Stuck on very simple power supply.

[Dubbie]

I am stuck with this power supply board.
It is supposed to output -5V which it does if it is unhooked from the other microcontroller board. If it's hooked up, it sags to ~3.3V

if I hook up a 15mA load to it without the other board, it sags to ~1.2V

the triac works fine. The sync wire is just rectified AC for the microcontroller to trigger the triac switching. The gate wire is the PWM coming from the micro to switch the triac.

I have checked or replaced all the components marked with check marks. I've got no idea where to go from here. I feel stupid to be honest! I thought this was going to be easy.

I've overlayed the front and back of the board so you can see what goes where easily.

Thanks in advance for any ideas.

[Dubbie]

Just to reply to myself.
I think it might be MC2 that is the faulty part.

I misread the label as 220nF and measured it at 220nF
That doesn't really make sense for the power requirements, and a 1uF would be much more likely.
The symptoms would match a bad cap there as well.

I will get a replacement part in there and hopefully all will be well.

[Ian.M]

Since you photoed it with wires obscuring the big cap, and didn't post the values of the other two yellow caps, we really cant be much help,   However for 15mA DC out, you need a dropper cap of at least 240nF (assuming full wave rectification, 230V RMS 50Hz), or 480nF if its half wave rectified.   1uF sounds excessive.  680nF would be more likely in a halfwave circuit and 330nF in a fullwave one designed to supply 15mA.

[Dubbie]

It didn't have anything besides 1.0K written on it. I take that to mean 1uF 10%

it's only half wave rectified. So it needs to be at least 480nf.

I will try a 680nF first and see how it goes.

I got the 15mA figure from running the other board off the lab power supply. It may need slightly more when driving the pwm.

[coromonadalix]

your board receive a gate from the main pcb and sync signal from the 4 pins connector,  is the second board ok ?? you mention 5 volts going into it ??

check r6 and r7 too

 i think the sync signal maybe the 50 or 60 hertz main frequency passing thru D3 for an half rectified / pulsed frequency ...

[Dubbie]

the sync signal is fine. So is the gate drive when I power the other board from the lab power supply.

[Ian.M]

I agree the big cap could be read as 1.0uF +/-10% (letter K for tolerance).   How certain are you of that 15mA load current?  As class X capacitors tend to loose capacitance with age (due to the self-healing process for a metal film cap), if its designed to function under low mains conditions, the max. load current is a bit higher, and there is an antisurge resistor in series with the cap, and its from before the EU mandated a 0.5W limit for standby mode, it could well be 1uF.   

[Dubbie]

All fixed.

Thanks everyone for your input.

a 1uF replacement cap (all I could find) did the trick nicely.
It's been a nice learning experience with this sort of power supply.

[cvanc]

Glad you got it fixed.

Do my eyes deceive me, or is this circuit board laid out by hand?  You've got some very wobbly traces on there...

[Dubbie]

No I scribbled around them in photoshop quickly to improve the contrast. That’s why they look a bit rough!

[Ian.M]

The fact the 1uF X2 capacitor lost over 3/4 of its capacitance indicates the PSU has been exposed to some pretty vicious HV transients.   Adding a suitably rated bidirectional TVS diode or varistor (but please put heatshrink over a varistor as they do wear out and when they die they make a mess), across L and N after the fuse and the common mode choke line filter, to clamp the transients would improve reliability   The varistor or TVS needs to be rated to stand off your nominal line voltage +10% + a small safety margin - 275V RMS (390V peak) is a good rating for 220-240V operation.  It looks like the top pads of the footprints for the not populated Y2-1 and Y2-2 caps would be a good place to fit the TVS or varistor across.

[Dubbie]

Thanks for the advice Ian.

You can't really see it in the photo, but right across the mains where it enters the board is a component that looks like a TVS or MOV. I'm not 100% sure what it is. Going off the image I would hazard a guess at a 7N471K MOV. I can't be bothered pulling it all to bits again to find out.

If it dies again. I'll add some more protection in there.

R

[Ian.M]

In that case, I'd replace the varistor (which should be after the fuse anyway as it will probably arc when it fails, so I hope there's a fuse upstream before the building's breaker panel! ) as it has certainly led a hard life and is likely to die with the usual drama sooner rather than later.    Covering the varistor in fire-retardant heatshrink is also advisable.