X32 stuck at POST with blank screen

[dr_st]

Unfortunately my new/old X32 is not in its best shape to say the least.

At some point it started with intermittent POST problems, and sometimes after POSTing it would not complete the boot (freezing / BSODing in the middle). Sometimes when it would complete the boot, it would work fine for several hours (including a number of stress tests), but eventually freeze or BSOD or reboot.

Most of the blue screen stop codes pointed towards memory or corrupt driver as the most likely cause, however, unfortunately it seems neither, seeing as:

* I tried 3 different RAM sticks in various single/dual slot configurations
* I tried three different OS installations (one XP, two Win7)

The problem did not go away and eventually got worse. Now the situation is as follows:

* Laptop won't POST at all. It gets stuck with a blank screen, power and battery LEDs on, and no beeps.
* On battery, laptop won't even turn on (no lights, nothing). Although, the battery is not dead - if the machine is turned on AC, and the AC is unplugged, it stays on.
* If I remove all memory and power on, I get the correct 1-3-3-1 beep sequence.
* Tried the famous reset sequence (remove battery - press power button 10 times and hold it for 30 seconds last time) - no effect.

Looks like something is not right with the board. My next steps will probably be a thorough disassembly of the machine, to inspect for loose connectors, broken parts, dirt, etc.

Thought to post it here first to see if anyone has some advice before I proceed.

[RealBlackStuff]

You seem to have checked all the usual suspects, maybe try another CMOS battery?
That should be near-dead after all these years.
And then there are bad wifi and/or modem cards...

[dr_st]

Will look into CMOS battery and modem card. WiFi card is pretty much out of question, since I swapped these too...

[misfit]

<snipped>Thought to post it here first to see if anyone has some advice before I proceed.

Hey Doc.

Did I give you the URLs to those Chinese sites that seem to talk about X31s and X32s and talk about replacing up to eight capacitors as well as some other SMDs? (My Chinese isn't great, or should I say my understanding of the translations isn't great.) From what I've been able to understand they 'blame' IBM for wanting the machine to run for as long as possible on the battery quietly. Consequently the embedded controller or BIOS was programmed to let the insides get quite hot before turning the cooler on, which didn't have good 'flow-through' pattern either. Hence certain components ran a lot hotter than they needed to. However, it should be possible to replace the components and run NHC and tweak the ACPI settings (as I do on all TPs anyway) and set it to run the fan most of the time.

It seems that X31s and X32s are, amongst certain circles in China, accorded the respect that I think they deserve. (Must get around to updating my sig, I have seven X31s now. I still haven't found an elusive X32 here in NZ at a price I can afford. :-/ The plan is to keep on using T60s for 'heavy lifting', dispose of all of the rest except X31s <sniff> and concentrate on X31s and hopefully X32s. IF we can get this major issue sorted!)

Anyway, let me search my stored URLs and see if you can make more sense of them than I can. Be sure to let me know if you get a good handle on what they're saying. It's times like this I wish I had a good friend who speaks both Chinese and Geek. Heck, I'd settle for just the former!

Here you go:
http://blog.163.com/bugoo_cat/blog/stat ... 133855472/
http://chs.mobile01.com/topicdetail.php ... st=7124852

Best of luck. (And if anyone reading this reads Chinese [or knows someone who'll translate it for them] please let us know. From what I hear there is quite a sub-culture in China around X31s and X32s, these are just two pages my below-par Google-fu unearthed, I hear there are more out there....)

[dr_st]

Well, no luck so far.

After disassembling and cleaning the board in an ultrasonic bath, I started putting the stuff back together and one of the small capacitors near the power jack circuitry fell off. Upon inspection it appeared undamaged except the broken legs, so we resoldered it back upside down.

I then proceeded connecting the bare minimums to start the machine (CMOS battery, RAM, keyboard, CPU HSF, power and subcard), but the symptops are roughly unchanged:

* On battery without AC - dead cold, no LEDs, nothing
* On AC, without subcard - 4x4 beeps (security chip) and no LCD
* On AC, without RAM - 1-3-3-1 beeps
* On AC, with RAM in 1st DIMM slot - no beeps, blank LCD (turns on but no image), keyboard responds (NumLk/CapsLk LEDs), but that's it
* On AC, with RAM in 2nd DIMM slot - no beeps, no LCD, keyboard doesn't respond

It is not clear whether I'm dealing with one or two issues here. I've seen cases described on the forums with similar no-post symptoms or with the laptop refusing to boot on battery, but not at the same time. Both issues seem to point to motherboard, but I wonder if it is possible for one faulty component to be the cause of both.

The only thing that seems is left for me to try before searching for a new systemboard is testing and possibly replacing the capacitors mentioned by Reagle in this thread:
http://forum.thinkpads.com/viewtopic.php?f=3&t=94160

I wonder what the simplest way of testing them would be, though.

[rkawakami]

The only thing that seems is left for me to try before searching for a new systemboard is testing and possibly replacing the capacitors mentioned by Reagle in this thread: http://forum.thinkpads.com/viewtopic.php?f=3&t=94160

I wonder what the simplest way of testing them would be, though.

That would be a capacitance meter . If you don't have one, then a regular ohmmeter set to the highest resistance range can be used as a very simple test. When "ohmed out", a capacitor should initially start out reading several Kohms and then increase in resistance. The larger the uF of the cap, the longer it would take for the resistance to read near infinity. Reverse the test leads and then you should see the same varying resistance. What's happening is that the meter is essentially charging the capacitor. It will start out at a low "resistance" and then when the cap is fully charged, the reading will level out at several Mohms. This is for a functional capacitor. If you see 0 ohms, then the cap is shorted or if there's no change from infinity when you connect the cap to the meter, then it might be open. This method can be used to check capacitors of 0.1uF or more and they do have to be disconnected from any other components.

[dr_st]

Thank you for the reply, Ray. I'd like to clarify a few things.

Basically to test it I have to unsolder the capacitor from the board first, and then use the ohmmeter - correct?

Assuming I do get my hands on a capacitance meter, what should the reading be for a functional capacitor versus a non-functional one?

[misfit]

Thank you for the reply, Ray. I'd like to clarify a few things.

Basically to test it I have to unsolder the capacitor from the board first, and then use the ohmmeter - correct?

Correct. However SMD tantalum capacitors can only handle a certain amount of heat (either over a long period or all at once, it seems to be cumulative). Therefore unsoldering and re-soldering is likely to push them over the edge unless you really know your SMD soldering.

Assuming I do get my hands on a capacitance meter, what should the reading be for a functional capacitor versus a non-functional one?

As Reagle said in that other thread, one is 330uF and the other is 220uF. At least that's for the two most stressed capacitors that he replaced.

As for a non-functioning one, it's hard to say. It might still read correctly on a capacitance meter but no longer be low ESR (so it would let ripple and spikes through, as mentioned by Reagle, which shuts down the control IC). It would be so much easier to just spend a couple $$ and get a couple new capacitors than mess about testing a ~$2 component.

Good luck, keep us posted.