Thinkpad T42 repeatedly needing hard reset.

[JPDyno]

Hi Guys,

Ive just joined the forum in search of some more information on a problem Im having.
I have a Thinkpad T42 2372-8YM (one of a fleet of about 18 thinkpads we have in our workplace)

In the last week it has turned off by itself, while plugged into an outlet, atleast 4 times. Each time needing a 'hard reset' (Im not sure what its proper name is, but its when you need to remove your battery and power cord and hold the power button down for 30sec or so, then plug the power back in to boot up)

As far as i can tell from the users (I have not been using the computer at the time of the crashes) The CPU or graphics hasn't been under load. It was running some milling software the first two times, The next two times it had been there sitting idle (powered on, most of our laptops spend alot of time powered on but idling). No one moved or even touched it. We just found it powered down and could not power it up again without a hardreset.

Ive had to swap the laptop for one of our spares, as we need reliablity when its running the milling software. However I cannot seem to recreate the crash.

Any idea of usual suspects when T42s go into this hard-protection mode??

The laptop is only just (ofcourse) out of warranty. If its a hardware problem i may still be able to get it fixed.

[zdriver]

It sounds like it could be an overheating problem. Run a program such as MobileMeter to monitor the temps and see if they are abnormal.
When was the last time the fan & heatsink assembly were cleaned?

Also, is the thinkpad running out in the harsh shop enviornment or in a relatively clean office area?

[JPDyno]

To be honest, i have no idea how long since its been cleaned.
Ive been here six months, and it hasnt been cleaned in that time.

Its a fairly harsh enviroment. Harsher than a normal office anyway. Not oily or dusty, but we have alot of clay around, and apparently some computers have had issues with the tiny amount of sulphur in the air previously.

thanks for the name of some temp monitoring software, I was just searching for something to do that.

EDIT: Also i thought it was worth noting, The area it works in is air conditioned, and the laptop runs in a free air position.

[richk]

Usually, overheating does not require battery removal.

[JPDyno]

Well I downloaded the mobmeter software,
On idle it sits at 70

I opened four antivirus windows to stress the system.
I managed to get it to 88, when it slowed the processor down to 1.4Ghz, and quickly cooled down. Still sitting on 86 now.

70 idle seems high, but it wont crash or shut down again.

EDIT: sooo... apparently 70 C is pretty darn hot. I booted another T42 and its sitting at 47 C on idle.
Will pull the hot one apart and see if its a mess. May change the thermal paste too.

[JPDyno]

Well I pulled it apart and removed the heatsink. It wasn't dusty or clogged up at all, seemed very clean actually.

Quick clean and some fresh thermal paste, put everything back together, and it still idles at 70 C

hmm...

[zdriver]

70 degrees C is very hot for idle, even under load. Should be somewhere in the 40 degree range (idle @ 600mhz)
Either the fan is not performing at the correct speeds or there is a problem with good thermal contact between the cpu and the heatsink.
Does the fan appear to be running at full speed?
What is the ambient room temp where the thinkpads are running, sounds like it is on the warm side if the other notebook idles at 47c.

I know it is a lot of extra work, but you could try installing the other thinkpads fan assembly in the problem unit to see if the temps drop to normal and help narrow down the problem.

[JPDyno]

Room isnt too warm. Climate controlled at 24 C.

The fan doesnt seem to run at full speed when the temp is at 70, But once it heats up even a little (say 72) it kicks in and seems to be working fine.

Im currently pulling apart an old T42 that the screen had died on and Im going to swap the heatsink/fan assembly.

I'll take the CPU out and sit it aside too, just incase its not the HS/fan thats the problem, I can swap that in.

I must admit, as i said earlier Ive only been in this job for 6 months and its the first time Ive had dealings with thinkpads (well, since i had a cheap secondhand one when i was 16 anyway) and Im blown away by how easy they are to pull apart, and swap out parts. Most laptops dont go back together real well after a couple of times. I know thinkpads have a reputation for being rugged, and well designed, but i didnt expect this, Im very impressed.

[JPDyno]

Now Im stumped.

something must be preventing the heatsink from mounting properly in the problem laptop.

Ive tried the heatsink/fan and CPU from a spare laptop, and i got 75 C plus. So i put that HS/fan back into the spare laptop and get under 40 C. Fan isnt even coming on. So that eliminates the HS/fan and cpy

Ive reinstalled the original HS/fan into the problem laptop...3 or 4 times... and i cant get any temps lower than 70 C.
The fan seems to be pumping out air, the heatsink is getting warm so it has some contact. But i cant figure out whats keeping it from making good contact.

[zdriver]

When you remove the fan/heatsink from the problem thinkpad, if you examine the underside small square area where the cpu makes contact, the trace or impression left from the conductant/thermal compound should give you some idea of what the problem is.
Are you useing a quality compound such as Artic Silver5, carefully applied (very very thin layer - almost just more than a mist)? Heat sink screws properly tightened? I assume you are probably good on the last two items because you are getting normal results fron the other thinkpad.
Is it the long fan saaembly, that also covers the ati gpu?

[zdriver]

When you remove the fan/heatsink from the problem thinkpad, if you examine the underside small square area where the cpu makes contact, the trace or impression left from the conductant/thermal compound should give you some idea of what the problem is.
Are you useing a quality compound such as Artic Silver5, carefully applied (very very thin layer - almost just more than a mist)? Heat sink screws properly tightened? I assume you are probably good on the last two items because you are getting normal results fron the other thinkpad.
Is it the long fan saaembly, that also covers the ati gpu?

[richk]

It is also possible that the thermal sensor has failed.

[JPDyno]

I disabled the fan on the problem laptop, and measured the heatsink temp of both the problem laptop and the working one (whose CPU is cool enough for the fan not to come on) using an IR thermometer. They're both within a few degrees of eachother.

thats making me think that the heatsink is making contact, and the only reason it was cooler to touch was due to the fan being on.

Which means, yeah, its possibly the thermal sensor.

Where is the thermal sensor for the CPU?
I notice there is nothing in the middle of the socket under the chip.

[zdriver]

Possible, but not likely since the pentium m has 2 thermal diodes integrated and both would have to be bad to tell the thinkpad to shut down when the max junction temp is reached. He has already switched cpu's and the problem followed.
It is looking like a motherboard problem because more than likely the cpu is contacting the heatsink reasonably well. Possibly a DA coversion problem with the board.
I would run the pcdoctor dos diags on the mainboard if it stays up long enough and see what it says.

[JPDyno]

Heatsink has a good impression of where the compound is. its a very thin layer, and Im using a thermal compound that our electronic instrument techs use in engineering. They dont normally skimp on quality of this stuff (I know, i used to work there. lol)

So the CPU temp sensor is in the chip? I thought it might be. couldn't find one anywhere close. Starting to think swapping the mainboard is the way to go.

I didnt think of PCDOCTOR. I'll do that now.
PC will stay on. It still hasn't shut down for me yet, Right now Im just assuming that since its reading this hot, that its likely the reason for the crashes.

[JPDyno]

PC Doctor diag says the thermal sensors (and everything else) are good.

Also has a "thermal shutdown by sensor" test, which does require removal of the battery to restart.

So yeah, from what I can tell, its all working fine except its reading the sensor high for some reason.

Looking more and more like the motherboard swap is the way to go.

[zdriver]

Maybe the best solution at this time is to utilize you spare T42 with the bad lcd screen and simply swap screens with the good one from the problem thinkpad (assuming they are both the same size)
Move over the drives, ram, & other accessories and you should be up and running without problems. Easier the switching motherboards and eliminates the serial number problems etc.
Then you can revisit the suspect motherboard at your leisure and when you have some downtime at work.

I give you alot of credit for the logical way you troubleshot the problem and the excellent feedback of important information.
Best of luck keeping things running....

[JPDyno]

I thought about swapping the screens over, however the spare T42 with the bad screen isn't in the best of conditions. It looks like its been through a war. chassis, case and the rest of it. *shrugs* both out of warranty anyway.

So i swapped Mainboards today. That was an experience. lol
Anyway, as suspected, the problem followed the motherboard. Good T42 is idling around 40oC now.

So the crappy spare parts laptop has just gotten a bit crappier, and now idles at 70.

Thanks for all your help. Pretty fast responses here, Im used to most forums taking days to help! As Im dealing with a number of these laptops now, I presume Ill speak to you again at some point!.

Now Im going home to have a bourbon.
cheers!