How much microstepping still makes sense?

Hi,

intro: my homemade CoreXY machine was doing well until I messed up the RAMPS when converting to 24V. Bummer. But also the trigger for an upgrade. So I got an SBASE board to run Smoothieware and I got two 400 step motors to replace the 200 step ones on the X and Y. I have 20T pulleys on the old motors but I also have some 16T ones, both standard GT2 stuf. I ran the RAMPS at 16 microsteps/step. So I did 80 steps per mm ((200 steps x 16 microsteps)/ 40mm). On the new setup I can go to 32 microsteps, 16T and 400 steps per revolution. Hence, 400 steps per mm.

Question; does 400 steps per mm still make sense? And how to I determine whether the MKS SBASE V1.3 can actually handle that at 200mm/s travelspeed (80.000 steps per second...)? The board has a 100M Cortex-M3 MCU-LPC 1768 processor and DRV8825 drivers.

Cheers,

Hugo

Edited 2 time(s). Last edit at 11/12/2017 09:08AM by HugoW.

It depends on how fast you want to print. If the MKS board, like the original smoothieboard, is limited to 100k steps/sec, you'll be limited to 250 mm/sec, which should be fast enough. But we're talking about extruding plastic out a nozzle, which shrinks and swells as it cools. I doubt you'd see any difference between an 80 step/mm print and a 400 step/mm print. Variations in filament diameter alone are going to limit precision and accuracy. Microstepping beyond 16:1 is useful for making motors quieter, not increasing resolution.

DRV8825s are know to have a problem at low speeds. See: [www.engineerination.com]#!

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Ah, thanks, good info on the 8825s, thanks. I am not really looking for better resolution, more quiet would be very nice. I also replaced my power source for that reason, it was really nice with an internal cooling system, but very noisy. I got rid of most fans on the machine by overdimensioning and adding heat sinks, it now only has a fan on the hot end and it will get one or two on the MKS board. I might even get away with no fan on the hotend since my print head as a whole is an alloy heat sink. But that's besides the point...

I'll just give it a shot, I print at a max of 80mm/s and travel at a max of 150 now, maybe 200 in the future if the board is up to it.

Cheers,

Hugo

even 1/16 microstepping isn't any good for increasing resolution. Look at the motor's spec sheet, some motors can be out by as much as a full step at any given time so expecting accuracy to 1/16 of a step is ridiculous, let alone under static or (worse) dynamic load. There was a HaD write up where someone did some empirical testing on this a while ago. Microstepping is really only good for reducing noise and vibration, not positioning accuracy. I wouldn't rely on anything more than halfstepping for accuracy at best.

[hackaday.com]

Quote
Trakyan
even 1/16 microstepping isn't any good for increasing resolution. Look at the motor's spec sheet, some motors can be out by as much as a full step at any given time so expecting accuracy to 1/16 of a step is ridiculous, let alone under static or (worse) dynamic load. There was a HaD write up where someone did some empirical testing on this a while ago. Microstepping is really only good for reducing noise and vibration, not positioning accuracy. I wouldn't rely on anything more than halfstepping for accuracy at best.

[hackaday.com]

Wholeheartedly agree. Typical accuracy of a stepper is 5% although it's non accumulative, so yes up to 1 step for any given position. Micro stepping simply divides full steps into small segments but does not improve the positional accuracy of the full step. Having said that, a 16 tooth pulley and 2mm pitch belt would mean 1 revolution is 32mm and 200 full steps (for a 1.8 degree motor) so 1 full step is 0.16mm. So positional accuracy would be +_ 0.16mm. Using 0.9 degree motors would halve that but they tend to have less torque. 2:1 gearing might be a better option as long as there is zero backlash. But given that our printers are basically robot controlled hot melt glue guns, if we want better accuracy, we probably shouldn't be looking at FDM.

Quote
Trakyan
even 1/16 microstepping isn't any good for increasing resolution. Look at the motor's spec sheet, some motors can be out by as much as a full step at any given time so expecting accuracy to 1/16 of a step is ridiculous, let alone under static or (worse) dynamic load. There was a HaD write up where someone did some empirical testing on this a while ago. Microstepping is really only good for reducing noise and vibration, not positioning accuracy. I wouldn't rely on anything more than halfstepping for accuracy at best.

[hackaday.com]

As you point out, using 16x microstepping does not increase accuracy, because of the limited accuracy of the full step positions. But it does increase resolution on many 3D printers, when the friction is low enough and the motor torque is high enough that each 1/16 microstep results in movement.

The main benefit of microstepping higher than 16x is reduction in the amount of noise and vibration generated by the motors and drive mechanics.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

In theory they do increase resolution, but in practice i don't think youll see the benefits since stepper motors arent accurate at that level of microsteping. They do considerably quieten motors though, like you said.

Dont mean to hijack the thread but I am planing to build a new printer in the near future targeting high accuracy/repeatability. The 5% error was mentioned above. Are there higher quality (higher $$) steppers that have less error? I am printing with a very small nozzle orifice (.08-.1mm) and will probably need the best accuracy possible within reason.

The specification of stepper motors I have seen quoted is that the positional accuracy in the full step positions is 5% of one full step, or +/- 0.09degrees. That's not too bad. For greater positional accuracy and higher torque stiffness, use a 0.9deg motor.

Edited 1 time(s). Last edit at 11/18/2017 03:03AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

If you have deep pockets, look up 5 phase steppers. They're typically 500 or 1000 steps per rev before applying any microstepping. The motors and drivers are usually sold as a set for $$$$. see [www.orientalmotor.com]

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]