Decimus has been running reasonably well and each competition sees the plucky little micromouse improving in some way or another. It is clear though that it is not going to be fast enough to compete in South East Asia with any great chance of success. There are at least two faults on it. One is a damaged motor and another, I think, is a dry-joint somewhere on the PCB. Every so often, it will just do something unexpected like turn through a completely wrong angle. When I made Decimus, I ordered extra boards. This week the new motors came so, with only a short time to the all-Japan contest, I am trying to put another mouse together….
The new motors are from EMS, the UK agents for Faulhaber. In a previous post, I had decided to use 1331 motors as these seemed to give the best combination of performance, mass and physical arrangement on the mouse. However, a bit more thought led me to consider the possibility of building a half-size mouse. The 1331 motor would be far too long for that so I have gone back to my other choice, the Faulhaber 1717T006SR with a 512 line encoder. These are wonderful little motors and should be small enough for half-size mice as well as full-size.
The motors are coupled to the wheels with a 12:48 gearbox using 0.5module gears. Unfortunately, this puts the motor farther from the mouse centre than I would like so the next version will have to use 0.3 or 0.4 module gears. There might also be some advantage in mounting the motor further inboard than its current position. That is, closer to the longitudinal centre line. This will further reduce the rotational inertia of the mouse but leaves battery placement a little awkward.
I will also slightly change the sensor geometry – something I have wanted to do for a while. The 45 degree sensors do not need to point that far ahead and the 90 degree sensors could usefully point forwards by about 10 degrees or so to allow detection of the posts a bit earlier without sacrificing too much in the event of a lateral error. Not having the sensors perpendicular to a wall is also a good thing to reduce the effect of having shiny walls or shiny patches on the walls.
There are lots of improvements to make for the next model but, for now, this version of Decimus may perform better than the original. If not, at least it will give me the opportunity to try different strategies with each mouse. The Japanese competition allows contestants to run more than one mouse in the heats but only one in the finals. This is a good idea since otherwise, there is nothing stopping a contestant from entering a whole stable of mice and, potentially, taking all the prizes. an alternative would be to allow several mice per builder but only count the highest-placed mouse in the prizes.
You can find out more about EMS here: