Q6600 60c idle - need help

[Guest]

Hello people, i have a Q6600 core2quad, with a stock heatsink. I get about 60C idle, and i haven't tried any games yet. The stock heatsink fan is running at 1548RPM, and the VCORE is 1.288V. I recently took the heatsink of, cleaned the cpu and heatsink first with baby wipes, then applied white spirit on both, and dried with a cotton swab. I then applied OCZ freeze compound on the CPU, and but the heatsink back on - still no improvement. Is there anything i can do to reduce the temp - ive heard people get abt 40C.

PS: Im using Nvidia System Monitor to see me temperatures. Thanks

 

[sportsfanboy]

The push pins probably aren't in all the way. It's a pain in the ars, but taking the mobo out and installing the fan out of the case is your best bet. That way you can look at the back of the board and confirm the pins are seated correctly.

Download Realtemp 2.75 and post your temps with that.

 

[w153r]

sounds like you need to make some changes to your fan settings (in bios), if not, use a windows based fan controller, "speedfan" works.

 

[MCMONOPOLY]

Using the stock HSF, you get about 60C @ idle, even with good TIM. I got that even after the 100 hours burn-in period for the AC5.

The included HSF that comes with the Q6600 is a POS. At full load on 4 cores, it got up near 95C..!!!!

I bought a AC freezer Pro 7, my temps went down to 36-39 Idle (depending on which cores), and I usually got around 50C at load at stock speeds.

I OCed to 3.6 using the freezer Pro, got around 51-53C @ idle and around 61-64C @ full load.

Just change your HSF for anything aftermarket and you should see those temps drop like crazy.

Hope this helps.

P.s. If you intend on using Speed Fan to monitor your core temps, make sure to add 15C to each core reading, since people calculated that there's about a + 15C difference from the COREtemp readings and what SpeedFan Reports...

 

[Grimmy]

I wouldn't trust core temp readings at all.

Just make things more simple, and go by the Tcase temp, which your MB utility application shows the CPU/System temp, and keep it under what Intel specifies. The thermal spec is 71C for Q6600 G0. The B3 is 62C.

Giving advice to offset the cores by 15C is ludicrous. If you don't know about tjmax values then you should try to refrain from giving advise on the core temps.

 

[Guest]

Using real temp i get the following:
Core Temp
31 32 29 30

Distance to TJ max
64 62 65 64

What does it mean by 'distance to TJ max'

 

[Guest]

Oh guys i just noticed that the nvidia system temp was giving temperature TJ (i couldnt see the TJ lol!). So i am 60C away from the max temp right ? Can anyone tell me the lowest voltage i can run this cpu at? thanks

 

[Grimmy]

Heh.. distance to thermal shut down.

Thing is, that tjmax isn't something that is sensed by the program. Its a value that can be changed. So if you wanted, you could change the tjmax to 100C to have the same temp readings on CoreTemp (default value).

Where you got RealTemp, the website should have the info, as well to understand how to calibrate your system.

 

[CompuTronix]

P.s. If you intend on using Speed Fan to monitor your core temps, make sure to add 15C to each core reading, since people calculated that there's about a + 15C difference from the COREtemp readings and what SpeedFan Reports...

MCMONOPOLY,

Respectfully, you are misinformed. SpeedFan has not required any +15c Core offsets for any Core 2 variants since 4.33. The current version is 4.35. Perhaps you would be better informed if you read my Core 2 Quad and Duo Temperature Guide. Click on the link in my signature.

Comp

 

[dallasjoh]

It's roughly the point where the CPU begins to automatically throttle itself. Which for a Q6600 is around 95c.

 

[Grimmy]

How quaint. 95C for a Q6600?

You know, I thought that was a good Tjmax for my P6N for my Q6600 G0. But since my P6N MB died, I switch to a P35 chipset by DFI.

That same 95C (as well using the same CPU on a different MB)tells me my core temps are lower then my Tcase. So again, I wouldn't really trust Core temps.

Edit:

Look at this screen cap:

My Tjmax is 100C. Now if I change that, you should know it will reduce my core temps down 5C. Those are load temps, but at idle, my cores would be idling 3C lower.

Edit:

Here's another screen cap when I was testing stablity:

This was when the ambient case temps were 83-85F. My Tcase registers 59C. My cores were 62C with tjmax of 100C. If I change that to 95C, it would make the cores that are 62C to 57C which is 2C cooler then the Tcase, not to mention the 2 cores that are 60C, which would make them 55C, a 4C difference. So what is more accurate? Whelp.. I dunno. :lol:. o O (all I do know is to keep the tcase below 71C)

 

[jimmysmitty]

I wouldn't trust core temp readings at all.

Just make things more simple, and go by the Tcase temp, which your MB utility application shows the CPU/System temp, and keep it under what Intel specifies. The thermal spec is 71C for Q6600 G0. The B3 is 62C.

Giving advice to offset the cores by 15C is ludicrous. If you don't know about tjmax values then you should try to refrain from giving advise on the core temps.

I agree.

Funny thing is that due to the summer heat here my idle temps jumped from 30c to 40c but when under load stayed at 50-55c. Goes to show what a good after market HSF like the Zalman 9700 can do.

For the OP, Core temp seems to be correct for a stock Q6600.

To add to your last post though Grimmy, I would think that a Q6600 G0 would be more correct at 95 since its a 95w TDP chip. But thats just me.

 

[Grimmy]

Heh, but I went from a NV650i to a P35 chipset. I had to take down screen caps to make room for new ones, as cox only gives me 10MB web space.

The only example I have from the P6N was when I had my Q6600 to its max (MB limitation), 3.375 ghz. It was an idle screen though. The screen cap showed:

Tcase:38C

C0 - 41
C1 - 43
C2 - 40
C3 - 41

The core temps were based on 95C tjmax. If you were to change it back to 100C, that would add 5C to the temps (46-48-45-46). So I hope you can see that it made sense to go with 95C on my other P6N 650i MB.

It seems to be different or have an opposite effect on my DFI P35, like I've shown above. If I stay with 95C, then my temps are equal to 3C cooler then the Tcase, which I am not used to seeing. Even the load temps act the same way to where the core temps would be reading cooler then the Tcase, which in my mind is wrong. The Tcase max value is 71C vs the tjmax at 95C. If they stay equal or if the cores are lower.. something is wrong with the delta part of it. Or distance to thermal throttle/shutdown. Not sure if I explained that right.

 

[Grimmy]

Hmm.. I had a crazy thought.

Maybe CompuTronix could answer or comment on my speculation. :lol:

Since the CPU chips with the core digital thermal sensors are calibrated from the manufacture, are they calibrated only for the stock speed? Or would that calibration be off, since I've changed the CPU speed, which in turn causes more heat, would that skew or mess up the calibrated distance it it was intended for the original CPU speed?

 

[CompuTronix]

Grimmy,

Since you've moved your Q6600 onto a different motherboard, Tcase has changed due to different BIOS. Use Sectgion 9 in my Temp Guide to recalibrate.

Comp

 

[Grimmy]

Heh.. well to tell you the truth, I didn't bother to calibrate the tcase on my P6N. However, since I do use a IR Thermometer, and a secondary multimeter that measures temp with a thermal couple, when I measured ambient temps at stock speeds, the Tcase read the same as both measuring tools. However it would be off by -1C or +1C.

I'd like to follow your guide more to a T, but at this time, I can't get 22C to conduct the test to a T.

For the hell of it, I changed my setting all to stock speeds, I did try to calibrate with ambient temps in 28-30C. Funny thing is, both the Tcase and external thermal readings read the same of 28-30C. And when I used prime95, small ftt's, my cores were 4-5C hotter then the tcase, which were:

45C - Tcase @2.4ghz
C0 - 50
C1 - 50
C2 - 49
C3 - 49

It even got to the point to where the Tcase went down 44C and the cores did drop 1C as well.

I have my system back to OC settings, and still at idle my Tcase is 32C, and the ambient in the case is 31-32C. The cores are 4-5C hotter, but this is based on tjmax of 100C.

So for now I really can't do the calibration to a T since I can't attain 22C or 71F. The only coolest night I had was 75F, which the above pic I posted was based off of, and it's core temps are 5C hotter then the Tcase.

I dunno... all I know, is to keep my tcase below 71C, or... get a stick for some marshmallows to toast them over the HS.

 

[gamerk316]

Please post Realtemp temperatures if you are still having issues. Intel (finally) released the values for Tj, and Realtemp was indeed right, and CoreTemp reads too high for some Intel CPU's.

 

[CompuTronix]

Grimmy,

Don't let the fact that you can't get your ambient down to 22c stop you. The Note in my Guide says "It is preferred that Calibrations are conducted as close to 22c Ambient as possible to allow for..." which is certainly not intended as a "requirement". Regardless, I know that you understand the calibration procedure, so you're OK.

 

[Grimmy]

Tanks Computronix.

Well.. I have some other questions burning in my mind at the moment:

Here's a simple 2 part question, actually for anyone.

Has anyone reach near the Tcase max? (71C Q6600 G0)
And if you have, were your core temps near 95C or 100C?

Well... Okay, my system isn't calibrated (all offsets for speed fan are 0), but it does reflect the current ambient temp.

So I decided to answer those question myself. I've turned down/off my case fans in attempt to get near 71C. I got to a flicker of 69-70C:

As you can see, this is no place for anyone to put their PC into. Now I'm having a really hard time understanding why a 95C tjmax is correct, especially for a Q6600. Speed Fan is configured for 100C tjmax. When my system hit 69-70C my core were the same. If I change the tjmax, the cores would reflect 5C cooler temps, just like realtemp shows.

Now if the 95C tjmax is true.. the only analogy I have to this, is it seems to me that the Tcase thermal spec, is like a fuel meter in your car. When it gets to a 1/4 of a tank, the driver will have the blues in refilling his/her tank. But, theres usually going to be more gas in the tank, then what the driver thinks. A safe guard to get to a gas station asap.

So to put that in perspective, I'd have to really offset my Tcase to the point to where the tjmax will actually get to 95C, which to me would have to reflect sub temps. So is the Tcase actually a safe guard to not get the CPU to the point of melting?

Another question. Why don't MB utility apps have offsets? I know my past MSI MB utitily app had no offsets. My current DFI MB utility app doesn't have offsets. The only offsets that can be done that I know of is speedfan, or the other apps that read cores. HD Monitor is the only freeware app I know it doesn't offer offsets.

. oO(this is driving me wacko... I think I need to go cool off.)

 

[CompuTronix]

As Intel has stated, Tjunction Max varies from part to part, so a particular Q6600 sample could be anywhere from 95c to 100c. Also, Core temperatures (Tjunction) are affected by "Slope Error" which introduces an additional calibration variable. Regardless, there should be a 5c gradient between Tcase and Tjunction at load, as illustrated by the following Intel graph:

As to why more utilities don't provide any means to enter offset corrections, who's to say? I'm just grateful that we have SpeedFan.

Comp