UPS Logic Power Issue?

[ImpulsiveJames]

Hi, I've been trying to teach myself circuit repair over the past few years and am making progress, but this 36v UPS has me stumped...

The unit will invert and put out 120v, but these components get very hot and am worried this is not normal behavior.

The main components that get hot are a:
7812ACT - 36v to 12v voltage regulator
DIODE ZENER 24V 1W   
BJT TO92 80V NPN 0.625W

Any tips or pointers would be apricated. I'm happy to provide any additional readings or pictures

[cruff]

Possibly normal for that unit's design. The 7812ACT is a linear regulator that has to drop at least 12 V (if fed from the zener) or 24 V (if fed directly from the 36 V supply) across it, so it will get warm in normal operation. Depending on the value of the resistor used for the zener current limiting, it will dissipate a maximum 1 W of heat unless it is designed to be more conservatively driven. Even at 0.5 W it would heat up a fair amount. Depending on what the NPN is being used for it could also warm up in normal operation.

Is there some other issue that causes you to think the UPS isn't working to spec?

[ImpulsiveJames]

Actually thank you for asking fairly simple question, the battery died and the unit wouldn't turn on to charge. After poking around looking for potential issue I found out that the whole unit is actually driven by the battery, not the AC from the wall. So you could totally be right and there might not be anything wrong with it. But when I use the thermal camera and saw it was reaching temperatures above 155F, I just immediately assumed there was an issue....

I'll get some new batteries attached to it, instead of driving it for my bench power supply, run some load on it and see how it does.

Thanks for the quick reply!

[cruff]

After poking around looking for potential issue I found out that the whole unit is actually driven by the battery, not the AC from the wall.

I found this to also be the case for an APC BackUPS 650 unit also when it's battery went flat after sitting unused for a year. Had to install a new battery to get any response out of the control board to verify it was otherwise functional.

[ImpulsiveJames]

Ok, so after measuring for a few mins, it looks like the 7812 starts leveling off around 210F. The data sheet says it is functional up to 257F (125c). You really believe this is how it is supposed to be? I assume they designed this unit around only being on battery power once a year for about 3 years, so I can't say I'm surprised, but... just skeptical, that is all

Another thing I noticed, but am not sure about is it looks like the circuit is not accurately measuring the 3 battery supply. It shows only 1% battery remaining while having 37.25v. I'm using car batteries, but I assume that should not matter.

I'll see if I can locate what chip does that measurement and see what I can find

[ImpulsiveJames]

I could also have this wrong, but a few days ago I removed the 7812 and measured it inline and with a 32v bench power supply the 7812 measured 22.5v in, 12.08v out with 60mA of current.... I probably should have checked the input amperage also, but..... did not

[cruff]

Ok, so after measuring for a few mins, it looks like the 7812 starts leveling off around 210F. The data sheet says it is functional up to 257F (125c). You really believe this is how it is supposed to be?

Without knowing more about how the rest of the circuit is designed, I can't say for sure if that is how it is supposed to be. It is, however, pushing that regulator pretty hard, which is wasting energy that could potentially be sent to the load instead. It could also be a poorly designed unit for efficiency.  Do you have a brand and model information? Part numbers for the board?

If there is another issue that you have not yet found with the circuitry being fed by that regulator, it could explain why it is getting that hot. Have you identified those components? Have you started making notes by sketching a schematic for them? Are there any other potentially failed components from a visual inspection? What about testing diodes or transistors having shorted with a DMM?