PS/2 Model 70 Power Issue

[oldpcguy]

A few weeks ago I posted an inquiry as to where I could obtain a floppy drive for this computer. As luck would have it a seller had a non-functional one and I purchased it for $50. The idea was to obtain the floppy drive, I didn't really care about the rest of the computer. I lucked out and the floppy drive does work so my original goal was met with this acquisition.

That said I would like to see about fixing this computer. So I powered it on and there was initial life and then it shut down. Thinking it was a problem with the hard disk I removed it and the system powered on, ran through its memory test, and showed the usual errors that come with such an old system. Then it just powered off.

I suspect the problem is in the power supply and, having a fully functioning system (except for the aforementioned floppy drive), I can swap in the known working power supply. Since the system is dirty I'll be performing a thorough cleaning but that will come over the weekend. In the meantime I thought I would post here for any suggestions. As I said the power supply is the first thing I suspect. I have the tools (DMM, scope, ESR meter) to troubleshoot it and the rest of the system. However I'll only be able to get to it as time permits so thinking I'd ask.

Oh, about the floppy drive. It's currently working but they're well known to fail over time. I think the capacitors leak on these. I've replaced them on the other drives I have so I know how to do so and have the experience. My question is: Should I? Or just use as is until it fails and then replace them? Leaking caps can cause PCB damage so I'm leaning towards proactively replacing them. But, given how difficult these are to come by, something about leaving something that works alone also sounds good. Advice?

 

[NeXT]

I forget if the power supplies for the model 70 were from Astec or Sony but preventative maintenance would absolutely be okay at this point.

 

[oldpcguy]

I forget if the power supplies for the model 70 were from Astec or Sony but preventative maintenance would absolutely be okay at this point.

Thanks NeXT. The preventative maintenance was wrt to the floppy drive. At this time I'm leaning towards a repair of the power supply to fix the power on issue.

 

[oldpcguy]

I began the disassembly of the system and discovered a broken DIMM socket retainer. I suspect this may be the reason for the system shutting down (it powers up, performs the memory test, display errors codes and then immediately shuts down):



As you can see the logic board is quite dirty. The DIMM sockets themselves are also quite dirty:



I'll be washing the board (as well as the case components) tomorrow since the weather will be nice. The only problem is I am unable to remove the logic board as it is stuck near the rear power supply. It's as if it is screwed in but there are not screws. I've removed the three from the back and, for good measure, the six nuts for the various connectors (I don't believe they need to be removed).

So, what's the secret to separating the logic board near the back of the unit?

This is the full logic board (as opposed to the short one).

 

[Unknown_K]

U115 looks blown or has some brown goo stuck to it, u126 has a spot that looks burned by one of the legs, the capacitor to the left of it needs replaced.

 

[oldpcguy]

U115 looks blown or has some brown goo stuck to it, u126 has a spot that looks burned by one of the legs, the capacitor to the left of it needs replaced.

Damn, great catches. I was so preoccupied with removing the logic board I hadn't even performed a visual inspection. Now that you've pointed these out I took a performed an inspection. U115 is fine, it's just goo. U126 looks to have burn marks, I'll see how it looks after cleaning. Two capacitors, including the one you referenced, are blown out. This is making me rethink the DIMM as being the problem. It looks as if something pulled too much power. I wouldn't be surprised if it were the two tantalum caps. Good eye!

 

[whaka]

burnt spot can be rest of the exploded capacitor, but definitely worth a check.
for the PSU, you can check capacitors at the output. but at your point, it might still be the blown capacitor shorting randomly.
simply cut it for the moment, the machine can perfectly live without one

i recently aqcuired a 70, and ran into power troubles after some hours of use.
when the floppy drive was accessed, it ended with a total crash of the machine with a ram code error, and... PSU whining.
no doubt i smelled capacitor issue, and it was... all capacitors at the power output were either bulgy or had leaked electrolyte.
i have an astec PSU.

 

[oldpcguy]

burnt spot can be rest of the exploded capacitor, but definitely worth a check.
for the PSU, you can check capacitors at the output. but at your point, it might still be the blown capacitor shorting randomly.
simply cut it for the moment, the machine can perfectly live without one

i recently aqcuired a 70, and ran into power troubles after some hours of use.
when the floppy drive was accessed, it ended with a total crash of the machine with a ram code error, and... PSU whining.
no doubt i smelled capacitor issue, and it was... all capacitors at the power output were either bulgy or had leaked electrolyte.
i have an astec PSU.

I'll definitely be giving it a thorough inspection once I've had an opportunity to clean it up. I hope to do that today if I can manage to figure out how to get the logic board out of the case. I did it on my other PS/2 but it has the short board (though I don't think that matters given where it's hung up).

 

[whaka]

yes i forgot about this point, you need to pull hard to get it out.
i have a short board too, and had exactly the same "problem", so seems normal :D

 

[oldpcguy]

yes i forgot about this point, you need to pull hard to get it out.
i have a short board too, and had exactly the same "problem", so seems normal :D

That's what I ended up doing. Got it out and all the parts have been washed. Unfortunately I wasn't able to wash the part which holds the hard drives, floppies, battery, and speaker as well as I would have liked as I didn't feel like removing the speaker. So I just dunked specific portions of it in the water making sure to avoid getting the speak wet. I'll be doing a "hand" cleaning on that area. I will also be pulling apart the power supply to clean it out but for the time being I just blew it out with compressed air. All parts are currently sitting in the sun ensuring everything is thoroughly dried (after I used compressed air to blow off the water).

Once done I'll be doing a detailed inspection of the logic board paying particular attention to the area where burn marks are suspect.

 

[oldpcguy]

All cleaned and ready for service. I need to purchase some 106, 25k tri legged capacitors. On initial search it looks like I may have to settle for dual leg. Here's a close up picture of U146 which looks as if the three pins are burnt. Everything is fine with it and also U115 (which is not pictured). There is a small divot between pins 1 and 2 but nothing that looks like an issue.

It appears all I need to replace, at least at the moment, are the two tantalum caps. I'll order some and, in the meantime, pull these and see if the system still powers down.

EDIT: Just pulled the two tantalum capacitors and will try powering it up.

 

[oldpcguy]

After partially reassembling the system it is still behaving in the same manner where it powers on, performs a self test, and then shuts down. If I attempt to power cycle it it immediately shuts down (this was the behavior I originally observed before I performed the cleaning). In a way it's not surprising as I suspect the blown capacitors were open and thus not causing the shutdown (but I was hopeful they were).

Looks as if I'll have to break out the DMM and start poking around. First test will be to ensure no short on the logic board between ground and any of the voltage rails. If that passes the next step will be to crack open the power supply and see what I can find.

 

[whaka]

you can suspect the PSU, because what happen after memory test is floppy drive seek and thoses drives pull on the 12V for their motors.
and if output capacitors are dead, this can cause a serious regulation issue between 12 and 5V, and trigger the shutdown.
it was exaclty what caused my 70 to crash, not shutdown, but crash when the floppy drive pulled on the 12V

 

[oldpcguy]

you can suspect the PSU, because what happen after memory test is floppy drive seek and thoses drives pull on the 12V for their motors.
and if output capacitors are dead, this can cause a serious regulation issue between 12 and 5V, and trigger the shutdown.
it was exaclty what caused my 70 to crash, not shutdown, but crash when the floppy drive pulled on the 12V

Thank you for the guidance, always nice to have a starting point. I am hoping to start digging into it in a few hours.

 

[oldpcguy]

Pulled apart the power supply and it was quite dirty. Here is a pic after I had previously blown it out with compressed air (prior to disassembly). Hoping, but not truly believing, a thorough cleaning might solve the issue I put the cover back on and attempted to power on the computer. I am getting mixed results. Sometimes the system powers up (and gives me a 201 error) other times the fan spins for a second and stops, and sometimes the fan doesn't spin at all. When it does power up it stays at the 201 code until I power it down (which has been maybe five minutes at most).

Looks as if I'll need to pull it all the way apart and start probing around. First step is the ESR meter to test for bad caps, then on to the DMM.

 

[whaka]

yep, astec PSU.
if you want to do the bare minimal, change the 4 brown nichicon capacitors. (2x 2200, 2x 220µ)
or do the full job, and change all capacitors. there's not a lot in this one

 

[oldpcguy]

By swapping in the working power supply from my original system I was able to confirm the power supply is the problem and, with a working power supply, the new system boots (at least into basic as I don't have the floppy or hard disk installed).

I attempted to power it on without the DIMM module in the damaged socket and I received the 201 error. Will the system eventually move forward without the module? Or do I have to reconfigure the system using the floppy? For my testing I just inserted the second memory module and held it in place.

As for repairing the power supply do you know if anyone is selling a recapping kit? If not what the list of capacitors are?

 

[NeXT]

First step is the ESR meter to test for bad caps, then on to the DMM.

Don't even bother with the ESR test. 30 year old capacitors in a power supply are basically preventative maintenance at this point regardless of brand.

Also you don't need tri-leg tantalums. IBM only used those because the component insertion machine couldn't install a cap backwards with tri-leg. Dual leg caps will work absolutely fine and cost less.

 

[oldpcguy]

Don't even bother with the ESR test. 30 year old capacitors in a power supply are basically preventative maintenance at this point regardless of brand.

Also you don't need tri-leg tantalums. IBM only used those because the component insertion machine couldn't install a cap backwards with tri-leg. Dual leg caps will work absolutely fine and cost less.

Since ESR testing can typically be done in circuit there's no harm in doing the test. It takes second per cap.

Unless I can find a recapping kit hunting down the individual caps takes time. I had the same problem with Apple II power supplies until I found a recapping kit. Thus if I can't find a recapping kit I'll perform the ESR test and then buy those caps first. Then assemble a list to buy over time.

Yes, I am aware I can use dual leg but I tend to like to keep things as "original" as possible. If tri legged caps are difficult to find or unreasonably high cost then I'll definitely consider dual leg.

 

[NeXT]

Another thing I've run into several times now is that "ESR" can be misleading in older switchmode supplies. That one catches the new folks off guard because you clean and recap a PSU perfectly with modern low-ESR's from DigiKey and now it's running even worse than before and all you did was change capacitors with identical value replacements.
I've recapped several PSU's using modern low-ESR parts and found that the new decoupling and controller input filtering caps worked too well and it confused the microcontroller. Switching to general-duty versions fixes it.