Thinkpad A31 doesn't turn on with power supply cord inserted

[mima85]

[EDITED: see posts below about strikethrough text]

Hi to all, thanks for having accepted me in your community. This is my first post. First things first, I'm not a native english speaker, so please excuse me if sometimes I'll slip with grammar and words

My name is Michel and I work in the IT field, my hobbies revolve around retrocomputing, electronics, music and related tech (especially synthesizers - better if vintage - and sound design).

Speaking of retrocomputing, I have a soft spot for older Thinkpads from before 2005, especially for models from the 90ies. Actually, I have quite a number of them in my collection, which I was lucky enough to get my hands on in times well before the retrocomputing mania exploded, together with all the collectors/ebay/price inflation circle of things, and those old PCs were just seen for what they actually were: worthless, disposable electronics. So about 15 years ago, I rescued from the junk yard a Thinkpad A31. It had the ultrabay slots unpopulated but it was in fully working condition (it even still had its original hard drive), and after having completed it with a CD and floppy drive and given it a good clean, I had a nice specimen to add to my collection.

I later used that Thinkpad as a test machine for some stuff I was doing at my workplace, where it ran flawlessly for months straight, always turned on and without never crashing or needing reboots. Not bad at all for an old boy like that, truly a testament of the build quality of the Thinkpads from that era. Then I finally retired it back in my collection.

In these later times I took it out to see if it was still working, and to my disappointment it wasn't. It would turn on, with the LCD backlight coming on toghether with the power LED, but the hard drive and CPU fan wouldn't spin and the machine wouldn't boot. After fiddling a bit and letting the battery get some juice, it finally turned on, and stayed on for a couple of hours, reliable as it was before (no crashes, etc...). Then, after power cycling it, again no luck.

I tried the usual stuff, removed/reseated all the drives, the expansion cards and the RAM, replaced the CMOS battery which was totally dead (of course...), then I completely disassembled the machine, gave a good clean to the motherboard which is a thing that I didn't do before, reseated the CPU and applied some new thermal grease and finally put it back together, but I still had the same behaviour: the machine would boot only after dozens and dozens of tries, but when it booted it stayed alive without any problem.

I fiddled with it again, and suddenly I discovered something strange: without the power cord inserted and only with the battery the computer would always turn on correctly, and at this time I could hook the power supply and it would work as intended, charging the battery and even keeping the PC on if I'd disconnect the battery. I tried several power supplies, all original IBM parts and with the correct rating, but with every one I had the problem. The power supplies, tested for voltage and ripple with my oscilloscope, were all in spec, actually with the other Thinkpads I have those work without hassle. I then tried with the power supply without hooking it to the mains but I had the problem again, so suspecting some weird behaviour due to impedance differences when having a power supply connected or disconnected, I tried with a plain cord which I cut from a defective power brick before and I have laying around, so it's literally just the power jack with a cable attached to nothing on the other end. And even with that inside the power inlet jack, the computer wouldn't boot.

So, suspecting some kind of bad connection on the power inlet jack when having the plug fitted in, I opened the notebook again and checked the pins with the tester for continuity, to see if there were short circuits somewhere. Both with and without the jack inserted I had the same, correct results, meaning continuity between the 3 ground pins (there's a center pin for positive and 3 leaf contacts for the ground inside the jack, and those 3 contacts are soldered to the same ground trace on the mainboard) and no short circuits between positive and ground. I also checked for continuity on the wire's other end and there was nothing strange there too, the positive wire would make contact with the positive pin's solder joint on the mainboard, and the ground wire would make contact with the whole ground plane, without any sort of short circuit between the two poles. Out of ideas I reflowed the power jack's solder joints, because hey, why not as I'm already here, but still no luck.

The whole thing just doesn't make sense and now, besides replacing 2 ICs which are close to the power jack (a Toshiba TPC8002 which has something to do with charging the battery, and an IC which is labeled "IOR 224N 5P4C F7416" about which I found nothing on the Internet and I don't know what it does) I really don't know what to think. I scrolled through the whole IBM service manual for that model, but besides replacing the mainboard it doesn't tell me anything useful.

So, before putting back the computer for the final time and accepting that I have to turn it on without anything inserted in the power jack, I would ask if someone of you had a similar experience and if there are some suggestions about things other things I may check or fix.

Thank you, and sorry for the long post

[RealBlackStuff]

Welcome to the Forum.
Assuming the CMOS is still alive, try this:
Remove main battery and charger.
Press and hold down the Power button for 20 seconds.
Insert only the charger, try to boot.
Any luck?

Have a look at this post about a T22, same era as yours: https://forum.thinkpads.com/viewtopic.p ... 48#p873648
Those same chips might be in yours.

Also, get the A31 schematics, they are almost identical to A31p.
And if you haven't got it yet, get the A30/A31 manual from the HMM link above.

[mima85]

Thanks for your answer RealBlackStuff, especially for pointing me at the schematics, those are really appreciated and will surely be useful.

So, I tried what you suggested, but unfortunately no luck. But while looking again at the mainboard, I noticed a thing which has previously escaped my sight: one of the two USB ports at the power jack's left side, the upper one to be precise, is broken. The pins are there and aren't touching anything hence no shorts, but the internal supporting plastic bit is missing, so for sure at a certain moment whatever was connected to that port took a hit and broke the connector.

By measuring the impedance between the broken port's pins and ground, on the DATA- line I had about 77 ohm, while on the good port there wasn't continuity between its DATA- line and ground. If I understood correctly how USB works its signalling is differential, the DATA- line is not connected to the ground because it carries the same signal of the DATA+ line but out of phase by 180 degrees, so there shouldn't be any connection between that pin and ground. Correct me if I'm wrong.

I tried to lift the decoupling capacitor C772 which is placed between that DATA- line and ground, and I had the same reading, so that capacitor shouldn't be responsible for the fault. And while I'm writing this post, I recalled that the USB ports actually never worked on this machine.

I guess that the incident that happened at that port caused a short circuit and toasted the USB controller, and this may have something to do with the problem I have now, because short circuits can cause a wide variety of issues down the line. Albeit I still can't figure why if I put a jack whose wire is cut and disconnected from everything into my A31's power connector I can turn on the computer only after dozens of tries, while if I pull it out I can turn on the machine succesfully from the battery every time I try to.

[RealBlackStuff]

I tried to lift the decoupling capacitor C772...

does not make any sense. Wrong C-number?
No idea how those USB electronics work, someone else chime in?

Also, are you sure the other USB-port is OK?

PS: my family used to have a house in Bissone, where we spent many great holidays.
But when they started to build the A2, we got out.

[ajkula66]

Albeit I still can't figure why if I put a jack whose wire is cut and disconnected from everything into my A31's power connector I can turn on the computer only after dozens of tries, while if I pull it out I can turn on the machine succesfully from the battery every time I try to.

I've had this problem on a T42 and a T43p, both times it was caused by moisture damage on the planar. YMMV.

Great to see that someone is still using an A3x system in any shape or form...

Good luck.

[mima85]

does not make any sense. Wrong C-number?

Yeah, noticed that too. On my board is marked as C772, others are C771, C773 and C774. C772 should be C825 or C827 on page 17 of the schematics, depending by what is considered for the upper port to be between USBP2 or USBP3 on the schematics. Maybe that mainboard has several revisions and the one the schematics refers to is not mine.

Anyway, no, none of the USB ports work. That's why I suspect the incident with the port fried the USB controller, and at this point I also suspect that the resulting current surge damaged something else on the mainboard.

PS: my family used to have a house in Bissone, where we spent many great holidays.
But when they started to build the A2, we got out.

Oh yeah, I know very well Bissone as I live nearby

They started building the A2 highway back in the 60ies, so it's quite a long time you got out of there.

[mima85]

I've had this problem on a T42 and a T43p, both times it was caused by moisture damage on the planar. YMMV.

There actually was a nasty spot of dust mixed with an oily substance under the fan, where the GPU and its RAM chips are located, which I had to clean away with alcohool and a brush because compressed air was not enough. And I found some traces of oily substance on the mainboard's bottom side, without dust this time. I don't know what that was, I just cleaned it away, but at least id didn't look like capacitor dielectric fluid nor smelled like that too. Actually it had no smell at all. Who knows...

Great to see that someone is still using an A3x system in any shape or form...

Good luck.

Thank you

Well, now actually I'm no longer "using" it, it just sits in my collection

But it's a nice specimen, by the way I was lucky enough to find one with the display having the best resolution of the series, 1400x1050. And apart the normal wear from use, cosmetically is in very good condition and the screen hinges are still very solid, a thing that unfortunately I cannot tell for some of my other Thinkpads from the 90ies. My 310E, 365XD and 390E were a pain in the a** to fix, and the plastics are still cracking away nonetheless. For the 390E I even found a way to reinforce them and I written a document on how to do that.

It's a shame that my A31 has this issue with the motherboard, but at least with some fiddling I still can turn it on and it won't crash or reboot. If I won't be able to fix it I'll just keep it as it is, until I'll eventually find another motherboard in working condition to replace mine.

It's definitely a keeper, especially considering that this family of Thinkpads found its use even on the Space Shuttle

[Gonzaleitor]

Hey, if it works i'll look at it as a win. Is a little wonky, but that's ok.
I have to fix my A31's screen (I replaced it from an R40 donor and the flex cable came out a little short). They are awesome retro machines!

[mima85]

Hey, if it works i'll look at it as a win. Is a little wonky, but that's ok.

Oh yeah, of course it is. It's just that sometimes I tend to be a perfectionist, altough I know that perfectionism and retrocomputing, where often you have to be happy just to have a given machine in whatever state it is, don't go well together. But I like to at least try to fix things, so here I am

Speaking about the USB problem, by taking another look at the schematics I saw that the USB ports are directly tied to the chipset, which has the USB controller integrated into it. There's no external USB controller IC to try to replace and I don't have the skills, tools and patience of dealing with BGA chips rework/replacement, especially if the component is a big one like an Intel chipset, so I'll forget about this specific issue.

And judging by where things are going now with my diagnostics - nowhere - I think that soon I'll put back together my A31 and I'll keep it as it is. In the meantime I'm still going to wait for some suggestions here, maybe someone on the forum was able to fix a similar problem in the past and at a certain moment will cross this thread, who knows.

[mima85]

An update: forget what I said about the different behavior when having or not something connected to the power in connector: it turned out to be just a coincidence. Today I worked again on the machine and in fact it can fail to turn on with just the battery too, so the thing is completely random. Which makes more sense actually, because after checking that the power supply connector itself doesn't have any short or defective contact, I couldn't think of any feasible reason about why the notebook shouldn't turn on when a jack who's not attached to anything is inserted in the power connector.

[SMA]

Mecanically defect usb ports are rather common on a30/a31.
Replacing these connectors are doable. I have been there myself.
The largest problem will be finding replacement connectors that will fit.
One option will be to get replacement connectors from an otherwise dead donor board.

Possibly you should not attempt to connect anything to a mechanical faulty usb connector.

If your usb ports has other faults than the mechanical ones, then most likely that will be the +5V missing.
Without having checked the schematics I will assume that there is a fuse for the usb power that may have blown in your tp.
Note that there are usb devices that can be powered externally so that they doesn't need 5v from the usb connector.
And easy way of checking the usb 5v without opening the entire thinkpad is to connect a "blank" usb male connector which will give
easy access to the pins of the usb port so that the 5v can be measured using a multimeter.

[mima85]

Mecanically defect usb ports are rather common on a30/a31.
Replacing these connectors are doable. I have been there myself.

Yes I did that myself too in other occasions, actually I have around some dead PC motherboards which I keep for parts, where I could scavenge an USB connector to replace the broken one. The fact is that the USB controller itself is toasted, I already tried to connect a USB device that was powered externally, but the computer didn't see it.

But the main problem with my A31 is the strange behavior at power-on. USB is a side thing, but I suspect the incident that happened at that port an the resulting short circuit may have something to do with the main problem, and my suspicions grown when I was checking the schematics and I found that the USB controller is in fact embedded into the Intel chipset's IC.

We'll see how long the mainboard will last.

[kfzhu1229]

An update: forget what I said about the different behavior when having or not something connected to the power in connector: it turned out to be just a coincidence. Today I worked again on the machine and in fact it can fail to turn on with just the battery too, so the thing is completely random. Which makes more sense actually, because after checking that the power supply connector itself doesn't have any short or defective contact, I couldn't think of any feasible reason about why the notebook shouldn't turn on when a jack who's not attached to anything is inserted in the power connector.

So I suppose the A31 is more likely to turn on properly with just the battery attached than with charger attached? And that you mentioned that when it doesn't turn on properly, the screen backlight turns on with the press of the power button and nothing else happens. I remember I had this with A31p motherboard before, I had a good A31p motherboard that would occasionally do this, and I remember if I press the power button either for longer or multiple times in quick succession, I would be able to get the laptop to switch on properly.
Can you check if the 3.3V power rail is working properly? A31/A31p motherboards heat up a lot more than its predecessor and I have had dying capacitors with A31p boards before, and the 3.3V main filtering capacitors on this laptop seem to all be Niobium capacitors that can slowly die by losing capacitance/increasing ESR like a regular electrolytic capacitor.

[mima85]

So I suppose the A31 is more likely to turn on properly with just the battery attached than with charger attached?

It looked like so initially, then after additional tries I realized that it was just a coincidence.

And that you mentioned that when it doesn't turn on properly, the screen backlight turns on with the press of the power button and nothing else happens.

Yes, exactely. Screen backlight goes on, keyboard LEDs flash for a second, power-on LED turns on and stays on, and the fan and the drives wouldn't spin.

I also add that, by checking without the heatsink in place, the ATI GPU starts to heat up considerely in some seconds, while at the same time the CPU stays cool, probably because it didn't start/initialize.

I remember I had this with A31p motherboard before, I had a good A31p motherboard that would occasionally do this, and I remember if I press the power button either for longer or multiple times in quick succession, I would be able to get the laptop to switch on properly.
Can you check if the 3.3V power rail is working properly? A31/A31p motherboards heat up a lot more than its predecessor and I have had dying capacitors with A31p boards before, and the 3.3V main filtering capacitors on this laptop seem to all be Niobium capacitors that can slowly die by losing capacitance/increasing ESR like a regular electrolytic capacitor.

That's a good advice, thank you! I'll check soon and report back here.

I'm thinking also about checking the reset signal. Years ago I had a similar problem, but it was on a musical instrument (a Roland D-50 synthesizer). Turned out that the reset signal wasn't being generated properly, due to a cracked solder joint of a chip on the power supply. After reflowing the solder joint the problem gone away. Maybe something like that is happening with my A31, albeit perhaps for a different reason instead of a mere cracked solder joint. Even though I think that in such a scenario the fan and the hard drive would spin, and there wouldn't be activity on the keyboard LEDs, but who knows.

[ajkula66]

but it was on a musical instrument (a Roland D-50 synthesizer).

Completely OT, but I haven't seen one of those in decades, and they were the cat's meow when I was young...good times...

[mima85]

Completely OT, but I haven't seen one of those in decades, and they were the cat's meow when I was young...good times...

Oh yeah, I'm into these things too. I also have other machines from the 80ies, including analog synths. If you search "mima85" on Youtube you're going to find my channel if you're interested

But now let's end here the OT, before we get some moderator angry

[mima85]

Can you check if the 3.3V power rail is working properly? A31/A31p motherboards heat up a lot more than its predecessor and I have had dying capacitors with A31p boards before, and the 3.3V main filtering capacitors on this laptop seem to all be Niobium capacitors that can slowly die by losing capacitance/increasing ESR like a regular electrolytic capacitor.

Did that, I checked the 3.3V rail on pin 28 of the MiniPCI connector, as in the meantime I put back together the machine and I didn't want to disassemble it another time. These are my findings:

1) At notebook off, I have 0V as expected.
2) When the notebook doesn't turn on, I have about 1.2V.
3) When the notebook turns on, I have the correct 3.3V.

I checked for ripple with my oscilloscope, and when having voltage in both cases I have about 5-10mV of ripple, which is normal so the filtering caps on the 3.3V rail should still be good.

At page 23 of the motherboard schematics that was linked before by another user, the power-on logic is laid out. At page 22 a GAL handling various stuff including power is being illustrated. Perhaps the problem could be in that area?

[kfzhu1229]

1) At notebook off, I have 0V as expected.
2) When the notebook doesn't turn on, I have about 1.2V.
3) When the notebook turns on, I have the correct 3.3V.
I checked for ripple with my oscilloscope, and when having voltage in both cases I have about 5-10mV of ripple, which is normal so the filtering caps on the 3.3V rail should still be good.

1.2V on the 3.3V power rail when the laptop isn't fully switched on? Then you might also wanna check the 5V rail (preferbly by the MOSFETs) since most probably the 3.3 and 5V are managed by the same IC. Guess the most unfortunate outcome would be if the voltage IC is dying.
Or maybe when it isn't switched on properly check for any hot components on the motherboard, it is possible for a dead component on the motherboard to be partially bringing down the 3.3V rail too (the most unfortunate outcome of this would be a partially shorted southbridge, which was the nasty surprise when I was fixing a badly beaten Latitude D800 motherboard)