R52 (hardware) bluescreens

[BlueScreen]

Dear Alll,

My R52 Thinkpad (ca. 3.5 years old, Windows XP Pro) has been suffering from random blue screens (with mains power or battery, idle or running programs). Was happening about once every 2 days or so, but getting a bit more frequent now.
I have checked the crash dumps (MS debugger) and there seems to be no particular pattern in the implicated process for each crash.

I ran tests on the RAM (MS DOS utility), hard drive (chkdisk, scandsk), and a full check with PC Doctor (Thinkpad version), but all tests passed (+ different virus scan programs).

To cut the story short (drivers, registry, etc.), after a few weeks of this I reset the whole system back to the factory state (with IBM recovery), i.e. format HD and original OEM Windows XP (which worked).
After a day or so, I got my next blue screen...

So, it seems that the problem is due to hardware, but not showing up in RAM/HD tests (nor any of the other hardware tests PC Doctor performs). (For what it's worth, the crashes happen both at home and work, i.e. with different USB peripherals.)

Could someone please suggest what it the problem might be and how to proceed?

A few notions of what the problem might be that I have collected while working on this problem:
- Faulty motherboard component (could this be found and repaired by service personnel...no longer under warranty...)
- Boot sector virus or similar too clever for my virus scan programs? (Norton, Sophos)
- Battery condition upsetting power (even when on mains adaptor)?
- Bad RAM contacts (should I pull these out and inspect them?)

Perhaps hopefully the blue screens will become more frequent: at present a service technician might have to wait 6-48 hours for the next crash!

Many thanks for any help (or sympathetic words!), MT.

[Neil]

The last system that I worked on with BSOD problems like you describe, belonged to my mother-in-law, and the problem turned out to be a faulty hard drive. It had a couple of bad sectors, and whenever the OS tried to access those sectors, it would crash. There is no guarantee your problem is the same, just sharing my experience.

[RealBlackStuff]

What is the 7-character TYPE no (on the bottom, looks like e.g. 1852-R5U)?
Do you have Intel or ATI graphics? If ATI, it could be the GPU coming loose.
That requires reballing or a new motherboard.
See also http://forum.thinkpads.com/viewtopic.php?f=2&t=73945
and http://forum.thinkpads.com/viewtopic.php?f=2&t=57021

[sjthinkpader]

You can check HDD health using HDTune. This is a freeware.

[BlueScreen]

Hi there,

Many thanks for the rapid and helpful replies. I really do appreciate the time taken and experience provided here.

So, while the tests with "HD Tune" are quite suggestive of a HD problem, I should perhaps address the other posts first.

RealBlackStuff: Type = 1847-CTO
Graphics adaptor: ATI Mobility Radeon X300
Depending on the next run of responses, I would have a few questions (regarding the GPU coming loose, what "reballing" actually is, or the cost-logic of replacing the motherboard out of warranty).

Anyhow, turning to the hard drive tests. Thank you 'sjthinkpader' for the tip to try "HD Tune" - seems to be a very handy utility. I got a few blue screens while using the its tools, but I don't know how systematic/conclusive this is.

Here are the results of my "HD Tune" tests (restarts and some more reboot BSODs omitted):
Drive shown as Healthy
Run Benchmark: Blue screen halfway through (besides a few dips in the speed...)
Run Benchmark: Success
Run Errorscan (full): Blue screen during
Run Benchmark: Success
Run Errorscan (full): Blue screen (when I came back...silly me to go eat)
Run Errorscan (full): No problems but I got chicken when HD temp reached 57°C and stopped it
Run Errorscan (quick): No problems

Assuming it is the HD, what would be the next step?
- Can one somehow locate and "omit" the bad sectors, or is a new HD required?
(I have no precise diagnostic results from HD Tune, except for "I had/had no blue screen...")
- Can I order and install that HD myself?
- Is there a way to copy/burn my IBM hidden recovery partition onto another drive to restore the OEM WinXP on a new HD?
- BTW: What might be the reaosn that chkdisk/scandsk do not suggest problems?
- Could the BSODs be coincidental/side effects?

Is that HD temperature hike during the tests normal...or even suggestive? (57°C == 135°F )

Many thanks again, MT

[Neil]

I've successfully recovered a couple of bad hard drives using HDD Regenerator. But first, to determine if the HDD is where you problem is, try booting from a live CD of some sort, and if you don't have any BSOD issues, problem must be with HDD. But, if you still get BSOD when the HDD isn't even running the trouble has to be elsewhere.

[BlueScreen]

Hi Neil,

I'll have a look for HDD Regenerator...wherever it might reside....

I think I have never had a crash running diagnostics from boot disks...but I'll try harder!

Thanks all again, MT

[BlueScreen]

Hello,

I scanned the disk using a trial version of HDD regenerator (DOS) and it found no errors.

So I am still left wondering what the actual problem is....

Regards, MT

[sjthinkpader]

Normally my HDD is running around 35 deg C. A spin down timer is set for both battery and A/C power sources. The key is the amount of reseek and read errors HDTune registers. If the amount of running hours is high, you should consider replacing it anyway.

Can you run PC Doctor from CD to see if it crashes? Are you using NHC to monitor temperature? Is your power profile set to Performance or Adaptive?

[BlueScreen]

Hi there,

Looking in the Health area of HDTune, the "Raw Read Error Rate" and "Seek Error Rates" both are listed as 100 (I assume this means 0%...as the results are "Healthy").

Tests using PC Doctor from CD boot disk passed fine, as per running under IBM Rescue and Recovery boot. I installed the latest Lenovo System Toolbox and everything passed as well.

I guess I have been using the Performance power profile (should have written this down before I started playing with that 2 days ago) - when I set it to Adaptive the CPU clock rate appears to go down to 800 MHz (using MobileMeter now to watch what's happening). Is that normal?
(BTW: I have to guess the English terms because this IBM is in German...)

I had started to wonder if it the crashes are somehow triggered thermally, although they don't seem to occur more often when programs are running heavily. When I first checked, my CPU was running at around 78°C without applications running. I left it in Adaptive with 800 MHz clock for 24 hours (CPU settles down to about 60°C), and no BSOD yet...
However, the statistics of these BSODs has been very erratic, so 24 hours is not convincing me yet...

For what it's worth, my fan never seems to freak out...

I also pulled out the "accessible" 512 MB RAM 2 days ago to eliminate that - still had a blue screen after that. Also ran memtest - no errors.

If it continues not to crash at low CPU and I get another crash in Performance, then I will open up the computer (warranty lapsed already) and look for excessive dust and perhaps put new thermal paste under the CPU. Do you consider that a sensible plan of action?

With kind regards, MT

[sjthinkpader]

78 deg C while idling is too high. You need to clean out the fan/heatsink and re-apply thermal compound.

Running "Adaptive" does not lose performance since the system will use max clock cycle mode when needed.

[BlueScreen]

Hi there All,

Well, since I have been using Adaptive CPU power settings, I haven't had a single blue screen.

I opened up the computer (R52) a few days ago in order inspect the fan and CPU, remove dust and apply new thermal paste under the CPU.
However, I couldn't remove the fan... but I'll start a new post for this new story... (so I just gave it a darn good vacuum cleaning with the fan/CPU still mounted)

Many thanks again for the helpful replies. MT

[sjthinkpader]

After removing the screws, grab top and bottom of the heatsink over the GPU area with your finger tips and squeeze. It should come off.