SMPS open circuit voltage OK, but can't supply current

[johnmeyer]

I have a small "wall wart" 15V 600mA switchmode power supply. It measures exactly 15 volts output with no load so I at first thought the problem was with the equipment (a TV modulator) that it powers. However, as soon as I attach even the smallest load (drawing only a few milliamps), the voltage dives and no current is produced.

One interesting note: if I cycle the input power on/off many times, it will sometimes get into a mode where it works, and it stays working (for months) until the mains power goes out.

The unit initially failed after a city-wide power outage. However, that may be a red herring because the unit may have been degrading/failing, and the problem was only exposed when it had to restart.

The circuit is extremely simple because it uses a THX203H controller that does most of the work. There is no MOSFET power transistor. I've attached the reference design for this chip at the bottom of this post. My power supply is very similar to that design.

Given that it works about every 40-50 times it is plugged in, I suspect the capacitors. There are three electrolytics on the input (10-10-47) and two electrolytics on the output (680-100). They do not appear to be swollen, bulging or leaking. I do get a capacitance reading in-circuit, but as is often the case, it doesn't match what is on the cap label. The diodes all check out OK. I only casually checked the optoisolator, but it appeared to be OK. The input fuse is obviously OK or I would never get the 15 volt open circuit reading at the output.

I guess I should just pull the three input capacitors and see how they test out of circuit, but I thought I'd first see if anyone had a few suggestions.

[Kim Christensen]

Capacitors can still be bad and have high ESR even if they look good to the eye.
Because of the startup problem, C5 would be my first suspect. But C7 & C8 are usually the first ones I check.
The input capacitor C1 is a lesser suspect.

[johnmeyer]

Because of the startup problem, C5 would be my first suspect. But C7 & C8 are usually the first ones I check.
The input capacitor C1 is a lesser suspect.

You are very, very good (see below).

I went ahead, before I read your post, and followed my hunch that it was the capacitors on the high voltage side. In the actual power supply, C1 was actually two 400V caps in parallel. I desoldered one leg on each cap and measured them. Bboth the capacitance and ESR tested OK. I then pulled C5 which, in the real product is 47uF at 25V rather than 16V. It measured only 22 uF, on two different capacitance meters, and its ESR was off the charts. I rummaged around through the parts I'd pulled from old power supplies and lucked out by finding two exact replacements. I took the better of the two, installed it, resoldered the legs I'd freed up on the other two caps, turned it on, and it works perfectly.

So, as these things go, a simple fix (although it took awhile to carefully cut open the welded plastic case).

Thanks for the help. You clearly know your stuff!

[abdulbadii]

Same https://www.eevblog.com/forum/repair/why-the-smps-gives-normal-24-v-if-q-is-used-but-35-v-otherwise/msg4693586/#msg4693586

Depends on its topology, so coukd please RE it and insert its schematic here ?