So I'm leaning against the cable/motor theory ....
Here's what I've done ... I started by completely cleaning up my work bench in the garage so I could disassemble it near all my tools ... YAY for cleaning up the garage shop, it really needed it. Then I took a long beer break ......
{whistles a lot while drinking a brew ...}
OK. SO then I removed the motor from the robot and decided to bench test it. The drive motor and water pump impeller both move freely and easily. No binding or grinding. Plugged the cable into the power supply and ran the motor on the bench. Everything spins up just fine and it passes through it's self-test cycles. Shut down the motor after a few minutes because you're really not supposed to run them out of the water for too long. While it was run I wiggles and bent the cable every which way to Sunday and at both connector ends. Motor never got interrupted. So I'm thinking all the connections are just fine and the motor is good enough at least on the bench.
So that led me to the wheels, brushes and treads ... and here is where I think the issue lies. Hard to explain verbally and I failed to take any pictures or video so you're going to have to put down the hot dog and beer, put on your imagination hat and really focus on what I'm trying to describe ... hopefully you haven't had too many beers ... or maybe it would be good if you did .... up to you.
When I manually turn the wheels/brushes and move the treads, they move freely for a length and then they start to bind up. If you watch the tread roll over their wheels (they have a gear and tooth design so that the treads are locked into the wheel) you will see the teeth on the inside of the tread pop up out of the gear teeth as if there is too much slack. This causes the wheels and brushes to bind up a bit until you force it to go further and then the treads fall back into place. It happens about twice during a full rotation of the main wheel. So, further disassembly revealed what I think is the issue -
Notice that gap between the front wheel and it's bushing (I don't think it's teflon, feels like nylon). So as the wheel rotates when the motor drives it, it's not a pure rotation but rather a rotation about an axis that is not coincident with the shaft on the robot body. The wheel essentially moves forward and backwards while rotating and it's enough to cause the tread to pop up out of its channels and bind.
The fix would be to get new wheels and bushings ... probably $40 worth of parts. The problem - Marina doesn't have them and I'm not seeing them ANYWHERE on the inter-webz. The part is -
Maytronics Dolphin 9993219 Wheel Kit
Also you can see how worn out the interior surface of the wheel is and the exterior surface of the bushing it rides on. Definitely falls under "wear & tear" ....