exxos's Atari, Amiga & retro forum

Hi all,

I am trying to investigate a possible fault with the back-plane of an SGI o2 machine where the power-on button stops working randomly.

The power on button is the second micro-switch from the left, the one next to the green LED (SW502) The first is reset.
I checked tantalum caps for shorts, but they all measure out of lock (infinite resistance?) I assume they're only bad if shorted.

The resistor networks at the top are to do with SCSI I think and all measure around 0.8ohms.
The micro-switch itself seems to test fine.

Is there anything obvious I can check? Or am I possibly testing the capacitors wrong?

How are you testing the caps?

To see if they are shorted, using a multimeter in continuity mode, put one probe on GND and the other to each one of the tantalum contacts one after the other, so + then -.

The - side should cause a beep, the + side shouldn't beep. If it's shorted, both sides will beep.

Just be aware that I have seen tantalums used in negative voltage circuits, in which case this can be reversed.

Wanted to get my hands on an O2 but ended up with 2 Octanes instead!

@rubber_jonnie Hey well if I get this fixed maybe you might want it, I'm currently in the process of restoring two.

Back to the tantalums... I'm just putting the multimeter across them (directly on them) I didn't think it'd matter which direction since they're not a diode. I thought, if it beeps, it means it's shorted. Your way sounds a bit more complex? Am I misunderstanding :)

Steve wrote: 13 Oct 2024 16:11 @rubber_jonnie Hey well if I get this fixed maybe you might want it, I'm currently in the process of restoring two.

Back to the tantalums... I'm just putting the multimeter across them (directly on them) I didn't think it'd matter which direction since they're not a diode. I thought, if it beeps, it means it's shorted. Your way sounds a bit more complex? Am I misunderstanding :)

Connecting straight across should work, it just sounded from the way you described it like you weren't doing it in continuity mode and were getting weird readings.

My way isn't more complicated, it's essentially the same, but if you have a lot of caps connecting one lead to the GND of the PCB and using the other to test for shorts is a bit simpler as you only have to wield a single lead to do all the testing.

For example on a games console with hundreds of tiny capacitors, connecting one lead to GND and then using just one lead for testing is way quicker IMHO.

And no, they aren't like a diode but they are polarised with the band indicating +ve, but that doesn't matter for testing for shorts.

Are you sure it isn't a problem with the powersupply?
Do you use this button to power off to?, does that stil work?

Check the soldering's, replace the button.

There are multiple ways caps can fail, measuring a short is only one way for them to fail...


No or wrong capacitance.
High ESR.
Abnormal behavior when cold/warm.
Leaking corrosive liquid.
Deformation.
Expolding.
....
....