OK, my mistake. Originally I thought that the resistor was joined to the red, but now it looks like it's connected to the pink, which doesn't make sense. it's really tricky figuring it out from a photo - much easier with a meter.
I can't understand it because normally, the 27v would go through a diode (the black thing near the red wire that shows the current direction as downwards), then through the big resistor to the "in " leg of the regulator, which is the one nearest the edge of the board, so everything looks right, but it doesn't tie up with your measurements.
My reckoning is that the 27v switches sides down the blue capacitor leg, then goes to the power switch. The switch latches the transistor, which then powers the panel and passes the 27v back down the pink wire for the controller. I think the latching is done by the processor via the small transistor. That's why you have to hold it for a bit to switch on. When you press and hold the power button again for a longer time, the logic in the processor detects it and powers up the small transistor that de-latches the main one - or something similar. It's possible that there'es something that I can't see in the photo, but I'm sure that my logic is correct. The pink can't be 0v because the diode would be the wrong way round.
Maybe it would be better to look at the controller rather than the panel to see where its regulator gets its power from. I bet it's the pink wire.