It seems we're getting somewhere. One point:
1. Around this point I'm not sure if the voltages are actually DC if in ALT mode. I haven't fully understood the circuit, but the circuit drawn in between both channels might control which channel is displayed. So if you select ALT you might see only one channel show any voltage while the other is disabled. This change every time the trace is fully drawn. Please use either A (if measure on Channel A) or B.
The problem is, that the path through the signal-path is current based. For example the current through R557 is supposed to be 4.2mA and 65µA/div. While the 65µA/div doesn't matter with GND at the input you can ignore that part. I hope your multimeter is doing good at low voltages. Can you measure the following points?
1. Setting: POS = middle; Input: GND, select Channel A --> Voltage across R556 and voltage across R557
3. Setting: POS = middle; Input: GND, select Channel B --> Voltage across R656 and voltage across R657
The voltage calculates like 4.2mA * 301
= 1264mV at this point. The POS-Knob offset is not applied to this. This will happen after the V513/514.
For the good channel I expect at point 2: 1264mV for both resistors. For the bad one it depends if the problem is left or right side. If the error is already there: about 1382mV. If the error is not present there in it will match the one on channel B.