Compaq Portable III capacitors repair

[vbug]

Hi, I powered my old compaq portable III for the first time for years, and bad surprise : electrical sounds and smoke... I quickly unplugged the machine power supply and have disassembled it : burned capacitors. So I'm now checking with a multimeter all capacitors on the mainboard in order to determine which one are working and which one should be replaced, according to their resistance. And here I would appreciate some help to be sure about what capacitor is defective. All values are measured with the orange power connector plugged (I noted that values can be different with or without that connector plugged). What I'm wondering on :

C2 (yellow) : around 2K ohm
C26 (yellow) : 0 ohm
C5 (orange) : 0 ohm
C6 (yellow) : 0 ohm
C43 (yellow) : 0 ohm
C109 (orange) : 0 ohm
C110 (orange) : 0.6K ohm (with the orange power connector from power supply plugged)
C24 (yellow) : 16 ohm (with the orange power connector from power supply plugged)
C37 (yellow) : 350 ohm
C52 (orange) : changing values
C80 (yellow) : 1.13M ohm
C90 (yellow) : 4.2M ohm
C94 (yellow) : 0 ohm

If someone could check those values on its own machine I would be very grateful. Thanks by advance

 

[schenkzoola]

I am waiting for some power supply capacitors to arrive from DigiKey, so I just happened to have one on the bench!

The following measurements were taken with the motherboard completely removed from the computer:

C2 - infiinite - C3, a tantalum next to it measured about 260 ohms. This is ceramic, so it's probably fine.
C26 - infinite - ceramic, on the 12v rail.
C5 - infinite - tantalum, so suspect, on the 12v rail.
C6 - infinite - ceramic, on the 12v rail.
C43 - infinite - ceramic, on the 12v rail.
C109 - infinite - tantalum, so suspect, on the 12v rail.
C110 - infinite - tantalum - on one of the negative rails
C24 - infinite - ceramic
C37 - I couldn't find C37
C52 - infinite - tantalum
C80 - infinite - ceramic
C90 - infinite - ceramic
C94 - infinite - ceramic, on the 12v rail.

I suspect C5 and/or C109 to be your issue.
Please unplug all the cables on your motherboard and re-measure. If you still have 0 ohms on the 12v rail, then snip one of the leads on C5 and C109. Measure between snipping each one so you can identify the culprit.

You could also have issues with your power supply. I load tested mine before I powered on the machine, and found its 5v rail to be unstable, which is why I'm recapping it. There are a couple tantalums in there as well, and I found a few of the electrolytics were well out of spec.

 

[vbug]

C109 is definitely burned, it is visually the one who smoked, and I suspect C5 too. One of my hypothesis is that I should change all orange tantalum capacitor without the same resistance than most of the others (around 250ohms). I ordered new ones, and some ceramic ones too, and I'll check the power supply as you suggest. Next update in few days! Thanks a lot for your help.

 

[schenkzoola]

I personally wouldn't change all of them.

There's probably no need to change the ceramics, they aren't likely to fail.

All the tantalums on my board have the same voltage rating, regardless of the voltage applied. I would consider changing the tantalums on the 12v rail, but not the 5v rail.
The capacitors on the 5v rail are not being pushed close to their ratings.

There is always the risk of damaging the motherboard when replacing things, so I think it's best to only replace what's necessary.

Also, please keep in mind that if the capacitors are electrically connected to something, the power supply, other capacitors, etc..., that will influence your measurements. To compare with my measurements you should disconnect all the cables from the board, including the floppy drives, power supply, modem, graphics card, etc...

 

[rimmeruk]

You should actually disconnect any component for testing out of circuit for a more reliable test. Many components are wired in parallel with other components which will affect the readings on your test equipment. You can do a quick short test around the board in continuity mode with your test meter. I guess you know the basics already but doesn't hurt to suggest it.

When it comes to caps I would certainly replace all the important ones and especially high voltage ones as suggested above. No real point in testing them, just replace them as you must remove them or at least ****-a-leg (excuse the pun) to test them correctly.

EDIT: Filtering wouldn't allow **** (c o c k) which I intended to be a phrase and not a obscene word

 

[schenkzoola]

When components have a reasonable amount of exposed lead, like tantalums usually do, I usually clip one of them about midway through if I want to remove it from the circuit for testing. If I want to put it back in circuit, I just move the cut ends close together and apply a little bit of solder to join. This doesn't risk the board.
When removing components from old boards, there is always the risk of damaging the through plating, or lifting a trace, or both. This is why I suggest replacing only what is necessary.

 

[vbug]

I finally replaced C109 only, and my machine is back \o/ I'm so happy ! Thanks a lot for all your advices and help <3