SMPS troubleshooting

I have an automotive battery charger like this one:

formatting link

I have not been able to find schematics, and the manufacturer ignores me.

There are two boards, one looks like the power supply section, while the other one contains buttons and LED display. The SMPS board looks like an ordinary flyback arrangement, with a transformer and opto isolator for feedback.

The symptom is that the 5 A input fuse blows as soon as power is connected. The fuse blows with a magnificent blue flash and produces an audible pop, so the overcurrent is considerable.

The SMPS is based on an ST 3845B, driving a 9N90C MOSFET. Across the MOSFET's D/S, there is a 470 pF / 1 kV capacitor, located very close to the MOSFET. The capacitor is split open and has spewed its guts onto the MOSFET. The capacitor measures open circuit with an ohmmeter. I do not have a megger available, so I cannot check the capacitor at high voltage.

I have unsoldered the MOSFET and the capacitor. Now the fuse does not blow. Between the MOSFET's G and S pads on the PCB, I can see a nice

27 kHz square wave with declining amplitude for 7 ms, and then nothing for 17 ms, before another burst of 27 kHz, so it looks like the controller is trying to start from its bootstrap supply.

The MOSFET tests OK on one of those cheap ATMEGA-based component testers.

What is the most likely fault scenario?

What should I check next?

Could the blown capacitor be the only problem, or is it only a result of the actual fault?

Reply to
HW
Loading thread data ...

The MOSFet will set you back a princely $4, including shipping. The cap will set you back, perhaps, a whole dollar ($1). The controller - perhaps one more whole dollar ($1).

Replace them all. But only *after* you have checked any and all diodes for open or short conditions.

Strange that these devices do not like surges of any nature, and yet they style themselves as jump-starters. Go figure.

Peter Wieck Melrose Park, PA

Reply to
peterwieck33

I would have bet almost anything the mosfet is shorted. Put it on an ohmmeter and check for continuity across all the pins.

Those tiny blue disc 1kv caps are known to both split and blow the device they're across. I stock the 331, 471, 681, 102, and 222 in 2kV

Reply to
John-Del

Reply to
Terry Schwartz

wire yourself up a dual outlet box with a line cord an on off switch and a duplex outlet.

Cut the tabs on the outlet and wire the two sections in SERIES.

Plug your UUT into one outlet and a 60 Watt incandescent lamp into the other outlet. Now you can troubleshoot. If there is a short, the light will limit the fault current.

Also beware using a scope that the ground in your circuit may be live. I would buy or make an isolation transformer as well for your safety.

Replace the FET and cap and turn it on with the 60Watt bulb in series.

m
Reply to
makolber

Sure. But since I had removed two components and the short went away, I felt it was worth checking them both for a short. It turns out none of them are shorted, which is where it gets interesting.

Some capacitors do fail short. In fact, some are designed specifically to fail short.

Reply to
HW

I actually have wanted to make one of those for a while. Maybe this is the right time. I like the dual outlet idea. That way, I can plug in different power bulbs depending on the type of device.

I recently got myself a differential probe.

That is also on my to-do list.

Transistor and capacitor set are on order.

Reply to
HW

The transistor responds just as it should to the ohmmeter. I have not tested it at high voltage or current.

In circuit, I believe the MOSFET (and the capacitor) sees about 640 V as it turns off. 240 V rectified is about 320 V, and I believe the transistor seed twice that at turnoff?

Yep, that's the capacitor type here. I have ordered a kit with 20 different values, 10 of each.

Reply to
HW

One of the goals with this repair attempt is to learn something about SMPS. So, before I start throwing parts at it, I want to scratch my head a little and try to go the methodical route.

OK, I will check the diodes.

Reply to
HW

There is always a balance when contemplating a repair between, for lack of better terms:

Accuracy - how suitable is the repair to the need? Precision - how few parts are necessary to complete the repair?

To which I would add "cascade artifacts": some part or component up-line fa ils or glitches such that one-or-more parts down-line fail visibly and spec tacularly. This will fool the typical tech into addressing the obvious fail ures, and not looking for the first-cause.

Hence: Check any and all diodes. If any zeners, special attention there! An d while in there anyway, caps and so forth, switches for dirt.... you get t he picture. General clean-up and policing.

Now, consider the analogy of a spun crankshaft bearing on a V6 engine. Only one bearing actually spun - but does a good tech replace only that bearing ? Not hardly. You had a spectacular failure (spun bearing) of a small part

- but are entirely unaware of the relative condition of the parts immediate ly adjacent to it. And no visible means to become aware.

Which is why, for an additional $4, it would pay to replace all the "bearin gs" as you have already torn down the engine in any case.

I am re-capping a vintage AR received I lent to a friend 20 years ago, and now it needs service. It really needs only two (2) small caps on the tuner board. But, as I am in there already, it will get new matched output transi stors, re-biased, and about another dozen or so electrolytic and small-valu e caps replaced. The cost is small, the time is not-much, and it will have another 20 years of useful life, at least.

Peter Wieck Melrose Park, PA

Reply to
peterwieck33

mmeter and check for continuity across all the pins.

My first rule of repair is that you'll never see everything in your career. In 50 years, I have seen just one high voltage high current semiconductor check normally on an ohmmeter but "short" under circuit conditions, then c heck normally again on the meter. They generally do you a favor and short themselves silly. The one I did find was an early 1980s RCA TV where the h orizontal/hv output transistor would go into full conduction with *any* dc on the base and blow the fuse. It checked fine out of circuit but a new on e stopped the immediate and spectacular blowing of the fuse. That's the on ly one I ever saw.

So, while you wait for your parts to come in, follow Peter's advice and che ck every diode you find in the circuit as that's the next part failure perc entage-wise that follows your kind of failure. Often, a zener will check O K front to back but will conduct significantly off it's printed voltage (ge nerally lower). Also carefully check all low value resistors.

Reply to
John-Del

visible means to become aware.

At work we had a motor speed controler to go bad. This was on about a

200 HP motor. Called in the factory repair man and he located 2 bad diodes in the 3 phase circuit. I asked him to replace the 3 rd one. He said it checked good and they were about $ 200 each. While it may or may not have been bad, he at my request replaced it. I told him it was costing us over $ 1000 an hour to hae the equipemtn down,so in the time it took him to get there and many ohter things, it was maybe a 50 to 100 thousand of lost time if not more. I felt that $ 200 was good insurance. While the diode was probably good, I felt that if 2 were bad, it may have weakened the 3 rd one.

Same as when I had a timing belt changed on schedule on my Toyota. The mechanic recommended replacing the water pump as it was in the same area of the tare down. Said no more labor,just the cost of the pump and he doubted it would last another 70,000.

Reply to
Ralph Mowery

I had a late 70s RCA console that would not start intermittantly out in the boonies but worked perfectly in town. The voltage at this farm varied between 110 to 115 volts. Replacing the horizontal output transistor repaired the set.

Reply to
Chuck

The capacitors arrived. Out of curiosity, I reinstalled the old transistor and tested the charger (without the capacitor). The charger started up normally, no fuse blowing. Then I added the new capacitor and ran the charger at full load for about 15 minutes with no problems.

In the meantime, I also bought a cheap megger and tested the blown capacitor at up to 1 kV. It does show some conductance, but it is in the hundreds of kohm range, so it should not be able to blow the 5 A fuse.

So, the charger works again. Although the blown capacitor seems to have been the root cause of the problem, the analysis does not quite add up. Maybe there is an intermittent problem somewhere?

Thank you everyone for your comments.

Reply to
HW

it may be that as the cap blew the fuse it also blew its own conductive path out

NT

Reply to
tabbypurr

Without a schematic you can't say for certain, but this is a possibility: with the bad cap in there it may be that the Mosfet was turned on 100% of the time, thus a dead short through the inductor to ground, instead of sending pulses to the inductor to charge it. The fuse saved the Mosfet and/or the inductor.

Ed

Reply to
ehsjr

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.