Re: [MakerGear] Min layer height for M2?

[Ketil Froyn]

From http://makergear.wikidot.com/m2-firmware:

#define DEFAULT_AXIS_STEPS_PER_UNIT {88.88/2,88.88/2,400/2,471.5/2}

So Z steps/mm is 400/2 = 200. That means one step on the motor changes
the height by 0.005mm.

I think the electronics does 16x microstepping, which I understand is
the maximum. But take that with a grain of salt, I'm way out of my
league discussing stepper motors and microstepping. This is based on
hearsay and assumptions. :)

Ketil

On Thu, Jan 3, 2013 at 5:47 PM, ddurant <ddur...@gmail.com> wrote:
> Somewhere else on the interwebs, I was babbling about min layer heights for
> some reason.. I didn't know what the theoretical max Z resolution on an M2
> (or Mosaic) was, though. A quick look through makergear.com didn't turn up
> an answer either.
>
> I know it's at least good for 0.02mm / 20 microns but what's the real max?
> Single step at the highest microstepping supported by the electronics..
>
> Yes, I understand that printing at the limits of the hardware is just silly.
> I'm curious anyway..
>
> Anybody?



--
-Ketil

[Joshua Wills]

Ketil's logic is sound, though our current revision of RAMBo only supports 8x microstepping.  At 8x microstepping, the minimum movement is one microstep, for 1/200mm=0.005mm, as Ketil said.  If you had access to 16x microstepping, that would be down to 1/400=0.0025mm.  That would also double the already ludicrous print time at 0.005mm. (At 0.02mm, Yoda takes 18 hours.  0.01mm would double that.  0.005mm would double that.  And 0.0025 would double _that_, for 144 hours...)

So to sum up - 0.005mm is the smallest possible Z movement on M2.  And yes, that is just silly X)
--
Josh
MakerGear

[ddurant]

Sheesh.. That's twice recently I could have figured something out just by looking at a config file.. Should have known that. Thanks!

 

Yup, Josh - it's definitely not a useful thing to do. I've printed around there, just to see if I could, and it's truely ridiculous.

 

My question came up via a discussion with somebody new to printing who thought layer height was tied to nozzle size - not an uncommon mistake for new people. He was surprised that machines could do better than "100 microns." I rattled off some max-specs I knew but could only say "at least down to 20 microns" for the M2. I just wanted to know what the real limit was.

[Tim]

Yup, Josh - it's definitely not a useful thing to do. I've printed around there, just to see if I could, and it's truely ridiculous.

Depends on your target application.  Five microns (or better yet, 2.5) is getting into the range of semiconductor electronics.  If you can dual-extrude a conductor and an insulator. . .  

[Triffid Hunter]

On Sat, Jan 5, 2013 at 4:38 AM, Brad Hill <uns...@gmail.com> wrote:

I seem to recall hearing that microstepping is mainly used to make motion smoother but that it doesn't really help with precision as much. Could be wrong though.

Precision is a measure of the size of a discrete measurement unit- a ruler with millimeter markings is more precise than one that only has centimeter markings. Accuracy is a measure of how close to the actual real-world value a measurement is. They are distinct, separate things, and we want both of them :)

See http://www.paduiblog.com/uploads/image/Harrisburg%20DUI%20Lawyer%20accurate%20and%20precise.gif for a visual explanation.


Microstepping increases precision far more than accuracy. The full step positions are fairly accurate, but other microstep positions are far less so- expect the motor shaft to be up to a half step away from where it should be if the load is high. This means that the accuracy goes up and down as the motor moves from one full step position to the next. When lightly loaded it should be quite accurate, but as the load increases, torque variations may occur and positional accuracy fluctuates more between one full step position and the next.

A stepper motor generates torque when the rotor is at a different position than the magnetic field - as the rotor moves into the commanded position, torque drops to 0. If the rotor moves beyond that position, the torque ramps up in the opposite direction, to pull it back in. This effect creates a mass/spring situation, and might be responsible for artefacts like https://picasaweb.google.com/lh/photo/vSqoSwRkacG4SiYpot10l9MTjNZETYmyPJy0liipFm0 although they're also likely to be related to the belt.

These effects occur with all stepper motors, however when they're connected to a leadscrew ala subtractive CNC, the effects are greatly reduced because half-step offsets in rotor position translate to significantly smaller offsets of the carriages. Belt drive gives a significantly larger linear motion per rotation so the effects are more pronounced at the carriage.

The primary reason we use microstepping is to reduce noise and eliminate mid-band resonance, both of which it is spectacularly effective at.

At 1/8 microstep we enter the region of diminishing returns. 1/16th is worth it for us, but don't think it's twice as good as 1/8th- it's not. 1/32 is only slightly better than 1/16 and I do not believe it is worth using for our purposes given that it stresses our firmware's ability to generate step pulses in a timely fashion.

[Tim]

How does microstepping eliminate noise?  Do you use it like a dither source?  My job is in electronics/signal processing, not mechanics, but there are many similaries between the two.