Sorry for the very late update. After making several changes, I no longer have the problem.
It seems that electronic noise from the motor interfered with the RC receiver, sending out
nonsense pulselengths. Since these were outside the limits I used to decide on the motor
speed, the situation was erratic. I was also having issues with the RC control code interfering
with interrupts for the motor encoders. My solution was to use a different micro (NanoEvery)
for RC control and the Mega for everything else. I may be migrating the Mega code over to a
Teensy 4.0.
Thanks to all,
John-