Ever disconnected from Samantha and burnt out motors? We from Say Watt?, along with other teams in our area, have experienced it multiple times. We saw that it was a big problem, so we set out to find a solution. After some digging, we discovered that the JoystickDriver.c that all teams using RobotC dutifully include in all of their code is the real source of the problem. We created an updated joystick driver that fixes this problem.

Advantages:
1) It is extremely simple for new teams to integrate. It involves changing exactly one line of code.
2) It picks up again the moment the robot is reconnected.
3) It refrains from adding additional tasks and calculations to the code, causing additional lag. It simply increments a variable.
4) It works for every connection type, even Bluetooth.

In essence, this is a very simple solution that can be put to use in less than a minute by teams of any level.

To see our solution explained or download the updated Joystick driver, go to our blog post.

I would correct you a bit and say that it's not always the real source of the problem.

We've seen numerous instances at tournaments where the comm count was incrementing (on the display) indicating communications were active, but the motors were running away.

Our belief is that this is due to a "feature" (it's not a bug, because it just wasn't planned for), that doesn't reset the tetrix motor controllers if they get brief disconnects. (Like if they get bumped during competition?) This causes the controller to continue with the last command (not the NXT). The RobotC firmware, and the HiTechnic/Tetrix motor and servo controller firmware need to be updated to correctly handle that loss of communication. Then we could have a solution that really solves the issue.

At our first tournament (at least half the times that I saw) we saw that communication was good, but the motor controllers were still running.

I don't think we've ever observed that at a competition. However, we do agree that this will not stop the robot in every single case. There are other cases in which we can't do anything, such as times when the NXT freezes (we've seen this multiple times).

Additionally, based on what you've described, wouldn't the Tetrix controller problem have existed with Bluetooth too? I don't think we saw more than a handful of "running away" robots when we were running on Bluetooth last year, compared to the dozens with Samantha this year.