1090t Running hot at Stock Speeds, but which temp is correct?

[Tat]

I have a problem...
1. I need to know my actual CPU temperature.
2. Assuming it is too hot, I need to cool my system down better somehow...

1.
As I read it, having read checked temperatures with both Everest, Core Temp, CPUID (my previous favourite) and PC Wizard at the same time, both at idle and when stress testing, I've come to believe that my CPU runs at approximately 40 degrees celcius at idle (38-41) and 59 at peak, without OC'ing at all ! ! !
Needless to say, this is FAR to hot and allows practically no headroom to overclock, my plans being to OC to 3,6Ghz, using the automatic clock boost to get to 4Ghz when using 3 or less cores.

I've taken two screen shots showing the various programs and their readings at idle and peak. I've outlined in red, the readings that I know are incorrect for sure, both temperatures and fan speed readings. My issue is, how to get a trustworthy CPU temp reading and vcore reading. Secondary would be to get proper fan speed readings...

Idle readings:


Power cut just stopped my stress test...
When redoing it, I don't get past 52 degrees, whereas before, I was at 59!
Also notice that where before CPUID could only register two of my four fans (I got 5, but only 4 connected to the motherboard - all 4 show correctly in the BIOS), though with ludicrous RPM, it now registers 4, with no RPM...
Peak Readings (was 59, but after powercut is 52 - how's that make sense?):


Before doing this, I did try a low OC, which crashed my computer after 20 minutes at an Everest CPU reading (doing a stress test) of 67 Degrees. Since I figure the CPU would shut down at 62 Degrees normally and since I figure it shut down at 68 (Everest degrees), with a 6 degree difference, does that mean my actual CPU temperatures at Peak, with stock speeds is 59-6=53? Of course the new reading is 52, but is that because the power cut and crash fixed/changed the temp reader?

My Bios readings:
CPU Temp 40
MB Temp 32
CPU Fan 1: 1520RPM
Chassis Fan 1: 746RPM
Chassis Fan 2: 745RPM
Chassis Fan 3: N/A
Power Fan: 1128RPM
Vcore: 1.272V
+3.3V: 3.360V
+5.00V: 5.040V
+12.00V: 12.408V


2.
Ok, IF my temps are as unseemingly high as I think they are, how do I get them down without buying a new cooler?
I've got the following setup:

CPU: Phenom II X6 1090T
MB: ASRock 890GX Extreme3
RAM: GSkill PC3-1800
GPU: Nvidia 460 with 1GB

Cabinet: HAF X
CPU Cooler: ThermalTake SpinQ VT
Thermal Paste: Tuniq TX3

Ambient Temperature in my room is 28-30 degrees, which I suspect to be a problem. I live and work in Bangkok, Thailand.

Inside of my case - I think it looks good - wires not impeding airflow:


I'm thinking of putting another fan in the top front, as I'm worrying the lower front fan is not enough with the disks in front of it:


From the outside:


So, I figure once I know how to get proper temperature readings from this CPU, I'd re-attach the cooler, with less paste, just to see if it makes a difference and then add good 120mm intake fan in the top front . . .
Any other suggestions are more than welcome

 

[boswell3782]

Might be a faulty temp reader or just try and buy some high spec thermal paste and try that.

 

[Tat]

Well...
I thought Tuniq TX3 was high spec thermal paste (bought seperately). I did try to get OCZ Freeze, but impossible here and for some reason Amazon doesn't ship this product to Thailand...

Seems that now my CPU readings are generally lower for no apparent reason...

Will do some stress testing and get my current temperatures, assuming they won't change, but instead of Everest, I'll try Prime. Then I'll re-attach the cooler, add a 120mm front intake fan and make the top fan work again (just realized it's not working.
After this, stress test again and compare...

If it's better - good... Then on to OC'ing again...
If not, I thought about reversing the airflow, as I think my G-Skill RAM might be in the way - not sure...

 

[Tat]

I do realize that ambient temperature means a lot.
In fact I would understand if AMD and Intel gave out lower warranties out here, as it's both more dusty, humid and hot than anywhere else I've lived and I had trouble with CPU's before, running at stock speed!

I installed OC Tuner from ASRock - since it comes with the board of course it's compatible and it actually shows not only my correct CPU temperature, but also correct fan speeds and voltages - finally a trustworthy temp reader

Anyways, I think the crazy temperature readings (as in very different, from test to test and not with regards to being too high) was from a thermal paste mishap, as there were fat leftovers on the side and so much of it that the CPU was stuck on the cooler and not to the board! When ripping it out with considerable force, a tiny bit got stuck in the pins - NOT cool...
With a needle, cotton buds and the cotton squares used for removing nail polish (so I hear), I cleansed first the CPU, while swearing a lot and then the Cooler. After I had done this I applied Tuniq TX3 thermal paste to the CPU again, though much less than last time, and realized it was much too fat and I do recall that being the issue last time as well. Impossible to work with. I wonder if it's supposed to be that fat...
So I cleaned both CPU and cooler again and used the Thermaltake paste that came with the cooler when re-attaching - much easier to work with.
Since I had the whole thing open, I added another fan to the front top, to make sure enough air gets pushed through.

CPU Chip Stuck to cooler:


Termal Paste in the wrong place:


Too much Paste:


New top front 120mm intake fan - notice how well it has been attached


Unfortunately my idle temperature is still around 40-41 degrees, so no change there ...
Stress testing now...

 

[Tat]

Now I'm getting temp readings that make no sense...

First I did an AUTO everything to check temperature max and it went to 67 degrees, but stabilized there (at an earlier run, it crashed, probably when hitting 68). Then I tried to manually set VCore to 1.237 (one step down) and it only went to 61 degrees, though I got readings of 1.29 and 1.27 (different programs). Then I tried Auto again and it only went to 62 degrees - did this, as I forgot what voltage it hit before. It went to 1.3 volts... Since I got a completely different temperature reading, I figured WTF and tried the other one again - that is manual voltage of 1.2375 and now the max temp seems to be only 58 degrees, BUT now the voltage readings are 1.312 and 1.29 ! ! !
For fucks sake - when the voltage is higher, so should the temperature be! Furthermore, at Auto it should hit the same voltages right?

The only explanation I can think of other than that the board or chip is funky, is that the thermal paste has to set or something...
Still - Looks like a return to sender and hope for a refund...

 

[Tat]

But...

I'm NOT overclocking!

If anything, apart from running everything at Automatic (BIOS default), I've tried UNDER clocking, so that it wouldn't get past 62 Degrees when running stress tests...
I figure running a stress test, before overclocking, helps me to know how hot my system is getting now, with no OC at all and since I'm past safety (62 degrees), I can't overclock at all . . .

I've seen plenty of people running cool at high OCs too - the ones I've seen though, have ambient temperatures a good 5-10 degrees lower than me though. Regardless, I know this CPU and MB can OC and run at very decent temperatures, but mine idles at 40+ and goes into 60+ when stress testing without the OC. 68 degrees last time I checked and well, that leaves no room for overclocking whatsoever...
The CPU cooler isn't the best out there, but it's still a lot better than the stock cooler and add to that the HAF X and I'm fairly sure that something's not working as it should...

 

[Tat]

Good point!

I found a Thai post where someone else has the same setup as me (MB, CPU and Cooler) and gets these temperatures:
Cool n Quiet on - normally it's 44-45C
Cool n Quiet off - is normally at 51C
(http://www.overclockzone.com/forums/showthread.php?p=20797626)

I assume those are idle temperatures or next to idle and they're even worse than mine!

I'll try the stock cooler and see what happens...

 

[mrface]

On Board Temperature diodes are known to be less than accurate. However, Since you are saying it is crashing on you, I would suspect this to being close to the actual temps...

From the pics, It looks as if your cable management is good So, I dont think airflow would be hindered by that. Have you tried checking temps with your side panel off? If it differs than when on, your issue is probably ventilation. You have big fans on there but the hot air needs to escape as well, otherwise you effectively have a convection oven... If this seems to be the issue I would dremel out the grates around the fan exhaust holes on your case, the metal "filler" (the mesh looking stuff) will adversely affect airflow.

Try the temps with the panels off and report back.
good Luck,
Cam

 

[mrface]

Did you even read his post?

As I read it, having read checked temperatures with both Everest, Core Temp, CPUID (my previous favourite) and PC Wizard at the same time, both at idle and when stress testing

Intake fans are less important than exhaust fans. If you have a ton of air coming in and no place to go it's creating negative airflow.

Yes and Yes, .

I agree that intake is less important than exhaust. That is why I suggested taking off the side panel.?.

 

[Tat]

Well...

With the TermalTake Tuniq SpinQ VT I got these much too hot results:


Witht he stock cooler, running rather hot, but much less so than the SpinQ:


So, I've already found a buyer for the SpinQ and am going to get the THERMALTAKE FRIO, which is the only cooler I've found out here that has been aprroved for the X6 AND has got high ratings in reviews - it's second on FrostyTech . . .

Really hope that'll solve all my problems

BTW I did try to take of the sidecover before - no discernable difference...

Will let you know how it all works out, but regardless, thanx for helping

 

[Tat]

I know the CoolerMAster Hyper 212+ and I've got it for my old E8500. Good and cheap... Dunno if it can do the X6 though.

I got the ThermalTake Frio, as it was the only one I've found that is both available out here and X6 rated (as stated before) and while FrostyTech isn't perfect, a cooler that's second in their list, can't exactly be bad either...

The Frio installed:


The Frio with both fans set at 1500TPM (minimum is 1200, but at 1500 they're pretty quiet already):


The Frio with both fans set at max = 2500TPM:



I now get MAX temps when running Prime95 (2500RPM) around 47 Degrees - incredible difference!

Tomorrow I'll start OC'ing

 

[Tat]

My OC:

Hardware:
Phenom II X6 1090T
ASrock 890GX Extreme3
2x2GB G.Skill PC3-1800 RAM at 9-9-9-24
NVidia GeForce GTX 460 with 1GB RAM
2x1TB WD Black FAEX SATA3 in RAID for better speed
HAF X Cabinet
ThermalTake Frio Cooler

I OC'ed to 3.9 stable so far, but at 4Ghz haven't had it stable (didn't try to hard though).
My moderate OC is 3.613Ghz - 16.5x219 with a Vcore set to 1.3125, which in Bios shows as 1.344 and in Windows fluctuates between 1.328 and 1.368.
My Northbridge is at 2409Mhz - 11x219 and 1.2125V.
My RAM is set to 876Mhz at 1.645V.
I've turned off Cool and Quiet, but all other automatic OC is on and I've seen it go as high as 3.95Ghz with Turbo Core...

Here's my Benchmark:


Any advice to get it to perform better is welcome

 

[OvrClkr]

My OC:

Hardware:
Phenom II X6 1090T
ASrock 890GX Extreme3
2x2GB G.Skill PC3-1800 RAM at 9-9-9-24
NVidia GeForce GTX 460 with 1GB RAM
2x1TB WD Black FAEX SATA3 in RAID for better speed
HAF X Cabinet
ThermalTake Frio Cooler

I OC'ed to 3.9 stable so far, but at 4Ghz haven't had it stable (didn't try to hard though).
My moderate OC is 3.613Ghz - 16.5x219 with a Vcore set to 1.3125, which in Bios shows as 1.344 and in Windows fluctuates between 1.328 and 1.368.
My Northbridge is at 2409Mhz - 11x219 and 1.2125V.
My RAM is set to 876Mhz at 1.645V.
I've turned off Cool and Quiet, but all other automatic OC is on and I've seen it go as high as 3.95Ghz with Turbo Core...

Here's my Benchmark:
[url]http://s2.postimage.org/oeHQi.jpg[/url]

Any advice to get it to perform better is welcome

Your temps are better than average, you have headroom to spare

Try raising your multiplier to 21x @ 1.45v (disable turbo)

CPU/NB frequency to 2600Mhz @ 1.25v

Leave your Bus (FSB) at 200Mhz (later on you can raise that if needed)

Not sure if you are using Afterburner or not for the GPU but if you bump up the clocks you will should get a nice boost (should get you at least 21k in 06)

Just giving you an example of what you can try, the voltages can be either lowered or raised if needed, keep the CPU under 60c at load (preferably under 58c).

Nice cooler btw, Enjoy ..