VLT110-4302 Quad Output SMPS Diagnosis and Repair

[bobthedespot]

Hi all, long time viewer of Dave's, first time poster.

I have a medical grade psu, the VLT110-4302, that has gone bad on me. After watching the combiscope diagnosis/repair, I was inspired to try to get it working again.

It's a quad output, with +/- 5V and 15V rails. The power good signal is used to drive a chassis fan, but I don't get fanspin or anything on the outputs of the supply.

The main input caps go to 160Vdc which seems sane, and all of the output caps seem to measure fine... no dead shorts at all. Two of the output caps are 6.3V, and I don't get a good reading on them, but I haven't removed anything from the board yet.

The main rail has a current shunt at 100ohms (4 digit coded precision), but reads slightly below at 85-90ohms... not sure if that's an issue.

If you have any recommendations for getting this diagnosed, I would be grateful.

thanks,
bob

[fzabkar]

A photo would help, but one common problem where the main bulk cap charges up and the supply fails to kick is an open startup resistor. This resistor connects the rectified mains (+160VDC) to the Vcc pin of a PWM controller IC. Resistor values are typically of the order of 100K or more.

Warning: make sure you discharge the bulk capacitor before working on the circuit.

[bobthedespot]

Thanks for the feedback fzabkar!

It looks like it's a KA3882 PWM controller. https://static.chipdip.ru/lib/249/DOC000249773.pdf

I am measuring Vcc on the KA3882 to the 160V as 187kohms, but only after a long settling time... it starts closer to 8Mohms.

As far as I can tell, the startup resistors are a series pair of 100k resistors, and I don't know why it take almost a minute for my fluke 116 to stop at 187kohm... maybe the fluke is bad? I've definitely abused it.

bob

[bobthedespot]

Would it be a terrible idea to solder test leads at Vcc input on the KA3882 and at the midpoint between the two startup resistors?

[fzabkar]

Did you discharge the 160V capacitor? I would hate to think that your Fluke was measuring the resistance of this capacitor when fully charged.

[bobthedespot]

Yes I shorted it with a paperclip once it was around 10V

edit: that's not to say i haven't done that in the past though, but I would be surprised if it still work at all in that case.

[fzabkar]

I would lift the end of one resistor and measure them individually.

Another common failure in these circuits is a small electrolytic capacitor across the Vcc pin. This capacitor is charged via a secondary winding on the switching transformer. When this capacitor dries out, the supply "hiccups".

[bobthedespot]

Ok, I'll break out the soldering iron tomorrow to double check the startup resistors.

I checked what I think is likely that small electrolytic... it's marked 150uF 25V, but it measures 30uF (haven't removed it from the board).

I'm no electronics wizard, but I can't work out how the capacitance would measure lower unless the cap is bad, right?

[bobthedespot]

Pulled the cap out, measurement across tantalum cap in parallel with the electrolytic fell down to 4-5uF.

Sample circuits for the KA3882 have the recommended values for these two caps at 100uF and 10uF for the electrolytic and tantalum respectively.

No idea if the 10uF is right for the tantalum cap on this board, but at least it seems that the 150uF rated cap needs to be replaced anyways.

I'll keep this thread updated with any news.

[bobthedespot]

Put a 140uF cap I had kicking around in place of the "150uF" that was onboard previously, and the measurement for this plus the 4uF tantalum immediately read at 144uF, so that was definitely an issue, if not "the" issue... gonna solder it all up and see if we get fanspin.