Uninterruptible power supply

[jheatac]

Hello,

I have a small 500va uninterruptible power supply labelled IBM OP500. I am wayyy over my head but if I let the smoke out it is not a big deal.

I've had this thing 20 years or so. It's on its 4th battery, I believe. The battery light started flashing a few months back so I ordered a replacement.

I put it in and gave it a couple days and the battery light still flashes bad battery. It does not carry a load when line voltage is removed.

I threw it on the bench yesterday and determined that it is charging - which I assumed because it got acceptably warm when I changed the battery.

My question is if anyone can give me a direction to a chip to look around.

I am certainly charging. How do these charge and sense battery voltage at the same time?

It does switch to battery when line voltage is removed but it isn't enough to carry a load. I had a 200w 12v power supply on the bench (idle) I used for a makeshift load and it was getting 24vac, I believe. So it is switching and trying, but not getting any current. I did not test output without a load.

My thinking is the same pathway that senses the battery voltage has a considerable voltage drop.

I don't have the board out to get pictures yet.

[indeterminate]

You need to get yourself a multimeter without one you are blind to whats going on
https://eevblog.store/collections/test-equipment
once you have a meter you can see if the battery is being charged at the wright voltage
and go hunting for the fault if it is not

[jheatac]

I'm going to file this under "Even a blind squirrel finds a nut every once in a while"

I separated the case and inspected the board. Everything looked perfect. Nothing hot. Solder connections all good. I checked the battery while it was charging and it was 13.6v to 14v

Being overwhelmed at where to start, I checked ohms on the power switch - SPDT slide. Really not on/off... More charge/and active UPS. It has a green LED regardless. When off, the charge LED blinks.

One way was 0.12ohm, roughly the test leads. When I checked the other, it was in the 100 ohms then kohms then if I wiggled the switch I got it down to 22 ohm. I took the switch apart and cleaned the contacts and the unit works again.

If only they were all so easy.

[MrAl]

I'm going to file this under "Even a blind squirrel finds a nut every once in a while"

I separated the case and inspected the board. Everything looked perfect. Nothing hot. Solder connections all good. I checked the battery while it was charging and it was 13.6v to 14v

Being overwhelmed at where to start, I checked ohms on the power switch - SPDT slide. Really not on/off... More charge/and active UPS. It has a green LED regardless. When off, the charge LED blinks.

One way was 0.12ohm, roughly the test leads. When I checked the other, it was in the 100 ohms then kohms then if I wiggled the switch I got it down to 22 ohm. I took the switch apart and cleaned the contacts and the unit works again.

If only they were all so easy.

So then maybe the old battery still works too.

[jheatac]

Yes, it does.

I ordered a new battery (~$30) without checking. Old one was 2018 so I just changed it out anyways

[AG6QR]

If only they were all so easy.

They aren't ALL so easy, but quite a few are.  You've learned a valuable lesson here: switches, connectors, and contacts are frequent sources of trouble in older electronics.  A bit of corrosion builds up, and then the pieces of metal no longer make a solid electrical connection when they touch.  A lot of failing older devices can be fixed by a generous spraying of deoxit contact cleaner, combined with working the switches back and forth to try and remove the corrosion from mating surfaces.  A device with card edge connectors can often be "fixed" by removing the boards from the connectors, cleaning them a bit, and reseating the  cleaned boards into the connectors.

Cables that flex are another common source of problems.

Electrolytic capacitors of a certain age will dry out, develop high ESR, and make devices behave badly.

After many decades, old carbon composition resistors increase their resistance by a lot, particularly if they have been running hot.

A good, experienced, repair technician has the skill to troubleshoot in detail, but also has the experience to have a pretty good intuitive feel for the most likely components to fail.  Depending on the situation, they might start with a quick and inexpensive fix like spraying deoxit on the switches before spending a lot of time with detailed troubleshooting.