Page 1:Power Supply Troubleshooting
Page 2:Background Story On Today's Patient
Page 3:Preliminary Testing: The 5VSB Rail
Page 4:Preliminary Testing: PS_ON# Behavior
Page 5:Words Of Caution
Page 6:PCB Analysis
Page 7:The Test Plan
Page 8:Setup And Safety
Page 9:5VSB Capacitor Replacement
Page 10:Auxiliary Output Results
Page 11:Auxiliary Output Capacitor Replacement
Page 12:PWM Bypass Capacitor Results
Page 13:Capacitor Postmortem
Page 14:Do We Have A Definitive Fix?
5VSB Capacitor Replacement
The first order of business was changing those 680µF 5VSB caps with some of my 1200µF FMs. Yes, that is overkill, but those are the smallest electrolytic caps I have in stock over 100µF. There is one little problem with this: they do not fit properly on the board due to how tightly packed other components are around them. The FM is 2mm wider and has 1.5mm-wider lead spacing.
Side by side, the size difference is hard to miss. I measured the old 5VSB Teapo SC capacitors after removing them. The rear one, nearest to the heat sink and flyback transformer, read 34µF instead of the expected 680µF. Yup, that is one sick puppy. The other one still reads 600µF, which is a bit on the low side but still 15x more usable than the first.
The FM I wanted to use was an impossible fit for the critical rear capacitor's location, so I moved the hopefully still-good (enough) old cap over there for the time being. Even fitting the FM in the front cap's location required splicing wires onto the nearby dummy-load resistor to move it out of the way enough so I could get the cap reasonably close to the board.
There's not a whole lot of change compared to the previous waveforms using an external capacitor on the 5VSB line. The switching noise is now almost completely gone and that is about it. The unclamped waveform still surges slightly beyond 8V and looks like it is going to settle around 6V. On the clamped waveform side, though, the surge now stops at 5.5V with the regulator shunting 1.25A before settling at 5.2V and 250mA. There are still some unexplained irregularities in the clamped 5VSB waveform that need to be looked into.
I did not take the time to fiddle with the shunt's set-point until I got a successful turn-on. All I wanted was to see what sort of effect the cap may have had on turn-on behavior. The 5VSB output is cleaner, but the main outputs' ramp-up is every bit as trashy as the previous run using the external capacitor. In other words: onward to step two of my repair plan.
- Power Supply Troubleshooting
- Background Story On Today's Patient
- Preliminary Testing: The 5VSB Rail
- Preliminary Testing: PS_ON# Behavior
- Words Of Caution
- PCB Analysis
- The Test Plan
- Setup And Safety
- 5VSB Capacitor Replacement
- Auxiliary Output Results
- Auxiliary Output Capacitor Replacement
- PWM Bypass Capacitor Results
- Capacitor Postmortem
- Do We Have A Definitive Fix?