28v SMPS repair - not starting up.

[src1989]

  Kinda describes how i'm feeling right now; i've got a 28v 10a SMPS from an intelligent moving head - a Martin Rush MH specifically - which is dead. i have confirmed all the processors/motors fed from this board still work normally with a bench power supply.

there are no visual clues to what's happened (no sooty skidmarks, no swollen capacitors, no solder blobs bridging out stuff, no burned components - other than a power resistor next to a superfast diode (which tests out like a normal diode) next to the inductor(190uH ?) on the primary side which tests about 2 megohms but has its colorbands faded off), the board does not spark, blow or trip anything on power-up, there is no fuse (there is a MOV that shows continuity). there is no output voltage - have tested with a load and open circuit. i have tested the diodes (including surface mount), resistors, full wave bridge, and they all seem normal (only tested in-circuit at this stage). the caps appear to hold a charge. i've reflowed the solder on all the major component leads (full wave bridge, caps, transformers/inductors, triacs, schottkys, resistors) - just not the surface mount stuff. the windings on the transformer/inductor pass continuity test.

i powered the board up from 240v mains via a plug-in gfi/rcd, and my meter alternates between "-00.0" and "00.1" - i have seen the same thing on a dead 24v halogen light transformer where the only thing on the secondary side was the secondary coil itself - so is it likely that the switchmode chip has failed? or is there some really simple solution i'm overlooking?


i don't have access to a thermal imaging camera, isolation transformer or a scope. i have some experience of electronics but nothing this advanced. thanks in advance for any advice.

[james_s]

That doesn't really tell you much. I would start by checking for any sorted semiconductors on the secondary side, then replace electrolytic capacitors. The big bulk filter cap(s) are usually ok but the others are fair game, especially small 10-22uF capacitors near the middle. Most of the time they don't swell or show any outward signs of failure, only really bad ones get to that point.

[floobydust]

It does have a fuse. The dark red rectangular part near the mains input connector.
If it's blown then the primary switch (mosfet) should be checked. Best to do out of circuit to see if it shorted. The part would be heatsinked.

I see two transformers so not entirely sure how this SMPS works, a better picture of both sides would help.

Check your confidence level as making measurements on the primary-side involves high voltage, which might be necessary. Sometimes you can blindly change a few parts, other times in depth troubleshooting is necessary.

Best to put the PCB on something insulated when powering it up, concrete is ground.

[eblc1388]

There is only one transformer. The other one is what is known as the common mode choke. Both the transformer and the choke seldom fail.

@src1989,
Look carefully at the board and locate an electrolytic capacitor with high voltage rating. The capacitance value should be from a few microFarad(uF) to about 50uF, 350V or more. Once you've done that, then feed a dc voltage from a few volts, upto 12V using a battery or from a DC power supply into the board ac connections. Polarity doesn't matter. Observe the dc supply current and measure the voltage across this high voltage capacitor. Reverse the dc polarity into the ac connections and measure again.

If you don't get any voltage, then there is an open circuit somewhere between ac input and capacitor. Looks for the cause and rectify.

If you get the same dc voltage regardless of the dc input polarity, then you can be sure that the rectifier and capacitor are in good condition.

If the voltage is significantly lower than the applied dc voltage, along with high dc current, then you have short circuit somewhere. Most likely would be bad semiconductor switching element, bad rectifier or bad electrolytic capacitor in that order.

[kripton2035]

close-up pictures please ?

[mikeselectricstuff]

The most common reason for smpsus to not start is a leaky cap on the primary startup supply. Look for a small low voltage electrolytic cap on the primary side and replace it.

[CJay]

The most common reason for smpsus to not start is a leaky cap on the primary startup supply. Look for a small low voltage electrolytic cap on the primary side and replace it.

And after that look at high value resistors that charge the low value cap.

[src1989]

i have added all electrolytics on board to rapid shipping cart, thought i might as well add the fairchild switching chip, caps and diodes while i'm at it and do a bulk/blitz replace.

the big power resistors seem fine, 2 low value one high value with bands burned off and charring where it made contact with the pcb, but as i say, it still tests 2 megohms - although what the original value was i have no idea. i'll add some resistors to my order, replace and report back.

[Seekonk]

The most common reason for smpsus to not start is a leaky cap on the primary startup supply. Look for a small low voltage electrolytic cap on the primary side and replace it.

And after that look at high value resistors that charge the low value cap.

A third thing to add, look at the voltage on the cap that resistor charges. Should be over 10V though it may only last for a moment.  This starts the chip and a diode and a winding running power. Pulsing would probably indicate the FET was bad as you don't see anything on the output.  Check for a small ohm resistor in the source of the FET.  The resistor may have failed open due to repeated pulses.  I got a really nice Xantrex supply for almost nothing because of that. For sure they weren't using cheaper parts than your supply.

[james_s]

i have added all electrolytics on board to rapid shipping cart, thought i might as well add the fairchild switching chip, caps and diodes while i'm at it and do a bulk/blitz replace.

We refer to that here as the "shotgun" approach to repair. I would encourage you to do some troubleshooting rather than just blindly replacing parts. Diodes are very easy to test, and the IC rarely fails unless something catastrophic happens. Each component replacement is a chance of getting something in backwards, wrong part or wrong place.

[CJay]

one high value with bands burned off and charring where it made contact with the pcb, but as i say, it still tests 2 megohms - although what the original value was i have no idea.

Which resistor in the pic was that one and can you post a pic of the underside of the board?

[floobydust]

There is only one transformer. The other one is what is known as the common mode choke.

Ahh it's two common-mode chokes and active PFC inductor, then the one transformer.
What is U2? that seems to be for the PFC section.

[src1989]

thought i added rear pics, sorry. went away to order my new components and my hard disk crashed....
will try measuring voltage of the wee bootstrap cap, ta.

the burned-but-still-measurable 2 megohm resistor sat right next to the big diode to the left of the letune inductor(?).

@mikeselectricstuff, thanks. love your youtube channel btw. helped me fix my panasonic microwave

[radioFlash]

What are U1 and U2? You may be able to find a data sheet wth a pin out or typical application schematic. I had a cheap 12v led supply that stopped working when a resistor to the controller chip’s vcc failed open.


Sent from my iPad using Tapatalk

[CJay]

From where it is, I'm thinking it's not meant to be 2MOhm, I suspect what you're measuring is an almost fully burned out resistor.

Can you see the markings on U1 and U2, I think one of them is the PFC controller

[src1989]

i've started thinking that myself; having recently taken delivery of a true RMS meter, it measures open...

u2 looks like a lnk302gn; assuming i've read that correct, the resistor in question may be a 8r2 Ω one.

[Vicus]

I recently repair a very similar psu, same manufacture but slight different design. I was not easy at. In the picture the top ic is the pfc controller, mybe L6562. It's powered from the middle ic, it's an lnk302. This buck converter provide power to the pfc controller than the main ic power up. In my case there was a low voltage (4-5V) on the output of the buck converter. Try to change lnk302, nothing. Test almost every part related (only few fortunately) and they measure fine. So I did remove the L6562 and of course the mosfet of the pfc blow out (no pulldown resistor and the mosfet turn on). But doing that I found that one fast recovery diode of the buck converter did get open. So the entire psu stop working by a small 1A smd diode.