SMPS blowing fuses, not semiconductors

[k8tek]

I have a SMPS from a Winix WAC9500 that keeps blowing the fuse. It had some obvious bulging 200 uF 200V main filter caps that were replaced with 250v ones, but that obviously didn't solve the issue.

The bridge rectifier was replaced (RS405M), the switching MOSFET was replaced (STP7NK80ZFP) and the fuse (a 2 amp slow blow fuse) is still blowing. It isn't taking anything out with it when it is blowing.

The MOSFET is controlled by a KA3842 PWM controller. Taking that out of circuit stops the fuse from blowing, so the problem isn't in the primary filtering, or the rectifier.

What the hell would cause the fuse to blow the way it did (it scorched the board) without taking another component as well. The MOSFET I took out was fine, I was just eliminating possibilities.

Thanks.

BD1 and Q1 were removed for testing, but they were fine (I still replaced anyways) thats why they are missing from the board. C32 was never populated.

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

That's rare, semiconductors are supposed to protect the fuse!

I guess you've swapped almost everything, all that's left are the 3842, PC1, and ...

pet peeve... your camera phone can be rotated 90 degrees and then you don't waste space on the picture for stuff that's not the power supply. The whole power supply may actually fit in one whole frame, and then we could see the whole thing without any gaps and inspect if there is a malfunctioning crowbar circuit or other short on the output (secondary diodes, etc.)... and perhaps even see a continuous picture of the solder side so we can see how it's connected without stitching pieces together.

Sorry to have to take it out on you, again it's me not you for not keeping up with the times. I just hate all these super-godly-camera phones and then people think it's the best thing in the world that solves all their problems, and forget that they can use the device in landscape mode and make things easier for other people...

[R_J]

Re: SMPS blowing fuses, not semiconductors

The ic could be defective and causing the fet to stay turned on instead of being pulsed. was the original fet shorted?

[budm]

Re: SMPS blowing fuses, not semiconductors

the ic could be defective and causing the fet to stay turned on instead of being pulsed. Was the original fet shorted?

+10

[kaboom]

Re: SMPS blowing fuses, not semiconductors

Always replace the driver and gate resistor when the MOSFET shorts.

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

The trouble is that he mentioned that the MOSFET didn't short... so I have a hard time thinking the 3842 was damaged too, hence have to consider secondary issues:

The MOSFET I took out was fine, I was just eliminating possibilities.

[kaboom]

Re: SMPS blowing fuses, not semiconductors

The trouble is that he mentioned that the MOSFET didn't short...

The original one? Or was there a second one, that tested OK, after parts replacement but when still blowing fuses?

The bridge rectifier was replaced (RS405M), the switching MOSFET was replaced (STP7NK80ZFP) and the fuse (a 2 amp slow blow fuse) is still blowing.

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

Well it was clear the OP did not need to replace the bridge rectifier, and I do wonder if the mosfet was even necessary to be swapped - the title mentions the semiconductors did not blow. Need for the OP to get back and be clear what was actually damaged and what wasn't -- because some of those 200V input filtering capacitors some manufacturers put domed covers on top which look like they're bad when you have to remove the cover to inspect the actual condition...

[kaboom]

Re: SMPS blowing fuses, not semiconductors

I see what may be a green varistor, right next to the fuse.

Not to be confused with the thermistor, black (TH1) two components to the right of the fuse.


Now a bad varistor would still blow the fuse w/o the rectifier installed.

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

Except it was demonstrated that with the transistor removed, the fuse won't blow with everything else installed:

The MOSFET is controlled by a KA3842 PWM controller. Taking that out of circuit stops the fuse from blowing, so the problem isn't in the primary filtering, or the rectifier.

Looking at the secondary is still necessary. Again I cannot rule out a burned 3842 but it's unlikely if the MOSFET didn't blow.

[k8tek]

Re: SMPS blowing fuses, not semiconductors

The pwm driver is removed and the fuse is fine.

[k8tek]

Re: SMPS blowing fuses, not semiconductors

The semiconductors I took out tested fine, the MOSFET wasn't shorted and the rectifier was fine as well.

The only semiconductors left is the triac driver the upper left hand corner, diodes and two voltage regulators (LM7805) (LM7815).

Tim

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

Just want to confirm, this looks like a 3-output PSU as there are three diodes. And that it will blow fuses with the PSU disconnected with the rest of the unit?

Like most multiple output PSUs it's group regulated, and the main regulating channel is the "big fat" diode D4, ensure this is still working as a diode. Note the feedback circuitry appears to be IC2/PC1/R15/R14/R18 though you may need to test these individually out of circuit.

Having IC4 or IC3 short is unlikely, but do diligence and test. BTW, are you using a dim bulb tester, for these experiments, blowing fuses can get expensive!

There is no OVP crowbar that I can see, so that throws out that idea. I don't think I see the triac but if it's cut off from the picture, it appears to be in the low current section and probably not the problem.

For completeness, things that need to be checked:
- IC1 ... though this is probably checked by swapping.
- D4
- IC2
- PC1
- Transformer If this broke, probably game over.
- IC3 (78xx I think) and its D5 though this probably would more likely just get hot but not burn fuses
- IC4 (the other 78xx) and its D6 though this probably would more likely just get hot but not burn fuses

I haven't discerned the function of the other photocoupler yet, cross that bridge later...

[redwire]

Re: SMPS blowing fuses, not semiconductors

I looked at the gate drive from UC3842 PWM output pin 6 and this circuit does not seem to match for IC1. C10 in series with R5 33R should not drive the gate?

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

Interesting, didn't notice... but I think it should still work but quite unusual, it's a high pass filter coupled with the 100K resistor R7.

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

Oh, now I see Q8 (TRIAC) and PC3 (ZCT isolator), which looks to be a remote control circuit / "relay" for CN5. I can't imagine how they would cause fuse overloads, the TRIAC shouldn't be able to pass more than an amp before frying. Again it would be interesting to know what was connected when testing this PSU, if the device CN5 is connected to is shorted and thus shorts the whole kaboodle with the fuse when the device is turned on. That device plus the PSU draw may be enough to also blow fuses.

I did notice a bit of heat discoloration by the thermistor? Or is it? Just as a sign of how much power this PSU had been passing, whether it was stressed from use or not...

Also when it blows the fuse (guess it's a slow blow?) is it instant blow or blow after several minutes?

[redwire]

Re: SMPS blowing fuses, not semiconductors

This air filter has a fan which I think is speed controlled by the little triac and opto.
The HV DC bus goes off on CN2 pin 1 and 3 to I guess the ion generator
OP are you running this with CN2 connected? If so, then that second "plasma" PSU might be shorted and actually the problem.

I'd use a series light bulb and save some $ on fuses.

Actually a UC3842 is OK but the circuit is different because the mosfet gate is AC coupled, hopefully to a nice square wave. Blowing fuses with the IC in, could check I_sense R11 for low value, but usually these fuse open.
Unless the unit powers up and turns on the plasma PSU and then pops the fuse?

[k8tek]

Re: SMPS blowing fuses, not semiconductors

CN2 goes to a multi winding fan motor. CN5 goes to the plasma generator. It is powered by the MOC3041 and a triac (Q8). PC2, the unknown photocoupler controls pin 2 of the MOC3041 i'd imagine since it's on the primary side.

[eccerr0r]

Re: SMPS blowing fuses, not semiconductors

Does the fuse blow when you have things disconnected?
Again try to set up a series light bulb "dim bulb tester" instead of burning fuses, it gets expensive to experiment with fuses. Right now all the data seen here does not seem to point at a fault of the PSU itself (though it's still inconclusive without the tests suggested).