On Friday I got the email that we’d been hoping for: we’ve been accepted into the beginner’s category for Pi Wars 2017. 🙂
After much insistence from my son that he couldn’t arrange the motors so that the body could go faster, he did a major redesign on Friday after school, and I have to say that it’s a very elegant design, and was indeed nice and fast when demonstrated using the Lego remote control.
So my job for Saturday was to get it working with the PS3 controller. Unfortunately that didn’t go as planned… turns out that the cheap imitation controller that I have uses a different chipset to the standard PS3 controller, and while it will work quite happily with a PS3 or Windows, it isn’t supported by the bluez libraries. So moral of this story is to make sure you can drive the hardware you want to use! So back to the drawing board on that one…
In the meantime, I hacked together a simple curses-based keyboard controller, so you can ssh in and just use the arrow keys to drive the robot. It’s very crude, just running the motors at a set speed: full forward, full back, stop right and forward left to turn right, and stop left and forward right to turn left. (I also tried it with forward right and backward left to turn left and vice versa for turning right.) And here we discovered a problem. It stalls on a turn with either of these settings. Our wheels are too grippy and/or motor not powerful enough. It puzzled me that son could drive it with the Lego remote until he admitted that to turn, he slowed down one side, so yes he can turn it, but with a large turning circle. And that isn’t good, because we’d like to to be manoeuvrable.
So this led to much more discussion with my son about the body design, and explanations about power and gearing, and him again insisting that there was nothing more he could do… and then an hour later coming up with a new design that was indeed more powerful (though much less elegant and a bit slow). Anyway, at least now he’s got the idea that he can redesign it further, and he’s working on a new design again.
In the mean time, I think I’ve already mentioned that we’ve previously played with a CamJam Edukit robot? Well, when we did that, the initial tests of the line follower worked, but once we’d mounted it on the robot it stopped working. I thought at the time that we’d damaged the hardware, but now we have some extra line detectors, so I decided to make sure… well turns out the hardware was working fine, but we’d mounted it too close to the castor wheel, which was causing it to see black all the time. Repositioned it and it now works fine. So that should be transferable knowledge for our Glitterator.