A new mouse is begun

By | July 22, 2009

Time to start serious thought about a new mouse. Decimus has been OK but has some serious limitations. Not least of which are a probable dry joint on the processor and a damaged motor. So what should the new mouse look like…

Not much different from the old one is the probable answer. I don’t have the means (financial or physical) to make a six-wheel micromouse much as I would like one. The new four-wheel mouse seen here:

http://blog.livedoor.jp/robolabo/archives/51284034.html

is very interesting but I can’t work out how to deal with the wheels scrubbing around corners. In the UK, we are all running full-size mice and the full-size competition looks like it will continue for a couple of years yet in Japan and elsewhere so there is no urgency to build small.

Thus it looks like a ‘standard’ two wheel mouse configuration but what can be done to improve on Decimus. Lots is the simple answer.

The most significant change will be the motors. Decimus uses Faulhaber 2224 motors. These are way over powered for the job and weigh in at about 46g each. Out of a total weight of 200g or so, the motors alone account for nearly half. After looking at the specification of many of the available small motors with built-in encoders, only the Faulhaber range have a good match. Of those, the favourite choices are the 1717 and 1331 size motors, rated at 6 Volts. Each of these has a mass of about 19g. That gives an immediate weight saving of 57g – about 29%.

Assuming some moderately high performance with acceleration near 0.5g, top speed of 3m/s and a mass of 150g, we get a peak power requirement of about P = mva = 0.15 x 5.00 x 3.00 = 2.25W or 1.125W per motor. Allowing a good margin of 2x, we get a requirement of about 2.25W per motor. The 1717 is rated at 1.96W and the 1331 is rated at 3.11W. Either will do although the 1331 is still somewhat over powered for the job.

The 1717 motor is short enough for the motors to be mounted coaxially, both on the same side of the wheels, allowing the battery to occupy the space the other side of the axle. However, this could leave the mouse unbalanced. Although there are some small advantages to having the centre of mass slightly forward of the axle when travelling in a straight line, there is a disadvantage when turning. On the whole, I think I prefer the centre of mass over the centre of rotation. That thought could change though.

There are two key numbers for the dynamic performance of the mouse – the rotational moment of inertia and the height of the centre of mass. Both should be kept small for best results. Solidworks can readily calculate these values if you have a suitable physical model. To compare the motor arrangements, I made a model consisting of just the baseplate and the two motors for a mouse. The motors were arranged symmetrically and placed so that, in each case, they had the outer face the same distance from the centre and at such a height as to leave 1mm clearance over the baseplate. I assumed that the rest of the mouse would not change with motor changes so those parts are missing from the measurement. Actually, the batteries will not fit in the same but let’s see how we get on. After choosing a material that had nearly the same overall density as a Faulhaber motor (Titanium), the three models were measured:

2224-test   2224 motors:
mass: 135g
Height of CG: 7.9mm
Moment of Inertia: 110,000 g/mm/mm
1717-test   1717 motors:
mass: 87g
Height of CG: 4.2mm
Moment of Inertia: 88,000 g/mm/mm
1331-test   1331 motors:
mass: 89g
Height of CG: 3.5mm
Moment of Inertia: 84,000 g/mm/mm
     

On that basis, the 1331 motors seem to be the best choice. Without going into the whole of the electrical analysis, they will consume less electricity and so run more efficiently than the 1717 motor and thus require smaller batteries, giving a further saving in weight. Since they are quite long, it won’t be possible to put the batteries in the same place as they would fit in Decimus. However, it looks like it will be possible to fit them upright, on their long edge between the motors. This will avoid raising the centre of mass too much and help keep the rotational inertia low as they will be directly over the centre of rotation.

All-in-all, it looks like the next mouse will be using Faulhaber 1331006SR motors. These can be had with suitable encoders attached  – either the IE-200 or IE-400 should do the trick.

8 thoughts on “A new mouse is begun

  1. Pranjal Chaubey

    Best wishes for the new mouse Pete!

    From where you are going to purchase Faulhaber 1331 or 1717 and the encoders?

  2. Pranjal Chaubey

    That 4 wheel mouse cant take turns without its wheels skidding……I gave a thought to building a differentially driven 4 wheeled mouse, my discussion in there in the forums. Finally decided against it because of the constant skidding on turns. The mouse in the video has wheels very close to each other to minimize skidding. The front sensor alignment looks interesting though…..

  3. Juing-Huei Su

    The young man who built the 4 wheel micromouse is also invited to join the 2009 Taiwan micromouse contest. His name is Yusuke Kato.

  4. Harjit Singh

    My calculations yielded similar results as yours. When I compared the 1331, 1717, 1524 motors indicated that a mouse with the 1331 would have a slightly lower CG:
    1331: 9.5mm
    1717: 10.0mm
    1524: 9.7mm

    Moment:
    1331: 137 kg mm^2
    1717: 143 kg mm^2
    1524: 140 kg mm^2

    Mass:
    1331: 130 gm
    1717: 134 gm
    1524: 132 gm

    Since the 1524 were in stock I went for those. Alas, this was eight months ago. I could have gone for the 1331 and had them by now but it appeared to be a wash.

    I’m sure other limitations will hold back the mouse more than these small differences.

    I expected your CG to be close to half the height of the height of the motor.

  5. Harjit Singh

    Oh, yes. The motor that the mouse turned out “best” with was a Portescap motor: 12G88-215E-MR2.

    Back in Dec. 08, they had just started producing them and didn’t have stock of the motor with attached encoder.

  6. peteh

    The figure shown was measured from the top of the PCB for comparison purposes only. A material for the PCB was chosen to give it about the same mass as the existing, populated board on Decimus. Either of these smaller motors would be a big improvement on the 2224s I am using now. While going to the trouble of getting new motors, I might as well pick the best I can. There are, indeed, many other factors limiting mouse performance. However, when all I am left with is the software, I will have a clearer idea who to blame 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.