The second thought:  That first thought was not very smart.

Nevertheless, we didn't get past the first thought, so we built our own speed controllers.  We decided to build the Open Source Motor Controller, rather than just buy speed controllers.  This was quite an interesting exercise.  We'd assumed that the authors had actually built and tested them before, but it certainly appears not -- if you decide to give it a try, be warned that the voltage regulation section of the schematic and board layout is completely messed up.  And so out came the first magic smoke of the project.

"Simple and elegant" is not our style.  Therefore, Daisy's control electronics are overpowered and complex enough to make Rube Goldberg proud.

At its core is a processor board (stolen from Bing) containing a 40 MHz PowerPC microcontroller (Motorola MPC555).  This runs a control program which decodes the incoming radio receiver channels and sends out PWM signals to the motors.  That board is in the rightmost box in the picture.

The leftmost box  (connected with a fistful of ribbon cables) contains a custom interface board which does further processing on the PWM signals, routes power to the various components, provides optical isolation between the cpu and the motor controllers, and (as detailed in a future build report) will provide current limiting functionality for the weapon motor.

So here is our second potential point of failure:  that something will go wrong with the dozens of connectors or hundreds of solder points, or the controller board itself, or the rather complex software, or something else.

However, this setup does give a great deal of flexibility for turning the radio signals into bot control, and eventually we want to add semi-autonomous behaviors and even completely autonomous operation, so we're all set for that.

We love it when this happens:  Instead of spending $10 for something, spend $2 plus many hours of labor.  Economical?  Eh.

We saw some wheels on American Science and Surplus's web site and they were way cheaper than those Colson things everybody seems to use, so what the heck?  Get some of those!

Of course, they are kind of flimsy, have no convenient way to mount them, and get terrible traction to boot!

To fix the first two problems, we filled in the hollow spaces in the wheels with casting resin (results shown in the picture at left), which made the wheels much sturdier and gave a nice solid space to attach wheel mounts (exact method yet to be decided).

The traction problem will be addressed by attaching nice grippy vacuum cleaner belts to the outside of the wheels.

The funnest part of an exercise like this is smiling enigmatically at the hardware store cashier while buying six different kinds of vacuum cleaner belts.

Using motors from cheap drills off an auction site takes care of the gearing problems (we'll be running them at their rated 18v instead of overvolting them because we're not that confident in the gearbox), so there are really only three issues left:

1) Attaching an axle.  The drills have threaded rods coming out of them so we figured that we would just drill a hole in the center of a shaft and thread that and voila.  We don't have a lathe, though, and six off center and crookedly-tapped shafts later, we decided to just grind the end of the motor shafts to flatten them and then stick them into a rectangular hole in the ends of the rods.  We do have a mill, so that seems to have worked acceptably.

2) Mounting the motors.  The picture at left shows most of the parts for this -- we have two pillow blocks to support the weight of the bot and a custom-made mount thing to grip the motor itself.  This mount was cut from 1/2 inch delrin and fits perfectly.  Note the section of aluminum angle stock which will attach the mount to Daisy's base.

3) Attaching the wheels.  Stay tuned.