Micronics Cyclone II 500W

[прямо]

Came across this famous Korean brand PSU (but made in China) with a strange problem that I have never encountered before.

It can start, and aside from 5VSB which is spot on, the rest of the output voltages are off the chart.

12V is 11.2V

5V is 5.9V

3.3V is 300mV (or you could also say no 3.3V)

primary caps (2x 820uF 200V connected in series), one reads 68V, and the other one reads 34V.

any idea what to look? obviously primary side has an imbalance and under voltage issue.

I wonder why the protection (GR8313) doesn't kick in and shut the power supply down.

[прямо]

A bit of an update:

Reflowed the solder points of all AC input, and that fixed the problem with primary caps not fully charged to 320V.

Can't do measurement in circuit for some components on the secondary side until I remove the dual rectifiers and their heatsink, but the thick ground plane without any thermal reliefs is giving me a very hard time in getting the solder to melt. I'm using the biggest tip I have and temperature is already at 420 C, but I just ended up making a mess of flux and cold solder blobs.

[temyth]

get easy on the output diodes, they are hard to get out and i've never saw one of them with some kind of defect. if you want to test for shorts in the secondary, inject the respective voltage of each respective line (12v, 5v, 3.3v) with another good PSU and see the consume. If too high (more than 0.5A), then you should take a look at the voltage comparators (normally some TL431).

[прямо]

Like what I did with the other PSU, because this one is also a half-bridge flyback design with UC3845B as the main controller:

Observed the PWM signal, it's fluctuating from about 2.7V to a little over 3V. Unfortunately, DMM says 0 switching frequency and 0 duty cycle. Doesn't make any sense.

VREF is correct at 5V. VCC is at 12.7V (chip needs 8.5V to turn on) and I forgot to measure the COMP pin...

PWM TO-92 switching transistor measured OK out of circuit. MOSFET too. RC snubber and Schottky diode as well. No ESR meter to measure the little electrolytics.

Interesting to note, I did noticed that the primary caps discharge went from really fast to a snail like pace after it has reached around 20V or so.

After it being off for some time, I injected 3.3V into 3.3V rail and measured its feedback circuit. Working normally as it should be.

I guess my next move is to replace the opto couplers and hopefully that will result in a total fix.

EDIT #1: nope, with new opto couplers, still the same.

Still don't understand why the supervisor IC is allowing the PSU to start. Maybe it has gone bad, and need to be replaced.

-12V rail is still missing. Could be why 3.3V rail is wonky.

EDIT #2: Found a dead 2A 100V Schottky barrier diode in the -12V rail.

[прямо]

Well, I replaced the diode with SS56. Had to cut a 5W resistor legs to use them with the SMD diode, and that is all it takes to fix the PSU.

All output rails are back to normal

[temyth]

Well, I replaced the diode with SS56. Had to cut a 5W resistor legs to use them with the SMD diode, and that is all it takes to fix the PSU.

All output rails are back to normal

good! guess i'm going to pay more attention on the shotkky diodes from now, i've never saw one shorted until today. living and learning
keep the good work!!!

[прямо]

good! guess i'm going to pay more attention on the shotkky diodes from now, i've never saw one shorted until today. living and learning
keep the good work!!!

Yeah, me neither, especially at -12V rail. That's why I didn't bother to check and I was looking at the resistors first because I thought maybe one of them has gone bad with open circuit.