#Mechanical issue/print artefact I can't pin down

336 messages · Page 1 of 1 (latest)

potent stirrup
#

I have a 350mm Voron 2.4 with a Galileo 2 extruder and one of Vitalii's CNC Taps. My machine has been printing for a good 21 months, but it has always had an issue where it adds VFAs as well as some wavy irregularities no matter what I do (see attached photos). Assuming that this is a motion system issue, I've thrown all kinds of money and time at this issue, but it persists. Things I've tried beyond deracking:

  1. replacing the belts (I've only ever used Gates),
  2. replacing all bearings in the gantry,
  3. replacing the idlers for genuine Gates ones,
  4. replacing the motor pulleys with genuine Gates,
  5. replacing the X axis LDO rail for a CPC with V1 preload,
  6. replacing the Vivedino y-rails for Trvis Z1 preload rails.
  7. replacing the Trvis rails for Honeybadger's newest H-class rails,
  8. Replacing my printed tap with a CNC aluminum tap,
  9. replacing my idlers with Rama idlers,
  10. reprinting and replacing all printed parts in the gantry,
  11. replacing the Moons' A/B motors with cheap ones (I switched back),
  12. Swapping TMC2209 stepper drivers for ones with less microstepping (I switched back).
  13. replacing my Dragon hot end with a V6 (I swapped back).
  14. completely removing, disassembling, and reassembling my gantry.
    I'm sure there are other things I've tried that I've forgotten about, but despite all of these changes, there has never been a noticeable change in print quality. The only parts of the motion system that have not been replaced since I started chasing this issue are the cable chains, the wire harness, and the 2020 extrusions. I feel completely lost and I'm sick of printing the same ugly test objects without making any progress. For your viewing pleasure, I have my latest ADXL results included below as well as an image of my machine.
Cube.png kShakeResults.png PXL_20231109_204357549.jpg
slate bone
#

Hi, judging from how corners look like, your PA may be too low? They are bulging.

potent stirrup
#

I agree, but do you think that is related to the artefacts on the print surfaces? I just bumped my PA to print a test cube.

potent stirrup
#

@slate bone no real noticeable difference. Bottom is new.

rn_image_picker_lib_temp_d630bb57-4a68-43a3-b1c1-c6ac737232f8.jpg
keen comet
#

Hey @potent stirrup.
Can you print the reducing tower from my repo please. Link is in my bio, but also please increase the speed by 1.618 times when you reach half way.

potent stirrup
#

1.618x from 70mm/s? So 113mm/s?

keen comet
#

Sounds good

#

I just love maths. So, i chose the least rational number, to try and avoid overlapping patterns if it is speed related or some sort of motor resonance thing. Resonant frequencies may have harmonics. This should avoid that. But it doesnt have to be exact. 🙂

#

Im gonna go to sleep. But i dont expect you think this will be a super fast process anyway. With any luck, we will get you some more information and ill come have a look in my morning to help interpret.
Please ping me whenever you got something, anything, as long as this thread takes. Im used to it and might forget this exists if you dont.

potent stirrup
#

@keen comet changes in speed didn't seem to change anything. Top side is 70mm/s, bottom side is 113.

PXL_20231109_234310847.MP.jpg PXL_20231109_234301922.MP.jpg
keen comet
potent stirrup
#

@keen comet both had gap spacing of 1.96mm (19.61mm over 10 gaps, and 27.5mm over 14 gaps for your stl and the cube respectively)

potent stirrup
#

Because anything close to 2mm makes me (and probably everyone else) think of belt path issues, here is my belt path:

keen comet
#

Sooo, if it were 2mm id be convinced its belts. If it were 1.414mm id also be convinced its belts. The way a motor moves the diagonal on a coreXY means belt rub in some places can show as √2mm of a simple right angled triangle. In others it can show at 2mm. That you measured accurately is great. And gave 2 measurements. I domt rightly know now.
So, how is your belt tension? Ive got a video here showing the ballpark of the place to be that might help you.
https://cdn.discordapp.com/attachments/678369247134613534/1112165819904954488/trim.55886A52-B05D-46EA-AD71-6DFB2E97A2AF.mov

▶ Play video
potent stirrup
keen comet
#

Thats good. Were these pulleys from your formbot kit?

#

For the AB motors

potent stirrup
keen comet
#

Hmmm, if we cant make it better with speed, can we make it worse?
Is there value in trying?

#

Can i get your printer.cfg please

#

A klippy.log also would be fantastic if you have included files.

potent stirrup
#

I will print a tuning tower with print speed increasing if I can figure out the syntax for that command. I'll send the documents once I'm home. Do you want included files as well? My macros are all separate from my config file.

keen comet
#

Printer.cfg and klippy.log will be perfect. The log file shows what gets loaded, so it sort of includes the included files, but its a total mess to read

patent sedge
#

have anyone suggested putting it on solid surface instead of carpet ?

potent stirrup
patent sedge
#

Hmm sometimes that ghosting can be caused by gears moving back and forward. Maybe check if grub screw is secured properly and shaft is not moving back forward when pushed on (add loctite on grub screw if u havent)

keen comet
patent sedge
#

Just replacing them means nothing if they move out of spot without loctite :d

scenic quail
#

@potent stirrup i read the whole thread and i can't think of a solution for you. But i really like your belt tensioners. Can you share a link or stl pls?

potent stirrup
potent stirrup
keen comet
scenic quail
#

Tnx guys!

potent stirrup
#

Is there a chance my issue comes from my Rama Idlers? I'm thinking about swapping out my A/B parts for ones printed with more perimeters and higher infill.

keen comet
#

Id really like to double check your printer.cfg still, and if there is an experiment to run, it would be any simple flat faced object with speeds of 40 and 133. 40 should be the uggliest it can get, and 133 should be pretty smooth. Could make for a nice test print for autotune too.

potent stirrup
potent stirrup
potent stirrup
#

42mm bottom, 108 top.

PXL_20231110_232709991.jpg PXL_20231110_232717376.jpg PXL_20231110_232722244.jpg PXL_20231110_232728712.jpg
keen comet
#

42 was meant to be ugly!

#

If anything i feel like 108 is the one with a pattern that stands out stronger.

#

I think this confirms it does change with speed, and is therefore motor vibrations. You have now entered the rhelms of wishy washy help, because this is certainly less well understood than i would like.

potent stirrup
#

The wavy bs in the 42 region is the artefact that I haven't been able to get rid of since building this machine. The VFAs are inexplicable, but there has been a time where I didn't have them. The problem is I have no clue when they appeared.

keen comet
#

My best lead here is the machine vibrations graph, and trying to find a low point on it. But we kind of tried that with 108 and it didnt really help

potent stirrup
#

If I print at 25mm/s, the artefacting is still visible, but easy to ignore.

#

Which matches that lead. That being said, I have no clue what I could possibly modify to reduce vibrations other than shortening all my extrusions and cutting my bed to make a smaller V2.

keen comet
#

Something in my head is saying "make sure the hotend isnt loose in the mount". But i dont know why. I really dont think thats it

#

I dont expect to see this from moons motors.

potent stirrup
keen comet
#

We should probably disable interpolate and set microstepping to 64 on X and Y tho

potent stirrup
#

Ok. I'm printing a cube without TMC autotuning enabled at the moment, then I'll print one with those settings.

keen comet
#

The settings from your klipper.log dont look like autotune settings. I suspect you could increase that current to 1.2 also if yoy have any airflow in your electronics bay

potent stirrup
keen comet
#

I havent messed with it myself. But i just told you the only differneces between our apparent settings for formbot kit XY motors.

potent stirrup
#

Gotcha. So if the autotune made any changes, they aren't apparent in the log?

keen comet
#

I think extra current could help here. Interpolate/microsteps probably not so big a deal. Although kevin recommends against interpolate.

keen comet
potent stirrup
#

Ok, to prevent multiple variables from changing, I'll do a 1.2a print, then a no interpolate print.

keen comet
#

That page lets you see it in a slightly more friendly fashion. The config page has headers hyperlinked

keen comet
potent stirrup
#

Ok, so microsteps: 64 and interpolate: false

keen comet
#

Thats it. Yeah, thats basically one change

potent stirrup
#

I don't have to make any changes to the jumpers on my MCU do I?

knotty belfry
#

Hello, been asked to pop my head in as I had a similar issue on my VT and it came down to speed selection and idlers in my case.

I ended up replacing my toothed idlers with bearing stacks which improved it on my VT (but made it appear on my V2 so that got genuine gates idlers instead)
Orcaslicer has a really nice VFA test built in to it so you can check if it is speed related and if there is a range you want to avoid, someone has taken that test and made a 3mf for use with PS/SS https://thangs.com/designer/dbsorkin/3d-model/K3D VFA Test-902194, it should change speed every few layers. If it is speed/ motor resonance related then you should see it change in intensity or spacing over the course of the test, it'll also show if its worse at different angles.

In my case, I found the ~2mm VFAs were worst at about 80mm/s on my 2504s running at 2A but completely absent <40mm/s and >120mm/s.

potent stirrup
minor zodiac
#

Has anyone suggested shake and tune vibrations calibration?

potent stirrup
minor zodiac
#

my two cents .

Can you run vibrations calibration on ab axis

image.png
#

are you using damping ratio in input shaper

potent stirrup
potent stirrup
# minor zodiac my two cents . Can you run vibrations calibration on ab axis

Vibration calibration? That must be new? I've never heard of it. Nothing is pulling on the toolhead. My reverse Bowden has lots of slack. I will look for fan vibration. Better yet, I'll run shaper calibration without my SB front attached to see if that line disappears. As for tension, my heat soaked tension is right around 115hz. Cold, the tension is right around 100hz.

minor zodiac
#

you ran vibrrations calibration.
It looks like Frix renamed it vibrations decomposition

potent stirrup
#

Gotcha.

potent stirrup
minor zodiac
potent stirrup
#

Like this?

Screenshot_20231111-094548.png
minor zodiac
#

yes like that.
And can you make the size 100

#

Also set your max accel in klipper to like 3000

potent stirrup
#

I suppose my best speeds are areas that are low on both vibration graphs (XY and AB)?

minor zodiac
#

so you see that you will have different speeds you can use on the diagonals vs the xy

potent stirrup
#

Does the vibration test have an "all" direction option?

minor zodiac
#

Try reprinting with the speed locked in one of the regions with the black arrow

image.png
#

I would try 100

minor zodiac
potent stirrup
#

I adjusted belt tension a bit (heatsoaked it was higher than I thought... hitting 120hz, so I dropped it to just above 110). These graphs are the three vibration tests (AB, XY, ABXY) with magenta lines representing the speeds that I'm printing four different tests seen below.

Vibrations.png
potent stirrup
#

Third from the left was printed as a control since it should be the worst speed for vibrations according to my graphs.

minor zodiac
#

I cannot tell at all which is the best. In your view point which is the best

potent stirrup
#

It depends on what side I'm looking at. 50 is best on a, 75 is best on x, 95 is best on b and +y, and 105 is best on -y.

#

Which is frustrating because 50 should be the overall best, yet it is dead last on several sides. 95 should always lose, but is best on two sides. 😮‍💨

minor zodiac
#

Are all the vibrations 2mm apart and vertical

potent stirrup
#

No. Some slant, some are closer than 2mm, some farther. Some don't go in straight lines at all. In general, faster speeds have wider spacing.

minor zodiac
#

Let’s do a fun little test. Can you loosen your belts to 100hz. And rerun the shape at 107 mm/s

#

Trying to see if we can reduce this by reducing belt tension. It will mean slower accel/speed

#

@keen comet @knotty belfry any suggestions?

potent stirrup
#

Ok. To be clear, you mean 100hz heat soaked, right?

minor zodiac
#

doesn't matter heat soaked or not. Just want to lower the belt tension

potent stirrup
#

Ok belts loosened to 100hz printed at 107mm/s. Photos are ordered: x, +y, b, a, -y.

PXL_20231111_193826566.MP.jpg PXL_20231111_193956967.jpg PXL_20231111_194013152.jpg PXL_20231111_194020282.jpg PXL_20231111_194032275.jpg
worldly sapphire
minor zodiac
potent stirrup
#

Or do you mean look at everything while moving?

#

I have checked belt paths many times in the past while running. I've also replaced every bearing/idler/pulley in the belt path with no changes.

minor zodiac
#

Can you swap a and b motors, then print another shape. We want to see if the nastiness follows the motor, or if it stays in the belt path. Please do this with power off. This will also require you to loosen and retighten belts

potent stirrup
#

@minor zodiac I can definitely do that. What nastiness are you looking at? Graphs or print surfaces? The ugliness in print surface shows up in both A and B moves. For some reason it is more prominent when those movements are in the negative y direction.

potent stirrup
#

Also, as was mentioned in the original post, I've swapped the motors for different motors with no apparent change in quality. I've swapped them for non-moons motors, and I've also swapped two of my z motors for the ab motors. I've swapped the a and b drivers as well. The only thing that hasn't been swapped at some point is the wire running between the mcu and the motors.

keen comet
#

spitballing. can you just try a print you have done recently, with a line width of 0.5

potent stirrup
#

My default width is 0.46, so I decided to do a 0.35, 0.4, 0.5, and 0.55.

potent stirrup
#

If it isn't clear from the photos, the surface defects look almost identical across the four prints.

minor zodiac
#

what external perimeter accel are you using

#

can you post your 3mf file if using prusa slicer

potent stirrup
minor zodiac
#

Change
speed
external perimter 75mm/s
perimeter 110 mm/s
Accel
external perimeter 750mm/s2
perimeter 1000 mm/s2

potent stirrup
#

+A, +Y, -X, -Y, -B

#

I took them slightly out of focus to reduce moire.

potent stirrup
#

This just finished printing too. What do you want to see from it? There are no sides that are free of defects, but higher speeds look better than slower ones. This photo is of the -Y side (which is tied for worst with the -B side). One interesting artefact I hadn't seen before is what I assume is the acceleration curve made visible by static acceleration with increasing max speeds. Of course photographing any of this is difficult...

PXL_20231113_225238401.jpg
minor zodiac
#

basically you are trying to find a area where there are the least amount of vibrations on each wing

#

Then use that as your external perimeter speed and reprint one of your models

potent stirrup
#

They look marginally better, but still not good in my opinion. My goal is not for this machine to print "ok." I would have bought an ender 3 if I wanted that. I want it to print like a well tuned Voron. Trying to avoid the symptoms of the problem by adjusting the speeds/accelerations will likely never satisfy me. I want the cause of the symptoms to be fixed. I've spent hundreds of hours and dollars trying to find the source of the issue, but still it eludes me. I don't think this issue can be solved with software or slicers.

minor zodiac
#

@worldly sapphire in your adventures what do you think

worldly sapphire
#

Well, to make a TLDR of my last two years of banging the head against a wall, i've been battling this issue on and off. I've rebuilt the printers many times, tried alot of hardware combinations to learn what each component does. Noisy motors certainly do not help. frix's vibrations measurements help decide if they are noisy or not. The bom OMC motors seem like a good common middle ground where things are less shit then many other options. I still haven't fully gotten rid of my artifacts, but they are a heck of alot less visible now that i changed perimeter width to be slightly wider.

#

So i don't have a good answer for you as to what is causing it or how to really solve it, as it's currently a very unsolved issue that seem to affect alot of people, some more then others.

#

This is where it all started for me, the faceted stealthburner body has alot of angles, which exaggerates the issue for me, so i've created a vase of the top of that part to use less plastic while hunting the cause.
So far the theory is that mixing A and B motion on diagonals isn't working perfectly on these angles, and there may be some calculation issues in the tmc drivers or something along those lines. We haven't really fully figured out what yet.
Duet and klipper behave much the same, that's also one thing i tried comparing in my long journey.

#

I can print a new stealthburner body with my current settings just to revisit where it currently stands. But it's not perfect, it still has some artifacts. But they are barely visible vs the picture above where they are deep groves you can feel.

# 
- 0.4mm perimeter thickness on all perimeters.
- Slightly underextrude rather then overextrude.
- 58-80mm/s perimeter speeds for the worst combination
- High belt tension rather then low. Low has a dampening effect.
- Have great rails with high preload (loose one dampens the effect slightly, but just slightly)
- Print the stealthvase (The 4 almost diagonal, but not fully diagonal moves have a weird motor mix that seems to create vibrations. This will definitely show up on your print if it's present.)

** Madness ensues **```
This is my summary/status so far.
Here is also the vase if you want to see the behaviour for yourself:
📎 stealthvase.stl
#

so recipe for less shit prints would be the opposite of the above, thicker walls, 130mm/s or whichever crossover is best with your motors for perimeter speed, not too high belt tension which will limit your top speed/acceleration.

#

And yeah, that small wall of text is the TLDR, more or less.

potent stirrup
#

I appreciate the write up. Sounds like some machines are destined to be B-class no matter what... Maybe it's the air in the spaces these machines live? 😜 I haven't yet tried swapping my AB motors for other reputable ones. I had always assumed that they couldn't be the problem since Moons' are generally considered upper tier. Though it aligns with your recipe for madness, I'm going to try some z2 preload y-rails because sloppy ones look bad, z1 looked bad (maybe a bit better) so maybe z2 will be the sweet spot? Or maybe it will kill my motors. Only time will tell. I figure deviating from your findings is ok since my artefacting isn't deep rippling but instead irregular diagonal waves.

worldly sapphire
#

yeah, there are some overlapping issues, and deviating is how you figure out what makes it better or worse.

#

that's some of the best looking i've had in a while

#

but it's still present

potent stirrup
#

I ran a test yesterday that is kind of fascinating. In 2014 I built a self-sourced Prusa i3. Over the years, I accumulated a stash of replaced parts, including NEMA17s. I removed the Moons' motors from my AB drives and swapped them for the generic steppers that ran my original i3's x and y axes. I then printed another test piece and had better (yet still far from perfect) print surfaces. I re-ran Frix's scripts and ended up with uglier graphs with more vibration and lower accelerations. It seems as if my adxl results are barely aligned with actual prints, and it seems like my Formbot supplied Moons' motors might be contributing to the issue. I'll be purchasing a pair of 0.9° LDO motors in the near future since that is the only piece of the motion puzzle that hasn't been truly swapped out yet.

keen comet
potent stirrup
keen comet
#

Yes. I have tagged you in an appropriate channel

#

I had consisdered the moons to be a pretty good motor tbh. I would also stay clear of the LDO or OMC speedy motors unless you are actually aiming for high speed. Hopefully @worldly sapphire might suggest a motor he felt best about.

worldly sapphire
#

0.9 is so damn loud it's not even funny

potent stirrup
#

I have only ever experienced Moons' and my no-name motors, and the surface quality with the Moons' is markedly worse. Could be coincidence, could be Formbot sent me fakes. I'll send a photo later

keen comet
#

I believe you. Might just be that they resonate with your machine badly for some reaaon.

potent stirrup
worldly sapphire
#

the moons motors i got was also noisy, but there are other models who supposedly are better. That goes for most companies i guess.

worldly sapphire
#

@civic sonnet can probably formulate why the 0.9motors are so noisy and how it affects things then me, i have only tested them minimally

keen comet
#

Im tempted to say OMC motors from the sourcing guide would be the best choice for normal/good speed printing. The newer low incductance stuff really struggles with slower pretty parts from what i understand. And printing fast requires a ton of tuning.

potent stirrup
worldly sapphire
#

the bom OMC is one of the better motors

#

2004 something something was the better of the ldo motors i tested

#

2004ac i think

#

but all my testing is very limited, so i have a hard time giving advise on anything other then omc because those are the ones i really know

keen comet
#

In the US i dare say you would be able to find some cheap on the marketplace.

worldly sapphire
#

for ldo i think it goes something like this:
2004 the best
2504 hard to tune
2804 an absolute pain to tune

#

but take all that with some grain of salt since i haven't done in depth testing on that myself

#

let's just do this, @amber ridge do you have a spare minute to explain stepper motors? you're the first person i come to think of when it comes to this

amber ridge
#

avoid 2504's, use 2004's with 2209's, use 2804's with 5160's on 24v if you want speedy motors

potent stirrup
#

Any opinions on high temp motors vs standard temp?

potent stirrup
minor zodiac
#

Is this pla?

potent stirrup
#

It's abs. Could be vfa, but in my experience, vfas follow the shape of the part. A rectangular part like this should have regular vertical vfas if it has them at all, shouldn't it?

minor zodiac
#

Was this a speed test print? Where different levels had different speeds?

potent stirrup
#

No, it was a test where different levels had different stepper current (0.6a, 0.8, 1.0, 1.2, 0.5)

keen comet
#

Honestly i think the curvy wobbly lines are cooling related but i could be wrong.

#

The underlying verticals are possibly motors. Possibly.

#

I say cooling. Single wall ABS prints arent terribly sturdy, and thats am artifact ive seen before. But it wasnt present on other prints. I wouldnt worry about that.

potent stirrup
# keen comet Honestly i think the curvy wobbly lines are cooling related but i could be wrong...

That's an interesting thought...I haven't tested the impact of cooling. I'll try printing the same part much slower with no cooling fan to see if that disappears. The only piece of evidence against it being cooling is how consistent the pattern is between prints (printing the same file and comparing waviness). Cooling also tracks when it comes to the waves only appearing in Y moves since SB part cooling fans blow air perpendicular to Y moves but parallel to x moves.

potent stirrup
#

Waviness still apparent on fanless prints, but I think I maybe have some ideas on where they're coming from. Without fans, vfas became more defined. I seem to have diagonal vfas on x moves, and diagonal/vertical hybrid vfas on y moves that end up creating the mottled look from the first photo. My assumption is that diagonal vfas being omnipresent suggests they are either belt related (I'm not convinced) or extruder issues (after a teardown/cleaning of my Galileo2 yesterday, I feel this is the one... Lots of friction in the gears). My vertical vfas on Y I blame on my Honeybadger rails which are quite crunchy. I'm replacing them as soon as Youmetong can get new rails to me.

rn_image_picker_lib_temp_e39b079c-8831-438a-a347-f1c2c72818a3.jpg rn_image_picker_lib_temp_c6a58705-31c7-4d3f-9b8b-fdc0a60d1ab6.jpg
keen comet
#

Should the VFAs be resultant from rails or bearings or belts (not belts if you cant find 2mm pitch). Then at the very least, any 45 degree angled parts should present at √2 ratio to some strait line part. That might be worth checking. Again, measure multiple gaps and divide for accuracy. If you can think it through, the larger spacing will be down the direction of the root of the VFA. So if diagonal, look at motor, if X or Y, look at rails.

#

I hope that makes sense. If nothing else, gimme the measuremrnts as best as you can get and tag me.

potent stirrup
# keen comet Should the VFAs be resultant from rails or bearings or belts (not belts if you c...

In the x direction, my VFAs are spaced by about 1.98mm. I took a lot of measurements across different peak and that was my average measurement (1.95 min, 2.00 max). My diagonal print was more interesting... My B motor moves were impossible to accurately measure, but my A motor moves showed two distinct patterns depending on the direction the toolhead moved. When moving in the +x+y direction, the peaks spanned about 2.76mm. When moving in the -x-y direction, the peaks spanned 1.42mm.

#

That seems like I get roughly √2 ratios between X and positive A movements, then I get roughly √2 ratios between negative A movements and X.

#

In other words...
√2=X/-A
and
√2=+A/X
...which means X/-A=+A/X

keen comet
#

That makes no sense whatsoever 😂

potent stirrup
#

Yeah... My machine is an enigma.

keen comet
#

Thats a fascinating datapoint. Most most interesting. Ill ruminate on that, it might be telling us something.

#

Can you like, take a photo of the part as it was on yhe buildplate, and label the walls with those wavelength dimensions please?

potent stirrup
#

Sure, I can do that.

#

The toolhead moved counter clockwise for each object

#

The VFAs don't show up very well, but here are my marks that I measured in the A directions.

PXL_20231118_164009116.jpg PXL_20231118_163949609.jpg
keen comet
#

Oh my days

#

Itll be the same speed. Itll be the same... everything. How can that be

minor zodiac
#

Can you take a look at your belt when you do a leveling routine, and when you move +y and -y and see if the belt moves up or down on each motor pulley

#

Want to make sure you are not hitting/rubbing the pulley flange

potent stirrup
#

Worth noting... This is the only place in the belt path where belts aren't centered on bearings

keen comet
#

The moves you have labelled A is actually a move made by the B motor with no input from A whatsoever if i understand correctly.

minor zodiac
#

Can you tell us which motors you are running. And where you got the kit?

potent stirrup
potent stirrup
minor zodiac
#

have you installed autotune?

potent stirrup
minor zodiac
#

you are using this one moons-ms17hd6p420I-04

#

in the auto tune

potent stirrup
#

I copied my motor P/N from my auto tune config

minor zodiac
#

okay

#

do you know if you have real gates motor pulleys or are they a chinese version

potent stirrup
#

I do have real Gates pulleys bought from filastruder

minor zodiac
#

Do you have real gates xy joint idlers

#

from https://ellis3dp.com/Print-Tuning-Guide/articles/troubleshooting/vfas.html this is what I am looking at for troubleshooting
Artifacts are most prominent in in B:

Bearings in B belt path
B motor pulley - confirmed Gates from filastruder
B belt - confirmed gates from fabreeko about 2 weeks ago
B motor - confirmed - moons-ms17hd6p420I-04 via a Formbot kit from late 2021

Artifacts are most prominent in in X:

XY joint idlers - confirmed Gates from filastruder
X linear rails(s) - V1 preload - CPC Chieftek, installed spring of 2023. Last lubed maybe three weeks ago
Ellis’ Print Tuning Guide
VFAs (Vertical Fine Artifacts)

Ellis’ Print Tuning Guide

potent stirrup
minor zodiac
#

okay who makes your belts and when were they installed

potent stirrup
#

Gates belts (from Fabreeko), installed two weeks ago.

minor zodiac
#

who makes your x-linear rail? when was it installed? and when did you last lube it?

potent stirrup
#

CPC Chieftek, installed spring of 2023. Last lubed maybe three weeks ago.

minor zodiac
#

I have not heard of that company

#

what is the preload on the rail?

potent stirrup
#

It is a very well regarded top tier brand. It has V1 preload.

minor zodiac
#

okay well that takes us through ellis's VFA with non-2mm spacing

#

is this a v2?

potent stirrup
#

Yup. I've done that a fair number of times the past year.

#

Yeah. 350mm

minor zodiac
#

That leaves 1) derack per Nero3D. including readjusting the rear extrusion, or 2) motors, 3) belt rubbing - you see any dust anywhere?

#

have you checked the belt tension on your z-belts recently? When was the last time they were checked? when were they installed?

potent stirrup
minor zodiac
#

See any belt dust?

potent stirrup
#

No dust. The bearings have had black greasy residue, but those have been cleaned in the past couple weeks.

minor zodiac
#

that black greasy residue is your belts wearing on the bearing surface

#

no buildup of that residue in the teeth of the motor pulley or in the teeth of the idelers.

#

if there is buildup there is needs to be cleaned

potent stirrup
#

Yup. It has been.

minor zodiac
#

has been cleaned? or has been builtup

potent stirrup
#

Cleaned. I've done just about everything that can be done to my gantry in the past month, including breaking it down to a pile of parts and rebuilding.

minor zodiac
#

so the only thing left is motors

#

have they been replaced, or a different set tested

#

what is your absolute belt tension

potent stirrup
#

I've swapped them for some astrosyn motors I had sitting around. VFAs stuck around.

#

My absolute belt tension is right around 112 when heat soaked

minor zodiac
#

can you reprint the two boxes but this time use 750mm/s2 accel for external perimetes and 1000 mm/s2 for perimeters, and use 50 mm/s for external perimeter speed and 107 mm/s for perimeter speed.

#

based upon this 70 is going to be hitting the noise on AB axis and 120 is going to be hitting the noise on AB axis

#

if you do not like those speeds, then you need to choose a speed which will be at lowest level in green in all three graphs

#

The accel was lowered as I have found that sometimes I need to lower it to reduce VFA's on the external perimeter

potent stirrup
#

Printing now.

potent stirrup
#

Well, after comparing the parts side-by-side, it looks like running slower had some improvement on surface quality, but in a confusing twist, the patterns described previously as being different on positive vs negative diagonal moves switched sides. The side that had 1.42mm VFAs now has wider 1.98mm ones and vice versa. Also, I think the 1.42 might be from resonance, because on these prints, it gets more faint as the printhead makes its way across the diagonal face. The same pattern appears in the +Y direction of my part that isn't angled 45*. The frequency is higher on the slower part, which seems to indicate it is resonance. I think I need to stop relying on my adxl graphs because they don't ever seem to help improve real-life prints.

potent stirrup
#

A thought: since building this machine, I have always needed to tighten my B belt more than my A belt to get the tension even. I've heard of three different priorities when tightening belts: 1) tighten each the same number of turns, 2) tighten them until they resonate with the same audible frequency, 3) tension them until the gantry has no belt-induced racking. In a perfect world, all three properties would be met at the same time, but equal turns results in a difference of 10Hz in plucked frequency. Tightening until they are the same requires turning my B belt nearly two full extra turns. It also creates racking on the A belt side. When the belts pluck at frequencies 10hz apart from one another, there is no racking. The belts are 100% tooth-for-tooth the same length. How is this possible?

hushed crag
#

have you tried the vfa test in orcasliser? Sometimes its not a mechanical issue, but when printer is in certain speed it just resonance and creates the vfa, try the vfa and change the print speed accordingly could help (especially dont slow down on outer walls)

potent stirrup
#

I have tried the vfa test. I saw some improvements in print surface with higher speeds, but it still didn't solve the issue I'm chasing.

potent stirrup
#

I think I may have found part of the issue in my machine, but I don't know what is causing it. I was printing some cylindrical objects today, and noticed that my machine overshoots on Y direction changes. In other words, when Y travel shifts from being positive to negative, the toolhead does not instantly change direction. Instead, it resists the change in direction, creating a slightly flattened bit on the cylinder, then creates a wave pattern after the toolhead begins to travel in the direction it should be travelling. I created a graphic to show what I mean:

Screenshot_2023-12-04_201255.png
#

It is exaggerated, but my machine is printing like the right shape instead of the left one. My toolhead does not seem wobbly, so I feel like it must be from the entire gantry rubber banding (my gantry can be moved a couple mm back/forth when the motors are energized). Is this inevitable? My machine seems to create the nicest surface quality when printing perimeters at 120mm/s, but it also creates this ugliness that cannot be fixed with input shaping.

worldly sapphire
#

@bleak marsh i'll ping you in for theories here, if you have any

#

that does seem to overlap some with the a/b mixing we discussed.

worldly sapphire
#

i'd expect it to be more like this

image.png
potent stirrup
worldly sapphire
#

huh, well okay then. I'm still very clueless, but that's an interesting datapoint.

bleak marsh
bleak marsh
potent stirrup
bleak marsh
keen comet
#

Somewhat unlikely, but it wouldnt be impossible for this to be a missed number in an input shaper run. Also, check your config block at the bottom to see if theres a saved value for shaper that you didnt intend to use. Just in case.

amber ridge
potent stirrup
#

Y isn't exact from my most recent chart, but those numbers are within rounding error.

hard cedar
#

That Y axis is rather low on the MZV recommended speed. I'm definitely suspicious of the bearing stacks after having encountered something similar myself

#

Also worth checking your Y MGN9s are parallel to one another. That there's no binding when the belts are loose

glass abyss
#

Just wondering since your graphs are super clean but the frequency is still kind of low…. I had similar VFAs after switching to RIDGA when my filament path was 0.5mm out of alignment with the drive gear

IMG_4884.jpg IMG_4885.jpg IMG_4783.jpg IMG_4731.jpg
potent stirrup
potent stirrup
hard cedar
#

Mine were 1.26

potent stirrup
#

I might disassemble my belt path and see if I can pair shims to create a consistent offset for each bearing stack. The shims are the only part of my belt path that haven't been replaced at some point.

hard cedar
#

Might be good to sand down the burring on them as well, that'd give you an idea of their true size

glass abyss
#

If you haven’t already, take your Extruder off and look down the filament path with a light on the other side and make sure the curve of the drive gear is perfectly centered in the hole

#

If it’s more to one side than another it can create some subtle and funky artifacts

potent stirrup
potent stirrup
#

I installed some OMC steppers (BOM model) for A/B. In addition to this, I swapped my Honeybadger Y rails for some Z2 preloaded Youmetong rails. My graphs now all look damn near perfect (albeit with low y acceleration), but my prints still look like garbage. I haven't pulled my extruder to check the filament path, but since it is a Galileo2, it isn't really adjustable. Here are my graphs (mostly so I have a history of how they've changed):

resonances_20231207_173342_Y.png resonances_20231207_173104_X.png belts_20231207_190952.png OMC_AB.png OMC_XY.png
hidden lake
hard cedar
#

Print quality has so many determining factors, it could be anything

hard cedar
potent stirrup
hard cedar
#

That banding is more likely belt tension on A/B

#

Seems odd because your belt graphs are good

#

The damping ratio is something that I haven't seen before. Can you comment it out and see what results you get?
Also might be worth switching to the pattern method for tuning pressure advance

#

Your A/B belts, what frequency are they tuned at?

#

I personally wouldn't recommend the 160Hz as mentioned above, they need to have some slack as pulling them tight can introduce bigger problems

#

Mine are at the suggested 110Hz over a 150mm span from the front idler to the corner of the X. It takes a bit of time for them to settle. Moving the gantry around by hand usually helps with that during the retuning

#

The peak of the power spectral density is also higher than I'm used to seeing. They could be in tune, but too tight perhaps

potent stirrup
#

I do use the pattern method for PA tuning. My machine produces that pattern no matter what tension my belts are at (I've tried everything from 100-165hz). The damping ratio is relatively new. It was added as an extra bit of data in Frix's new scripts, but it made no perceivable difference to me.

hard cedar
#

Okay, so the other thing to look at would be your M5 bolts holding up the gantry at each corner on the Z rails

potent stirrup
#

I think the bulging at corners that appears to be a PA issue is actually related to my bumpy cylinder issue. I think my machine has some resonance that my adxl is not seeing that causes rubber banding at the nozzle.

hard cedar
#

And yes the bulging is PA. The other artifact on corners is your seam line in your slicer profile

#

If you can attach your ADXL to the nozzle that might give you better results. What ADXL module are you running? I'll see if I can dig out a mount

#

Also what hotend

potent stirrup
#

I'm saying the corners is not PA. Almost certainly. My PA is set higher than it should be based on Ellis' corners test. The bulging is coming from my hot end (dragon) overshooting the corner. I didn't show a photo of my seam corner because I know why that is ugly.

hard cedar
#

That's odd.. it looks like a PA tower?

potent stirrup
#

I've tried nozzle mounts and the stealth burner mount, but results are the same

#

It's just a single walled cube. I've stopped printing substantial test prints because of the obscene amount of plastic I'm wasting

hard cedar
#

With the dragon then it could be loose heat block screws that go into the heatsink. If those are loose, that can cause this. I added loctite to mine

#

They're extremely delicate m1.3mm so definitely be careful

potent stirrup
hard cedar
#

Yeah you never know. Mine rattled out after 500hr printing

potent stirrup
#

I also have some BOM shims on order from China, so I'll be able to swap those in late December...

hard cedar
#

Awesome. In the meantime so long as your belts aren't drifting off your bearing stacks I think you're good

#

Ask me how I know 😂

#

That's belt dust

#

Anyway let me know how things go when you're done at work and such. I'm going to be slow at replying; moving house tomorrow

visual dew
#

Just another spitball, but did you check the belts? All teeth look good, not malformed, correct tooth pitch for the pulleys, etc?

potent stirrup
# hard cedar Anyway let me know how things go when you're done at work and such. I'm going to...

I just finished checking the dragon's screws. No budging at all. I even took the entire hotend assembly and chucked it into my bench top vise then pushed on the tip of my nozzle with a steel tool (my nozzle is a Triangle Lab PC diamond tipped nozzle) and there was no wiggle at all. I have a V6 in its own SB parts that is currently printing the same gcode, but my guess is it comes out the same. Thanks for your idea anyway!

potent stirrup
visual dew
potent stirrup
#

Also, for whomever mentioned filament path, it looks good to me... G2 isn't really adjustable, but the parts are designed well enough that it doesn't appear necessary.

rn_image_picker_lib_temp_74e5fd92-e62f-4260-99da-3721f27e0a02.jpg
glass abyss
potent stirrup
#

Well, I've checked everything that I can think of checking, and I still have a visible bump on the front edge of cylinders, as well as around corners that require a change in Y motion. I'm completely lost.

PXL_20231210_214240221.MP.jpg PXL_20231210_214203892.jpg
hidden lake
# hard cedar That would make little difference in reality, I don't run pins mod and get lovel...

I don't think you should dismiss the idea that quickly. Mechanically it is the only thing that hasn't been upgraded or changed. A rapid change in either x or y could easily displace a pulley if it is not fully constrained by it's axle and considering the accelerations of the motion system. I don't run pins mod either ... yet. I am still getting to know my machine and I do have vfa's, but I also haven't done input shaping yet.

hard cedar
potent stirrup
#

I do have the pins mod on my xy joints. I have had pins on my AB mounts, but had issues with it that I don't recall...

elfin storm
#

not sure if this is relevant or not but saw this in an old post and thought of you

#

"Just wanted to shout out into the void how awesome Orca is. I went through so much tuning on my 350 2.4 and was still getting garbage prints. Orca, pa calibration, and I’m getting same quality as my prusa"

potent stirrup
#

I haven't spent much time with Orca, but maybe I should? My issues still persist, though I've proven a few things as non contributors to my problem. Thanks for the tip!

bronze spoke
#

@potent stirrup i was just going to ask if you figure it out i have the same thing so i was looking for your help . im going to put a lgx on it to see if it helps

#

mine looks the same as yours

potent stirrup
#

Unfortunately no. I'm still trying to figure it out. At this point based on swapping parts, it cannot be my extruder, my AB motors, or my belts. Swapping any of those had absolutely no effect. I don't think it is tap related since it happens with any tap I've tried (CNC, original, or Flex). I've swapped all electronics, all rails, extruder motors... Nothing. Like a crazy person, I keep looking for a source, but I don't know what I haven't checked.

gusty plume
#

i perused though some of this long ass thread, I saw at one point you were messing around with belt tension, but you were using the frequency method. IMO I could never get the ambient noise around my printer to be low enough for me to actually tell what freq it was so I looked for other methods. I found this: https://github.com/Diyshift/3D-Printer/tree/main/GT2 Belt Tension Meter
in my opinion its a much more objective and easy to use tool for your belt tension than using freqs. Its easy to make.

GitHub
3D-Printer/GT2 Belt Tension Meter at main · Diyshift/3D-Printer

Voron Stuff. Contribute to Diyshift/3D-Printer development by creating an account on GitHub.