Why does my print have tiny zebra stripes?

[neotko 1,417]

For what It understand the microsteps ain't done fully, so for a full step on 1/16 or even at 1/8 the stepper doesn't complete the turn completely and the remaining step to complete it it's what does the zebra skin.

Imagine that something like if a Quartz could stay on a 0.99 for each hz, and when it goes to 1.01 it does the full pass completely. Because the mechanical nature of the motor and the way the wave makes the stepper stay between each step with magnets. I really didn't explained myself...

This it's easier to understand from the website he linked

 

Later Pololus and some StepSticks use the A4988 chip. That has an interesting section in the datasheet: -

Low Current Microstepping. Intended for applications where the minimum on-time prevents the output current from regulating to the programmed current level at low current steps. To prevent this, the device can be set to operate in Mixed decay mode on both rising and falling portions of the current waveform. This feature is implemented by shorting the ROSC pin to ground. In this state, the off-time is internally set to 30 μs.

.

[gr5 2,224]

Hi @SandervG. Microstepping is very complicated even for an electrical engineer like myself. You have 2 sine waves going to the different halves of the stepper and they are phase shifted but for some reason I don't understand the amount of current going to each half is different and changes as the phase changes. Because the current/power is changing evidently some of the step positions are using less current - so low that it's not enough power to move the stepper that additional microstep. This is what I read but doesn't agree with other information I read about how it all works. Anyway then what happens is on the 17th step the power is back up and finally the stepper moves again. As a result only 14 or 15 of the 16 microsteps actually work and when you hit those "weak" steps you get a bump in the part (the zebra stripes).

Of course the bump is so tiny it's hard to see on a low quality printer.

But if you have almost (but not quite) vertical walls you can see the bumps as lines.

[DrR1pper 4]

An observation that may or may not be useful here, i've since seen in person benchy prints on both a zortrax and um2+ at different z angles on the print bed and i could not see any zebra stripes on the zortrax printed benchy.

Edited May 27, 2016 by Guest

[SandervG 1,521]

So depending on the inconsistency of the current it may lead to a missing step down the line?

Like if it does 0.8 instead of 1 for 5 times, you are missing a full step?

There is probably plenty of which I don't understand, but what struck me as weird (and maybe it is not weird at all so that is why I'm asking) is that he said it also happened on his home build 3D Printer. So that leads me to think it is a stepper motor problem? Or stepper driver?

But it can be fixed by whatever he added to his electronics, right?

So initially I thought hm, maybe we missed this small thing in our electronics design, but if it is in more printers, and can be 'relatively easy fixed', how come the error still exists?

Doesn't really make sense that is why I think it is more complicated then I just wrote down here.

[DrR1pper 4]

Wouldn't it be weird if both machines, from different manufacturers show the same occurrence?

If i understand what you are saying/asking Torgeir correctly, the stripes also appeared on my wanhao duplicator 5s mini and old prusa i3. The only prints i've seen that don't seem to have zebra stripes are from an m200. But this is just my limited experience with a handful of prints from different printers that i've checked for zebra stripes on.

Here are two prints i have from these printers with as close to the same settings and print time as possible. Am i right in assuming those are zebra stripes on the um2+ print?

UM2+

m200

Edited May 27, 2016 by Guest

[SandervG 1,521]

@DrR1pper, is this the same material for both?

How would you explain M200 doesn't have them?

[DrR1pper 4]

and the same part printed on a d5s mini for comparison (that i think i printed way too quickly, lol.)

[DrR1pper 4]

@DrR1pper, is this the same material for both?

Ah, yes. So sorry, forgot to state material. The um2+ was printed in PLA and m200 in ultrat (supposedly an abs blend with something else) i believe.

 

How would you explain M200 doesn't have them?

 

I originally thought it might be a difference in the motor choice because as i understand from the specs i can find online, the m200 has an x/y precision of 1.5 microns whereas every other 3d printers specs that i have looked at shows 12.5 microns (including the um2+). However, Torgeir's post might have just killed that theory.

[SandervG 1,521]

Well it could still make sense, and if it prints with smaller micro steps there could still be missing steps except so small you would barely see them?

And the choice in material could also influence the outcome. well.. it won't affect the motor of course but the level to which it shows in the surface might.

[DrR1pper 4]

Well it could still make sense, and if it prints with smaller micro steps there could still be missing steps except so small you would barely see them?

I think so too. Here are pictures of a vertically flat surface printed on both (top is m200, bottom is from um2+). Same materials as before and both vertical flat surfaces were printed along either the x or y axis only (i.e. not diagonally printed across the bed).

 

And the choice in material could also influence the outcome. well.. it won't affect the motor of course but the level to which it shows in the surface might.

 

I agree material choice can hide or amplify print surface details and sometimes to quite an extent but i don't think it plays a significant role in this instance with the artifact we are observing and comparing.

[DrR1pper 4]

I'm not sure if these are that helpful to you or not but some more close ups with different objects and differently lighting. Sorry for the shoddy camera work...best i can do.

The more matt looking ones are from the m200.

Edited May 27, 2016 by Guest

[Torgeir 259]

Hi SandervG,

Thanks. Well, by sharing this I'm trying to give a little back to the forum and Ultimaker for the veld of information there is open for everyone to see and use for their ideas.

Yes it was the same printer that printed the Bency-boat and the Tail of the A-380 that I've now modified. (With shafts from scrapped scanners... Well.)

I don’t think so, because they use same firmware, same slicer and same stepper driver.

Hmm, a really tricky question.

When this stepper driver is doing micro stepping, this stepping is made with very low current, typically around 0.1 Amp (or 100 mA).

When stepping in this mode, the feedback (EMF) is monitored by the stepper circuit in order to keep the current steady.

In some application where micro stepping at low speed is necessary, a lack of EMF pulses can cause the current to increase rapidly in the actual stepper winding, -resulting in missed steps cause it will jump two steps instead of one.

At such low current micro stepping, it can be the stepper motor, the impedance or to long feeders (increased resistance), or to low «stepper voltage» that is «creating» this problem.

A fix might be to move the stepper drivers closer to the stepper motors. Might work, but...

I just changed the mode of operation on all stepper drivers from; «automatically-selected mixed decay» to «mixed decay set to be active 100% of the time». This will increase the current, the noise and the ripple current (more EMC radiation).

Thanks.

Torgeir.

[DrR1pper 4]

Sorry to hijack the thread with this slightly off-topic question but does anyone know if it is possible to significantly improve the z-wobble type artifacts present on this um2+ print (printed with 0.15mm layer height)?

would be nice if you could get this level of layer upon layer uniformity/precision:

Thx

[Labern 775]

You might want to start a new topic as there is quite a few reasons and the discussion could be lengthy.

There is a lot of topic already on z banding so you might want to do a site search on Google.

[Torgeir 259]

Hi DR1pper,

I'll think this is caused by the same problem, the missing step.

Thanks

Torgeir.

[Torgeir 259]

Hi SandervG,

Just a note here, the most used driver for stepper motors is the one from Allegro A4988.

Also in the application data sheet, they advice this:

By pulling the ROSC pin to ground, mixed decay is set to be active 100% of the time, for both rising and falling currents, and prevents missed steps as shown in Figure 3. If this is not an issue, it is recommended that automatically-selected mixed decay be used, because it will produce reduced ripple currents.

 

-

This might be the reason, why there is so many printers make those line patterns.

Just to mention this.

Thanks.

Torgeir.

[valcrow 145]

Wouldn't it be weird if both machines, from different manufacturers show the same occurrence?

 

I have also tried this on a UP plus, Upbox, and the problem occurs on there as well. I also tried a UMO(non+) which has a different board and it happens there as well.

I made a few specific test pieces for this issue specifically https://www.youmagine.com/designs/curve-test-print

[DrR1pper 4]

Hi DR1pper,

I'll think this is caused by the same problem, the missing step.

Thanks

Torgeir.

 

...interesting.

[cloakfiend 996]

Dont worry @DrR1pper i get that as well. on all my prints regarless of layer height.

Edited May 28, 2016 by Guest

[DrR1pper 4]

Dont worry @DrR1pper i get that as well. on all my prints regarless of layer height.

 

Oh right, ok, thank you for letting me know. So this is the typical print surface quality on the um2+ and it can't be improved?

[Torgeir 259]

Hi DrR1pper,

Well, there is "no problems" -but solutions...

I can improve the resolution a lot... But the speed of printing may suffer.

The Ultimaker is “one of a kind” when we talk about “performance adaption”...

This is open source -and it will improve, very much.

How fast can your M200 print compared to a UM2+?

Thanks.

Best regards

Torgeir.

[Torgeir 259]

 

@DrR1pper, is this the same material for both?

Ah, yes. So sorry, forgot to state material. The um2+ was printed in PLA and m200 in ultrat (supposedly an abs blend with something else) i believe.

 

How would you explain M200 doesn't have them?

 

I originally thought it might be a difference in the motor choice because as i understand from the specs i can find online, the m200 has an x/y precision of 1.5 microns whereas every other 3d printers specs that i have looked at shows 12.5 microns (including the um2+). However, Torgeir's post might have just killed that theory.

 

Absolutely not, you're right here I'll think.

Thanks

Torgeir.

[DrR1pper 4]

...

How fast can your M200 print compared to a UM2+?

...

m200 has two only two speeds. "Normal" and "High", lol.

"High" speed prints don't look as good as "Normal" and since i am only interested in the best quality prints, i would only ever use "Normal".

With the same infill percentage, and infill pattern (m200 only does rectilinear) and the um2+ using s3d (default speed 41.7 mm/s, outline underspeed speed 65%, solid infill underspeed 70%, x/y axis movement speed 100mm/s), print times are comparable. um2+ was also set to 800mm/s^s acceleration and 10mm/s jerk. Not sure what values are most optimal for the um2+ yet.

I was suprised to find out that with the above settings, the m200 was printing the outer perimeter 50% faster than the um2+. Ofc the um2+ can be made to go quicker but need to run more tests first to find where what the speed/acceleration limit is where it starts to noticeable affect print quality.

btw, the motors on the m200 are quite impressive. They're so silent (can easily sleep with them running in the same room) and the motors are barely even warm to the touch during a print unlike the um2+ which are ridiculously hot to the touch. There is something very special about them. Shame the printer doesn't print pla though.

Edited May 29, 2016 by Guest

[neotko 1,417]

Slow acceleration doesn't mean better quality. At 800 (um2 default it's around 5000) you will have some disadvantages. I think you really should make a post about your problem to debug it since you adjusted too many parameters to just pinpoint the issue to one think like @labern said

[Torgeir 259]

Hi Folks,

Just made this last picture of two of my "3D Bechy Boat" print, the boat to the left is printed before I made the configuration change to the main PCB. (It's not a hack, just hardware configuration.) The boat on the right side is made after this change.

It is the same gcode file that is printed as close as possible to the front of the printer, some vibration here when the extruder accelerate/decelerate.

Anyway, as you can see -there is no zebra stripes anymore.

I also printed Valcrow's «curve-test-print», to see how this print will be. However, I never printed this before this change, so I have no idea what to expect.

https://www.youmagine.com/designs/curve-test-print

I made 3 picture of this print, so here they are for you to see.

Thats all.

Thank you.

Torgeir.

Edited May 29, 2016 by Guest