Hi,
I'm designing a delta printer for high precision printing and I'm wondering if adding a planetary gearbox with ratio say 100:1 which decreases the stepper angle of a NEMA 17 motor of 1.8 degrees to 0.018 degrees will benefit the print quality dramatically? Please share some examples if you have any.
Would using a gearbox require reconfiguration of the printer firmware? (would like to use Repetier)
Would using a printed planetary gearbox be suitable?

example: [www.phidgets.com]

Thanks!

Current designs have a max speed well under 1000mm/s. This means that with a 100:1 reduction you will never break double digit speeds. If that is acceptable then you have to examine the introduced backlash from the gear box. I suspect it will be on the same order as the resolution for current setups.

What do you mean when you say high precision? Current designs can't get much better without having finer more accurate control over the rate of plastic deposition. Filament diameter variance causes 5 micron errors in extrusion width. What I am saying is positioning resolution is already pretty good. There are reasons we don't really care to get better than 20 microns.

I for one would like to break that barrier but we need some advances in the deposition control department.

Hi Minlin

I understand stepper motors can be microstepped. Not an expert you understand. This would appear a more promissing aproach in that you would not experience any slowing or gearbox backlash. One thing I have read is that there is no holding torque on a microstep. I asume therefore that it would be better to hold on a full step while operating other axes. This might complicate the firmware. You could simply use a motor with a smaller stepping angle, 1.8 degrees is nothing special they are available with much smaller steps.

Rory

@Rory166: Time for bad analogies. Imagine a circle of 200 fence posts. Now tie an energetic 6-year-old boy to one of the fence posts with a bungee strap. This is not unlike a stepper. The boy has some wiggle room but can't get to the next post. When you go to microstepping you don't change anything about the setup except add more posts. If there are 3200 posts, it still takes the same amount of force to break the bungee and the boy still has the same amount of wiggle room. The difference is that the wiggle room is bigger than the distance between the posts when you add the extra posts. (The force to get to the next microstep is less but the force to break the bungee is a constant.)

My point is that there is no need for the complicated control system. The holding torque between a full step is just as good. I think what you are referring to is that in microstepping mode you can't actually hold the stepper to that particular microstep. That goes whether you are on the native full step or not. The wiggle room is the same but the distance to the next microstep is just smaller.

Additionally, I have done the calculations for a few designs and the 0.9 degree steppers never actually helped. In fact they are usually more expensive for little to no gain.

I guess what the real question here is what are the needed specifications.

Additionally, print quality and resolution have surprisingly little to do with each other. (Give me the crappiest printer and with the right tweaks it will make beautiful prints.) Of course you have to have a certain level of resolution but beyond a point you have to address other things such as flex in the system, print settings, cooling, etc.

But then again, you have to keep in mind that the bungee boy has a lot less room to dance around when you use 0.9 degree steppers versus 1.8 degree steppers. A big problem for many printers is ripples after corners, and having finer resolution steppers helps reduce the ripples. The stepper motor's axle will still resonate a bit after sudden changes of speed, but the amplitude will be smaller with finer steppers.

Here is a graph about the problem: [www.linengineering.com]

Edited 3 time(s). Last edit at 12/29/2013 07:12PM by jkoljo.

My point is that in most setups a 1/32 microstepping 1.8 degree steppers or 1/16 0.9 degree steppers are almost the same. Granted the 0.9 degree stepper will be marginally better (IMO, better in ways most people won't notice.) and will be better still if you bump up to 1/32 microstepping. However, to really see improvement you will probably have to sacrifice speed.

A semiheated debate on the matter.

IMO you would be better off designing a motion system with less backlash before worrying about stepper resolution.

---

WWW.ZATOPA.COM - Your Place for high quality 3D Printing Filament and accessories

Steppers are only "truly" accurate to half stepping. So the stator will be able to resonate between half steps, but not cross them, that would be skipped steps. That is why it makes sense to have 0.9 degree steppers, because as I said, the amplitude of the resonance will be lower. You cannot really make a comparison of accuracy between 1/16 and 1/32 ustep, because microstepping does not bring accuracy, only smoothness of motion.

The resonance of stator is backslash by itself.

The lin engineering graph is quite artificial, they are doing doing a half step in a half second, not a realistic speed. Obviously they are trying to sell their dampers, so have picked a something that looks bad on paper even if it is not typical.

The Lin dampers are presumably mechanical dampers, microstepping is effectively the same thing but done electronically by the driver.

Obviously, if microstepping increases smoothness, then by definition it is reducing resonance - resonance is lack of smoothness.

It would be interesting to see some realistic tests done with typical speeds and microstepping, then you could perhaps see if there is a any difference.

---

What is Open Source?
What is Open Source Hardware?
Open Source in a nutshell: the Four Freedoms
CC BY-NC is not an Open Source license

... I've tested some 3-phase-steppers from Precitec (NEMA-23 types with 2Nm torque) first set to 1000 steps per turn in fullstep-mode -- and then with 10x microstepping.

They were pretty accurate in both modes with the test load below 1Nm and I could push the max. stepping frequencies up to 20kHz in fullstep- and slightly above 200kHz in microstep-mode until they start stalling.

With higher loads in the range of 2Nm on the motor axis this could change, but they are 5:1 geared, so enough power for my needs even in the 'safe range'

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Quote
jkoljo
Steppers are only "truly" accurate to half stepping. So the stator will be able to resonate between half steps, but not cross them, that would be skipped steps. That is why it makes sense to have 0.9 degree steppers, because as I said, the amplitude of the resonance will be lower. You cannot really make a comparison of accuracy between 1/16 and 1/32 ustep, because microstepping does not bring accuracy, only smoothness of motion.

The resonance of stator is backslash by itself.

@jkoljo: Everything you said is technically true but the meta message that 0.9 degree steppers will give people an automatic improvement over 1.8 degree steppers is a bit misleading.

To be clear, a situation that ends in a resonate response with a 1/2 step amplitude is rare. Moreover, you should design so that you avoid those situations. To hit this resonance you have to have an undamped system that is taking a sharp corner with a speed and jerk setting that is just low enough to prevent a missed step. To avoid this situation you have to decrease your speed, lower the jerk setting, and/or dampen the system. However, all other things equal, 1.8 degree steppers would be easier to get into this situation and would have about twice the amplitude.

Also, microstepping can help with your positional accuracy to some degree. I agree that you can't just keep getting improvements by making the microsteps smaller and smaller. You need to do the math for your setup to see how good or bad it is.

Thanks for the insights. Looks like there are no silver bullets for higher quality prints. I'll just have to look at properly tuning the printer settings first before tweaking it for SHQ prints.

@nicholas.seward Yeah, good point, it does not automatically increase print quality. I have noticed significantly lower rippling with 0.9 degree steppers, though. That is why I consider it a good upgrade, especially when one is buying steppers, and has an option to go with 1.8 or 0.9 degree steppers

By the way, I think that the steppers don't need to even actually resonate, stopping the stepper at high speeds seems to cause some oscillation in the axis anyway. With 0.9 the oscillation is smaller.. Which makes sense, considering the finer step size. I would love to hear experiences from other 0.9 reprappers.

Hmm, I have seen this rippling effect in prints, I didn't think this was the cause.

But, looking at this a different way: at 80 microsteps per mm (@16 microsteps) an error of +/- 8 usteps (half a full step) is 0.1mm. That would surely be visible.

Although microstepping smooths motion, it is still open loop positioning. If an axis comes to a dead stop, which often happens as we print rectangular objects, then that axis will tend to wobble more or less, depending on the deceleration curve. Obviously to get higher speeds, we dial up the acceleration.

I guess that the belts have some elasticity as well, which would make the effect worse.

So I think I am changing my mind on this...

---

What is Open Source?
What is Open Source Hardware?
Open Source in a nutshell: the Four Freedoms
CC BY-NC is not an Open Source license

Guys,
When I look at all these round rod printers with rectangular frames my brain just screams "ITS NOT STIFF!!!" I wouldn't waste any more time trying to improve the steppers until you eliminate backlash and stiffen the frames. Those motions which you can hardly see are likely to be huge compared to a microstep. Get yourself a cheap dial indicator and set it up against your frame or the rods while printing and see how much deflection you are getting there. I think you'll be amazed.

Well, the point is that our printers' steppers are not accurate to a microstep, only to half a step to be pessimistic. For example MendelMax produces the same ripples as other designs, and I would consider it quite rigid. I still think that the ripples are caused by magnetic backslash.

Edited 1 time(s). Last edit at 01/05/2014 05:38AM by jkoljo.

I totally agree that stiffness and stepper accuracy has low priority for most builders.
My first printer is not finished yet but I have spent many years im developing test
equipment with lots of moving parts including stepper motors.

Hard data on accuracy are hard to find, probably because the truth doesn't look too good in data sheets.
This is actual test data: http://www.euclidres.com/apps/stepper_motor/stepper.html
The braincell has vague memories of even worse figures, errors of several percent for the nominal step size.
The errors change for each step too.

Some steppers don't respond to microstepping in a nice way, current vs. angular position is not
neccessarily perfectly linear. I think it's the shape of the many poles that has to be really accurate and
designed to give good microstepping.

Three thin vertical pillars with no cross-bracing will move a lot for torsional loads to the top plate/structure.
My own first attempt to build a printer will use THK linear actuators in what will look lika a portal mill.
Overkill perhaps but I won't have to worry about unwanted movements.

Hi,
if you are looking for an extruder with a great reduction ratio, I'm putting up with this.

[www.youtube.com]

---

Enrico

[www.dapa3dservice.it] [www.studiodapa.it]
repstrap dApA, Ramps 1.4, MarlinKimbra / scanner 3D Cubify Sense / Formlabs Form 2