I built a cheap Delta printer with ATMega board and 1.8° motors. The micro stepping is limitied to 1/16. Beside a decent print quality, I observe a moire effect on flat surfaces. The moire is clearly caused by a combination of both, the 1.8° motors steps and the low microstepping.

I thought about using 0.9° stepper motors together with a combination of board and drivers which support < 1/32 micro stepping. Is there a comparison somewhere illustrating potential quality differences on larger delta printers and is this the way to remove the moire effect? For cartesian printers I would not bother using such motors, but I noticed that a higher holding torque at smaller steps is desirable for delta printers.


Not one of my prints, but this is how it looks enter image description here


Delta bots always need all motors to step to maintain a straight level. Microstepping, is not magic, the incremental torque decreases per step so that you will be more likely to miss a few micro-steps. Furthermore, the signal that creates voltages for the micro-step positioning is usually not perfectly sinusoidal (pulse-width voltage modulation is used to achieve micro-stepping by controlling the current; the driver sends two voltage sine waves, 90 degrees out of phase to the motor windings), micro-stepping drives can only approximate a true sine wave. This means that some torque ripples, resonance, and noise remains and hence resulting in odd stepper behavior, like seen below from this ref. (after the half step the stepper jumps to the full step and maintains that value for a while):

enter image description here

This is seen as a Moiré pattern in your printed products. As an example, if the head is moved in Z direction by micro-step, you will almost certainly notice that the head doesn't move on every micro-step, but only every 3rd or 4th micro-step (as an example). When using higher resolution steppers like the 0.9° stepper motors, you will still miss micro-steps (e.g. the same, so also on every 3rd or 4th micro-step the head moves), but as the micro-step is half the size of that one of a 1.8° stepper motor, the accuracy as in precision and resolution is higher.

In that sense, if you change your stepper drivers for higher micro-stepping drivers (from 1/16 to 1/32 as you mention), it will not help you improve the resolution much because the incremental torque from one to another 1/32 micro-step is lower than for 1/16 micro-steps as can be seen in the figure below (taken from this ref.).

Incremental stepper torque as function of micro-step size

So, using 0.9° motors (and keeping 1/16 micro-stepping) improves positioning accuracy as described above, it will also reduce the noise, because the torque per unit angular error is nearly doubled. Also remember that if you are using 8-bit electronics (you hint to an ATMega board), then even 1/32 micro-stepping burdens the processor to achieve reasonable travel speeds. With 8-bit electronics, it is usually suggested to use 1/16 stepping.

Upgrading an existing printer from 1.8° to 0.9° stepper motors is probably not worth for the majority of users (note that the maximum allowable speed also reduces when using 0.9° stepper motors). Unless you are designing and building a new delta, or aren't on a tight budget you could opt for the additional costs of buying 0.9° stepper motors.

Note that updating to higher micro-stepping values not necessarily implies that the quality of your products also increase. See e.g. this reference.

  • $\begingroup$ Exactly, my idea was to make a new printer with 32bit electronics and 0.9° steppers. Regarding the moire, I was not sure whether it is the result of missed micro-steps or low angular accuracy. Thanks for explanation. $\endgroup$ – dgrat Dec 4 '18 at 9:43
  • $\begingroup$ @dgrat Your question did not mention the use of a 32bit board. Still, would you go beyond 1/16 steps with such a board knowing that the incremental torque from 1/16 to 1/32 is halved? Upside is that you can use better stepper drivers that can handle more torque (Amps through the driver). $\endgroup$ – 0scar Dec 4 '18 at 10:12
  • $\begingroup$ High microstepping is always better. The motors run much smoother and quiter. The torque per microstep goes down, but the worst what can happen is that the motor jumps to the next full step. $\endgroup$ – dgrat Dec 4 '18 at 11:44
  • $\begingroup$ I tested to reduce the microstepping from 1/32 to 1/2 on a different printer. The motors became with each iteration louder and vibrated more and more. The print quality went down too. $\endgroup$ – dgrat Dec 4 '18 at 11:57

What you're seeing there is commonly referred to as "salmon skin" and isn't a result of the motor stepping, but due to power backfeeding from the motors. Install a set of flyback diodes on each axis (you can buy these premade specifically for printer motors, normally in sets of 8 diodes per motor) and you should find the issue either minimised or eliminated entirely. It's a very cheap upgrade and I'm surprised it's not a standard feature on printers that have affected drivers these days.


I am looking at the example print you supplied and trying to understand the defects. Was the object printed vertically, so that the object Z-axis is running from top to bottom in the picture?

If so, then the horizontal defects look like resonance artifacts, which could be caused by the issue you raise of the stepper motor not responding to small movement commands. The only thing that will prevent the motor from responding is stiction, or difficulty starting a movement. Stiction is a non-linear aspect of friction in which the "starting" friction is higher than the "sliding" friction. It might be worth checking for surfaces which must slide against each other, and assure they are lubricated with a dry, non-tacky lubricant.

A typical delta machine behaves the same way on each layer. The equations which govern how high on the tower the top of the delta-leg should be for each x,y, and z point is linear in Z, but non-linear in X and Y. With that in mind, I have no hypothesis to describe the defects that would be described as cathedrals if they were wood grain.

Or, perhaps the object was printed so that the surface shown is parallel to the bed? If so, the defects looks like more than a couple of microsteps. Is it possible that one drive is not working correctly? If so, I would suggest that the motor on the tower pointed to (or pointed from) by the cathedrals is not working like the others. It could be binding too tightly to the tower, or the pully may be loose, or the driver isn't working correctly for one of the two coils.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.