The Bad News: Very, very small gearhead motors are not going to be particularly cheap. As the size goes down, the precision involved in making them goes up, as does the price. About the most affordable micro-motors I know of are available from Solarbotics, at almost $10 a pop. I should also point out that these also don’t come as continuous rotation, and need to be modified before they can be used to drive wheels or tiny little tracks.
The Good News: If by “the pico servos”, you mean the B-8 micros that you recently got for $9 each, it looks as though they can be modified for full rotation, provided you are willing to get a little bit “creative” in the conversion process, but only at the price of having to do a whole lot of modification and re-engineering, in order to end up with something that will probably only be marginally capable.
I just opened one up to have a look, and aside from the case screws that are literally smaller than straightpins, had few problems getting into it. The circuitboard is no larger than the 16-pin surface-mount chip soldered to one side of it, and is floating inside the case, held in place by the wires. As with most servo designs, two go to the motor, three to the feedback pot, and three lead out to the pigtail. Aside from the small size, access to the wires should be no problem, as the circuitry is not encased in anything at all. When I rewired my Lynx-6 arm’s servos, I was hindered by the fact that the signal wires are encased in a blob of epoxy as a strain relief, but this is not the case with these little servos - the only strain relief here is the slight pinch-point where the wire enters the case.
The Bad News Abut The Good News: The final gear and output shaft are press-fit onto the shaft of the feedback pot, which acts as the “inside” support point for the shaft. There’s another gear that shares this shaft, spinning freely on it. In order to bypass the feedback and make it a full-rotation unit, you’ll need to replace the pot with a resistor bridge or external pot, and cut down a couple of plastic stops in the body of the feedback pot so that the shaft can spin freely through 360 degrees. There will then be a question of the wiper contact having three “speed bumps” around the perimeter of the track, where the three wires connected to the pot’s resistive element. I suppose that it could be done, but it would involve extremely fiddly work on very small things - about 1mm or less.
Strange Gears: If that weren’t enough, the physical stops that keep the output shaft from rotating freely are made up of a combination of the case and the final output gear. While the gear itself has teeth running all the way around its circumference, half of it has teeth that run all the way up its height of about 4mm, while the other half has teeth that are only 1.5mm high. This allows that part of the gear to pass under the stop, while the taller teeth contact it and stop its rotation. If you were to trim off the rotation stop on the case (about 1 x 2 x 2mm, recessed into the case by about 8mm - you’ll need a fine burr on a Dremel tool, or a lot of fiddling with an x-acto knife), you’d still only get full-height teeth for half the output shaft’s rotation, and 1/3 height teeth for the other half, dramatically reducing the amount of abuse that they can take without stripping gears.
In Other Words: you could convert them to continuous rotation, but you probably wouldn’t want to.
Good luck finding small motors that fit your needs - you may well find that there’s a very good reason why many people stick with standard-sized servos and the like, only building super-mini versions of their robots after several larger proof-of-concept models: Small is usually expensive.
, I’ve got a set of photos of the dissection of one of those little B-8 micro servos. I can send the photo links to anyone who would like to check them out, or if there is interest, I can post them to a new thread in the General Electronics forum.