PSU repair- Not main caps or diodes, now what ?

[lordvader88]

I'm very new to working on these

I'm working on a TX650W that cycles on/off repeatedly, as per the inner timing of the PS229 supervisor IC.

Everything seems more or less fine when I've poked around with a DMM. With it hooked to a PC, it appears to hit all the correct voltages. I've used single capture triggering to look at most everything on the 14-pin DIP PS229 and the secondary side of the 3 opto's. Nothing is horribly wrong, there's been some spikes I wonder about but IDK.

So I went ahead and desoldered all the main output caps, and they all read fine with an LCR meter, and DMM. They are all 2200u and 3300u.

There's a 100u cap that might be near the ESR limit, depending on the chart I look at. It was 100u 25V with an ESR of 0.134

The big bulk 390u 400V is great.

So I desoldered all the main sec. rectifier diodes, and they are perfect, nicely matched

So I checked some sec. SMD caps and I think it's ok. (2 looked funny and I broke them getting them off, but they 'were' ok, the LCR meter wasn't reading the caps in-circuit)

I've mapped most of the PCB, but I've not tried to calculate hardly anything, or measured every little voltage yet. I want to tho.

The LM393 fan controller had power, and the PTC (or NTC) is ok. There's a few TL431 circuits I've mapped, and they seemed ok on the surface to me anyways.


I'm assuming the pri side is fine, IDK.

I'm having fun but I'm really walking in the dark at this point. I should start injecting voltages and calculate a bit to look at some networks, while the diodes are out.

[shakalnokturn]

+5Vstb is good and stable?
How about supply (Pin 13) to PS229 itself?

If all else failed I'd toss the ICE2A056 out, having seen enough of them with erratic behaviour...

[jaunty]

what voltage are you reading across the reservoir caps?

[lordvader88]

The 5Vstby seems fine, it's 2200uF and 1uF electro's have good ESR, the power good out to the opto circuit works, the TL431 works

The bulk cap hit's 320V


The 3.3V rail is regulated with a TL431 and shunt pass transistor, it works fine with DC I hooked up.



When I measured some spikes on the outputs, using a good scope, I drew it, and did some calculations and they should be within tolerance. I have a few datasheets these type's of 4-ch supervisor chips. They are all very very similar. I haven't found or figured out the exact timing on what the chip actually does if a UV/OV/OC happens.

In this case it starts, then runs for standard ~380ms, during which time it decides something's wrong and then tells the primary side to shut down, while it resets, and tries again.

If a UV/OV spike it doing this, IDK if it would stop immediately or wait a moment.

I'll poke around a bit more on the output side.

Soon I'll have to put it back together and check on the pri.side too