Is something wrong with my T60's cooling system?

[syncrotic]

First, tpfancontrol tells me that my GPU is at a temperature of 72C when I have no 3D apps running. It won't go below this temperature no matter how long I leave the fan running at maximum speed. If I start up google earth and leave the globe rotating, the GPU temperature will reach 95C; again, this is with the fan running at maximum speed. Question One: given that tpfancontrol is made for the T43, does the GPU sensor readout actually reflect the GPU temp on a T60?

Second, the CPU temperature can reach similarly dangerous heights under load. Running two copies of prime95, one on each core, the CPU temp hits 96C... that's with fan at full speed, and that's not a maximum; I merely shut down the test. So, Question Two: are other people seeing these sorts of temperatures under load?

Question Three: Is my T60 slowly frying itself?

If so, what can I do short of sending it back to Lenovo? The way I figure, the problem can be one of the following:

-Defective fan not pulling enough air.
-Poor contact between CPU/GPU and heatsink.
-Defective heat pipe.

How much surgery is required to get at the heatsink/fan? Is it relatively easy to remove and inspect?

By the way it's a 2623-D6U - that's the 14.1" model with a 1.83GHz CPU and ATI X1300 / 64MB. The machine is sitting on a wooden surface with the air vents unobstructed. As a final note, the CPU temperature normally doesn't go below 60C at idle, though it will get as low as 55C if I force the fan speed to max.

[christopher_wolf]

First, tpfancontrol tells me that my GPU is at a temperature of 72C when I have no 3D apps running. It won't go below this temperature no matter how long I leave the fan running at maximum speed. If I start up google earth and leave the globe rotating, the GPU temperature will reach 95C; again, this is with the fan running at maximum speed. Question One: given that tpfancontrol is made for the T43, does the GPU sensor readout actually reflect the GPU temp on a T60?

Second, the CPU temperature can reach similarly dangerous heights under load. Running two copies of prime95, one on each core, the CPU temp hits 96C... that's with fan at full speed, and that's not a maximum; I merely shut down the test. So, Question Two: are other people seeing these sorts of temperatures under load?

Question Three: Is my T60 slowly frying itself?

If so, what can I do short of sending it back to Lenovo? The way I figure, the problem can be one of the following:

-Defective fan not pulling enough air.
-Poor contact between CPU/GPU and heatsink.
-Defective heat pipe.

How much surgery is required to get at the heatsink/fan? Is it relatively easy to remove and inspect?

By the way it's a 2623-D6U - that's the 14.1" model with a 1.83GHz CPU and ATI X1300 / 64MB. The machine is sitting on a wooden surface with the air vents unobstructed. As a final note, the CPU temperature normally doesn't go below 60C at idle, though it will get as low as 55C if I force the fan speed to max.

You got it nailed on the first guess I should think.

Given the reports of really high GPU temps knocking the upper end of the specturm, I would say that the sensor codes changed. Since it was designed for the T43, it was labelled as such along with the temp fields so it is quite possible it is reporting something different.

Yet if the CPU is *that* hot with both cores going, I would give serious thought to calling up IBM/Lenovo.

[sugo]

What is the ambient temperature where you use your T60?

I don't think you are alone:
http://forum.thinkpads.com/viewtopic.php?t=25304

[own6volvos]

While the GPU might be running a tad hotter than one might like, remember that GPU's tend to run on the hot side to begin with. My 6600GT AGP card in my desktop idles around 61C or so, and goes upwards of 80 during heavy gaming. It also has a pretty big fan and heatsink on it.

Also, while the insides of the t60 are pretty similar to the z60m, and while they practically share heatsinks and whatnot, the T60 reported temps close to 20-30C higher than the z60m.

When my t60p arrives, I will be using a infrared gun on the heatsinks to see if they are anywhere near the temps they are reporting as.

[Badger]

does your computer feel hot(ter than normal)? maybe the sensors are being misread.

[syncrotic]

I don't really know what's normal for a T60 - I know that the dual core prime95 test definitely made the underside of the machine uncomfortably hot to the touch.

The ambient temperature isn't an issue - it's 25°C at most in my room. One possible consideration: i'm at about 1700m elevation, so the air is a bit thinner up here... less effective air cooling? Anyone from Denver want to comment on whether this is significant?

From the previous thread on this subject, I see that it's not uncommon to have GPU temps in the 70s... and that 90 degrees under load is acceptable for a GPU... on the other hand, no one mentioned CPU temps getting to 95°C, idling at 60... it seems there aren't any easy answers.

I'm tempted to try re-seating the heatsink and seeing what happens. Any other suggestions? Comments? Anecdotes?

[own6volvos]

Just remember that reseating the heatsink is very risky unless you know what you are doing. Any damage from broken plastic or chipped pieces will void your warranty.

Also, unless you plan on cleaning the heatsink, and maybe applying artic silver, reseating it will probably do more harm than good. A heatsink compound application is only good for one press down.

[Anthony S]

...I would say that the sensor codes changed. Since it was designed for the T43, it was labelled as such along with the temp fields so it is quite possible it is reporting something different.

I think this observation is very pertinent. I'm inclined to agree that the readouts from tpfancontrol are misleading for the T60.

My T60p always runs cool (never blows hot air) and the over-active fan really doesn't need to run at all most of the time. Maybe the fan controller is also relying on inaccurate readouts?

[christopher_wolf]

...I would say that the sensor codes changed. Since it was designed for the T43, it was labelled as such along with the temp fields so it is quite possible it is reporting something different.

I think this observation is very pertinent. I'm inclined to agree that the readouts from tpfancontrol are misleading for the T60.

My T60p always runs cool (never blows hot air) and the over-active fan really doesn't need to run at all most of the time. Maybe the fan controller is also relying on inaccurate readouts?

Well, that's the thing; I don't know what the EC code does with the fan for a T60. I already looked through the microcontroller code for the H8S; does the T60 use the same EC?

But before we worry about any of that, the sensors have to be re-mapped for the T60 if they did, indeed, change the sensor codes (which they probably did). That requires opening up the T60 for the harder-to-find ones, although those were probably the ones most likely to have their codes changed.

[rvacha]

While the design is of course totally different, it is interesting to see that Macs run typically lower - but not always. See this page for an enterpising guy that has set about to log Mac Core Duo temps:

http://www.intelmactemp.com/list

Note that you can filter the list by notebooks only

[archer6]

While the design is of course totally different

The statement above says it all, they are different. Therefore the comment below negates it all.

it is interesting to see that Macs run typically lower - but not always.

What these two have in common is the fact that they are both notebook computers, however beyond that it's two completely different machines. Void any comparison.

[rvacha]

I wouldn't have posted this unless I thought there were some conclusions that might be drawn, mostly that the Macs may have better thermal designs than Lenovo. A second conclusion which I think is valid is that regardless of the design, there seems to be a lot of variance in measured temps. This could indicate that either the temp sensors on board the CPU and GPU are rather sloppy and inaccurate, or there is a lot of variance in MoBo mechanical assembly, or both. While my machine runs high (supposedly), it is not high enough that the CPU is thermal throttling to keep from melting, at least according to RightMark CPU Clock

[own6volvos]

I wouldn't have posted this unless I thought there were some conclusions that might be drawn, mostly that the Macs may have better thermal designs than Lenovo. A second conclusion which I think is valid is that regardless of the design, there seems to be a lot of variance in measured temps. This could indicate that either the temp sensors on board the CPU and GPU are rather sloppy and inaccurate, or there is a lot of variance in MoBo mechanical assembly, or both. While my machine runs high (supposedly), it is not high enough that the CPU is thermal throttling to keep from melting, at least according to RightMark CPU Clock

I am about to fall out of my chair, as Apple of late has been horrible with cooling systems on laptops. Macbooks and Macbook Pros idle on the thermal limit temps, and start to throttle under partial load.

Just google up the heatsink compound issues and see the mess those things have caused.

Lenovo and practically every other maker is worlds above Apple.

[young guy]

Is anyone using Fan Control and stop the over-active fan? I'm curious of the results if anyone has tried this It just doesn't seem right to have the fan running all the time even when my computer just sits there idle for a period of time.

JD.