Digitech 5207 UPS Repairs

[Lumpy20]

Hi,
This is my first post so would appreciate some help.
I have a Jaycar (Australia) Digitech MP5207 1500 UPS that has failed, the Inverter transistors literally blown up, twice. Initially the unit would not turn on after a short power outage. The batteries read OK unloaded but one battery would not hold load. We replaced the batteries with two new ones and allowed them to charge overnight.  The unit was turned on and when the main power was turned off to test the Inverter the power transistors failed, as per the photos. We replaced the FP50N06 inverter transistors and the LM317, LM7812 regulators and tried again. The unit would run on mains and happily supply a light load (40W bulb). However, when the power was turned off for the second time the inverter transistors failed again, as per the photos attached.
Does any one know how to fix this, a circuit diagram or suggestions would be appreciated. Regards, KH

[ion]

Perhaps look for any shorted components?
Too much current through the transistors when they're turned on is a good way to blow them up.

Not knowing your background, I'd make one more suggestion in the interest of safety - troubleshoot with the unit unplugged and batteries disconnected as much as possible.  A UPS can pack quite a punch.

[TheMG]

Try it with all the inverter MOSFETs removed, and use an oscilloscope to check for the correct gate drive waveforms.

This looks like a typical half-bridge "modified sinewave" (fancy marketing word for a square wave with dead time to achieve the correct RMS voltage) inverter.  So the expected waveform would be alternating pulses at 50/60Hz between the two sets of transistors, with some dead time in between.

This may be difficult to observe as the UPS will probably fault off after just a few cycles when it detects there is no actual output voltage (a DSO in single-shot capture mode helps in this regard).

What you're looking to check is mainly that first of all, only one set of the transistors should every be "on" at a time, and also that none of the transistors are being turned on "hard" (ie: not pulsing, but flat-line constantly on). Either of those situations would result in excessive current and release of magic smoke. Also check for correct frequency.

[Lumpy20]

Thanks TheMG,
I will give that a go. It is the same set that blows each time. Like you said, could be a timing or hard-on issue. I will post back in a few days. Regards, KH

[Circlotron]

What was the UPS powering when it blew up the first time? Those standby UPSs are okay for powering a switchmode PSU or a resistive load, but no good at all for a transformer load. What happens is when the mains stops and the inverter comes on, the inverter generated waveform will almost never continue where the AC mains left off. Totally random. You may for example get a positive half cycle from the mains, mains goes off, then another positive half cycle from the inverter. This will make the transformer load saturate and draw a big heap of current and probably blow something up. I have done it myself.

[Lumpy20]

Ho Circlotron, The MG

Thanks for your suggestions. I hope this is not too confusing (or dumb) but I think this is the process we have been going through for the past weeks, at least four.

The first time the UPS failed (stopped working) was after power was restored following a 30min power shutdown.

a. Initially with power restored the inverter would try to fire up but would drop out with no load.
b. I disconnected the original batteries and powered up the UPS and immediate failure of the mosfets followed. I later discovered that one of the original batteries was faulty, 12.75v no-load but 2.3v with 100 ohm (10W) resistor.
c. We replaced the mosfet transistors several times and with new batteries installed the unit would power up and supply load to a 60W bulb and a small fan. We also discovered that some of the spade connectors were loose and replaced them
e. Disconnecting mains power the UPS would drop out.
f. This time we replaced the LM317, LM7812 regulators as well as the transistors and the UPS would power up and supply a 100W light bulb and small fan?
g. When power was disconnected we had immediate mosfet failure.

I know this seems very strange but I think if I had replaced the old batteries at first the unit might still be OK. Replacing the mosfets and regulators and powering up and off just repeats the same failure of the mosfets so I think the suggestions of the oscillator circuit - control of the mosfets might be the problem but without a circuit diagram I don't think that it can be fixed.

I came across a post a few years ago about the DigiTech MP5207 UPS which said that a few have caught fire, which we experienced a few times, and that there was a design fault with this UPS not protecting the mosfets from 'full on' state?

I will try and borrow an oscilloscope to have a look at some waveforms with no mosfet transistors installed?

What seems strange is that each time we do the repair, transistors and regulators, the display comes up as normal and we can power down the UPS with the front panel button without any problems.  It only fails, catastrophically, after replacing the mosfets and regulators when mains power is removed.

I am not sure that we have tried a power-down with no load connected?

Regards, KH

[Lumpy20]

This link at Whirlpool might be helpful? https://forums.whirlpool.net.au/thread/9k8p5mn3

[TheMG]

That UPS looks like a line-interactive model, in other words the inverter is not active when on AC mains power, simply powers the load directly through a relay. That would explain why it only blows up when you disconnect it or the power goes out.