I noticed my Power General open frame 40W PSUs doing something really peculiar: It's really well below rated voltage. It was dumping out around 4V on the 5V line, and 9.8V on the 12V line. It has an adjustment but even when turned up all the way, it still would not hit more than 11V on the 12V line.
So I added another resistor in parallel with one of the voltage sense resistors (which is used as the input to a TIL431). With this resistor (I soldered it away from heat sources - more later) I was able to now adjust the PSU to get a full 5V on the output.
However I now noticed something else that was peculiar : the voltage changed as the PSU apparently warmed up. When stone cold it was producing 4.5V and when it warms up, 5.35V(!) Likewise the tracking 12V went from 11V to 12.3V. These are no-load voltages (but the PSU was rated that it should be fine running under no load.)
It appears the supply is using the TIL431 to regulate. The TIL431 is what drives the optoisolator that tells the UC3844 to lower the voltage.
The "critical cap" that filters the UC3844's supply was replaced as it did fail. Since Vref of the UC3844 seems to be a solid 4.98V I think it should be OK, but I should go scope it and the power input just to doublecheck.
I'm at a loss at what could have such a temperature variation. It has a couple of metal film resistors to set the voltage so those should be fairly stable. The only things I'm concerned about -
the TIL431 being bad and temperature sensitive... ? I tried heating it with my solder iron and did not see a rapid voltage increase but I may not have heated it enough (plus it's too cramped in there...)
other resistors (and the potentiometer) being temp sensitive after all these years?
temperature capacitor leakage? There are no electrolytics I can see on this feedback circuitry, just some yellow inline "drop" capacitors (monolithic?). I can't imagine these to be under significant stress to fail... I might have to replace them with ceramics to see if it helps the issue.
Anyone seen something like this and debugged it? I would suspect this is an age related problem, this PSU is around 25 years old now I'd guess. Unless it's one of those monolithic capacitors which I can't really test, the ESR of the other large electrolytics seem fine...
So I added another resistor in parallel with one of the voltage sense resistors (which is used as the input to a TIL431). With this resistor (I soldered it away from heat sources - more later) I was able to now adjust the PSU to get a full 5V on the output.
However I now noticed something else that was peculiar : the voltage changed as the PSU apparently warmed up. When stone cold it was producing 4.5V and when it warms up, 5.35V(!) Likewise the tracking 12V went from 11V to 12.3V. These are no-load voltages (but the PSU was rated that it should be fine running under no load.)
It appears the supply is using the TIL431 to regulate. The TIL431 is what drives the optoisolator that tells the UC3844 to lower the voltage.
The "critical cap" that filters the UC3844's supply was replaced as it did fail. Since Vref of the UC3844 seems to be a solid 4.98V I think it should be OK, but I should go scope it and the power input just to doublecheck.
I'm at a loss at what could have such a temperature variation. It has a couple of metal film resistors to set the voltage so those should be fairly stable. The only things I'm concerned about -
the TIL431 being bad and temperature sensitive... ? I tried heating it with my solder iron and did not see a rapid voltage increase but I may not have heated it enough (plus it's too cramped in there...)
other resistors (and the potentiometer) being temp sensitive after all these years?
temperature capacitor leakage? There are no electrolytics I can see on this feedback circuitry, just some yellow inline "drop" capacitors (monolithic?). I can't imagine these to be under significant stress to fail... I might have to replace them with ceramics to see if it helps the issue.
Anyone seen something like this and debugged it? I would suspect this is an age related problem, this PSU is around 25 years old now I'd guess. Unless it's one of those monolithic capacitors which I can't really test, the ESR of the other large electrolytics seem fine...
Comment