A31p locks up (bad system board?)

[mediasponge]

I've seen several posts on similar topics. I'd like some advice, and possibly help figuring out what my options are. (First post, be gentle )

System described in my sig line is A31p 2653-N5U, 1GB IBM Ram, 120GB aftermarket (Fujitsu) HDD, XP-Pro SP2, Fedora Core 6.

Symptoms: Machine locks up hard at some random time after booting up, maybe minutes, maybe hours. Fan OK, All PC-Dr diagnostics pass. When this happens, the screen and cursor just freeze where they are, no funny graphics glitches, no blue screen, no black screen, nothing in the system logs (event viewer). The only cure is hold down the power button to turn it off, then turn it on again, when the cycle may repeat at some random interval. It seems like it never locks up when booted from a CD, only when booted from the HDD, which is new.

What I tried: Remove all daughter cards (wireless, BT, ethernet), CDRW, swap Ram cards, try one at a time, etc. No luck. BTW, it's not a Windows problem, because it happens in Linux too. I stripped the entire system down to bare mobo, cleaned it, put new thermal grease on the fan, put it all back together OK. (I'm good at this. I even got all the screws back in the right place ) The only thing I found amiss was that the thermal tape over the ATI chipset was squashed and oozing slime. The tape is apparently to make thermal contact to the metal baseplate under the mobo, which it was not doing. I "shimmed" the baseplate and used another type of thermal tape to make a thermal contact for the ATI chip. For a while, I thought this cured it, but the problem came back.

From what I've read, this points to a bad system board. The IBM FRU numbers for this model are 26P8232 or 26P8399. The only suppliers I've found want $$$$ for it, such that a new system may be a better idea. There's a seller on eBay offering a 93P3563, which he claims is for an A31p 2653, but I can find no such FRU on the Lenovo site.

Questions: What alternative system boards will work? It looks like a 26P8430 from a 2653-Rxx would work. What the devil is a 93P3563? A 93P3542 is for a 2652-Dxx, which has different graphics. Any other options I should consider?

Thanks for any advice.

[ajkula66]

Welcome to the forum, first and foremost!

You've done your research and then some, and there's really NOT a lot that I can't tell you that you already don't know or suspect...

a) Your problem DOES sound like a MB issue. Although I've never had it personally on any of my machines, I've heard similar stories and complaints.
b) Motherboard for the 2653-Rxx seems to be your ONLY solution, and these will not come cheap. My suggestion is to find an A31p with a bad LCD and do the transplant from there...


The best of luck.

[fasterbybike]

Would mobo P/N 26P8430 not work ?
This is spec'd for models 2653-R3x, R4x, R5x, R6x, RNx, RPx.

Searching Ebay for (26P8399, 26P8232, 26P8430) shows boards selling recently for as little as GBP57 plus GBP15 shipping, or $85 in the US.

HTH

[mediasponge]

Thanks for the comments. It looks like I can get an exact replacement for around USD $229 + shipping. I'm reluctant to source one from the UK. The same UK seller that had the "cheap" ones available has more, but at a higher price, naturally.

What bothers me is that I never got PC Dr to point out any failure, and if it's actually the CPU that's bad, then I'm into the zone of spending more money on it than the thing is worth. The 1600x1200 display is so crisp, tho. Besides, I'm just stubborn enough to keep working on it until it submits.

[ajkula66]

Testing out another CPU on that board would be a smart idea...

And personally, I'd probably keep on pushing it until it dies myself, so I know where you're coming from...

The best of luck from one A31p buff to another...

[rkawakami]

I would run PC Doctor in a looping mode on several tests, at the same time flexing the chassis (lift up on front right corner and push down on front left corner and then do the opposite). To place PC Doctor in a looping mode, once you have selected the diagnostic test to run, press the F2 key and then select the "Set Pass Count" option. Change the value from "1" to "10" and that will cause the test to repeat that many times. Run the following tests while flexing the case:

CPU/Coprocessor
Systemboard
Video Adapter

The Memory Test can be run without changing the pass count since it takes quite awhile to run one pass anyway. If you set the pass count high enough, you can run the system overnight and see if you have a temperature related issue (i.e., hotter temp might cause a problem).

I had a problem with a T23 which showed up as a "freeze". Flexing the system and running PC Doctor revealed that the RTC Clock test in the Systemboard diagnostic was intermittently failing. This was confirmed by opening up the system and checking a couple of crystals. One of them had two legs disconnected . A quick re-solder job and all was well.

[mediasponge]

Yes, if I had another CPU to test, I certainly would have done that. There is a large surplus store near work I drop into frequently, but they don't have anything as good as an A31p or a 2.0 GHz P4M.

Flexing the system, eh? Worth a shot. I have run PC Dr many, many loops on most tests. The only time I got anything was a failure on the interrupt handling test while violently wiggling a USB mouse around. I assumed that was just port overload during the test.

If I find something easy to solder on mine that cures it, I will be a happy camper.

[phool@round]

Meh....you could buy another A31P parts box. A parts laptop would cost somewhere @ the price of another system board, probably less if I'm doing the shop'n! Something sans hard disk, memory, ultrabay BUT is tested working....... might get some upgrades in the deal too eh?

[proaudioguy]

Have you tried substituting the RAM?

I have found "mostly good" RAM to be the cause of many problems. IOW, it's mostly good, but has one bad spot. When the computer pages to that spot, it crashes. It could be weeks, days, hours, or during boot, but eventually it tries to use that section of the memory.

[mediasponge]

Having one bad spot in a ram implies it's in one chip on one DIMM. I have 2 DIMMs, and ran the thing with one, then the other, then both with no difference in the symptoms. I don't have any spares to try, but the odds are that they wouldn't both be bad. I did run memory tests for hours and hours, with no errors.

Update to thread: I have a new mobo (exact replacement) on order, should arrive soon. If this doesn't fix it, I will be at the proverbial wit's end...

[rkawakami]

Going back and re-reading your original post, this strikes me:

It seems like it never locks up when booted from a CD, only when booted from the HDD, which is new.

You may have a bad sector on the disk drive. Have you tried running PC Doctor's Fixed Disk diagnostic? I've been testing many disk drives over the last month or so and the Linear Verify test always seems to catch the problem. The other PC Doctor tests will pass, but for the random sector failures, the Linear Verify test will find them.

Also, since it appears you have a Fujitsu drive you can try their diagnostic software:

http://www.fujitsu.ca/support/hdd/download.html

The PC Doctor Fixed Disk tests will NOT alter your drive's data. I'm not familiar with Fujitsu's software so be sure to read all of their documentation before running any of those tests.

[mediasponge]

Yes, I did run the PC Dr disk tests. The drive is new. I cloned the old one using Acronis, with full verification and error checking. (Took forever.) Besides, it does this in both Windoze and Linux, booted from 2 separate partitions.

Good suggestions from the group, which I may have to resort to if the new mobo doesn't fix it. (Fingers crossed, it hasn't arrived yet.)

[phool@round]

I hope it does fix it too, I feel a wee bit guilty suggesting the board is the problem........my fingers crossed with ya. I'll be the first to *not* suggest anything if it doesn't.....promise.

[mediasponge]

Update: New mobo works!

Well, the new mobo i ordered seems to have fixed the problem. (Please, computer, don't make a liar out of me, OK?) This time when I dismantled it, I actually followed the sequence in the maintenance manual, it made things a bit easier. I also blew the dust out of the fan (cough, cough).

I did have a bit of a scare when it booted and complained about a security chip tamper detection, but there was an option in the BIOS to clear this. I had to select it twice, but then it booted right up. I ran a few passes of PC Dr, then booted XP. Linux is happy now too, but downloading updates was so slow, I deferred that till later.

Thing is, the new mobo has a pretty old BIOS. It has 1.02 (1GET30WW) with ECP 1.00a. The one I replaced was on 1.13 (1GET41WW) & ECP 1.05. Is it worth updating, or should I leave well enough alone? Which one are you supposed to update first, the ECP or the BIOS? Inquiring minds want to know!

[goofyGAguy]

According to the Lenovo website, the BIOS should be updated before the Embedded Controller.

[mediasponge]

Interesting. Where on Lenovo did you see that? In the BIOS readme file I found:

<1.13 - 1GET41WW>
Note: This BIOS version will only work with Embedded Controller(ECP) version 1.04
(1GHT17WW) or later.

This kind of implies that the ECP has to be installed first for the BIOS update to work, hence the confusion.

According to the Lenovo website, the BIOS should be updated before the Embedded Controller.

[goofyGAguy]

This is stated on the Downloads and Drivers page for my R51. Quote:

" If you need to update the Embedded Controller Program as well as the BIOS, update the BIOS first. "

[jrv331]

Dont know if its the same or not but when I updated my A31 BIOS and ECP I had to do the controller one first. BIOS would not update with out it.
John

[madmaven]

I was wondering if the motherboards are interchangeable on the A31s? I have 2 that are A31' but different types will that work one has a good display but bad mobo and the other has no display but boots. THanks

Also I have a A31p and on the center front edge of the lid is a bracket with some gold contacts where on the A31s there is just a cover type thing a blank i guess ANy ideas?
thanks again
Madmaven

[mediasponge]

On the motherboards, the only safe way is to check the list here: http://www-307.ibm.com/pc/support/site. ... R-39574#23
If the model types use the same FRU, they are directly compatible. Mine is an A31p 2653, which has the UXGA graphics, so not compatible with most of the others.

I didn't even notice that connector on mine. According to Access Thinkpad, it's an UltraPort (tm) to which you can connect an UltraPort camera, flash card reader, or digital array microphone (?), all purchased from IBM, of course. Mine was under a rubber pop-out.

I was wondering if the motherboards are interchangeable on the A31s? I have 2 that are A31' but different types will that work one has a good display but bad mobo and the other has no display but boots. THanks

Also I have a A31p and on the center front edge of the lid is a bracket with some gold contacts where on the A31s there is just a cover type thing a blank i guess ANy ideas?
thanks again
Madmaven