Why is this laptop overheating?

[shamrin]

I've got a Vista-era Toshiba laptop here that appears to be overheating. The customer brought the unit in because it was shutting down spontaneously on him. It sounded like a heat problem and tests like a heat problem. If I stress the CPU and watch the temps (using Speedfan), as soon as they hit 82C the computer shuts off.

• This unit uses an Intel Centrino Duo. Intel chips should slow down if they are overheating not shut down.
• Even when the temps are showing hot, the exhaust from the heatsink/fan never feels hot when you put your hand by it

I've re-seated the heatsink, it looked fine when I was working with it but the re-seating does not seemed to have helped. I have also tried stressing the GPU but that doesn't seem to make the computer shut down.

I'm looking for ideas here...

 

[iisjman07]

I'm assuming you've cleaned out all the dust and replaced the thermal paste?

 

[FoolishTech]

if he has done that, my best guess is the heatpipe has an internal fracture and needs replacement.

Seen this a lot with heatpipes (the pipes that direct heat from the cpu to the actual heatsink/fan assembly)

You can always tell with a temp probe - point it at the plate on top of the cpu and it reads hot as hell, follow down the pipe and just before the pipe connects to the heatsink you would see the temp drop dramatically.

In a working heatpipe the temp should only drop a few degrees at most between the CPU plate and the bit just before the heatsink itself; my guess is on your rig it drops a lot more than that.

 

[SilverLeaf]

Could very well be ACPI related (maybe disabled in the bios). Usually ACPI would control the temp via fan speed, clock speed, etc. If ACPI is not functioning, the next line of defense is hardware based, and will shut down the system in a severe thermal event.

 

[NYJimbo]

We are getting tons of the L640/L645 series coming in with overheat issues. Cleaning the fan and regrease/pad everything seems to work but then the customer comes back. Replace fan and that seems to help more, but the original fans are always spinning as fast and are almost perfectly clean.

In my case it seems that many Toshiba's are "on demand" fans and the fan does not spin up quick enough to make up for the sudden increase in heat. Also I have read for this series that sometimes when coming out of sleep you will get a "system cooling has failed" or something similiar and wants to shut down.

So what I am finding is that even though the fan looked ok, replacing it made a big enough difference to keep it from overheating.

What model is that laptop ?

 

[shamrin]

if he has done that, my best guess is the heatpipe has an internal fracture and needs replacement.

It never occurred to me that if the pipe and heatsink looked OK that you could still have a problem. But with the exhaust not being hot, it does feel like we're not actually extracting heat from the CPU. This seems worth a try if the customer is game.

Could very well be ACPI related (maybe disabled in the bios).

No ACPI in the BIOS. I do have "Dynamic CPU Frequency Mode" which, if I disable it, seems to keep the CPU in some low-speed mode and prevents the computer from overheating. It also reduces the performance by around 25%. (IM me if you want a tip on a certain construction company in Danville that you want to be careful in doing business with)

We are getting tons of the L640/L645 series coming in with overheat issues. ...

What model is that laptop ?

This is a Satellite A305. It does have that on-demand fan. I don't detect that it cranks up as the computer approaches thermal meltdown but it's going at a reasonably high rate right along. It occurred to me that maybe it's not being controlled properly but the fact that the exhaust isn't hot anyway sort of trumped that observation.

 

[NYJimbo]

It occurred to me that maybe it's not being controlled properly but the fact that the exhaust isn't hot anyway sort of trumped that observation.

Just another note. We had a laptop in on Monday that was overheating. Temp were going crazy but output was not really hot. Turns out the GPU was not making enough contact with the heatsink and that was part of the whole cpu/gpu copper assembly. The pad looked ok, but really was not. So the heat from the GPU had no where to go.

Venting temps were low but the heat remained on the GPU and radiated to the CPU internally so everything overheated. This can also happen just with the CPU, temps are high internally but venting gases are not hot because of bad contact with the copper cooler.

Always regrease/repad/clean vents or those bastards will come back to haunt you.

 

[technibbling]

if he has done that, my best guess is the heatpipe has an internal fracture and needs replacement.

Never thought of that one! Thanks!

 

[technibbling]

Always regrease/repad/clean vents or those bastards will come back to haunt you.

Fully noted 20 characters

 

[shamrin]

Defective heatsink, who would have thought?

if he has done that, my best guess is the heatpipe has an internal fracture and needs replacement.

Spot-on FT! There were no apparent problems with the heatsink and heatpipe but the replacement heatsink has solved the problem. I'm sure, if you're set up to do it, measuring the temps along the heatpipe is a good diagnostic method, I think the symptom of overheating and having the exhaust stay cool is a good indicator to consider this solution as well.

I did put a copper shim on the GPU while I was in there so I'm sure that helped as well but I would not have thought to change the heatsink on this one, thanks!