I'm totally puzzled by this power supply!

[Maxxarcade]

I'm working on a Powerwel PW-1450 Rev. E power supply from a plasma TV.

What kind of circuit would have two big MOSFETS connected across the ground and B+ lines, so that they effectively short the supply when they are turned on?

The voltage comes off the bridge rectifier, goes through a choke, then a TO-220 diode, then into the bank of filter caps and the primaries of the transformers for the various supplies. The pair of MOSFETS in question are FQA24N50, which are right after the choke and before the series diode. The sources and drains are connected in parallel, and the gates are connected to separate circuits.

I powered it up with the FQA24N50's and series diode removed, and they are getting gate drive pulses before the supply shuts down. If I install them, the supply blows the input fuse as soon as it is powered up. It usually takes out the series diode and both FQA24N50's when it goes, unless I use a light bulb current limiter.

I could not find any shorts farther down the line, so I'm really baffled as to why it behaves like this.

[Toasty]

Re: I'm totally puzzled by this power supply!

Push-pull converter. Both Source to ground(-). Drains to opposite sides of center tapped transformer and CT to B+.

Sounds like PWM circuit is faulty or transformer has shorted. B+ voltage with no FETs?

Original component failures?

Toast

[Toasty]

Re: I'm totally puzzled by this power supply!

SMPS Push-Pull

[Toasty]

Re: I'm totally puzzled by this power supply!

Duhh....

Forgot to say, check the output rectifier(s). 3-legged dual-Schottky devices, sometimes back-to-back on heat sink. Case markings may show a 2 diode symbol with cathodes joined.

One of the input caps may be weak or open and may cause this also.

Time for bed.

Toast

[Wizard]

Re: I'm totally puzzled by this power supply!

That's called totem transistor topography. Very common design.

Um, The IC that "runs" these dies if transistors shorts out.

Best to replace whole board if you are stuck as transistors runs high side of expense.

Cheers, Wizard

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

SMPS Push-Pull

That's not how this circuit is wired.

Drains and Sources for both FETs are connected to the same pieces of foil, like they are being used in parallel. Drains go directly to the output side of the choke that comes off the + on the bridge. Sources go to ground from the - on the bridge, through a .22 ohm resistor.

These FETs feed the filter caps that provide input B+ for the rest of the supply, which works like a normal SMPS.

[Toasty]

Re: I'm totally puzzled by this power supply!

@ Wizard - I thought Totem was Drain of #1 to Source of #2 with load (transformer) connected there? These are paralleled with both Drains & Sources tied and separate gate drive.

Can't see the area where I want to due to glare on bottom side shot. The diode is going, so check those 3 big caps. Open or shorted.

It looks as though you've checked the little cap above to the left. With ESR meter? If not with ESR, sub it with another equal. It looks as though that's the drive to the gates. Both those transistors check out and the 2 diodes?

Toast

[Toasty]

Re: I'm totally puzzled by this power supply!

Also, what are other items at upper right that are removed? Are those bad also?

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

Also, what are other items at upper right that are removed? Are those bad also?

Those are the FETs for another supply circuit. I used them to try and get another board going. I have three of these boards that have the same issue. That circuit IS a push pull type, but not the one I'm trying to fix.

I did look at the components in the gate drive circuit, and replaced the cap. As I mentioned in my first post, the gate drive signals look good on a scope.

I'm suspecting the 3 big caps as well, since they are cheap caps. They do not measure shorted on my ohm meter though.

Still can't understand the circuit though... Why would they put the sources and drains across the outputs of the bridge? I'm going to remove the choke tonight and look underneath it, but I'm pretty sure the traces are connected the way I described.

I can try and get better pics tonight if it would help.

[kc8adu]

Re: I'm totally puzzled by this power supply!

that a pfc section.ring that inductor.

[Toasty]

Re: I'm totally puzzled by this power supply!

Dang it. Beat me to it.

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

I don't have an inductance meter... Is there a way to do it without one? I do have a scope and DMM.

Also, any idea where I can read up on how a circuit like that works? I've never dealt with a PFC circuit before. The 3 caps that it goes into are connected in parallel, and are 220uf 450v each. from there it goes to the rest of the supply circuits.

I've been using a 75W light bulb for a current limiter, and so far it has kept the FETs from blowing, but the diode keeps shorting out after a couple attempts to power it up. I even tried removing the load (caps, transformers, etc.) from the other end. I'm assuming it needs the caps to work though.

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

I see what it is now. A "boost converter" type PFC circuit. Will it operate enough for testing with just the caps on the output side? That explains the 450V caps. And I want to say that the voltage did shoot up around 400V before it shut down a couple times. When I tried again, the diode shorted.

I wonder if there is a way to test the rest of the supply, by injecting DC voltage where the PFC circuit normally would. I want to make sure the other transformers and such are good before I mess with the PFC circuit too much.

Now I can understand why it pulls so much current even without a load... It has to charge the inductor. And if the inductor is shorted, the FETs are just going to ground out the rectifier. I do have a working supply that I can compare to, but what I would really like is a schematic...

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

Just another update- I figured out a really dumb mistake on my part for the diode...

The original is a F10L60U, and the replacement I was using was an STPS10L60D. The F10L60U has a PRV of 600V, and the STPS only 60V

So I need to at least find a TO-220 diode rated for 600V before I can do too much more. I'll order some FFPF10UP60STU from mouser to try out.

[Toasty]

Re: I'm totally puzzled by this power supply!

Not to mention the 1A vs 10A rating...

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

Not to mention the 1A vs 10A rating...

They are both 10A forward current according to the datasheets.

[Toasty]

Re: I'm totally puzzled by this power supply!

Yup, got the numbers crossed between sheets. Forward on one and peak reverse on the other. AAARGH!!

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

Well, I tried using a 1N5408 (1000V PRV, 3A) with just a 220uf cap for a load. It still blows the diode on power up.

Does it absolutely need to be an Ultrafast recovery to work or something? I figured the 1N5408 would be enough for testing purposes and no load.

The FETs are not blowing, so at least the expensive parts are still good.

Using my Simpson 260 meter set to AC volts and connected to the input of the diode, I can see it go up to about 200v before the supply shuts down. I figure if the diode would hold and charge those caps, it would be up to the 350-400V DC that it should be.

[Toasty]

Re: I'm totally puzzled by this power supply!

MUST be fast diode. Datasheet even says "superfast". 25nS recovery time. The 1N's don't even list a recovery time.

Diode is 10A to handle the surge at startup.

>>with just a 220uf cap for a load<<
Do you mean you took all the other caps out of circuit and used a single 220uf @ 450v by itself?

Toast

[Maxxarcade]

Re: I'm totally puzzled by this power supply!

MUST be fast diode. Datasheet even says "superfast". 25nS recovery time. The 1N's don't even list a recovery time.

Diode is 10A to handle the surge at startup.

>>with just a 220uf cap for a load<<
Do you mean you took all the other caps out of circuit and used a single 220uf @ 450v by itself?

Toast

Yes, I left the cathode of the 1N5408 off the board, and connected it to the voltmeter and a 220uf 450v cap. Normally there are 3 of the 220uf in parallel, but I figured one would be enough for testing if there is no load.

I did some reading and I can understand why it needs the ultrafast, but I'm still starting to doubt that a proper diode will make this circuit work...