Here’s a good explanation of Stepper motor torque capacity drop-off and lag versus microstep resolution.
This came to mind when I noticed that a recent 3D printer on KickStarter was bragging up to 128 microsteps
http://www.micromo.com/microstepping-myths-and-realities.aspx
Very good article indeed.
If you have ever tried turning a stepper with your hand while it was powered up, you’ll probably have noticed that you can turn it a bit, and the driver-induced noise increased as the motor coils move from their ideal position. No matter if you’re in the middle of a microstep or in a full-step position, the relative amount you can turn the motor shaft will always be the same.
What would be a definite upgrade in positioning accuracy would be a shift to 400 step/rev motor - if you can find a type that produces the same torque as a corresponding 200 steps/rev one.
@Thomas_Sanladerer the 400 step/rev Nema 17s that Sparkfun sells are awfully close to the Kysans in strength, and an upgrade in accuracy. IIRC, whosawhatsis uses them.
Be careful increasing your microstepping and/or steps/rev. I calculated that my bukobot with 200 steps/rev motors and 1/32 drivers (112 steps/mm) maxes out at about 350mm/s. Marlin can’t pulse the drivers fast enough to go any faster, even though the machine is capable of moving faster. I do like those 400-step/rev motors, but I’d have to go back to 1/16 to use them on that machine. Looking forward to the move to faster processors.
With the DRV8825 drivers and 32x microstepping i, too, managed to get capped at far less speed than with 16x ms, since my bot has pretty small pulleys for maximum acceleration. A 400step/rev motor would only be feasible with 8x ms for me.
I’m playing with Pololu DRV8825s being driven from a Teensy 3.0. It’s running a 32 bit Cortex M4 processor at 96Mhz. Should be able to drive at step rates over 1 Mhz. So far the only 3D printer firmware port to it that I know of is Teacup, but I’m having tons of problems getting it to compile correctly.
long time ago when i first dabbled my brains into steppers (i think im still miles away from knowing enough btw), and that was back when reprap forum was all i knew of, nophead did an explaination which was iirc … kind of the same as the site you linked. but in my mind, to me those were really math stuff calculating what to expect on a physical motor. and when i did actually put my hands on them, the other real issue is to match the right motor with 2 major types of stepper drivers --> either the 49xx/8825. once you can get a stepper with pretty nice characteristics 1/16 is going to be a breeze. i even tried a 400 step motor, IT WAS awesome, but sacrifice it for speed. i think our real limitations are the drivers (vs the flood of motors). i believe someone also did a DIY L297 series driver, and that was fantastic too. i think somebody did try to adapt the L6xxx series SPI to “talk” to arduino, i cant rem more details. maybe somebody here can fill in those? i think the ST micro L-series drivers have much potential to be explored. (i found my old video of the 400step stepper test, just stick some kind of pointer on it and watch how consistent are the u-steps. http://www.youtube.com/watch?v=JXErUddVG2A … i know, kinda like being a little obsessed like camera freaks looking for the smallest perfect pixel rendition haha, there was a previous idea to stick lasers and plot them on wide walls just to be really sure they be totally equal spaced steps)
i think 1/8 u-step on close scrutiny, has the best repeatable … errm stepping. however if the same test is conducted with a straight move length and stop, and move back (as in A move to B, B move to A) the large movement will defeat the smaller “sticking” errors (and maybe individual u-step errors). what i think would be fascinating is, are the slicers or software compensated for this? as in purposely having to drive steppers in larger chunks of movement to overcome sticking? or drive them slightly faster/slower?
(on a sonic level, i really like what 1/32 does to the silence of the drives)