My daughter had her UPS die on her recently. She brought the unit to me to look it over.

On opening it up, I found that the main PC board is single sided copper with surface mounted parts on the bottom and a lot of wire jumpers to make up for the lack of copper on top. The output of the board feeds a transformer which then connects to the output outlets. The quick rundown of looking for burnt or exploded parts or bulging capacitors revealed nothing, but turning the board over I found issues with the solder one of the MOSFET's pins. They are all TO-220 package 3 lead parts. This particular MOSFET had no solder on one of its pins. It looked as if someone had used a solder sucker on it as one pin was completely clear of solder. Checking continuity on it revealed that it was toast. All pins had low resistance between them. The rest of the MOSFET's showed varying continuity.

My question for the forum is whether the unit will work if I simply clip the leads of the dead MOSFET leaving the other 7 in place? I would have tried it, but I don't have the batteries on hand. It has a pair of gel batteries to run off of 24 volts.

The MOSFET's in the unit are SM6002N which are a Chinese brand made by Sinopower. I haven't yet found a distributor in the US for them. If I do replace them I'll replace all 8 at once. I'm hoping to try the unit with the dead MOSFET removed. If it works that way then I'll assume the dead one is the only problem with the unit and I'll press forward with finding a source for new parts.

 

Worth a try. If all the leads are shorted, it may well be shorting out the gate drive for the other ones.

Edited to add: Of course, you should not use it near its rated output.

 

Both sides of the bridge / Transformer must have the same number of transistors. So a balance is maintained.

 

Meh. With nothing to lose and the batteries in hand, I tried cutting the leads on the damaged FET. It still doesn't work.

I haven't dug into reverse engineering the PC board, but it looks like there's a section for a 60 Hz oscillator that drives the power FETs. The FETs feed into a beefy transformer to get the line voltage AC. From what I could tell that should be always outputting AC and if the incoming AC goes away, the load is switched by relay to the output of the transformer. There was no output from the transformer at all. I may dig further, but the root problem could be a bad transformer or that the 60 Hz oscillator isn't present at the FETs.

 

Here is a warning, a bit late, though. MANY component failures do not produce any visible damage at all. So while an inspection will probably spot most burned parts, it is likely that another part failure caused the burned part to be overloaded.
Itis quite unlikely that the output transformer is the failed part. A simple resistance check can verify that. The fault might possibly be a relay, or more likely an electronic component.

 

Both sides of the bridge / Transformer must have the same number of transistors. So a balance is maintained.

sorry, why ?
will the remaining fets not share the load?

ok, yes , working with one fet missing , is going to put strain on the remaining, leading to progressive failure,
but I assume this is not the permanent sollution !

 

...this is not the permanent sollution !

Correct. It was a test to see if a single FET was bad and nothing else was wrong with the board.

As it turns out, something else is wrong. I've got a bit of sleuthing to figure out what.

 

Probably it is only at higher loads that all of the devices are needed. It might be that other devices have also failed, or more likely that some other component has failed.
Given the number of scrapped UPS devices I have seen, they are the most disposed of electronic equipment around.

 

Given the number of scrapped UPS devices I have seen, they are the most disposed of electronic equipment around.

Indeed, I'm not likely to spend a whole lot of time on this unit. I'm tempted to design my own circuit to drive the big transformer (it's a 1500 VA UPS) but with the pitiful amount of time available right now, I don't see it happening soon.

As well, my daughter has taken the batteries back for use in her own projects. She's already bought a replacement UPS.

 

I have found the relays from a scrapped UPS circuit board to be quite handy for other projects.. AND there are other useful parts, plus the outlets are often good quality.
Now I am wondering if the UPS output stage, driven by an audio amplifier and oscillator, could serve as a useful higher frequency mains supply, within some limits.

 

sorry, why ?
will the remaining fets not share the load?

ok, yes , working with one fet missing , is going to put strain on the remaining, leading to progressive failure,
but I assume this is not the permanent sollution !

It is not critical for small loads. But it becomes critical when nearing full loading.

 

It is not critical for small loads. But it becomes critical when nearing full loading.

As mentioned , it will put more strain on the remaining , but having same number of fets is not a requirement for it to work ,
Just clearing up in case anyone is designing and thinks it is a requirement to have same number

 

Do not try to run the unit from its gel batteries with the FET removed/clipped. The batteries can supply heavy fault currents that will likely create new damage or a fire risk.

 

Always unplug the batteries when working on the UPS, that avoids the hazard. An easy safety trick. AND if the user has already got the replacement, there may be useful parts to salvage. Possibly the remaining FETs are good, and can be used for experimenting. Certainly there are useful relays, and some heat sinks. And some useful mains outlets. AND an interesting big power transformer.