OK, you are missing the main supply from the channel closest to the large connector.
The pins on this connector go like this:
Violet?? (not sure) = temperature sensor, connects to violet wire across the driver board. Voltage seems OK
Black = power ground left channel = 0V OK
Black = power ground right channel = 0V OK
Brown = main power supply = 47.8V OK
Orange = positive bias supply = not OK, should be round +80V. See note below.
White = output left channel = not OK, should be 0V, see below
Red = output right channel = 0V OK
Blue = negative bias supply = not OK, should be around -80V. See note below
Green = speaker protect out = voltage is ok but it should not be present because left ch output is not 0V
Violet = main power supply = not OK, should be -47V
The 6 pin connectors on the sides of the board are wired in parallel, that means pin 1 on the right connector to pin 1 on the left, etc, until pin 6 on the right to pin 6 on the left. Your previous measurement shows that there is something wrong or you have skipped a pin.
Here is what I do know:
1) The 47.5V positive main voltage is missing from both channels or is intermittent. The wire going to the bottom pin on the long connector on the side is probably broken or has a bad contact. you are measuring 19.7 or so V and it should be -47V. Check that the wire from the power supply board to the driver board is working, or that the connectors have good contact - this comes from connector (51) pin 6 on the rectifier board (mounted on the two large main capacitors).
2) The bias supply is working correctly but because the main -47V supply is missing, the circuits are pulling it down. This can be seen from the fact that the difference between the bias supplies is still around 160V (37.9V+131.3V) but it is not symetrical.
3) The speaker protect circuit (driver board, Q501, Q502) is not functioning properly. It should not let the relay click. Either there is a bad contact on pin 1 of the large connector on the side of the driver board or some of the components in the circuit are faulty.
4) The LEFT channel (the one with 18V on the output) is likely working properly. The right channel, the one with 0V on the output is not. The circuit cannot output 0V with the negative power supply missing.
5) C307 is likely installed backwards. Also check that R319 (1k) has not smoked. Also check C357 and R369 on driver board (same function as C307, R319 but other channel). It is VERY strange that C305 is damaged, SOmething may be wrong with other components around it, including transistors.
6) C408, C416 may be installed backwards but also their failing may be because of the boad voltage overload. Check R414 for signs of overheating! Also, there is a design problem with the bias supply in that the positive and negative bias supplies influence each other depending on load. This can be fixed by replacing diode D405 with two diodes of the same type in series, and connecting the connection point of the two to ground.
7) does your amplifier use wires wrapped directly to the connectors on the sides of the driver board, or does it use white female connectors? At some point it will be required to isolate the power supply for the right channel. The 6-pin connectors on the driver board are a 'bridge' that can originally be removed, early versions had female connectors here. However, it was later discovered they can be unreliable and if the contact of a bias supply fails, destroy the amplifier, so they started using directly connected wires. Tell me what is on yours and I will see how this can be solved.