ThinkPad A31 Mail-In Repair

[Virtual_Law]

I have a non-functional ThinkPad A31, and a some what functional ThinkPad A31. I am wondering if there's anyone I can send my ThinkPads to on this forum for retro repair, to atleast get one of them working. I can't do it myself, I'm more of a desktop/workstation user. I'm willing to pay if necessary.

[fegabi]

Can you please describe the symptoms? It might give us an idea how far they are gone.

[Virtual_Law]

Can you please describe the symptoms? It might give us an idea how far they are gone.

One of them boots, but nothing appears on the screen. Not even a BIOS screen, just blank/dark/empty, etc... The other doesn't even boot.

[Gonzaleitor]

Can you please describe the symptoms? It might give us an idea how far they are gone.

One of them boots, but nothing appears on the screen. Not even a BIOS screen, just blank/dark/empty, etc... The other doesn't even boot.

Did you try swapping the RAMs? Maybe they are dead and that's why it won't show image.

[TPFanatic]

Unfortunately GPU death is very common on A31.

I had A31p with GPU death, absolutely on video output to the internal LCD. Complicating matters further, the internal LCD was also dead. However it would output to an external monitor. The machine failed in this way back in 2005!!

Sadly these are just not reliable models. Nor am I particularly sold on them anymore. Dual Ultrabay, while cool, is barely actionable, especially with PATA-based drive accessories being practically unobtanium now.

I wish you the best of luck in having your machine back to operating condition.

[astral]

dosdude1 runs DosLab Electronics and accepts mail-in repair: https://doslabelectronics.com/
If there's a guy who can fix your machine, it's him.

It's likely as others have said that your dead A31 has suffered from GPU failure. If you're willing to pay for it, he should be able to replace and/or reball the GPU to fix the problem. As the A31 failures were caused by defective lead-free solder that cracked easily, a reballed A31 should actually be decently reliable.

[ajkula66]

As the A31 failures were caused by defective lead-free solder that cracked easily, a reballed A31 should actually be decently reliable.

It's a bit more complicated than that.

GPU has no active cooling in the A3x systems, just the "sponges" that dry out over time.

Once reballed, the machine should be kept running 24/7/365 even if not in use, with the LCD turned off as needed. An upgraded fan from the later version of A31p is also recommended, along with running TPFC.

Otherwise the constant on/off cycles will cause the GPU detachment again, and quite fast.

The youngest A3* is turning 22 this year. There might be other multiple points of failure that can't be fixed by re-balling.

My advice is not to sink any $$$ into repairing these systems, but try to locate a working unit from a reputable source such as oldtimers on this forum.

Good luck.

[astral]

The Mobility Radeon 7500 in theory should be just fine without a heatsink at all - many laptops including the ThinkPad T30 didn't even use passive cooling. I don't think the T30 is *especially* prone to GPU failure, although I have heard of it happening - usually the RAM slots get them first. The reason why the A31 was hit so hard is because of the lead-free solder plague, made worse by the lackluster cooling. It should have been fine in theory, but when the solder is more susceptible to heat cycling, then things suddenly don't go so well. I also think the GPU was on the bottom of the board on that laptop? If I'm right then that couldn't have helped. When it comes to ATI 7500 laptops, the most reliable ones (like the Latitude C640) are so because they have far better cooling. But they would have been fine if the solder wasn't defective.

That's all to say - it should be far more reliable when reballed with leaded solder. Plenty of modern laptops that run like a furnace with lackluster cooling work fine for years with no BGA problems. Usually it's the VRM that kicks it on those.

I'm sure that age is going to make them more failure prone then they were when new, but for the sort of casual non-daily driver use that an A31 is going to see these days, a reballed one is probably going to be plenty reliable. I'd imagine that the wear on the VRM, fan, and other hardware from keeping a "couple times a month" laptop running 24/7 is going to be more harmful to it than the heat cycling would. That might have made more sense if it was 2005 and you were daily driving one.

This is all for the ones with the ATI 7500 graphics - no clue how much hotter the FireGL in the A31p runs, or, for that matter, if the VRM hardware in these laptops is also prone to failure.
I'd assume that OP's unit with a no power condition probably has either a failed VRM or bad DC/DC hardware.

[ajkula66]

I also think the GPU was on the bottom of the board on that laptop?

Correct.

I I'd imagine that the wear on the VRM, fan, and other hardware from keeping a "couple times a month" laptop running 24/7 is going to be more harmful to it than the heat cycling would. That might have made more sense if it was 2005 and you were daily driving one.

And you would be very, very wrong with that presumption.

[Gonzaleitor]

The Mobility Radeon 7500 in theory should be just fine without a heatsink at all - many laptops including the ThinkPad T30 didn't even use passive cooling. I don't think the T30 is *especially* prone to GPU failure, although I have heard of it happening - usually the RAM slots get them first. The reason why the A31 was hit so hard is because of the lead-free solder plague, made worse by the lackluster cooling. It should have been fine in theory, but when the solder is more susceptible to heat cycling, then things suddenly don't go so well. I also think the GPU was on the bottom of the board on that laptop? If I'm right then that couldn't have helped. When it comes to ATI 7500 laptops, the most reliable ones (like the Latitude C640) are so because they have far better cooling. But they would have been fine if the solder wasn't defective.

That's all to say - it should be far more reliable when reballed with leaded solder. Plenty of modern laptops that run like a furnace with lackluster cooling work fine for years with no BGA problems. Usually it's the VRM that kicks it on those.

I'm sure that age is going to make them more failure prone then they were when new, but for the sort of casual non-daily driver use that an A31 is going to see these days, a reballed one is probably going to be plenty reliable. I'd imagine that the wear on the VRM, fan, and other hardware from keeping a "couple times a month" laptop running 24/7 is going to be more harmful to it than the heat cycling would. That might have made more sense if it was 2005 and you were daily driving one.

This is all for the ones with the ATI 7500 graphics - no clue how much hotter the FireGL in the A31p runs, or, for that matter, if the VRM hardware in these laptops is also prone to failure.
I'd assume that OP's unit with a no power condition probably has either a failed VRM or bad DC/DC hardware.

The GPU is next to the CPU socket here. I put a 2mm thermal pad on top of the GPU die so it makes contact with the heatsink. I don't know if it's the best solution, but works great for a little cooling and haven't had a problem in 3 years.

[Saucey]

If I were to reball a GPU, what solder balls would be best to use?
Will it also be good with T4x machines that have the dreaded flexing southbridge?

I eventually want to get a reflow station, currently only have a heat gun.
Don't worry, I won't harm the A31's I have, nor my precious A31p, I have a few MacBook Pros to experiment with.

[astral]

I I'd imagine that the wear on the VRM, fan, and other hardware from keeping a "couple times a month" laptop running 24/7 is going to be more harmful to it than the heat cycling would. That might have made more sense if it was 2005 and you were daily driving one.

And you would be very, very wrong with that presumption.

What makes the A31 any different from the dozens of other laptops affected by the lead-free solder plague that can be made reliable via BGA reballing?

[ajkula66]

What makes the A31 any different from the dozens of other laptops affected by the lead-free solder plague that can be made reliable via BGA reballing?

The fact that it doesn't hold on them. Proven by experience with several different high-qualified individuals both in the U.S. and the UK reballing them and the systems failing within ~18 months.

Having seen that you're somewhat new to this place I'll forgive you for being unaware of the fact that I've had my hands on/in more A3x systems than anyone you will ever come across. If you feel like picking arguments over theoretical stuff you'll have to find someone else and likely on another forum.

Have a good day.

[astral]

Whoa - I wasn't trying to pick an argument, I was just asking out of curiosity. I wasn't intending to insult your experience or anything of that sort.

In any case - if you're correct that they fail regardless of whether they have fresh solder or not, that's very curious. It is still the GPU failing in these cases and not another component? If so then strange - the ATI 7500 chips themselves aren't failure prone as far as I'm aware, so it shouldn't be a bumpgate situation like the T61 is. Do you know what temps the GPU reaches in an A31? It certainly isn't the only laptop that stuck the GPU on the bottom of the board from that time - the last generation iBook G3 comes to mind there.

[ajkula66]

Whoa - I wasn't trying to pick an argument, I was just asking out of curiosity. I wasn't intending to insult your experience or anything of that sort.

All is well. Sometimes stuff written on the Internet doesn't translate to writer's intended thoughts. Quilty of that myself in many cases.

In any case - if you're correct that they fail regardless of whether they have fresh solder or not, that's very curious. It is still the GPU failing in these cases and not another component? If so then strange - the ATI 7500 chips themselves aren't failure prone as far as I'm aware, so it shouldn't be a bumpgate situation like the T61 is. Do you know what temps the GPU reaches in an A31? It certainly isn't the only laptop that stuck the GPU on the bottom of the board from that time - the last generation iBook G3 comes to mind there.

It's the GPU that fails with those lovely messed up colours all over the LCD...

There's no GPU temp sensor on the A3x series so I wouldn't know what the temps are, but the cooling on the entire generation is far from adequate IMO/IME. They do run hot in general. P4M was a failure in its own right but that's a subject for a different discussion.

I've never opened a Mac so I'll bow to whatever your experience says about them...

[kfzhu1229]

In any case - if you're correct that they fail regardless of whether they have fresh solder or not, that's very curious. It is still the GPU failing in these cases and not another component? If so then strange - the ATI 7500 chips themselves aren't failure prone as far as I'm aware, so it shouldn't be a bumpgate situation like the T61 is. Do you know what temps the GPU reaches in an A31? It certainly isn't the only laptop that stuck the GPU on the bottom of the board from that time - the last generation iBook G3 comes to mind there.

AFAIK the Radeon 7500 on the A31 is not a flip chip design with the underfills that thermally gets stressed, unlike the Radeon 9600, X300 FireGL T2 V3200 etc that has a silicon die. It is rather that the chip when laid bare heats up to a certain high temperature and stays at that temperature basically most of the time, and then it's the repeated power on/off that the thermal cycling itself makes it go bad.
If underfill were to fail, pressing on the chip or twisting the motherboard for example won't temporarily change anything.
The Radeon 7000/7500 on the ThinkPad R40 has a basic heat shield cooling with thermal sponge, and even that is not enough in itself to keep the GPU functioning reliably. the same Radeon 7500 also goes bad on T40/T41/T42 but that's more to do with the chip conveniently located where the PCB bends the most, made worse by again the chip being laid bare and 1st gen lead free solder.
Most other implementations of Radeon 7000/7500 I've seen all have cooling of some sort. All Dell Latitudes have it and its own dedicated heatsink/fan, and even the Acer Travelmate 650 has it.
My Radeon 7000 is now sitting under a stack of copper shims with proper thermal paste that make direct contact with the heatsink, and that seems to wick heat away from the chip reliably enough.

[dr_st]

Most other implementations of Radeon 7000/7500 I've seen all have cooling of some sort. All Dell Latitudes have it and its own dedicated heatsink/fan, and even the Acer Travelmate 650 has it.

I guess the Evo N610c cooler covers the ATI 7500 chip too. Haven't heard of these failing en masse.

[astral]

I believe the N610c is prone to this failure - at least, I'd sure think so after owning one. That thing had the single worst Pentium 4M thermal design I've ever seen. Loud, droning fan that was on at all times, and yet it ran like an absolute furnace. The hard drive is right next to the CPU and heatsink as well, so it gets so hot that CrystalDiskInfo will throw a big red warning due to the high temperature warning. I wouldn't be surprised if a lot of those suffered from premature hard drive failure due to that.
(And yes, it was repasted and cleaned).

As for the GPU specifically, they stuck it right under the keyboard with a crappy thermal pad that's supposed to transfer heat to the bottom of the keyboard. I can't imagine that does much. How reliable they are is going to depend on the solder quality.

[dr_st]

The hard drive is right next to the CPU and heatsink as well, so it gets so hot that CrystalDiskInfo will throw a big red warning due to the high temperature warning. I wouldn't be surprised if a lot of those suffered from premature hard drive failure due to that.

The hard drive in those idles at 50C. It's horrible. That's the big fault with these models. Mine went through three hard drives. Would have gone through more but it stopped being useful in general and went into a closet.
If I had some twisted need to revive it now, I would have used some sort of SSD with an IDE adapter. These probably wouldn't mind the temperature as much.

As for the GPU specifically, they stuck it right under the keyboard with a crappy thermal pad that's supposed to transfer heat to the bottom of the keyboard. I can't imagine that does much. How reliable they are is going to depend on the solder quality.

Interesting, so the heatsink does not cover the GPU? I may have been mistaken.
I would have just taken the keyboard off since it's so easy on these models, but the machine is miles away from me at the moment.

Oh, and the keyboard itself was pretty bad. Mushy as hell. BUT... it has the traditional Thinkpad 7-row layout, complete with Fn on the bottom left, and a Synaptics trackpoint compatible with the Thinkpad caps! Which is great because the original black caps are awful.
And it has USB 2.0 as well unlike IBM's contemporary solutions. Compaq integrated a NEC controller into those models, since the Intel ICH3 only supported USB 1.1. With a swappable drive bay, optional WiFi... Wouldn't have been so bad if it didn't keep cooking those HDDs!

[kfzhu1229]

The hard drive in those idles at 50C. It's horrible. That's the big fault with these models. Mine went through three hard drives. Would have gone through more but it stopped being useful in general and went into a closet.

Looks like that is the reason why the Dell Latitude C600/C610/C640 have the C-bay and battery bay swapped compared to the C8xx series and earlier models so that the hard drive resides on the left hand side, far and wide clear of the heatsink on the right hand side

As for the GPU specifically, they stuck it right under the keyboard with a crappy thermal pad that's supposed to transfer heat to the bottom of the keyboard. I can't imagine that does much. How reliable they are is going to depend on the solder quality.

Yeah I saw that, being just a naive thermal pad, not like the C840 which has a copper heatpipe and coldplate strapped under the keyboard which wicks heat away from the GPU chip in that manner.
The Acer Travelmate 650 has the Radeon 7500's heatsink as a part of the PCMCIA cage, also on the far left side like the N610c. I don't think that cooling is any better than the cooling on the R40's Radeon 7000/7500, but being positioned on the edge of the motherboard probably makes it being subjected to physical stressed far less, while the thermal pad technically doesn't have much cooling performance, but will contribute significantly to cutting the spike on thermal loads compared to leaving it bare.
I have yet to gotten my hands on any of the Compaq business class laptops before the HP Compaq nc6xxx/8xxx era

[astral]

The hard drive is right next to the CPU and heatsink as well, so it gets so hot that CrystalDiskInfo will throw a big red warning due to the high temperature warning. I wouldn't be surprised if a lot of those suffered from premature hard drive failure due to that.

The hard drive in those idles at 50C. It's horrible. That's the big fault with these models. Mine went through three hard drives. Would have gone through more but it stopped being useful in general and went into a closet.
If I had some twisted need to revive it now, I would have used some sort of SSD with an IDE adapter. These probably wouldn't mind the temperature as much.

As for the GPU specifically, they stuck it right under the keyboard with a crappy thermal pad that's supposed to transfer heat to the bottom of the keyboard. I can't imagine that does much. How reliable they are is going to depend on the solder quality.

Interesting, so the heatsink does not cover the GPU? I may have been mistaken.
I would have just taken the keyboard off since it's so easy on these models, but the machine is miles away from me at the moment.

Oh, and the keyboard itself was pretty bad. Mushy as hell. BUT... it has the traditional Thinkpad 7-row layout, complete with Fn on the bottom left, and a Synaptics trackpoint compatible with the Thinkpad caps! Which is great because the original black caps are awful.
And it has USB 2.0 as well unlike IBM's contemporary solutions. Compaq integrated a NEC controller into those models, since the Intel ICH3 only supported USB 1.1. With a swappable drive bay, optional WiFi... Wouldn't have been so bad if it didn't keep cooking those HDDs!

I've found that the mSATA SSDs actually run pretty hot, so I'm not so sure how much that would help.
Wish I had known about the USB 2.0 when I owned one. You must have had to install some driver to get it working as it only ran at 1.1 speeds when I had it. I vividly remember waiting ages for CD-ROM disc images to copy over...
It definitely is 2.0, the specsheet from Compaq said so.

[astral]

Looks like that is the reason why the Dell Latitude C600/C610/C640 have the C-bay and battery bay swapped compared to the C8xx series and earlier models so that the hard drive resides on the left hand side, far and wide clear of the heatsink on the right hand side

Actually, the earlier models have it in the same place as the C6xx models. At least, the CPi does.

[dr_st]

I've found that the mSATA SSDs actually run pretty hot, so I'm not so sure how much that would help.

That is correct; I observed it in my T430s, where the mSATA SSD idled at 44C, while the SATA SSD in the main bay was at 28C at the same time.
There are also "true" IDE SSDs, but they are rare, crappy, expensive and buggy, or so I'm told.
But anyhow the question is not so much whether they run hot or not, but whether they can cope with this level of heat for prolonged periods without failing. I guess only an experiment could tell, and I'm not sure who/if has done one.

Wish I had known about the USB 2.0 when I owned one. You must have had to install some driver to get it working as it only ran at 1.1 speeds when I had it. I vividly remember waiting ages for CD-ROM disc images to copy over...

I believe early versions of WinXP (the kind that would have been shipped on the original installation CDs, pre-SP2 or maybe even pre-SP1) indeed required a separate driver/hotfix to enable USB 2.0.

[astral]

I was running fully patched XP so I should have been fine - I think I installed RTM but ran the service pack installers - or so I remember, it's been a couple years.