HP8510p thermal sensor malfunction?

Hey guys!

I've got a serious problem with my 8510p.
I've replaced the thermal paste about 3 months ago and recently tried overclocking the GPU (ATIHD2600) to core 640 and mem 780 (Destiny's sig says his 8510p ATI is working, so I gradually oc'd there myself).
One of the temps in Hardware monitor idles at 30°C (86°F) but goes all the way up to 100°C (212°F) under load. I assumed that was my GPU, so I repeated the paste replacing (I used AS5 btw, and got big temp drops the first time, so I presumed I did everything ok).
I even bent the heatsing to increase pressure on the GPU.
No changes.
Then, I discovered, that the ATI GPU overclock tool also has thermal sensor monitoring. I tried that and...The GPU idles at 34 and only goes up to 38 under load.
(I'm using Far Cry 2 benchmark btw)
So now I'm stuck.
I've no idea what that sensor is.
I'm pasting a screenshot of hw monitor for reference:

Also, I'm pasting a temp log of ati gpu tool:
ATI GPU OC Tool LOG

Oh, did I mention that if I OC my gpu for 50 MHz my computer will restart within 15 min?

Many thanks to anyone who responds.

EDIT: Also forgot to mention that the temp swings wildly. I've set the PowerPlay in Catalyst Control Center to Maximize battery life and the temp is swinging from 30 - 60 °C

 

Yeah the 8510p's GPU sensor is abit funky, for many people, we get a very low temperature reading. <20C.

But every GPU is different, mine goes up to max 640/750.

 

Under what sort of load does the temp goes upto 100*C ?

Try running Orthos (Small-FFTs) and monitor the temps.
Then run rthdribl, and monitor the temps. (for the GPU as well)

You could run PC Wizard or Everest and view the GPU temps in them as well.

 

I'm pretty sure that the restarting is due to overheating...But what overheats??
As I wrote before, the temp rises to 100°C under gaming, running at factory clocks (500/600).
I'll run rthdribl and report with the results, but I doubt that they'll be any different than what I have now.

Something just popped into my head:
Could that temp sensor possibly be the one covered by the third branch of the heat pipe?
If you don't know the insides of a 8510p:
The heat pipes cools 3 cores (or what-nots). CPU, GPU, and a third one on the motherboard. The third one has a thermal patch on. Could it possibly be, that the thermal patch got damaged a bit when I took the cooling apart?
Doesn't look very damaged, but since the GPU isn't heating up...ok, real stupid method, but I guess it should tell the truth - if I put my hand underneath the laptop, the place where the CPU is, is slightly warmer than the location of the GPU - which would fit nicely with the temps I measured (CPU 60°C, GPU 40°C).
Is there anyway I can replace the thermal pad with AS5? Would just bending the sink in place work?

Thank you for your quick replies!

 

Wherever there are thermal pads, don't mod them. Playing too much with this stuff, could void your warranty. Apart from the CPU, GPU, the chipset is covered by the heatsink, which actually doesn't generate much heat.

Either temp sensors are screwed up, or HWMonitor is screwed up. Try updating the BIOS to the latest version, that might help.

 

I wouldn't say that hw monitor is wrong. Because whenever that temp rises to 65°C or more the fan starts to spin faster. My warranty has probably been voided, since I removed the heatsink. AFAIK that voids it (I think I even asked an HP rep thru chat about this, not sure though).
Anyway, the rthdribl thing runs at 90~110 FPS at 16x multisample and every other option I could find maxed out, the GPU being OCd to 640core/780mem. According to ATI GPU Tool, the GPU temp is 45°C. And I believe it.
I know, the chipset shouldn't get to 100°C EVER!

Well, out of curiosity then, anyway to replace thermal pads with thermal paste?
(I will do a search right now I promise , but if anyone knows of a good thread...)

 

Not sure, but if you replace the thermal pad with just thermal paste, there will be some space between the die and the heatsink surface - which will not be good. You could replace the pad with some other pad, or Cu sheet, etc.

There is a thread in the Dell XPS forum, about a M1330 GPU heatsink mod....check that out.

 

Yes...there was a 8510p bios released recently (decemeber). It is version F15. While you are at it, have a look at NBR Cooling Central.

http://h20000.www2.hp.com/bizsupport/TechSupport/DriverDownload.jsp?lang=en&cc=us&prodNameId=3355680&taskId=135&prodTypeId=321957&prodSeriesId=3355678&submit.y=0&submit.x=0&lang=en&cc=us

 

Worst effing thing ever!
I flashed the bios with the latest version and now I get a bluescreen everytime I boot. It comes up right after the vista load bar and my laptop restarts to another bluescreen.
Last good configuration does nothing.
Oh, and the HP bios flashing program (on a bootable usb key in freedos) won't let me flash with an older version, supposedly because my it's not compatible with my hardware.

I've done a search about downgrading the bios, but haven't really found any usable answers, since no laptop died like this after a flash...

Any ideas would be much appreciated.

 

................>.>

Well I would sent it back to HP.

 

I just tried to run a live distro of Ubuntu and....IT WORKS!!! So I presume nothing is physically wrong with my laptop.
Something messed vista up bad... I'm gonna try that repair thing that they have on install cds.

 

oh i see...yeah the recovery discs should work.

 

no no...not recovery discs. I purchased my laptop without any installed OS, hence no RECOVERY discs.
I meant the repair option that vista has on the installation dvd...
haven't tried it yet.
I'm in ubuntu right now.
Wish I could find all the programs that I need and just switch OSs... Curse you, M$!!

 

Yeah, you could run the Repair thing in Vista. Do you know the STOP code shown in the BSOD ?

 

Oh I see.......~_~

Oh and as for the CPU temp. I'd try undervolting to reduce the heat.

 

repair thing didn't work.
back in linux.
dammit.
the cpu temps are ok. as are the gpu ones.
overything's fine. except for that one sensor that goes haywire everytime I run a game.
and I still don't even know what it measures.
I did undervolt the cpu BTW. lost about 10~15°C.

i'm just frustrated now, because I have to rescue what files I have on the vista partition and format. the way I see it now, it's the only solution. cos nothing is physically wrong with the computer. it's just that vista won't boot with the f.15 version of the bios.
ubuntu however... 8.04 sometimes had this weird video driver bug, that is now nonexistent. so I guess I can say that f.15 is a good upgrade.
unlike f.14 that tripled my boot time (something about quadruple checking of credentials - i use a fingerprint logon method).

will post back, when I solve something.

if anyone has any more ideas, keep 'em comin'! ...please...

 

Weird because I dont have a thermal zone 0 and HWmonitor shows my GPU temps. Also F14 worked fine for me with finger print.



Tell Andy what he needs to know, he is the phonejacker ad he demands answers.
\/

 

Whats the STOP code shown in the BSOD ?

 

can't really say. it goes away even before the whole lcd has a chance to turn completely blue.
is there any way I could find out?

 

Ok, does anyone have any ideas how I could fix this without having to reformat my HDD?

 

I reformatted. Since then, Windows7 Beta and currently Vista x64 have been working well.
Still don't know why all OSs crashed after that BIOS update...
I tried to revert back to the old bios, with no luck, since the flashing tool deemed my hardware incompatible with it.
But I don't really care anymore, anyway, since I'm out of the red...
Thanks for all your help!!
Much appreciated!