On Saturday, I took grisbot to the Seattle Robotics Society meeting and showed a demo video of the light-follow mode. After the main presentation, I got out grissy and tried to set up a real-time demo in case someone wanted to come by and talk.

Well, after all these months, suddenly the robot decided not to work. Every time I tried to transmit data from the screen to the robot, I would get a solid flash on the LED — the signal for a data transfer error.

And as luck would have it, somebody did come by to watch. But I had nothing to show him except one failure after another to do the most basic function of the robot. Is there a jinx or what?

Tonight, back home, I tested out a theory, and discovered the following:

To put in words: the data transfers fine unless I’m using an old battery and have the computer unplugged. Then I learned the hard way.

And at that, while I have identified the physical cause of my problem (ie, old battery plus computer on battery power), I still don’t know the systemic cause (ie, why that causes transfer error).

On Saturday morning, I had left the apartment and realized in the car that I had forgotten to take a back-up 9v battery with me. I stopped to buy stuff at a store but forgot to buy a back-up battery. I stood outside in the parking lot and thought about it but decided not to go back. Nah, what are the odds?

So, Boeing, you have my sympathy.