Notifications
Clear all

[Testers Needed] Inconsistent Extrusion  

Page 8 / 15
  RSS
raspeitia
(@raspeitia)
Eminent Member
Re: [Testers Needed] Inconsistent Extrusion

or since they are still there at higher power, if software related, can it be some sort of mis-timing in the extruder pwm?

is there a way to test the print output by varying the extruder pwm across a face like this?

Posted : 20/04/2018 5:47 pm
Kwaad2
(@kwaad2)
Honorable Member
Re: [Testers Needed] Inconsistent Extrusion


So doesn't this point to the extruder needing to be geared as others have mentioned?

Geared extruders do not have this issue. (at least it's not visible)
The 3.5:1 that I was running was flawless. (I went back to stock, to verify my issue was not caused by the gear reduction extruder. (it wasn't)

Will be rebuilding again tomorrow... uggg.

Hi, I'm Sean. I used to work on CNC machines.
I try to not make mistakes, but the decision is YOURS.
Please feel free to donate to my filament/maintance fund.

Posted : 20/04/2018 6:35 pm
matthew.m34
(@matthew-m34)
Eminent Member
Re: [Testers Needed] Inconsistent Extrusion


I think I have definitive proof of this issue. Attached is a single wall test object printed at 30mm/s

About every 5mm I increased the amperage to the motor from about 300mA up to full power at around 950mA. (By the settings I used, 20, 30, 40, 50, 63) EDIT: The object is about 30mm tall.

Anyway, I want you to notice that with the "lower" amperage lines (towards the bottom) they are "smoother" but they are still rather visible. However once I get the power cranked up to wide open, the lines look BINARY. In fact, it is SUPER easy to see, the exact "over" "under" extrusion. And it's not a "fade" like I was thinking that it would be. It's very much a "hard switch" between the two.

Let's zoom in a little bit.

Ok, now what we are looking at is the top left corner of the test object. (same photo)

I want you to notice how it looks "woven". You see that every other layer, is either "over" or "under" extrusion. (visible as a reflection, or a shadow)

using my layer lines as a measuring tool, I measured 1mm of vertical height, and then took that measurement to find that the "pattern" repeats every ~4mm.

My "paper math" said 3.5mm, which I assumed was high, and I called 3.0 to 3.3.

My guess is that every ~4mm of linear travel will yeild 1 "step". (a full 1/200th of a rotation step)

Something is happening at some point in the step to "increase" and "decrease" the power, and it directly related to the location of the step, within the step.

So under a microscope, there would be 16 steps on both "under" and "over".

I don't even know how to properly explain the problem.

But basically, there's I'm gonna say "two" modes for microsteps.
"hard" steps, that it steps "harder" and "easy" steps, that it steps easy.

The "Hard" and "easy" are configured wrong.


***

EDIT: You should rename this issue to "consistent extrusion issues". 😈

Do you think maybe it might be caused by different step modes on the TMC drivers? I know they have a couple different modes they can go into when stepping...

Also, for some reason, the pictures didn't show up for me.
EDIT: That's because of my network.

Posted : 20/04/2018 10:20 pm
Kwaad2
(@kwaad2)
Honorable Member
Re: [Testers Needed] Inconsistent Extrusion


Do you think maybe it might be caused by different step modes on the TMC drivers? I know they have a couple different modes they can go into when stepping...

Also, for some reason, the pictures didn't show up for me.
EDIT: That's because of my network.

Devilhunter said on GitHub that he thinks it's the "Driver Profiles". That's what I think it is too. (I just don't know the name of this stuff) 😐

Switching the modes does not "change the pattern" in the sense of it's "Frequency".

Switching to say, "stealth mode" does change the way it looks, but the "frequency" is the same.

This is the exact same theory as I had when changing the amperage from "low" to "very high".

The "Frequency" remained constant... just the "sharpness" of the differences, changed.

Basically: There are 32 "microsteps" per "full step". Due to the way microsteps work, some have more torque than others, so you have a program that will adjust the power on certain microsteps, to give them more, or less power, so they all have equal torque.
This is configured wrong.
And no-one I know of, except maybe metacollin knows how to configure this. (And I kinda doubt he knows, because if he did, he would have fixed it for us I would think, as he is suffering the same issue)

Hi, I'm Sean. I used to work on CNC machines.
I try to not make mistakes, but the decision is YOURS.
Please feel free to donate to my filament/maintance fund.

Posted : 20/04/2018 10:44 pm
michael.a35
(@michael-a35)
Trusted Member
Re: [Testers Needed] Inconsistent Extrusion

The thought really comes to mind though... what does Prusa do differently that others don't? TMC drivers work great on other printers.

Posted : 20/04/2018 11:16 pm
brad.h10
(@brad-h10)
Active Member
Re: [Testers Needed] Inconsistent Extrusion


The thought really comes to mind though... what does Prusa do differently that others don't? TMC drivers work great on other printers.

I'm not sure if this is worth looking into or not, but one major difference is the use of the Ultimachine Einsy. I believe the Einsy uses 220ohm sense resistors which would limit it's current compared to other controller / driver solutions.

Another thing is that the EinsyRetro a similar product at launch suffered from a bad 5V power source that caused erratic issues, my thermistor readings were unstable for example (particularly at voltages higher than 12v). The Retro has since been revised to include a different 5V supply and my problems have gone away. Is it possible the Einsy has a similar issue? Probably not, but who knows. Maybe someone can try vanilla Marlin with TMC2130's in a different controller on their MK3.

Posted : 20/04/2018 11:55 pm
matthew.m34
(@matthew-m34)
Eminent Member
Re: [Testers Needed] Inconsistent Extrusion



Do you think maybe it might be caused by different step modes on the TMC drivers? I know they have a couple different modes they can go into when stepping...

Also, for some reason, the pictures didn't show up for me.
EDIT: That's because of my network.

Devilhunter said on GitHub that he thinks it's the "Driver Profiles". That's what I think it is too. (I just don't know the name of this stuff) 😐

Switching the modes does not "change the pattern" in the sense of it's "Frequency".

Switching to say, "stealth mode" does change the way it looks, but the "frequency" is the same.

This is the exact same theory as I had when changing the amperage from "low" to "very high".

The "Frequency" remained constant... just the "sharpness" of the differences, changed.

Basically: There are 32 "microsteps" per "full step". Due to the way microsteps work, some have more torque than others, so you have a program that will adjust the power on certain microsteps, to give them more, or less power, so they all have equal torque.
This is configured wrong.
And no-one I know of, except maybe metacollin knows how to configure this. (And I kinda doubt he knows, because if he did, he would have fixed it for us I would think, as he is suffering the same issue)

Ohh! I get it! I know some of the microsteps would have less torque, so it makes sense that it would be compensated for - otherwise you would have inconsistent extrusion. But somewhere that's configured wrong and it's adding/subtracting too much, and that's causing the issues?
In that case, it would be software, right?

I have a RAMPS with 2130s I may end up trying sometime in the future to see if I can duplicate the issue or not.

Posted : 21/04/2018 2:37 am
michael.a35
(@michael-a35)
Trusted Member
Re: [Testers Needed] Inconsistent Extrusion

Has anyone reached out to the facebook Prusa owner groups? There might be some capable people there than could lend a hand. Not sure if they are aware of this issue.

Posted : 21/04/2018 3:44 am
Nikolai
(@nikolai)
Noble Member
Re: [Testers Needed] Inconsistent Extrusion


Has anyone reached out to the facebook Prusa owner groups? There might be some capable people there than could lend a hand. Not sure if they are aware of this issue.

Github is the right place and it was addressed: https://github.com/prusa3d/Prusa-Firmware/issues/602

Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram

Posted : 21/04/2018 3:50 am
digibluh
(@digibluh)
Reputable Member
Re: [Testers Needed] Inconsistent Extrusion

similar issues:

http://www.soliforum.com/topic/10108/extruder-pulse-moire-at-132-microstepping-causing-diagonal-lines/
http://www.soliforum.com/topic/4759/newly-installed-e3d-problems/
http://www.soliforum.com/topic/9347/solidoodle-press-print-quality-and-how-to-fix-it/

if it's not voltage or misconfigured software drivers then it's probably an enisy board issue/stepper motor mismatched for trinamic or they're not up to spec from factory and impedance wasn't checked in QA? so some are ok, others not.

Posted : 21/04/2018 3:50 am
michael.a35
(@michael-a35)
Trusted Member
Re: [Testers Needed] Inconsistent Extrusion


similar issues:

http://www.soliforum.com/topic/10108/extruder-pulse-moire-at-132-microstepping-causing-diagonal-lines/
http://www.soliforum.com/topic/4759/newly-installed-e3d-problems/
http://www.soliforum.com/topic/9347/solidoodle-press-print-quality-and-how-to-fix-it/

if it's not voltage or misconfigured software drivers then it's probably an enisy board issue/stepper motor mismatched for trinamic or they're not up to spec from factory and impedance wasn't checked in QA? so some are ok, others not.

Good work Eric. Very nice find. That first link is an especially interesting case.

The E3D specific stuff is also interesting. Will be following along closely with how this changes things.

Posted : 21/04/2018 5:12 am
raspeitia
(@raspeitia)
Eminent Member
Re: [Testers Needed] Inconsistent Extrusion

Very good find!

I'd be curious to know what option the prusa firmware chose to set the driver resolution and mode to.

Found an interesting article on the trinamic drivers on hackaday. Even shows a tuning flow chart.
https://hackaday.com/2016/09/30/3d-printering-trinamic-tmc2130-stepper-motor-drivers-shifting-the-gears/

The page also has a link to the Trinamic data sheet that shows a bunch of modes available starting on section 1.4
https://www.trinamic.com/fileadmin/assets/Products/ICs_Documents/TMC2130_datasheet.pdf

Posted : 21/04/2018 6:04 am
Kwaad2
(@kwaad2)
Honorable Member
Re: [Testers Needed] Inconsistent Extrusion


similar issues:

http://www.soliforum.com/topic/10108/extruder-pulse-moire-at-132-microstepping-causing-diagonal-lines/
http://www.soliforum.com/topic/4759/newly-installed-e3d-problems/
http://www.soliforum.com/topic/9347/solidoodle-press-print-quality-and-how-to-fix-it/

if it's not voltage or misconfigured software drivers then it's probably an enisy board issue/stepper motor mismatched for trinamic or they're not up to spec from factory and impedance wasn't checked in QA? so some are ok, others not.

I was not aware this issue pre-dated the Mk3 in such a way.

This is both wonderful, and horrible news.

Let me expain this.

The trim-pot, adjusts the amperage/current on the "cheap" style drivers. It is controlled via software on the Mk3. This is GOOD, as it means it should be easily fixable by Prusa, and not require "fine tuning by every mk3 owner in the world".

The bad news is... when is Prusa gonna fix this, becuase with it being software... Unless we get someone from Trinamic, or someone who knows their way around the Trinamic drivers... We just gotta wait for Prusa to fix it. (And I don't have a single doubt in my mind, they WILL get it fixed... someday)

So, let me make a hypothesis based on looking at those posts.

Rather than it being an issue with the microstepping, I think it has more to do with "over-current".
The Driver in the mk3 is "smart" in the sense that if it moves easily, it produces low torque, if it moves hard, it produces more torque.

If for whatever reason, it's producing TOO MUCH torque. (at ALL "current settings") That would explain this issue. I will browse the code looking for this tonight, but tomorrow, I'm "out" regarding testing again, as I'm going back to my geared extruder.

Hi, I'm Sean. I used to work on CNC machines.
I try to not make mistakes, but the decision is YOURS.
Please feel free to donate to my filament/maintance fund.

Posted : 21/04/2018 6:10 am
JLTX
 JLTX
(@jltx)
Reputable Member
Re: [Testers Needed] Inconsistent Extrusion


Very good find!

I'd be curious to know what option the prusa firmware chose to set the driver resolution and mode to.

Found an interesting article on the trinamic drivers on hackaday. Even shows a tuning flow chart.
https://hackaday.com/2016/09/30/3d-printering-trinamic-tmc2130-stepper-motor-drivers-shifting-the-gears/

The page also has a link to the Trinamic data sheet that shows a bunch of modes available starting on section 1.4
https://www.trinamic.com/fileadmin/assets/Products/ICs_Documents/TMC2130_datasheet.pdf

This is good info. I scanned through the data sheet and poked around in the FW and it is all laid out pretty clearly. Changing configuration registers to enable, disable, tune features looks straightforward. I don't have access to my printer this weekend, but would be willing to try.

Posted : 21/04/2018 4:10 pm
Brigandier
(@brigandier)
Reputable Member
Topic starter answered:
Re: [Testers Needed] Inconsistent Extrusion



Very good find!

I'd be curious to know what option the prusa firmware chose to set the driver resolution and mode to.

Found an interesting article on the trinamic drivers on hackaday. Even shows a tuning flow chart.
https://hackaday.com/2016/09/30/3d-printering-trinamic-tmc2130-stepper-motor-drivers-shifting-the-gears/

The page also has a link to the Trinamic data sheet that shows a bunch of modes available starting on section 1.4
https://www.trinamic.com/fileadmin/assets/Products/ICs_Documents/TMC2130_datasheet.pdf

This is good info. I scanned through the data sheet and poked around in the FW and it is all laid out pretty clearly. Changing configuration registers to enable, disable, tune features looks straightforward. I don't have access to my printer this weekend, but would be willing to try.

If you're looking at the TMC defines in Configuration_prusa.h, watch out. Lots of those are defined there but not used elsewhere in the firmware. Do a global search on the define you're editing to make sure actually changing anything. 🙂

Good example: INTPOL defines aren't used that I can tell. That feature is hard coded to 1 and has to be changed in tmc2130.h I believe.

I am betting they intended to tie the defines to things later on, but forgot.

My MK3 Parts: [Bowden] [New Shoes] [TPU Micro Springs]

Posted : 22/04/2018 3:58 pm
thrawn86
(@thrawn86)
Honorable Member
Re: [Testers Needed] Inconsistent Extrusion

argh, just when i thought i was making progress! my first layers with the RC are pretty good, but extrusion is as bad as everyone else's. No idea what the big ridges are. working on printing some smaller cubes on both the RC and final firmware.

Posted : 23/04/2018 3:53 am
Kwaad2
(@kwaad2)
Honorable Member
Re: [Testers Needed] Inconsistent Extrusion


So I printed it with the 3.5:1, with amperage cranked up. The pattern is TOO tight now. 0.4mm or 0.5mm. That's about 1/8th the stock extruder, with 3.5:1! It should have been ~1.0mm.
EXACT same STL file.
I'm stumped with this result. This result does NOT fit into ANY concept I've seen, or had myself. This result is damning IMO. It means everything we are looking into/trying is WRONG.

Pattern changed. Not XY
Occurance rate changed. Not based off "steps" on E.

Before anyone says "Durrrr, you need to ____ to make better quality." (Cause someone does every time.

Same object, just lighting that makes it look better.


***

Recall that time I said "Be careful messing with those settings, someone's gonna burn their printer down."
Yeah. I'm an idiot.

It was 96c when I first took the reading before I dropped the current. I was more worried about saving my printer, than taking that picture though. So best you guys get is 92c!

How my extruder didn't melt into a gooey mess, I have no idea. I would NOT trust even trust PETG at that temp. Honestly, even ABS is gonna start getting soft around there. Unless it was PC or Nylon.... Yeah. That's toasty.

Hi, I'm Sean. I used to work on CNC machines.
I try to not make mistakes, but the decision is YOURS.
Please feel free to donate to my filament/maintance fund.

Posted : 23/04/2018 4:19 am
Kwaad2
(@kwaad2)
Honorable Member
Re: [Testers Needed] Inconsistent Extrusion

Has anyone tried.

M221 S100

And ensure the Slicer profile is not changing this on you. I started a 10 hour print an hour ago, so my printer is occupied for a bit.

Hi, I'm Sean. I used to work on CNC machines.
I try to not make mistakes, but the decision is YOURS.
Please feel free to donate to my filament/maintance fund.

Posted : 23/04/2018 2:13 pm
JLTX
 JLTX
(@jltx)
Reputable Member
Re: [Testers Needed] Inconsistent Extrusion




Very good find!

I'd be curious to know what option the prusa firmware chose to set the driver resolution and mode to.

Found an interesting article on the trinamic drivers on hackaday. Even shows a tuning flow chart.
https://hackaday.com/2016/09/30/3d-printering-trinamic-tmc2130-stepper-motor-drivers-shifting-the-gears/

The page also has a link to the Trinamic data sheet that shows a bunch of modes available starting on section 1.4
https://www.trinamic.com/fileadmin/assets/Products/ICs_Documents/TMC2130_datasheet.pdf

This is good info. I scanned through the data sheet and poked around in the FW and it is all laid out pretty clearly. Changing configuration registers to enable, disable, tune features looks straightforward. I don't have access to my printer this weekend, but would be willing to try.

If you're looking at the TMC defines in Configuration_prusa.h, watch out. Lots of those are defined there but not used elsewhere in the firmware. Do a global search on the define you're editing to make sure actually changing anything. 🙂

Good example: INTPOL defines aren't used that I can tell. That feature is hard coded to 1 and has to be changed in tmc2130.h I believe.

I am betting they intended to tie the defines to things later on, but forgot.

No, I was looking only in tmc2130.cpp. Lots of things to tune in there, though there are some defines from elsewhere.

Posted : 23/04/2018 2:26 pm
JLTX
 JLTX
(@jltx)
Reputable Member
Re: [Testers Needed] Inconsistent Extrusion


Has anyone tried.

M221 S100

And ensure the Slicer profile is not changing this on you. I started a 10 hour print an hour ago, so my printer is occupied for a bit.

Interesting thought. We’ve all been using S95 to fix over extrusion, but what if the FW is getting math precision errors while scaling on-the-fly. I’ll try when I get my printer back.

Posted : 23/04/2018 2:30 pm
Page 8 / 15
Share: