NEC PowerMate Portable Plus: Debugging fault in ISA/screen connector board

[mikerofone]

Hi all,

TL;DR:

I need to identify a fault in a connector board that seems to only house connector headers and passive components: capacitors, resistors and some mystery things I need help identifying. What are the things marked "F1" through "F12" on the silkscreen? I'd have guessed fuses, but they have three pins. The outer ones measure as direct shorts, and to the middle pin they both measure as open lines.




The long story

I've got two NEC PowerMate Portable Plus units recently (not much information online, but there's this). After recapping the power supplies, which both had very very failed caps (there was a puddle of electrolyte in the PSU housing) they both worked fine for a couple hours.

Then after turning one on after a few weeks of it being stored in my apartment, it started fine for about 5-10 seconds but then made a somewhat wet, smoochy sound (really hard to describe) that wasn't the usual crackle/sparkle that I'd associate with a blowing tantalum capacitor. You could see that there was a short power drop (HD spun down and back up, backlight flickered) and I shut the machine off. I found some liquid in the PSU that was stuck in the fan header. Not sure where it came from (I didn't wash the boards, so maybe more electrolyte was hidden under components that managed to flow into there. I cleaned it out, and the PSU works again. But now the internal screen no longer shows an image, even though the backlight turns on and there is an image on the VGA output.

Aside from the screen, the machine is working fine again. I started swapping parts between the working and broken machines and verified the screen and mainboards both to be working. There is one connector board that sits between the screen and the mainboard, and houses the ISA ports as well as the connector for the front panel contrast/brightness knobs (see first image). I've isolated the fault to this board: If I use the one form the bad machine in any screen / mainboard combination, the internal screen illuminates but stays blank. Swapping for the good one makes any combination work well.

The top edge connector plugs into the mainboard, the screen into the connector on the opposite end with a ground wire screwed to the ground hole next to the connector. On the right, the front panel with the contrast/brightness controls connect.

I've verified the capacitors with an ESR/Capacitance tester and they check out good and with close to identical values on both boards. So I guess maybe these "F" components might be bad, but I don't know what they are or how to test them. The "B" ones seem to be simple wires ("bridge"?). Any suggestions?

I'll start beeping out traces to compare the good vs bad board, but I don't see any damages with the naked eye there. Hints as to where to go next are welcome!

Thanks for reading!
mikerofone

 

[DeltaDon]

Does the good board have different resistance readings for the "F" devices? If so, then there's your problem. As you stated "F" usually means fuse on PCB's, but a 3 leg fuses is something I've not seem myself. Also, it would not be normal for all the "F" things to blow out with a normal short on a single trace. So unless someone applied power to all of the traces you would expect only one or two of them to be defective. Lastly, there appears to be on "F13" that is bridged with a wire.

 

[mikerofone]

Thanks for the reply, DeltaDon! To my great surprise both boards are working again after reassembling the parts on my bench for taking some live measurements, even though I'd swapped all components a bunch of times before and thought that I had consistent behavior. Maybe a dirty contact? Very strange, but for now I assume it's "fixed" (just waiting for everything to start failing once it's properly assembled).

I'll give measuring the F things' resistance a go before I button everything up, I had not tried doing this before. The sparse markings on them didn't make any sense to me either, but I just got a component tester today. I might desolder one and stick it in there, maybe that'll shed some light. If I do that, I'll try to remember posting an update here.

Thanks again!

 

[mikerofone]

So I finally discovered the issue. It is the 35V rail in the power supply that failed, it's 0V on one PSU and, well, 35V on the other one. The fault follows the PSU for any mainboard, riser card and display unit, which makes a lot of sense and is a big relief, as that should be the easiest issue to fix.

I misdiagnosed this so grossly since these machines are quite annoying to take apart, with many screws and mounts to undo to get to the PSU, at which point assembling the mainboard, riser, display and PSU becomes hard since you can't lay everything flat. And the riser is mounted on top of the PSU, so while "swapping the connector board between the two machines" I also swapped the PSU. *facepalm*
I have to yet take the PSU apart to understand what exactly has failed. But since it sounded like the PSU made the unhealthy sound, plus the fact that there was liquid in there, at least paints a very coherent picture. There was also a lot of corrosion on some of the beefier traces, maybe one of these gave in. I hope to be able to investigate this weekend, and will report back.

I still haven't figured out what the "F" components are. I failed to desolder any, and when measured in-circuit my cheap component tester registers only one of the outside to middle pins as ~300-500pF capacitors, even if I connect all three pins simultaneously. My Peak ESR70 tester doesn't detect a component at all. But since I no longer need to figure it out, I'll leave it at that.

Cheers
mikerofone

Thanks for the input

 

[T-Squared]

If I had to guess, those are line-filter capacitors that serve the same purpose as the infamous RIFA caps that tend to blow in more-vintage computers. I have similarly-shaped ones (and replaced them) on my old Sanyo tank computer.

 

[mikerofone]

Thanks for the suggestion! They are on the low voltage side though, outside of the PSU. Are filter caps used there too? I don't know too much about how circuits work, but I'd assume that the voltages are already "filtered" after the PSU.

Cheers
mikerofone

 

[DeltaDon]

Good news that you have a working power supply to use for known good readings if you need to troubleshoot past obvious defective components. The bad news is screws, screws and more screws.

Some old laptops I've worked on have 8 or more different screw sizes plus were used by the pound, it seems. I've made maps of all the screw locations so I didn't wrong screw them putting them back together. One PITA laptop has that issue and needs to have the lid assembly and the top plastics removed along with the optical drive to just change out the 2032 CMOS backup battery. It takes nearly two hours start to finish just to stick a new one in. Dumb.

 

[mikerofone]

Finally got around to checking out the PSU. I couldn't find any visibly blown components in the PSU, and the 35V section seems to be entirely separate from the 5 and 12V section. To illustrate, check out this picture of the two boards in the PSU. The left board does 12+5V output, the right one 35V and mains filtering and rectifying. (Not sure what level the link from the right to the left board is running at, I don't really know how power supplies work in detail.)
Since I luckily have a working PSU to compare, I measured capacitors, transistors etc and didn't find any real deviations. The transformers seem fine too.

The only difference I found was on the DIP-14 chip in the top right. It's a Mitsubishi M51996P "switching regulator control" chip. I measure a near-short (11 ohms) between its GND and CLM+ pins (12 and 13), whereas it's closer to 100 ohms on the working PSU. The 12+5V section has the same chip, which measures ~62 ohms on both units.
The 11 ohm reading is unchanged after desoldering the chip, which makes this seem a plausible culprit. Since desoldering was a bit of a mess (the chip got very hot, and I destroyed two pads) I will not transplant a working one to see whether that fixes it. But that gives me a problem:

Where does one find replacement M51996P chips? I could only find offers directly from China on ebay/aliexpress, but I would prefer to buy from Digikey, Mouser and the likes.

I assume that maybe there is a more modern replacement that could be used instead, but I have no idea how to go about finding this out. (Googling "M51996P replacement" was not fruitful.) Any suggestions appreciated!

Cheers
mikerofone

Edit: Forgot to comment on the "screws screws SCREWS" point, my apologies. I've treated myself to iFixit's magnetic project mat, which keeps the screws mostly in place and allows me to sketch where they belong. I've let a project sit for weeks and was still able to put stuff together correctly once I got back to it. It's fantastic, can strongly recommend it!