After few months I notice a strange behavior in one of the units, it become that one of the ceramic capacitors was damaged by the electrolyte leak, either I didn't see it, either didn't clean the board very carefully after I replaced all the electrolytes. never the less this was the reason to start digging in to the schematic. Beside the issue the service manual calibration process looks strange and didn't give a good results, so I wanted to find what exactly those trimmers are doing. The 251 and 210 was obvious set the gain and the offset, but 430 is not well described. So after spend two days reversing the schematic, testing and calibrating i get to the following calibration process that is the only one really working. 

1. You need to open the amplifier, you will work only with A10 board. You need to have an access to the TP300, everything else is a compromise. 

2. Let the board works for at least 20 min, longer is better, I think the optimum is 30 -60 min. It's important, calibration cold unit wild not work.

3. You should select 1V , DC range and have 0 V input signal

4. Adjust R210 until you get few mV on TP300. The perfect will be to have +/- 2 mV but you should keep on mind the next step.

5. Switch to 100 mV, DC range and check the offset, if the offset larger than +/- 40 mV you should try to adjust it in that range. 

6. You should do few iterations between steps 5/6 until you get a  reasonable offset on bot ranges. The worst board i have is -30mv at 1 V and + 40mv at 100 mV range.  Those values are still in the range the autocalibration can solve with a good margin left. 

7. Connect your probes to +5 V. Please not that if you do not use the original probes but 1:10 instead, you should add 655k resistor in series, and 5-40 pf capacitor to compensate the input. 

8. Using R251 you should set the TP300 to be 500 mV. 

9. Set the input to 0 v and double check the offset,.

10. Assemble the amplifier, and let it works for another 20-40 min. 

11. Set the range to 1V DC, set the input to 0 V and the output, keep on mind that the output should be a 50 Ohm terminated. 

12. With R430 adjust any offset you have. 

13. Switch to 100 mV DC and check the offset, if is more than +/- 40 mV try to adjust it using R430 and distribute it equally between the 1V DC and 100 mV DC range. 

14. Set to 1V DC range and connect the input to a 5 V

15. Using R251 adjust the output reading to 0.5 V plus the offset you have. For example if your offset is 10 mV you reading should be 510 mV, if your offset is -10 mV your reading should be 490 mV. 

16. At 1V DC range connect the input to a 1 Khz 10 Vpp source, and the channel output to the scope.

17. Compensate the probe, at this point you may not be able to have it fully compensated.

18. Compensate the amplifier input with C100, if you with standard 1:10 probe this will not be enough for proper compensation. 

19. Adjust the trimer capacitor in your extra compensation network /655 k + 5-40 pF/

20. You may need to do some iterations between 17-19 until you get the max signal flatness. 

21. Press Ch1 down + CAL to initialize the memory

22. Press CAL to start the calibration. If all step correct the calibration should pass successfully. 

Step 22 calibrate only the offset and the gain for 1V and 100 mV ranges. Further you may perform a full calibration CH2 DWN + CAL.

R430 set the LED current, since the amplifier use amplitude modulation, the LED current on A10 appear as an offset on the A20 output. You want to have the A20 output with small DC offset when the A10 is properly balanced. There is no way to do this without opening it. 

Also on two of my amplifiers i notice an issue with the trimmers, mainly R210  / 430. They are very sensitive to a touch, it's very hard to calibrate those amplifiers, especially one of them is impossible, the moment you lift the screwdrivers it changes with 20-30 mV, some times more, so they need to be replaced.