Restoring a Compaq Portable III

[IcedMuons]

First time posting to the forum. I had a Compaq Portable III 386 back in the late 90s and wrote my undergraduate honours thesis in chemistry on it. Sadly I made the mistake of parting with it back in 2007 after not using it for several years. I'm now very keen to get one of these back up and running to use as a head node for our home computing as well as authoring some manuscripts on again for fun.

I have a Compaq Portable III 286 that's booting and has a Compact Flash to IDE card. It's too gutless to really do anything serious so I picked up a second Compaq Portable III off eBay for a very good price, with the only down side being that the seller said "Power supply is not attached"... actually they just meant there was no power cord. Plugging it in I could hear the fan engage for a fraction of a second and then everything stopped. Opening up the back and powering on again I can also see the red LED come on indicating a fault. Given how quickly everything stopped I suspected a short or something similar. Luckily I didn't have to investigate long and spotted the culprit as soon as I got the metal system board cover removed:

There's not much left of C47 and this is a little beyond my usual skill range - am I correctly identifying this as a tantalum capacitor? I think this would explain the power issue as I don't see any other obvious signs of capacitor problems. What the heck part to I order to replace this? Does anyone else with a Compaq Portable III know? There's not even enough left to read a part number off it!

There's signs I'm not the first one in the case as the screws were way over-tightened, a couple of the internal screws weren't put back but I can see signs of wear from when they were there. It's got a modem as well as the memory expansion board, except the board has no SIMMs at all (queue the hunt). Sadly it's missing the 32-bit interface board that bridges the memory expansion board and modem with the main board. Luckily I can salvage the interface board from my 286, as its interface board just sitting doing nothing as it has no additional expansion cards.

... it looks so empty!!

The main board has these SIMMS, but I'm not having any luck finding something equivalent on-line yet (except some site called Allegro.pl, but they don't quite look the same as mine):


I also have this empty slot for another chip on my board (pic below) - does anyone know what this is and what component I should be looking for? Is this a co-processor, maybe?


It also came with the expansion unit, which I've never had, but there's absolutely nothing in it except the board that connects to the expansion port.


Anyone have suggestions for things I should be considering for this?
I have a Gotek floppy emulator but I can't get my 286 Compaq Portable to recognize it regardless the jumper settings I use, but I'd ideally like to get this 386 setup with:
1. Booting off a Gotek or some other internal drive
2. USB memory support for transferring files if the Gotek floppy emulator isn't possible
3. Expand the memory up to the maximum possible
4. Internet access - telnet/ssh via Windows 3.11 would be sufficient. (I saw folks talking about a parallel port wi-fi adapter??)
5. Make use of all that space in the expansion unit - possibly a sound card, additional storage, or something else I haven't considered.

Any suggestions and info from the community here would be greatly appreciated - especially with identifying that capacitor
Thanks for any help you can lend!
~IM

 

[pbirkel@gmail.com]

There's not much left of C47 and this is a little beyond my usual skill range - am I correctly identifying this as a tantalum capacitor? I think this would explain the power issue as I don't see any other obvious signs of capacitor problems. What the heck part to I order to replace this? Does anyone else with a Compaq Portable III know? There's not even enough left to read a part number off it!

Yes, it's a dipped tantalum decoupling/smoothing capacitor on the adjacent power-input rail. Lacking a schematic I'd (1) look at the markings on the nearby looks-the-same tantalum capacitor and determine what the voltage is on its associated power rail, and (2) determine the voltage on the blown-tantalum power rail. In general the voltage rating for tantalum capacitors should be at least 1.5x the rail voltage; 2x never hurts! You may be able to analogize from the other tantalum for rating and capacitance given the supply rails. Lacking that I'd start with 47 uF; lower may be fine, higher is unlikely. Examples: https://www.mouser.com/c/passive-co...um-capacitors-solid-leaded/?capacitance=47 uF

This one might be a good bet (unless that's a 12 Vdc rail!): https://www.mouser.com/ProductDetail/KYOCERA-AVX/TAP476M016SCS?qs=IU2FXLSM7M%2BysEL7/EpgXw==

 

[IcedMuons]

Thank you. This was helpful.

It also helped to find the second exploded tantalum capacitor:


Oof. I carefully checked each of the other caps on the board and they do not appear to be damaged, so it's just these two now. This one is adjacent the power connection to the drive bay. The initial capacitor I spotted in my first post was next to the PSU connection to the motherboard.

The neighbouring capacitor to the blown one next to the main power connection is this guy:


I appreciate the polarity marking on the appropriate side, but would the "10-15" be an old manufacturer code or human-readable information? Could this be 10 uF, rated for 15V ? Looking in the diagnostics manual (which does not have board schematics or much detail in regards to identifying pins, etc) I see the PSU supplies +5 VDC, +12 VDC and -12 VDC

Checking out my old Compaq Portable III 286 (which is quite different from the 386 board, actually) I see the same location next to the power, marked C34, has:



Hmmm, 91-01??!? What a mystery this is. LOL... orrrrr 10-16 So again, 10 uF and rated for 16 V? or am I just out to lunch making the wrong assumptions about these markings? From the page capacitor-codes-markings it would imply that 10 is the capacitance value and 16 is the working voltage. As suggested by pbirkel@gmail.com above I'm expecting 47 uF will be as good or better than the 10 uF cap. Is this likely just stabilizing the voltage on the board and minimizing noise/ripple? It has been 28 years since I last cracked open my copy of Horowitz and Hill in university, so this is all fuzzy hobby-level understanding and recollection at this point.

~I

 

[IcedMuons]

I may only be illuminating myself here, but in case anyone stumbles upon this thread and wonders, I determined that the empty slot in the image above is for the 80387 math coprocessor which provides faster instructions for floating-point operations as well as some more complex operations. As can be seen on the current CPU in the picture, I have a 386 20 MHz CPU, best option would be an 80387 20.

On a related note - due to the significant mother board differences between the Model 2660 (286) Compaq Portable III, and the Model 2670 (386) systems I can confirm that the 32-bit memory/modem interface boards (sometimes referred to as the backplane) that serves as a bridge between the motherboard and these extra cards that get snugged up underneath the system, are not interchangeable.

While it does look like the connections should all line up between the memory boards and the 32-bit socket on the motherboards, I tried the interface board from the 286 on the 386 board. Although the modem and memory expansion board fit back into their slot, the placement causes the interface board to block the LED/Speaker and Keyboard cables, which could be re-routed, but the bigger problems are that the 32-bit connectors do not align, and the board clashes with several of the standoffs as well as the IDE cable - so the hunt is on for the correct part (No. 107707-001).

 

[pbirkel@gmail.com]

I appreciate the polarity marking on the appropriate side, but would the "10-15" be an old manufacturer code or human-readable information? Could this be 10 uF, rated for 15V ? Looking in the diagnostics manual (which does not have board schematics or much detail in regards to identifying pins, etc) I see the PSU supplies +5 VDC, +12 VDC and -12 VDC

Modern capacitor markings can be inscrutable; so little surface space for marking. Yes, that would be 10 uF maximum 15 Vdc -- perfectly reasonable for a 5 volt rail.

Hmmm, 91-01??!? What a mystery this is. LOL... orrrrr 10-16 So again, 10 uF and rated for 16 V?

Yes. Hopefully another 5V rail. There ought to somewhere be some capacitors with higher voltage ratings on the +-12V rails; may not be tantalum.

or am I just out to lunch making the wrong assumptions about these markings? From the page capacitor-codes-markings it would imply that 10 is the capacitance value and 16 is the working voltage. As suggested by pbirkel@gmail.com above I'm expecting 47 uF will be as good or better than the 10 uF cap.

Yes. 47 uF was my "worst case" expectation. 10 uF is a more "moderate" value and completely appropriate in these circuits.

Is this likely just stabilizing the voltage on the board and minimizing noise/ripple?

Yes. The other MLCC capacitors (shorter, yellow) are also there to reduce high-frequency switching noise; they're located electrically closer to the ICs rather than adjacent to the SMPS input lines.

 

[IcedMuons]

Modern capacitor markings can be inscrutable; so little surface space for marking. Yes, that would be 10 uF maximum 15 Vdc -- perfectly reasonable for a 5 volt rail.

Yes. Hopefully another 5V rail. There ought to somewhere be some capacitors with higher voltage ratings on the +-12V rails; may not be tantalum.

Yes. 47 uF was my "worst case" expectation. 10 uF is a more "moderate" value and completely appropriate in these circuits.

Yes. The other MLCC capacitors (shorter, yellow) are also there to reduce high-frequency switching noise; they're located electrically closer to the ICs rather than adjacent to the SMPS input lines.

Thanks for all of this clarification! I'm out of my element here with this stuff - as a chemist by training you can imagine how unsettling that must be, haha