One of the tricks that Adam’s team used to trace this to a USB problem was listening to the tones generated by Windows 7 when connecting and disconnecting a device from a USB port. When the extension cable was in use – we could hear the “ba-dink” and “ba-dunk” sounds as the USB connected and disconnected when the PC was plugged directly into the USB extender with an A-A cable. This suggested that the A-A cable should be plugged directly into the Cortex, which worked perfectly, with no USB dropouts. Adam then plugged VEXnet keys in and the robot worked.
The moral of the story is not to simply blame whatever you don’t understand (like “the field” or “VEXnet”) and keep looking for a real problem.
Adam’s robot was also making some weird mechanical noises which the team thought was normal, until an hour of debugging found a missing screw, some loose screws, and an axle shaft that had slipped. When everything was tight, aligned, and replaced the robot is now much quieter and works better. I think they are ready for the BC Season Opener this Saturday.
I guess the other advice is that mentors really can be helpful sometimes. 