Need help with switching power supply repair

The cap that filters the supply to the pwm chip on the primary side, maybe ? It's pretty common on most designs of switcher. Work on the thing on an isolation transformer if at all possible. They are potentially very very dangerous if you are not fully competent with them

Arfa

See if you have an optical coupler in there. I've seen these used as part of the regulation and it's possible it isn't working any more.

Also check for bad caps, ripple could be causing a reg issue.

Jamie

"senator richards"

You real sure of that measurement?

Check for open resistors or diodes in the supply chain to that IC.

Any yellow glue in sight ?

.... Phil

Try to load both sides with about .1amp, low load on a switcher can confuse the hardware sometimes.

On Wed, 21 Sep 2011 13:31:46 -0700 (PDT), senator richards put finger to keyboard and composed:

Just to add to what Arfa said, the supply to the PWM chip often starts with a simple resistor feed from the +170V rectfied mains voltage, and is then augmented by a regenerated supply from a tap on the switchmode transformer. If this is the case, then there will be a small electrolytic filter capacitor (~10uF) that often develops high ESR.

- Franc Zabkar

"senator richards"

.... Phil

try to check the main capacitor with an esr meter, or change it ! build yourself easy an esr meter grab any you want on my web site below regards,

e ?

Thanks for the help. I borrowed an ESR meter and checked the cap that supplies the pwm chip. The cap is a 47uf 50v and esr measures .5 which appears to be about right.

I don't have an isolation transformer but will look into getting one.

Thanks for your help.

-R

There is an optical coupler. Its a CNX82A chip. Is there an easy way to test it? I looked at Mouser and the part is discontinued and NLA.

I've checked all the caps for low ESR and they all check out OK.

Thanks.

-R

Hi Franc,

I found the 10uf cap you mention and measured the ESR. It was very high (around 50), so i replaced it with another cap i had around. it is not a low esr cap and only an 85 deg one, but figured it would be ok for testing. it measures about 4.5 on the esr meter which is still a little high (should be closer to 2). I still get the same result when its hooked up.

-R

I actually have one, but didn't have it with me the other day. I have it now and have gone through all the caps. The only one that measured bad was the 10uf 16v cap that Franc mentioned above. I changed it with another cap and it didn't appear to fix the problem.

The main filter cap (68uf 400v) tests ok for ESR.

-R

A photo of the PS would be helpful. Mostly, I'm looking for what manner of regulator chip is being used and whether there's a transformer or optoisolator involved.

If you have +30VDC on an output circuit that is suppose to only tolerate +15VDC, it's possible that the output caps are now dead, especially if this PS has been run for quite some time. The output caps are probably rated at 25WV, which doesn't apprecitate the 30V (or more if there are voltage spikes on the output).

Yeah, that's the usual problem.

Fluctuations like that are sometimes arcing in the xformer or spurious oscillations in the regulator chip. High output voltage usually means a failure in whatever is regulating or sensing the output voltage. As others have suggested, get an isolation xformer before you kill yourself. Then, attach a scope probe to the output lines, regulator pins, and optoisolator, and see what's causing the fluctuations. That's not normal.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

determine what output rail the supply regulates from. then check it's filter caps for ESR. I've seen where the filter cap(s) on that rail had a high ESR,and the PWM control tried to drive it back to the nominal voltage,and the OTHER output rails climbed high because of that. And that caused high dissipation in other circuits that eventually burned a hole in the PCB.

you could also have a bad electrolytic on the reference output on the PWM chip. or a reference divider resistor changed value.

Find the datasheet for the PWM Ic and learn the inputs and what levels they should be at,that will tell you how your supply is supposed to work.

I've checked all the caps. i replaced all the output caps initially before I had the ESR meter. Now with the meter i've gone through all the caps (there are 8 total on this smps) and only one tested bad. It was the 10uf 16v that Franc mentioned previously. Replacing it didn't seem to fix the problem. I have the datasheet on the pwm. It is a UC3842BN I checked the input voltage and it was erratic as mentioned previously. I checked the two resistors, diode, and filter in the supply and they all seem ok. Here is a link to some pictures I took with my phone:

A brand new cap might be a little better than that, but certainly right ball park at 0.5 ohms, and would not be an issue at that figure. About the only other thing that you could try, assuming that it uses a startup resistor from the 170v, is to disconnect the self-feed diode from the switching transformer. That way, you will force it to run from the startup supply only, just in case the self-run voltage is fluctuating, and interfering with your reading on the pwm chip's supply. Bear in mind though, that you can't run it for too long on the startup resistor, as it will get quite hot, being normally intended to supply current to the chip, only for as long as it takes the supply to fully start up and settle. If the voltage supply to the chip still jiggles around when it is only being fed by the resistor, then after the 47uF cap, which may yet be faulty, but not in an ESR way, the next prime suspect would have to be the chip itself.

Arfa

I've never seen a chip 3842 dead... but the small cap 10uF beside it has really often been solving the case !!!!!

And I do few SMPS repairs.

.... Phil

"Arfa Daily" wrote in news:bPQeq.11661$ snipped-for-privacy@newsfe24.ams:

IIRC,the values on the DSE meter are MAX values,so IMO,if a cap is reading what's listed,it's marginal.

usually,the startup resistor doesn't supply enough current to keep the IC running,just enough to start,that's why they have the housekeeping supply from the XFMR. Otherwise,you get a "burst" or "chirp" mode of operation;the PS starts,then dies because the IC isn't getting enough power. Often,it's the housekeeping supply electrolytic cap that's gone bad,as you said.

I've also seen where the start resistor goes high in value,even open,preventing starting.(but not this guy's problem)