T21 - crash sometimes after boot, sometimes during boot

[bayankaran]

A week back I was decoding/encoding a DVD with an external DVD writer connected to my thinkpad T21. The DVD came out fine, but immediately the system crashed with the screen going ghostly white/whitish.

The CPU was working 90-100% capacity for almost 2-3 hours while doing the decoding/encoding process.

At times the system boots and works fine for 5 minutes, sometimes it hangs right in the middle of the boot process before the login screen appears.

There is not time to install a new version of the OS as the system crashes during the installation process.

I tried the following:

- tried another hard drive
- disabled lots of devices, especially USB ones which might cause such issues
- uninstalled video drivers
- opened up the system, cleaned up everything
- removed the fan/CPU case and cleaned up the accumulated dust
- put the system on the side, switched on the table fan and let the air blow to the CPU case and switched on the system

Still the same old problem - crashes while booting or 1-2 minutes into booting.

Does anyone has any idea what could be the issue?

[Orevin]

If you have a modem/ethernet card (mini PCI) installed, remove it and see if the laptop boots fine.

[bayankaran]

If you have a modem/ethernet card (mini PCI) installed, remove it and see if the laptop boots fine.

Took out the mini PCI and still the same problem. The fan starts in the beginning but immediately switches off as the whole thing hangs.

If the fan runs when you press the powerswitch means fan/cooling is okay?

[Orevin]

If you can hear/feel the fan spinning, it should be fine. Did you renew the thermal compound between fan and CPU?

[Thinkpaddict]

If the fan runs when you press the powerswitch means fan/cooling is okay?

Not necessarily. You could have the fan running like crazy, but if there is not appropriate contact between the fan sink and the CPU it won't do much good.

Your problem sounds a lot like the same problem I kept having with a T23 (random freezes). After a while I figured out that if I ran the laptop on a scheme with low CPU speed it would never hang. I then replaced the thermal paste in this machine with Arctic Silver. It seemed to help, but after a while I kept having the same problem, so I went back to running the computer with low CPU speed.

I think this problem might be related actually to thermal cycling in the CPU (large changes of temperature within relatively short periods), which cause the CPU to actually lose contact temporarily with the socket. I haven't been able to prove my theory though.

If I were you I would first try and run the computer on low CPU speed, and see if that ameliorates the problem. Or go ahead and reseat the CPU and replace the thermal compound between it and the fan.

I know you say that the computer sometimes hangs during boot, so it would seem unlikely that this problem is caused by thermal issues. However, I wouldn't discount it as a potential problem, especially if your CPU is not seated well in its socket it wouldn't take much to throw the thing off.

Good luck.

[bayankaran]

If I were you I would first try and run the computer on low CPU speed, and see if that ameliorates the problem. Or go ahead and reseat the CPU and replace the thermal compound between it and the fan.

Let me try adding thermal paste and reseating the CPU. Will give you a feedback once I do that. Many thanks for your suggestion.

[bayankaran]

While at the shop for buying thermal paste, I found a 2.5 inch HDD enclosure for $9. I took the T21 hard drive out, fixed the enclosure and connected to another computer.

Initially the drive came alive and I could see the folders. While trying to copy/backup the message "Delayed Write Failed" appeared. I reboot the machine and now the hard drive is not even getting registered as an external storage medium. It simply does not show up in Windows XP - My Computer.

Googling "Delayed Write Failed" does not show relevant information which could be useful.

[bayankaran]

The hard drive works fine...I could connect the drive to an external enclosure, copy all the data.

Opened the laptop, took out the CPU case, put thermal paste and connected everything back together. Now the laptop does not even do POST...nothing comes on the screen...you can see some small flickers after a few minutes. The fan works fine and I can hear it.

Any idea if its the motherboard or the CPU?

[bayankaran]

I bought thermal paste and applied and reseated the CPU. The first two times the machine would not start...did that once more and it started - apparently the screw seating the CPU to the socket was not tightened properly.

The machine works fine now...I have kept it running for about 2 hours...I am yet to enable all the devices. I put the wireless card for internet and that works.

It took me exactly one week to get it back to working.

Many thanks for all the thinkpads.com forum members who chimed in with suggestions and ideas.

What precautions I should take henceforth?

[Thinkpaddict]

Glad to know you have your Thinkpad back working.

As far as suggestions for the future, I would recommend that you download MobileMeter:
http://www.geocities.co.jp/SiliconValley-Oakland/8259/
Use this to every once in a while monitor your CPU temperature and make sure it stays within reasonable values.

Take care.

[bayankaran]

....monitor your CPU temperature and make sure it stays within reasonable values....

Really good utility...it says 46 degrees for the CPU...I guess it is a reasonable temparature!
http://www.pantherproducts.co.uk/Articl ... ures.shtml

I am also using SpeedswitchXP to reduce the clock speed to 650-675 from the regular 800.

[Thinkpaddict]

Hello,

That temperature (46 C) sounds about what my T23 gets idling after some time of being plugged in. Try to stress your CPU and see what is the maximum temperature you can get. (To stress the CPU you can use something like Prime95, or if you have Photoshop open a picture and resize it really big, something like 1500x1500 pixels, and then apply a radial blur filter at maximum quality repeatedly).

Out of curiosity, what advantage does that application that you mentioned have over using the regular speedstep schemes?

Please keep us updated as to the progress (whether your Thinkpad has fully recovered or not, hopefully yes )

Regards.