Okay I think we fixed it finally. A while back I noticed that one of the wiring bundles that goes down the arm to the center rack had been damaged. Looks like it had snagged on the frame when the rack was lifted up. Ripped a couple of wires that had been then been spliced back together before we owned the machine. The splice looked okay to me, though. However I traced the bundle and it was the main AIM command Pro harness that went from the Raven rate controller ICU near the pump all the way to the back control box. So it became my prime suspect.
Long story short, the wiring harness must have still had some breaks in the wires that intermittently either added noise or dropped out the signals. Case wanted $2600 for a new harness, which is insane! So I made up my own harness for about $50 and haven't had a pump cut out in a couple of days now, knock on wood.. Details below.
I got to thinking about some of the weird things I'd seen such as the rate hunting occasionally when speed was steady, and often times that seemed to precede a pump cutting out. With Aim turned off, I still occasionally saw rate bouncing around but the pump rarely shut off. Tracing the AIM command harness, I could see that it essentially inserts itself between the sprayer's sensors and the Raven rate controller. Most of the wires pass through from the Case harness plugs to the Raven controller without even going to the Aim Command Pro controller, but 5 wires did: The flow sensor wires, and the wires that control the hydraulic water pump (AIM calls this the "servo"). So in the AIM Command Pro harness, 5 (well 4 actually, one green wire is unused by Aim Command Pro for whatever reason) wires take these signals to the controller, and 5 wires (white striped) take them back. These 9 wires plus the main battery power wires make up the entire wiring harness. If any of these wires had a problem that could cause the symptoms I was seeing.
Thinking $2600 was insane for a new harness, I bought some 18-guage stranded multi-conductor wire from my local electrician and ran some new wire from the controller on the back to the Raven unit underneath the tank. 9 wires in all. I happened to have a supply of Deutsch connector hardware (DT and DTM, both of which are used in this harness), including the pins, which is really handy for farm machines since nearly every machine uses Deutsch connectors. I pulled the suspect wires out of the Deutsch plugs on the harness, taped them out of the way, and put my wires in their place, replacing them pin to pin with no splicing except in one case when I was short a DT female pin, and in that case I soldered them together.
A final multimeter test, a check of my notes and photos on the old wires, and plugged it all back in. And it seems to have worked. I'm much relieved.