help with SMPS

I've got a couple of faulty power supplies, 12V/20A that are sold as power supplies for LED stripes. Both exhibit a similar failure but I cannot figure out how to fix it.

Both provide 12V without load, but as soon as load if applied (even if only 100mA load) the power supply will cut down the power, it will to restart, and the cycle will repeat, as if it was overloaded (the led in the power supply will flicker).

From that I guess that high-voltage switching transistor is ok, regulation is ok as voltage is stable as while load is ok. Output diode is ok (I have already tested one of them after taking out out the circuit).

What else could be causing this failure mode?

Thanks!

Those are a sort of smps idd, but are not voltage mode. They are working in sort of current mode, because thats whats relevant for leds. Usually there is a "led driver ic" that does this current mode regulation, and is specifically made for application in led power supplies. Check its datasheet online, and app notes.

So these are supplies for power leds, e.g. probably set to about 50-100mA just to have an idea of the scale. And those leds need to have that much current in them and no more, regardless of anything else. Since they are leds, if current gets more than that, the leds would burn. So thats what it does, it keeps the current to that low level, its job is not to keep the voltage at any level because the voltage is irrelevant.

I would say thats not a failure mode, but probably its actually working as intended.

I guess I should have been more specific (many LED stripes do include groups of serial leds plus resistors designed to be powered at 12V or 24V).

This type of supplies [www.ebay.com] is what I am dealing with for RepRap use. Most of them work ok for me, but from time to time I get a defective unit. When that happens I try to repair it if possible. If a component blows up then it is easy to spot, main switch or dead output diode are the usual suspects, but this time all looks nice and clean.

Oh ok, now i am getting it finally. You do have a smps that is meant to output 12v, and can be used for anything, incl leds that have resistors in series with a value suitable for 12 supply. Sry i didnt got that from the first time, its coz this is very terrible efficiency wise, e.g. 3-3.5v led drop means ~75% of output is wasted on series resistors, not to mention smps losses, so probably just 10-20% is actual power on leds. This is why when you said "led supply smps" i just skipped and imagined a setup of a "led driver" ics that does just that without series resistors - which is something else. Sry for the previous then and my bad.

Then you have a regular smps with a single 12v output. It does make 12v with no load, and then voltage falls down under load.

Perhaps the ic that regulates the duty cycle fails to increase it as response to load changing. If the duty doesnt increase then ofc the voltage falls like that. Sry i have never dealt with that kind of smps, so cant say much, just speculation. But that could be a valid culprit. The ic that gets the feedback from output, and then it outputs the duty cycle for the switch. This one could use checking, e.g. check its duty output when there is no load, then put a load and see if its duty changes, e.g. if it reacts to load changing then it works, otherwise it doesnt. Also if the load changing does produce a change in feedback level for that ic to see it in order to react, e.g. the ic might be ok but it doesnt register a feedback path change. Could see its duty cycle with a scope. Perhaps with a true rms dmm, those readings even might be incorrect but at least should show a change if there is one. Or if you get some supplies like that which do work, try replace that ic with one that is confirmed to work, and perhaps do that with other parts. Btw, if you do anything while its powered up, do put a light bulb in series with the mains voltage, that would limit the power into the smps and might help in case of unfortunate shorts.

I would try to scope out the signals on good vs bad power supplies to see if you can narrow down where the problem is. The symptoms you give are similar to the smps reacting to an over-current situation, many have a output resistor with a voltage comparitor on the output someplace for current sensing and feedback through an opto-coupler into the PWM circuitry to disable the P/S if the drop is too high. I would look for that part of the circuit and scope it out compared to a good supply to see if that is being triggered. Without a scope, temporatily disabling the feedback may help you test if that is a factor.

Yes, next step is to scope both the good and the bad one, as primary inspection of current sensing resistor and related components render showed no error. Thanks again.