BFL,
I was out on a trip for 10 days with no electronic communication so I came late to the party on this thread. The discussion is interesting but I didn't understand a couple of your examples quoted below. I tried to send this as a PM, but since you don't accept PMs, here goes.
"If the battery is too full, then I run some lights to get enough "room" for the amps to show. EG, Isc is 8.2a. Batts are near full so only accepting 1a. Turn on four lights (12a) see -11a on Tri. (+1 for battery) Hook up the solar. Now see -3a. So 8a is reaching the batts from solar."
Any time you see negative on the trimetric, no amperage is going to the battery. If you turn on four lights and see -11a, 11 amps is coming from the battery. If you turn on the solar and now see -3a, the panels are now supplying 8a of the 11a demand. 8 amps is not reaching the battery; 8 amps is being produced by the panel and 3 amps is being given up by the battery.
A later post in this thread says:
"I did do AH tests for the PWM by putting only solar on the Trimetric shunt. I had to leave lights and fans on all day so the batts would keep accepting the full solar."
My understanding of what was happening here is that you had to keep lights and fans on to keep the panel production at max. However, the lights and fans were consuming this power, not the batts. They were probably only accepting the same amount they would have been had you turned off the lights and fans. . . Unless you meant to say that the load kept your controller in Bulk or Abs so that more might be available to the batts.