Interesting experiments this evening...

The other day I found these: http://www.aliexpress.com/item/Great-10pcs-Speed-motor-MABUCHI-Standard-130-motor-encoder-photoelectric-encoder-DC-motor-high-quality/1903700249.html

Actually, I was looking to see if prices had come down on the higher-quality motors with attached-encoders... I had no idea I'd basically find the wonky setup I've cobbled-together for extreme-worst-case-testing as a prebuilt item! OK, the higher-quality motors are running about $30 apiece, so that's not really an option. But these are only $1.50! They look to be darn-near identical to my setup, right down to the encoder-disk and quadrature detector which look pretty much identical to those found in old ball-mice. Kinda funny.

Actually, the motor I was using (described in detail in the last log) is a little bit better-suited for position-control than these, I think. My experience with these style of motors is that they're well-suited for high-speeds, but have very distinct snapping-action when rotated by hand.

So, I had to try it...

(Clarification: I did this experiment with a similar-looking motor I had on-hand, and the mouse's encoder/disk, *not* with the motors shown in the picture, which I do not yet own).

The current setup is such that the motor attempts to resist any external force; tries to maintain its position. Sure-nough I can actually drive 'em with the feedback loop! There exists that distinct snapping action, which means the motor seems to snap between -10 and +10 encoder "ticks" away from where it's intended to be, when external force is applied. (The setup is kinda hard to describe... I'm rotating the motor while the encoder stays stationary, and the disk tries to stay stationary with the encoder. The motor's shaft/disk "snaps" as the motor is rotated, about 1/10th of a turn). But, really, that's minor, especially considering these have worm-drive. (I wonder if I could laser-cut a matching gear for that... maybe right on the drum-pulley?).

And on to the crazy motor driver (L9947, described in a previous log, as well)

So, the motor I've got seems to be working well at about 1/4 to 1/3rd power... which kinda makes sense... alot of these motors are designed to be used *in toys* which are usually powered by a couple AA's... (3V... and I'm using a 12V supply)

And then something hit me... If these things work fine at 1/3 power... then the "crazy motor driver" might actually be usable to its full-extent... It's designed to handle *3* motors, just that only one can be powered at a time. So, if I do some time-division-multiplexing *of* the (essentially) PWM signal... Then maybe I could actually run all three axes "simultaneously" from this one (crazy) motor-driver chip that appears to have not been designed for PWM nor simultaneous motion in the first place!

There're a few caveats to be considered... the motor-driver is heating up quite a bit with this style motor, despite running at only 30% power... I hear they've got nice inductive-spiking abilities, so maybe it's not because they draw a lot of current, but that the motor-driver's diodes are heating up... or maybe because of slow switching-times on its mosfets (though that would've been a problem with the other motor as well, right?). Possible it's the diodes, anyhow... and kinda preferable, because the "crazy motor-driver" is rated for 3A on one motor, but only 0.5A for the other two motors.

Regardless of the motor-driver, these cheap/small motors with attached-encoders (and worm-gear) could be quite handy for this "tiny" system, even if they tend to snap to 36.5-degree increments.

Not sure what to do about "homing", though... which was part of the reason I was considering those multi-turn servos, but those, too, seem to run about $30 apiece, which is way out of my budget. Don't *really* want to install limit-switches/opto-sensors on each axis of this tiny thing, but it may have to happen.