Okay so now I'm back to thinking abut my selectable 120V parallel and 240 V connections.
To reiterate, two panels. Each panel would have a recessed male 4 prong connector to connect to the generator.
The generator would have a cable output with a mating 4 prong female connector. It would bring out U1 U2 U5 and U6.
Each panel would shunt the leads together properly internally to configure for its output voltage, 120 or 240.
Originally I thought to put the AVR in the 120 panel, and leave the 240 panel unregulated.
But:
In researching genhead prices last night (when I thought mine might be bad) I found Tom Osborne's Ebay store, and noticed I could buy a replacement AVR for $49.
So perhaps it would be possible to have two AVR's and put the original AVR in the 120V parallel panel, and the second one in the 240V panel.
Not sure if it would be possible to do because both outputs must be connected to the stator -- they might interfere with each other. (The input lines of one AVR would always be disconnected, however.)
This type of AVR has output leads that are typically spliced (according to Tom's docs) in series into the Z1 (or Z2) stator lead.
http://www.centralgagenerator.com/documents/avrpics/AVR110-Silver,-Blue-end.jpgWith two AVRs I'm not sure if you would need to connect their outputs in series with each other, or in parallel. My guess is series and that the AVR simply acts as a variable resistor.
I'm also not sure if the input lines of the unused AVR need to be shorted together or whether an open is an acceptable input state. I suppose I could test that on a single AVR with its output connected normally.
Note, the AVR output leads would be hardwired from the panels to the generator. They are not part of the cord with the connector.
I'm not planning a dual AVR experiment (yet), and will probably just set up panels with a single AVR on the 120 parallel as explained earlier.
But if I got more information about the AVRs operation and it proved do-able, dual AVRs is how I'd ultimately like to set it up.