Q9650 DTS faulty?

[ArchangeL1979]

Hi guys. Have a bit of a problem. Friend picked up a Q9650 with Asus Maximus II mobo and it's great. Just one problem though. I have a feeling the DTS sensors on 2 of the cores might be knackered as the reported temp on 2 of them is much much higher than the other 2 (and much higher than what they should be).

As you can see cores 1 and 2 report an idle temp of 67oC which is high (but both the cooler and cpu itself are not even warm (checked this by touch when I tried reseating the cooler). Another point to note is that when I reseated the cooler the pressure pattern in the thermal paste indicated there was full contact on the cpu. Cores 3 and 4 seem to be reported correctly. I have applied some

I'm using a Xigmatek HDT S1283 cooler with arctic silver.

So my concern is that when I'm playing games, even at stock, there may be times when the incorrectly reported temps exceed the TJ Max value and it starts throttling unnecessarily which is going to hit me on performance. Further to that it doesn't leave me with much headroom for overclocking.

Anyone have any good advice other than to live with the fact I'm not going to be able to get top performance out of this chip?

 

[Scon]

Have you checked the temperature in the BIOS?

Anyone have any good advice other than to live with the fact I'm not going to be able to get top performance out of this chip

Looks like the way it's going to be THB.

 

[ArchangeL1979]

Temperature in the bios is low. Something around 28oC to low 30s.

 

[ChrisLandin]

Are they just stuck sensors???

Try running prime or IBT and see what happens? If they move up over when prime is running but then just back to 67 after its finished you know they are just stuck which isnt a huge worry as you know they will report temps over 67 degrees which is the most important really as it helps you know how far you could push the chip! My q9650 has a stuck sensor on core 3 and reports 41 degrees always until its stressed and then does move with heat just never reports a true idle temp.

Hope this helps

Scratch it. Just looked at the pic again. Im a dunce. Does seem like an issue somewhere with the chip

 

[ArchangeL1979]

I dont think the sensors are stuck as they keep fluctuating around 67-72 at idle.

 

[Scon]

If the BIOS says that then that is the most accurate temperature so the CPU shouldn't be hitting any dangerous temps i don't think there is anything you can do about the throttling, i'd just keep an eye out for another cheap second hand one somewhere.

 

[ArchangeL1979]

Bummer. Is the throttling something that can be done via BIOS settings, or is it hardcoded into the chip instructions?

 

[Scon]

I'm pretty sure it's the latter, could be wrong though.

 

[El Jimben]

The sensors on core 2 and 3 are knackered and unfortunately this is a known problem with Intel 45nm CPUs. As you'll see here from a previous thread mine is the same. I have no experience with your cooler but hitting nearly 70c under no load isn't good so either it's not up to the task or it's not making good contact with the IHS.

 

[Big.Wayne]

Hi ArchangeL1979

• Intel® Core™2 Quad Q9650 @ 3.37GHz
• Asus Maximus II mobo
• Xigmatek HDT S1283 cooler with arctic silver

I have a feeling the DTS sensors on 2 of the cores might be knackered as the reported temp on 2 of them is much much higher than the other 2

Hmmm, I've worked with a lot of Intel® Core™2 processors and have run into a few DTS anomalies before . . . I've seen sensors that won't sink below a certain point making it hard to gauge idle temps . . . I've seen sensors that are capped above a certain point making it hard to gauge load temps and I've seen sensors that are just plain stuck and make it hard to gauge any temps!

What I haven't seen before are DTS that just spit out erroneous temp readings like what you are getting in the screenie above?

My hunch here is that your heatsink isn't fitted correctly? . . . . I've used a couple of those Xigmatek HDT S1283 before and while they are good coolers the flaw with them is they use the dreaded Intel® LGA775 Push-Pins which are notoriously difficult to fit . . .

Temperature in the bios is low. Something around 28oC to low 30s.

The BIOS CPU readings are taken from the socket IIRC and are not read from the DTS . . . .

The cooler the pressure pattern in the thermal paste indicated there was full contact on the cpu

While this does lead one to believe that the cooler was fitted correctly it is also possible that the "pressure pattern" in the TIM was made while you were attempting to fit the cooler and was applying downwards pressure . . . of course as soon as the downwards pressure ceased and "If" the push-pins were not engaged correctly you would be getting the scenario you are experiencing above . . . . uneven and ridiculously high temps due to the coolers push-pins not being engaged correctly . . .

both the cooler and cpu itself are not even warm (checked this by touch when I tried reseating the cooler

Hmmm . . . well having a heatsink not getting warm/hot while the CPU is loaded means either its a stunning cooler that is dissipating the heat evenly across its mass *or* it means the heatsink isn't making good contact with the processor? . . . you say you check the actual CPU temps by touching the processor when you reseated the cooler but in the time it took you to power down and dismantle the the HSF the chip could have easily cooled down . . . . DTS @ 88°C could drop to 34°C in about 10 seconds . . . and these are "internal" temps which do not reflect the same as Tcase . . .

My conclusion is that "If" you are indeed using the Intel® LGA775 Push-Pins then probability dictates these high temps are likely being caused by an incorrect fitting of the Xigmatek HDT S1283 over having faulty DTS? . . . The way you could test this would be:

1. Apply even/downwards pressure on the heatsink and see if this affects the temps
2. Check the reverse of the mobo to see if all four black pegs are poking out the back?