unless I read it wrongly, the 10Meg resistor is in entirely the wrong place, (reference to the drawing) surely? -The schematic shows one end of it being PERMANENTLY connected to point 20... not what the layout above shows. So -yes- I see a difference here... the resistor is effectively wired in parallel across the opening contact of the 4:1 switch, and it should be across the opening contact of the 20:1 switch.
HOWEVER... -Not that I can immediately see how that would kill the operation: it appears to be there as an "all-switches open" terminator, and effectively inactive when the unit is in operation.
If you have a 2-channel scope and a PAIR of 10x probes, I'd use a pair of 10X probes (in order to keep the impedances over 10Meg instead of 1meg...) and read the voltages during compression at the 150Ω/470Ω junction, and the other channel at the 20-wire. Are they the same? -You SHOULD read a descending control voltage as you move from 150-470-560-1.5k resistors. Any failure for this voltage to reduce would mean no increase in ratio. On the other poles of the switches, the resistor ladder (56k-68k-56k-56k) just reduces the sidechain signal level by a few dB (about 3,6,12 etc) as the ratio increases, to help reduce the interactive fiddling. You should easily be able to read the audio signal at wire 23 gradually becoming quieter as you switch up through the ratios... this just moves the threshold by a few dB (I estimate about 12dB total between all 4 ratios, without resorting to any serious calculation...) Sounds like this is not your problem, though.
One final thing...
Measure resistance with the power OFF between pad 20 and ground... you should read 150Ω. -Actually do this first. -You should read 150Ω to ground from the blue "20" wire in the diagram above. If not, your control voltage is not being attenuated to change the ratios.
Keith