T22 dies after a few minutes

[sperla]

Hi,
I have a T22 that seems a bit flaky. It boots up fine and then dies without any error message - the LCD just goes blank. I connect it to an external display and there's nothing there as well. The fan does not seem to be running properly - I could hear it 'creaking' - it seemed like it was trying to run. I suspect the laptop might be overheating. Is there a thermal sensor that might have gone bad ? Or is it just the motherboard gone bad ?

Sometimes, the laptop dies while booting itself with a couple of beeps. The HMM indicates the system board might be bad, but since this doesn't happen always, I'm not sure if I should go ahead and replace the motherboard. Is anyone familiar with this sort of problem.

thanks,
sp

[rkawakami]

Searched your earlier posts about this problem and your use of MobileMeter to determine the CPU temperature. If MM still is reporting sub-65 degree temps when the laptop shuts down, then there may be a couple of problems.

Refer to the following thread where I posted information on where to download PC Doctor and how the run the fan and thermal sensor tests:

http://forum.thinkpads.com/viewtopic.php?t=31668

If both of these tests report as passing, then I would assume that there is something wrong with your CPU and/or motherboard. You can try running several of the other tests in PC Doctor (CPU, systemboard, memory) to see if you can pin down what/where your problem is.

[sperla]

Thanks for your quick response. Yes, I did post a while back, but since then, the laptop has been dying faster prompting me to think about the motherboard replacement.

I will try your suggestions to see if I can narrow down the problem and report the results.

[sperla]

Hi,
Sorry for the delay in updating this post. But my laptop refused to stay on longer than a few minutes at a time for me to do a diagnostics test. So, I gave up and decided the system board had to be replaced. I hunted around Ebay and bought a system board.

I was going to disassemble the laptop when I thought of giving one more shot at the diagnostics. And guess what, the laptop booted fine and stayed on. In fact, I restarted quite a few times while I fiddled around with making a bootable disc (using another laptop) and changing the boot order. Finally, I got PC-Doctor to run and here are the results:

Thermal sensor test : Passed
Fan test : Passed

I then went into the Systemboard menu (under Diagnostics) and ran all the tests (F3/F5 I think). All of them passed.

Is there some other test I should run ?

Now, I'm in a dilemna - do I go ahead and replace the system board or do I keep using the laptop till it fails and then replace it. The place I bought on ebay has a return period of 2 weeks.

I have a feeling it might fail again (just because it has in the past), but I want to be sure replacing the system board would fix it. The last few times it crashed, it died with the one long, two short beep sympton

If I do need to replace the system board, is there anything I need to be aware of or prepare for ? I've added/replaced cards on my desktops, but have never opened up a laptop. I did get a couple of useful pointers from these boards regarding the serial numbers and the backup battery and it looks like I don't need to worry about either.

The other thing I realized while reading the HMM regarding FRU replacement was that I needed thermal grease to put the fan back on. Is this necessary ?

The HMM says the screws cannot be reused. Is that to guarantee their reliability standards ? Can we reuse the screws if we can live with the 'lower reliability' or is there something about them that prevents them from being reused ?

The HMM is pretty detailed, but is there any guide on system board replacement - thought I saw one on these boards, but can't find it anymore.

Appreciate any comments and/or suggestions.

regards,
srinivas

[rkawakami]

Finally, I got PC-Doctor to run and here are the results:

Thermal sensor test : Passed
Fan test : Passed

I then went into the Systemboard menu (under Diagnostics) and ran all the tests (F3/F5 I think). All of them passed.

Is there some other test I should run ?

As long as you either run the "Normal Test" suite or manually select the CPU/Coprocessor, Systemboard, and Other Devices (AC adapter, Battery, Thermal Sensors and Fan tests), then I would say you've covered most of the hardware. What's left, hardware-wise, is a test or checkout of the power systems. I'm assuming that you're using an AC adapter to power the laptop. Can it be powered from just the battery alone? A sudden loss or reduction in power could cause some of the symptoms you describe.

Now, I'm in a dilemna - do I go ahead and replace the system board or do I keep using the laptop till it fails and then replace it. The place I bought on ebay has a return period of 2 weeks.

I would swap out the system board just to see if the eBay board is functional. If it runs fine for the next week or so, then I would say you fixed your problem. If it doesn't you can return it with a clear conscience that it didn't work as advertised. Did that board come with its own CPU and fan?

I have a feeling it might fail again (just because it has in the past), but I want to be sure replacing the system board would fix it. The last few times it crashed, it died with the one long, two short beep sympton

That beep code is normally what you would hear during POST if there is a systemboard or memory problem. If there was a sudden, complete loss of power, you shouldn't be hearing anything. If your "new" motherboard came with another fan, try replacing that first before tearing down the entire system.

If I do need to replace the system board, is there anything I need to be aware of or prepare for ?

Lots of screws! Prior to disassembly I would get something that will keep all of the screws that you remove for each FRU separate. Either paper cups, a desk organizer tray, a partitioned utility box, etc. As you follow the HMM, put the screws into a marked cup/area with the FRU it's associated with. Makes it easy to re-assemble later and you shouldn't have any "extra" screws when you're done .

The other thing I realized while reading the HMM regarding FRU replacement was that I needed thermal grease to put the fan back on. Is this necessary ?

While it's not absolutely required to get the laptop up and running, you are better off if you can find some of this stuff before the teardown.

The HMM says the screws cannot be reused. Is that to guarantee their reliability standards ?

Personally, I think it just IBM's (and eBay sellers) way of trying to extract more money . I've taken apart a fair share of laptops and re-assembled them. Up to now I haven't noticed any more "looseness" to them. Yes, the screws would now have a greater chance of vibrating loose but I'm not subjecting them to anything that extreme.

The HMM is the only thing I've seen which details the inner workings of the laptop. I'm not aware of any other guide/manual outlining a motherboard replacement. Take your time, work slowly, carefully document where each group of screws came from, and you shouldn't have any problems. Try to be aware of any static charge that might build up and dissipate it before touching the laptop. That means don't shuffle your feet along the carpet or slide off a fabric-covered chair and then immediately grab the motherboard. Touch a grounded piece of metal (computer case) or a nearby family member (!) first.

[sperla]

Thanks for your detailed response, Ray.

In response to some of your questions:

rkawakami>>Can it be powered from just the battery alone? A sudden loss or reduction in power could cause some of the symptoms you describe.

The battery holds very little charge, so I generally have it connected to the power supply. It used to fail even on power supply.

rkawakami>>Did that board come with its own CPU and fan ?

No, it was just the bare board.

I ran all the tests available in PC-Doctor and all of them passed.

Now, on to today's experiments:

Yesterday, when I ran the diagnostics, I hadn't run the mobile meter again to watch the temperature. I tried that today and am not sure how to interpret this:

1. I first booted up the laptop and started mobilemeter. It showed the temperature to be around 47C. This is without any other programs running. But, surprisingly, it showed the frequency to be 700Mhz. I checked in Control Panel->System and it showed it to be around 696MHz as well. I am fairly certain this is a 900MHz laptop. I then started uninstalling some programs that I didn't use (in the hope that there might be some hidden program that was consuming cycles and contributing to the laptop heating up). All the while, the temperature hovered around 50C and the frequency at 700MHz.

One of the uninstalls required a restart, so I rebooted the laptop.

2. On restarting, I started mobilemeter to keep track of the temp. But now, it showed 900MHz and Control Panel->System also showed 900MHz. However, the temperature starting to rapidly go up. The fan kicked in at some point above 60C, but seemed to have no effect as the temp. kept rising. I was thinking of shutting it down (to prevent further damage), when at 84C, it dialed down to 700 MHz and the temp. started dropping as fast as it had risen. It went all the way down to 41C and then stabilized around 45. Control panel->system still showed it to be 900MHz. The fan also turned off.

I continued with the uninstalls (temp. moved up to ~55C) and at one point up, opened up Windows explorer to view some files. The temp. started rapidly rising and the frequency also bumped up to 900MHz. Task mgr indicated explorer.exe was taking up 99% of CPU. At about 84C, I closed the explorer window, but the process was still running (shown in task mgr) and still consuming 99% CPU. The temp. rose up to 90C at which point I killed the process (ignoring the dire warnings about killing processes). Temp. immediately started dropping (as did freq to 700) all the way down to ~45C.

The menu/toolbars all disappeared (prb. due to killing explorer.exe) - so I restarted explorer.exe (File->New Task on Task mgr). Menus reappeared and everything was fine. Started Windows explorer and same scenario above repeated (freq up to 900, temp. rapidly rising, explorer.exe taking up 99% CPU). Killed explorer.exe at 75C itself and waited for the temp. to go down. One difference between the first and second time was that explorer.exe showed up in 'caps' in the task mgr the second time. Not sure if that is significant (?)

Repeated this one more time and same results. This time, I noticed the heating up started not as soon as I opened Windows Explorer, but only when I accessed a particular directory (that had some huge movie files). One possibility is that these files are heavily fragmented and this caused repeated hard drive accesses (hard drive temp. did not go up - assuming HDD#0 is hard drive temp in mobilemeter)

Now, the questions:

1. Is the freq behaviour related to Intel's speedstep or is there some other overclocking/underclocking going on ? I haven't fiddled with any of these things.

2. This seems to point to the fan not being effective, but I could hear the noise of the fan - so am not sure.

3. It also points to explorer.exe getting 'stuck' and consuming all CPU cycles - any clues ?

4. It seems to indicate the laptop is heating up when running at 900MHz. What is the steady state temp. at 900MHz ? Is there a way I can keep the freq. at 700MHz (I don't mind that if it means I can prevent overheating). Does this have anything to do with XP ?

5. Yesterday, my hypothesis was that though PC-Doctor indicated the thermal sensors were working, it could only have checked if the thermal sensors' reporting circuitry was working (by triggering some software bit that triggers a response). So, if the actual sensor wasn't working, it wouldn't detect rising temperatures (or atleast not effectively) thereby causing the laptop to overheat. But today, since the fan was running above 60-65 (didn't notice exactly when it started), I'm not sure about this.

My laptop config:

T22, 2648-E7D, P3 900MHz, 256MB, XP.

Thanks for reading the long post (I figured more detail was better than less). Any comments/suggestions welcomed and appreciated.

thanks,
srinivas