NEW T60 Temperature Thread - Especially high GPU Temperature

[Kira]

Thanks for the advice, Troels. The GPU also runs unusually hot (mid 70's) when idle, but it only ever goes to shutdown levels when running graphical stuff. Does this still sound like a contact issue?

I'm planning on picking up a can of compressed air tomorrow, and I'll give that solution a try. What idle and loaded temperatures can I reasonably expect as an indicator of success?

[RealBlackStuff]

If you use the compressed air, make sure you hold the fan itself in place with your finger or other implement, to stop it from turning!
Without holding it, you'd over-rev the fan motor and destroy the bearing.

[Troels]

Thanks for the advice, Troels. The GPU also runs unusually hot (mid 70's) when idle, but it only ever goes to shutdown levels when running graphical stuff. Does this still sound like a contact issue?

Ah ok, mid 70's when idle - that's a little high, but unfortunately not far from norm, so i'd say you can probably decrease this by compressed air. Since th fan is a brushless type, it would be hard to damage any part of it by compressed air. but by holding it still, you atleast blow off any dust from the fins which would reduce fan noise by a little.

My absolute maximum temperatures at full load with a lot of 3D is 94 deg.
Mine operates at 75 degress at idle, but it's due to the fact that Tpfancontrol does not turn on the fan before that. The hardware fan control would otherwise keep it at about 68-70 deg i think.

If the heatsink didn't make proper contact it would go above the mid 70'es at idle.

[Kira]

Thanks everyone for the suggestions. I disassembled the machine, looked at the GPU, and found that it was basically clean. There was a nice, full imprint of the die against the heat pad, so the connection can't have been that bad. Any more ideas for what could explain it going to shutdown heat levels after a minute or two of running 3D?

[Troels]

One last idea (after all dust has been cleared and the overheating still happens): The heatpipe is not working as intended, i.e. lost its vacuum so the coolant has evaporated. Even a small microscopic crack is enough. Any visible damage to the heatpipe connected at the GPU section and going to the fan area?

[Binh]

Thanks for the advice, Troels. The GPU also runs unusually hot (mid 70's) when idle, but it only ever goes to shutdown levels when running graphical stuff. Does this still sound like a contact issue?

When the keyboard is opened, try to run a GPU stress software and use your finger to feel the temperature of heat-pipe above the GPU. If GPU temperature is above 80C then the temperature of heat-pipe should be 60C or higher, what is very hot for your finger.
If the heat-pipe is only warm, then the problem is the contact between heat-sink and GPU.

If the contact is good, try to feel the temperature difference of the GPU heat-pipe (small one, the big one is CPU heat-pipe) between areas around its two ends. If the temp. difference is large then the problem is the heat-pipe itself, as pointed by Troels.

[RealBlackStuff]

Since cooler/fan seems to be OK, I would suggest to replace the thermal paste.
It has probably dried out with all that heat (and how old is it?).
Clean both CPU and cooler with Isopropyl alcohol (91% or higher).
Apply Arctic Silver 5 (from e.g. The Shack).

[Kira]

Cleaning out the guts made it so that the GPU no longer regularly shuts down, but it's still running around 100 deg under load. I'm not thrilled with that, but I'm running out of ideas; looks like I'll just have to be careful how much load I put on it. It's too bad, I was really hoping this thing would be a worthy replacement to the T43p, which never gave me any issues until I got hit by the GPU Reflow Issue.

Thanks to everyone who offered advise, it's highly appreciated.

[RealBlackStuff]

Out of curiosity, are you both talking about the same temperature scales?
Troels is in Denmark/Europe and using Centigrade.
Kira is in USA and using Fahrenheit.
Compare temperatures here: http://www.infoplease.com/ipa/A0001731.html

[Troels]

Hehe, you do have a point there, RBS!
But 100 deg. F would be rather "cold", unless Kira is converting a reading in Fahrenheit to Celsius and converted it incorrectly - but everything's possible.

Kira - was the heatpipe working as intended (based on the measuring method proposed by Binh) ?
On what type of surface if the laptop placed when it overheats?
What software do you use to stress test the GPU?
Is it still under warranty?

[Kira]

Troels -

The heatpipe seems to be good to go, although I'm not sure that my measurement technique was perfect. The laptop is usually in my lap when I'm using it; it hits 100+ just playing Audiosurf for 3-4 minutes, which is a simple game and generally shouldn't cause massive overheating.

Thanks, as always.

[RealBlackStuff]

If Kira means 100 deg. Fahrenheit, then she has absolutely nothing to worry about.
The CPU can run up to ca. 160 deg F without shutting down by itself.
It would still not be overheating at that temp!
Having said that, if you wear a skirt and have the laptop on your lap, your skirt-material might close off the vents in the laptop, which WOULD cause overheating very quickly.

[Kira]

RealBlackStuff - I wish it was farenheit. Sadly, tpfancontrol seems to report celsius, in which 100+ is very hot indeed.

[RealBlackStuff]

This is the latest version: http://staff-www.uni-marburg.de/~schmitzr/tpfc_v062.zip
which can display both C and F.
See how that works for you.

[wujstefan]

Hello,

I have swapped my beloved but pretty worn T60 14,1" for an almost new 15,1" SXGA+ Flexview T60 with ATI X1400 graphics.
What I have observed is an enormous increase in idle temps comparing to 14,1" I still own.

14,1" with T2400 and X1300 idle temps were about 45/55C (fan lvl1) and 65/75 on stress
15,1" with T2500 and X1400 idle temps are about 46/65C (fan lvl1) and 75/90 on stress
Ambient temp ~25C

Isn't it way to high? I have reseated the fan and applied a AS5, also the blown air is hot when the fan is working.

[RealBlackStuff]

X1400 runs hot by nature, probably worse than the T61p.

[wujstefan]

Then I believe that I should't be worried with 90C in stress as it rather won't harm the system?
I thought this temperature is way too high...

[JoeHallenbeck]

Hello,

Reporting same problem. X1400 core 2 duo t7200 and on the battery GPU is on 80, writing this text. I changed last night heat-sink suspecting tube failure and the teperature is even higher because I replaced old used temp pads with new thinner ones.

During my first power on test keyboard bezel off and palm-rest off I touched the main pipe leading from CPU, right next to the top exhaust (literally the end of the pipe) and noticed immediate hot temperature during BIOS post start. The other tube was cold. Since the GPU is heating much faster than CPU it should be opposite way.

Now when the GPU is 81 degrees bios is blowing air through heat-sink but it is not hot but cold. CPU is 40, so there is definitely problem with proper contact of the copper plate to the GPU chip. I was considering the 20 degrees offset mentioned here before but even if it was not 80 but 60, it should be still hot because if I test the CPU on 60 degrees I can feel by holding my fingers next to the exhaust that the air is hotter at 60 degrees.

Next step is to take it apart again (probably 10th time) and check the aligning. That is why some people are reporting that polished coins or re-applying the thermal pad for thicker or thinner is working for them but not for everyone. I think that the alignment is the culprit of the problem since every possible heat-sink I saw dismantling T60's is different.

[JoeHallenbeck]

First attempt to resolve my GPU temperature seems to be working.

I noticed that the copper plate is not touching south bridge properly. After applying thinner thermal pads on the plate I tried to settle the plate correctly. But since south bridge is a bit lower than GPU it was not seated properly. I was expecting higher temperatures on the south bridge but since that silver metal clip has those two spring clips holding coper plate down it caused that the plate was touching GPU by its corner only.

After I dismantled it again, I removed all thermal pads and started to align the copper plate accordingly to the CPU leaving south bridge unaffected. I tried several times to sit the whole heat-sink assembly on "dry" on both chips the CPU and GPU. When I was okay with the result I choose a different approach to connect the platter to the GPU. I used instead thermal pad a thermal paste on the GPU. Spread evenly as I would on CPU. Then I took old used thermal pad, cut it with sharp knife to approximately same size as south bridge die, no bigger as it is on default and place it on the die, not on the copper platter. Than I made little "tunning" to that metal clip which is holding whole south assembly on the both south dies by lowering tension on south clip and increasing tension on north clip.

Before all screws tightened up I twisted the whole assembly around its center to spread the thermal paste even more.

Now after whole thing is assembled together TP-fan is reading GPU at 68 degrees as I'm writing this text and CPU at 50 degrees on smart switch number 3 and 3442 RPMs.

AC is in and set to power source optimized which was set up by me practically all things on maximum performance when plugged in.

Idling it takes 68 degrees on GPU and 50 degrees on CPU. GPU-Z showing 0-1% of GPU activity and CPU 0-1% activity. The best thing is that holding fingers to the exhaust the air coming out is much more warmer than before indicating that there is good contact with copper plate and GPU with CPU together.

Before this modifications using Google sketch up 8 and walking inside the default modeled house used to spike temperature of the GPU to 101 degrees before I stopped walking in the house to decrease temperature. I found a spot in the model where GPU is used by 100%. Now walking around gets about 78 degrees GPU load 100% RPMs around 3800 smart mode. Then the machine locks its self up after while, so there is definitely driver problem as well.

I'm worried about the driver because I found in event viewer that "atikmdag" crashed flashing screen and driver recovered and some BIOS ACPI EC problems. I have to find the culprit there...

[JoeHallenbeck]

Lock ups solved.

I was not able to walk in default house model in sketchup pro and play few seconds of Call of duty 2 even the GPU temperature was low. I uninstalled and cleaned entire package of ATI drivers and perform registry entry delete and installed all things one by one, first drivers from microsoft update catalogue and then catalyst center from lenovo support center. It stated that driver was incompatible with CCC and it was obvious from poor FPS in Call of duty 2. The driver version was 8.561.0.0 and lenovo CCC setup installed 8.383.1.1000 and now I have clear smooth FPS in Call of Duty 2 even during high smoke combat...

I hope it stays this way because I was so frustrated, I was thinking of reselling the x1400 motherboard but now it looks like it can be saved and turned to nice decent working T60...

Now I need SSD disk, old one in to the ultrabay and it will be best laptop for the money.