I tried doing the layered transparency trick, but it didn't really work out very well for me. Everything just looked all muddied up and I couldn't tell traces apart from components. (Keep in mind I'm working with an awful camera for this sort of thing.)
But I gave the PCB another really thorough go-over, and I more closely looked at the datasheet for the optocoupler in particular and quadruple-checked how the optocoupler was connected into the circuit, and this is what I came up with a second time around. Now I'm even more confused by how the heck this thing is supposed to work, but somehow it works and puts out a steady 9V DC under no load and about 8.2V under load.
That 10uF/50V electrolytic off the negative end of the primary coil also kinda bugs me a bit, because it's not in an AC signal. It's in a full wave rectified DC signal. So I don't think it would be acting on the circuit with capacitive reactance, would it? Especially with that 2u2F/400V electrolytic across the terminals of the diode bridge smoothing out the rectified AC?
Here's links to the optocoupler datasheets that I found:
(I couldn't find a direct link to any OEM datasheets. Even the NJE3003B is a defunct part)
Another friend seems to think the feedback coil of the transformer bootstraps the 2 transistor oscillator and gets it kickstarted, and the zener diode somehow plays a part in all this. I'm guessing when the output voltage rises above 8.2V, the zener clamps it down and shuts off the oscillator.
Anyways, here's more detailed pics of the PCB, top and bottom, and a new sketch of the circuit. Maybe that can help a bit? ^^;
EDITIiiii made a mistake in both circuits. The little signal diode ST4148 is
not connected to the base of the transistor NJE3003B. It's connected to the node between the 4u7F film capacitor and the
other end of the feedback coil. NJE3003B is connected to nothing else but one end of the feedback coil. I think that's part of what's been messing me up this whole time.
TIDE