Failed 24V Switchmode PSU [Repair Complete]

[jaxbird]

It's served me well for a few years, so I thought worth having a go at getting it up and running again, but I have little experience with common faults, so all suggestions are very welcome

It failed totally in a non spectacular way, just stopped working.

Fuse is fine and only about 0.5V on the output without load.

I found a very bad capacitor (pretty much failed open)



Didn't have a size that would fit, so I bodged this one in for now:



But still very dead. I measured about 145V across the new capacitor.



The output capacitors also look a bit dodgy, but still measure ok, so I'll leave them until I know if I can get it running again.

More detail pictures coming...

[jaxbird]

More pictures:













Much appreciate any suggestions on what to measure.

[SeanB]

Look at high value resistors on the primary side ( anything over 47k and over 0.5W) and look around the controller for small low voltage caps that are either shorted or open. Your secondary side likely has bad caps as well.

Look up the TL494 datasheet and apply power to it and see if it is oscillating and driving the transistors with the board not connected to the mains. That way you see if the chip is faulty and if the transistors are working. properly. Most likely it is a faulty resistor that is not allowing it to start oscillating on the primary side, or a shorted capacitor. If no oscillation and all the resistors and low value emitter and base resistors along with the diodes you have low gain output transistors, or a shorted driver transistor on the small transformer.

[tautech]

I fixed a similar PSU that the chip V+ supply 22uF E-cap had drifted to just a few uF and high ESR.
In short the chip did not have a to spec power supply and that was all!

[jaxbird]

Thanks for the suggestions 

I've been poking around a bit more and one of the two switching transistors appear to be faulty. I assume they are transistors as the pins are marked B - C - E.

So I desoldered them and using DMM diode measurement, it appears C and E are shorted on one of them (0.08V in both directions). The other one measures as I would expect from an NPN.

I believe they are NPN transistors, but I really have no idea what the specs are, no luck on Google searching for TK100A relating to a transistor and no brand markings I'm familiar with.

[SeanB]

Grab the nearest PC power supply and use the transistors from it, they will work. Just look that they are transistors and not mosfets.

Looked at the closest donor unit and it uses mosfets, khd 5D0N50F, so I cannot give an equivalent at the moment, but any SMPS primary side driver in a totem pole arrangement will work. Just reuse the ferrite beads on the B and E leads.

[jaxbird]

Grab the nearest PC power supply and use the transistors from it, they will work. Just look that they are transistors and not mosfets.

Looked at the closest donor unit and it uses mosfets, khd 5D0N50F, so I cannot give an equivalent at the moment, but any SMPS primary side driver in a totem pole arrangement will work. Just reuse the ferrite beads on the B and E leads.

Excellent idea  thanks.

Found an old PC supply, with a pair of ST13007

http://pdf.datasheetcatalog.com/datasheet/SGSThomsonMicroelectronics/mXuqrys.pdf

They are TO-220, but that will probably be fine.

I'll try get them soldered in and see what happens 

[nctnico]

What often helps to diagnose SMPS without the danger of getting a shock is supplying the controller chip and the primary part from a DC power supply. That way you can see if the controller chip still oscillates and the transistors get driven. Also: you can put a probe and ground clip wherever you want.

[jaxbird]

What often helps to diagnose SMPS without the danger of getting a shock is supplying the controller chip and the primary part from a DC power supply. That way you can see if the controller chip still oscillates and the transistors get driven. Also: you can put a probe and ground clip wherever you want.

Thanks, yeah, SeanB did share that neat little trick earlier. So I am injecting 12V from my bench supply to the PWM controller as well as the rectifier:




The PWM controller looks healthy on both outputs:

[jaxbird]

Had a bit more time to work on this, so as everything looked fine with DC bench supply, I hooked it up to mains, and success: 




So it was 1 capacitor and 1 transistor that had failed. Replaced both transistors, and I will also replace the output capacitors. Plus probably both primary side capacitors when I find some that will fit properly.

Big thanks for all the helpful advise from everyone, especially SeanB. 

[jaxbird]

And just to wrap it up, new output capacitors vs old ones:



Pretty sure they are not really Nichicon, but then again I'm not sure the old ones are Rubycon, although they do look slightly more authentic. Anyway, the new ones got slightly lower ESR.


Leftovers for the bin:




And a PSU ready to go back into duty