I wonder if you get a chance to read that page for me as I'm not sure if I've determined a fault or not. I re read it and it's unclear if the manual is suggesting that it should or should not be greater than 17V on the collector, as it's only taking me to further tests in that area if it's not greater than 17V.
Perhaps this part is ok.
The manual suggests it's OK if it's greater than 17V. If it isn't you're supposed to check #11
Anyway, by passing #6, you should have jumped right to #16, so it wasn't required to do #10. Look into the Yes/No column what step is next.
Did you measure and check all output voltages according to table 4-11? If they all are OK, you wouldn't do the troubleshooting sequence anyway.
From your description, I'd locate the fault within "AC BUFFER" (see block diagram fig. 3-4) or "AC CONVERTER (same diagram). These both are part of the control loop for AC but not for DC. Study these diagrams in chapter 3 to get an idea of the different modes.
Then locate the signal "CONTROL" and open the loop - that makes troubleshooting easier. Drive Q106 (as seen in fig 3-4) from an external source (study the schematics to locate the signal and what source will be appropriate). Watch what happens at the output of A1, input and output of AC buffer, input and output of AC converter.
That would be my approach from your description of what doesn't work, beware there alway may be something I missed or you didn't test / describe that would point me somewhere else.
I'd suppose to roughly do all the performance checks to see what ranges / modes are working and what isn't working, then go trough the block diagrams to see which circuit blocks are within the signal path for the failed modes and not for the working modes.
Edit: Often I just don't follow all the troubleshooting tables in the manuals, taking shortcuts like above - but one needs to have a good understanding of how the device works to do so. Luckily it's quite well documented for the 5100 series.
Edit 2: Read your first post again: Check AC oscillator ouput voltage at different frequencies, might be just too low.
Edit 3: Just looked into the performance tests - don't be scared by all the instrumentation that is supposed to be required. Do the test roughly, just use a DMM to verify all the different setpoints, don't check the accuracy to ppm level.