys are right. The problem looks fairly common, especially on the E8400's which is what I have. The only way I can get it over 40c is by overclocking to 3.6ghz or above and stressing with Prime95, so I'll try that and see if the sensors fluctuate. No biggie I guess, I'll just have to keep an eye on my overall CPU temp when o/c'ing to be safe. Thanks for the help fellas!

 

[newtekie1]

The stuck sensors will work as temp goes up, they just get stuck going down. So it is fine to use them to judge temps, as 40C is still safe and if the temps go above that they will be accurate.

 

[Marineborn]

the sensors are diffrent then the die, the die is your lower one, the die is the one you have to watch

 

[fizur2002]

is this a really common problem or is it just totally sporadic?

 

[erocker]

It's in between common and sparratic. RealTemp has a temp sensor test.