Saturday, 19 November 2011

Bluebot-L the retired veroduino robot

We learn by our mistakes, but it can be cheaper and less painful to learn from the mistakes of others. Here's one of my mistakes; I hope the story will save you time and money.

A few months ago I started work on a wheeled robot based on Veroduino - an Arduino clone build using Vero strip board.

I had a wheeled base which contained a couple of geared motors. They drew a fairly low current so I decided to use a TC4427A h-bridge controller on a shield-like board that sat atop the ATMega326 micro controller. I decided to use separate power supplies for the controller and the motors;  I bought a couple of battery holders, each taking 4 AA cells.

I used foam plastic for the base; it's easy to cut and drill,  and it's lightweight, but strong enough for the job (so long as it isn't crushed). Construction was rapid and I soon got to the point where I could try it out on a simple line-following exercise.

Disaster! The batteries at each end contributed to a really high moment of inertia (MoI). When the robot turned, the high MoI meant that it over-corrected and hunted around the line without ever settling down. It swung rapidly from side to side like a boat in a raging stream.

The hunting was made worse by the length of the base; a small turn to the left or right moved the light sensors by a couple of inches, so it was surprising that it followed the line at all. Sometimes it didn't, and raced off looking for something new to follow.

Fun to watch but not very satisfactory.

I could have compensated for the MoI by tuning the software, but it was obvious that this was a flawed design. So Bluebot-L has remained in a project box, waiting until I have time to sort out a better designed base.

Meanwhile I have moved on to Trackbot. I'm pleased with the first iteration for Trackbot, but I still have a couple of unanswered questions about its capabilities, and more experiments to do.

I'll post more about Trackbot later this week-end.