This was a long-winded response to a comment/question by @alpha_ninja regarding the precision attainable with sled-mechanisms driven by DC motors, rather'n steppers. It seems rather irrelevent, now that I've discovered that most DVD (vs CD) drives use steppers... but I'm still pro making use of those mechanisms for something...
This is based entirely on my own observations, and could be completely flakey... frankstripod's link looks to have some good general guidelines.
The HaD blog link (http://hackaday.com/2014/05/31/poor-mans-3d-printer-looks-rough-prints-great/) uses drives with steppers and avoids motors from older drives.
These notes, below, are also regarding actuators based on
DC-motors (vs steppers).
*Some* drives in my collection have various position-detection schemes (some have none), apparently generally attached to the motor-itself. One has hall-effect sensors that appear to be in quadrature. Another has an optical encoder wheel (but no opto-electronics were installed!) with ~50 slots. Most don't seem to have any position-detection on the linear-actuator, so I assume they must use positional-feedback based on the CD/DVD's "tracks". Their read-heads generally have a voice-coil that can move radially (in the direction of the linear-actuator) at least a couple mm in each direction, which suggests the linear-actuator can be expected to get to a position within *at most* a few mm. That's pretty low for most 3D/2D systems we might be interested in (laser-etching, *cough* 3D-printing *cough*). Add to that: The motors are usually gear-reduced *dramatically*... some seem to have a series of gears, others seem to have worm-drives. The former (gear-reduced) are likely to have quite a bit of "slop"/backlash. OTOH, they also often seem to take that into account by using measures such as spring-loaded double/stacked gears... With so much gear-reduction, even low-precision position-*detection* *at the motor* c/would result in pretty high position-accuracy (vs precision) at the actuator.
As for *precision*... that remains to be seen... *early* experiments with a *really low [positioning] quality* DC-motor shows it's definitely possible to get positional-precision control of 1/6th of a rotation. Gear-reduce that as much as these actuators do, and it might be plausible to expect fractional-mm precision/accuracy. Add to that the fact the motors used in these actuators are quite a bit more positioning-friendly (magnetic poles at 1/6th turn?! BAH!)... I think it's possible. *I* (whoever I am) think it's possible to squeeze quite a bit of precision out of these things.
We're talking (DC) motors likely rated for something like 3V... Lets control 'em with 12... Those motors with strong magnetic poles(/pulls) at every 1/6th turn... those poles could likely be compensated-for by pumping up that voltage AT those poles. (see: early-early experiments at: https://hackaday.io/project/6892/log/24550-pwm-vs-pulse-density-modulation-power-supply) treat those poles a bit like a stepper-motor in microstepping-mode. It'd be hard on the motors and hard on the motor-drivers, but likely doable with today's [non-bulk-cost-optimized] technology.
(Note that manufacturers try to shave *every penny* in such things made in the millions, whereas nowadays the technologies available to the general hobbiest are generally less-optimized to the specific purpose, and therefore often of higher-general-quality to appeal to a wide-range of projects, while still remaining within a "hobbiest budget." E.G. manufacturer plans 5KHz PWM to drive a motor: uses H-bridges that have switching-speeds no higher than 10KHz... Hobbiest wants "PWM" [motor, speaker, LED, switching-power-supply, who knows] expects it to "just work like <so-n-so> did" gets "H-bridges." They happen to handle switching-speeds of 50KHz+)
So, yes, hard on the motors. Yes, hard on the H-bridges... in the sense that it might be smart to add a heatsink where it wasn't required before, and probably won't last for 100,000 hours of usage (but what DVD drive does?)... But, more importantly, *doable* and probably not much more (if at all) expensive than the over-spec'd devices already in wide-hobbiest-use. E.G. Again. (I assume) microstepping is used on stepper-motors used in most 3D printers... Microstepping requires fast switching speeds. Implies Dual-H-Bridge-chips (rather'n individual MOSFETs). Well-suited for TWO DC-motors.
SO, precision: Surely doable... Might need some feedback systems not originally-installed. Could probably be acquired from other "junk" laying around (ball-mice encoders...? Maybe even optical-mice?)
And, let's also take into account *other* potentials: E.G. say we're making a simple 2D laser-etcher using an old DVD-burner laser... say we leave it mounted in the original head-mechanism... which... just happens to have a high-precision voice-coil positioner... our motors might only get 2mm accuracy, but that voice-coil could do the rest. A bit of feedback... maybe hacking into the pick-up and some interferometry (since our DVD doesn't exist to read the tracks for positioning)... Absurd, but plausible. Or, what if we mounted the etcher/burner-head face-down, and attached another head facing the other direction with... A DVD. Used for nothing but positioning of the head(s).
Maybe not *cough* 3D printing *cough* (or maybe?). But definitely some amount of amazing positioning precision + accuracy.BOOM.