DDEX print issues and tuning log.
Now that I've got RatOS playing nicely and the mesh in a good place I can finally start one of these threads in here.
Similarly to other threads I'm just kinda posting a log of my changes and troubleshooting in the hopes that others might come in with advice as I make progress.
99 Replies
I believe that the z-stacking issue is either a belt tension or an extrusion issue.
I'm using tr8*2 lead screws, and the 'waves' don't line up every 2mm (closer to 3), so this is not an obvious z axis issue, though I'm not immediately ruling that out
very consistent
I might have been wrong, this is much more consistently every 2mm than I originally thought.
I'm thinking this is a z issue
I like... am not sure where to begin troubleshooting this. I was under the impression that the v-core bed system with oldhams would essentially negate any z-wobble whatsoever.
brought the bed down to 50mm less than total z volume. heard some odd creaking/squeaking in some spots. loosened all the oldhams from the arms and couplers from the rods/steppers. retightened all of them. when I raised the bed back up to 0z there was no additional noises this time.
homed a1nd z-tilted, now printing.
still a long way to go but that was faster to fix than I thought
man I cannot get rid of this reverse elephant's foot
this might just be a PLA thing i should stop chasing and worry about it when I get to real materials.
dropped the bed temp down to 54c, very very minor improvement so I don't think it's that.
realized I haven't done a pid tune on the bed or extruder, did those. no difference. also changed how/when the fan comes on in the first few layers, no difference.
bleh.
The z banding comes back pretty fierce the more I go down. Might have to try re aligning then again. Honestly something in the the z axis sounds kinda 'sticky'. Not sure if it's the pom on the threads or the pom against the brass coupling. I'm tempted to lube them but I know that goes against my better judgement.
Pretty bad hull-lines (but that is expected from classic slicing, no Arachne yet)
Overhangs and cooling obviously need more tuning as well. That 4028 is so scary sounding I don't have it ever go over 30%
lots of artifacts that will clear up once IS is run as well
Oof
The unsupported overhangs are entirely on me. I knew it would be rough but didn't expect it to be this rough.
Those layer stacking issues higher up could just be caused by the air printing but hard to tell.
Good reference for post tuning, but I think I'm gonna go back to test prints until then
It really does help if the grub screw in the pulley is fully tightened.
1st, 5th, and 10th cubes.
The 10th again at .2z, and a new one at .125z
Looks like I really just need to do some more general tuning and input shaper at this point
That 2mm wobble is back
old-apricot•6mo ago
Filanent dry? Like, have you dried it?
Could either be you assembled decouplers incorrectly or the rods are reeeeeeally bent, any visual cues when jogging bed up and down?
it's just PLA, there's no other symptoms to assume this is a material issue so I haven't bothered switching materials yet.
sat and stared at the rods while moving the z axis, nothing immediately jumped out at me. I dropped the bed and loosened/retightened the couplers on the steppers/rods once more to no major change.
I do notice that the back z rod sits further forward within the cutout in the arm compared to others which is more centered.
I loosened the coupler and pom nut to try and center it a bit more but it settled right back into the same forward position.
Kinda want to remove the top bearing constraint just to see if it is that pushing it forward.
old-apricot•6mo ago
With this you should be able to see bed moving when moving up and down, like you can put a ruler on the edge of it and move it cliser ti the side extrusion, you should see the gap appear and reappear
If it doesnt, z is not the problem
Top bearing restricts the presumably bent rod from introducing angular forces to the bed arms
Unless it is pushing the rod forward
old-apricot•6mo ago
Youd see it on your print
Anyways i'd dry the filament first, not take another spool, dry this one
Judgung from the first pics in this thread it is/was very wet
I am very clearly seeing a 2mm wobble on my print, as shown in the images above.
If it were a material, or extrusion, issue it would not be at a consistent 2mm spacing along the z axis on such wildly different geometries.
If more symptoms present themselves to suggest it is a material or extrusion issue I will go down those troubleshooting paths. However, as of right now, I'm unconvinced that this isn't an issue with the kinematics of the z axis.
Again, this is just PLA. I am not looking to troubleshoot stringing/etc when I have barely scratched the surface of print tuning.
I am only interested in focusing on the 2mm spaced artifacts on the z axis. Which there is nothing to suggest this is a material issue.
old-apricot•6mo ago
Ok
Pics of printer: https://discord.com/channels/582187371529764864/844170779717402664/1183619974644899881
So you're using the new decouplers... they'll only be effective if they have freedom to slide. If the lead screws are "bumping" against the side of the ability for it to slide, they'll cause artifacts
The scale of the ridges you're showing, you also likely could see the bed arms bump around a bit if you went from high to low Z (command a -250mm move or something and watch closely)
removed the top constraint and when I jogged the bed up and down there was some very obviously wobble. like the threaded rod wasn't just sitting forward it was coming in contact with the two screws that hold the oldham into the printed arm piece as it wobbled back and forth.
totally disassembled the back z system. the oldhams were very 'sticky' still. I actually inverted the pom piece and reinstalled it into the two brass pieces, moved them around a bunch until they started to move smoothly with little friction. (if this is the root cause I'm gonna blame this cheap 3.0-3.1 upgrade kit they came in)
also inverted the rod too for good measure.
This is definitely getting worse, the rod is now rubbing against the screw when rotating.
I'm gonna sleep for now
@lemcott my wild guess is you are binding somewhere ie, rail is not going straight, the rod might be bent, or the rod might go on an angle as well.
that's the idea
but I'm not exactly sure where to start first so I'm kinda trying everything
Take out the leadscrew
Roll on table
If it rolls smoothly, then its not the lead screw.
That would be #1 for me.
yep, tried that. lead screw looks fine
I removed the pillow block to allow for some more movement in where the motor is mounted and there is a slight reduction in wobble but not all of it.
What about your top screw holder?
Is the rail extrusion a bit angled?
The middle as well can be angled left and right as well
Is the motor aligned well with the rigid coupler, and is the leadscrew set inside correctly?
I did mention that I thought the top constraint was pushing the screw forward.
This didn't really pan out, the constraint was doing it's job and the screw was now wobbling worse and coming into contact with the oldhams screws.
As far as I can tell I am installing everything correctly, how I prove that I don't know. Obviously something is incorrect but I don't think it's the standard things you find on any printer like stepper-screw couplings or else I'd see this issue on all my machines.
For now I'm troubleshooting the lower motor mount itself, I'm seeing the greatest reduction in wobble so far when adjusting the pillow block and mount
This is without the (back) pillow block entirely. Still visibly noticeable, not really affecting the geometry too much but hard to tell.
This rolls perfectly flat on a granite countertop
so I'm assuming there's gotta be something forcing this out of alignment.
But I've reset the top captive bearing and entire motor mount now
I see the wobble there in the lower middle... if that lead screw is straight, the coupling below is absolutely not. Swap it if you have extra, or with another lead screw's coupling and see which of the two is wobbling.
Or take out the lead screw AND coupling, and roll them on the edge of a table and see if the coupling wobbles
Short of buying replacements for the screw and coupler everything I can do to tell if they are flat says that they are flat.
Wouldn't it be that even if they weren't the constraint and oldham would counteract that effect?
Just to absolutely rule out an extrusion issue.
Also did a one wall cylinder right on top of the arms (or as close as I could get to the right one with my print head).
Wobble is seen most in the back and left arms, right arm is slight it is barely noticeable.
(note that there are a couple of really bad layers here and there, that is me actively fucking with one of the arms during the print.)
(and yes, retraction still needs to be tuned better. Not my worry right now)
It looks like that either the lead screw is bent or the coupler is not good, or its inserted a bit on an angle.
Totally replaced the back coupler (stole a known good off another printer) - no change
This probably should've been a fix my printer thread. Originally I just thought this would be slicer tuning.
I guess my biggest question is... Even if the screw and coupler are bent, so what? Wouldn't the oldhams counteract this? Why am I seeing wobbly motion on the lead screw below the bed arm but I'm just seeing any adjustment in the oldham?
Ordering another leadscrew will take time, but there's a chance that won't do anything because I feel like there is something else at play here
Yes but you are expecting too much of it, as if its bent, you have binding, and you have rotation that is not proper, so not only does it bind it can move less more just that tiny bit as well.
If its all on 90* and no binding then the coupler can move properly and counteract the leadscrew position when it moves so its always "flat"
Indeed, it is definitely a mechanical issue. Those oldhams can only correct for so much -- I think there is a 12mm(?) hole through the center, so if the lead screw with more than 2mm off where it was wobbling, it'd certainly be hitting (as it is)
Honestly, I'd start from the parts, inspect, and continue from there.
If it helps at all, you can see how I tested couplings for straightness on their own here -- https://www.youtube.com/watch?app=desktop&v=ZVf-GGLvtQQ
8Complex
YouTube
3D Printer 5x8mm Coupling Study
I was having a bunch of Z ringing and knew that my lead screws were straight (rolled them on a surface plate before assembling), so I was fairly sure it was my couplings between them and the stepper motors. I ordered up a few different types from Amazon with the intention of finding the best ones possible. Yes - they're likely cheaply made overs...
You can also check for things like how straight the stepper shaft is with a square/straight-edge. Once you install the lead screw, do it without an arm or the constraint at the top, then see if it is parallel to the linear rail, also if it wobbles
Then just add 1 part at a time to deduct where the issue begins (likely the culprit of the issue)
Really, the most important factor with wobble is that the lead screw is straight and parallel with the linear rail. Any deviation there is an issue. Even go so far as to put the bed arm on without the oldham and see if the lead screw is centered in the hole there. It could even be as simple as a twist in the arm where you attach it to the linear car
Okay, yeah, these lead screws are not bent. I don't have as fancy of a setup to confirm this but I have spun them up without the arm or constraint installed and rolled them on granite, no wobble. At the very least it's certainly within any variance that an oldham (should) be able to compensate for.
Once a screw is fully installed it wobbles while rotating.
Harder for me to confirm anything is wrong with a coupler, but I've switched out the ones I was using with ones from a known good printer to no change.
This is going to take fucking forever to troubleshoot if no one who's experienced this before can shed any light on it.
ugh. why does force_move always shutdown ratos. first command works like normal and then second shutdown the mcu and I have to restart firmware.
I mean, something is out of line, you just need to narrow it down to that one piece. From what I'm seeing, I'd remove the upper constraint and see where the lead screw is (centered in the arm hole or not). I'd then remove the arm altogether and see if the lead screw is wobbling out of the coupling.
Since there is a rotating angular misalignment, my money is that it's in the lower coupling, or even something as crazy as a bent stepper shaft. I really doubt it'll be in the arm or oldham as the angle would remain constant.
If the stepper was mounted on an angle, that angle it was bending the lead screw would be constantly one direction. Same thing with the nut/oldham -- they don't rotate, so the angle would remain constant.
This. Pretty much somewhere along the line its happening. If you remove the ARM on top, does the screw bend as well like that?
If not then its a frame issue as the arm goes too much back.
If not then its eithe the leadscrew, the printed part, or like @8complex motor itself shaft.
I am dying
Had to remove the entire bottom panel to get out the motor mount
I don't even know what to do any more
Every single test I have for the individual components show they are fine
But the second I put shit back together nothing is straight
I'm just so tired
from my best judgment
this is what I think is happening.
I think the back motor mount itself is tilted forward
fully assembling the lead screw, coupler and motor none of it wobbles.
installed in the arm and the screw gets pushed forward within the oldham where it is almost nearly touching the front screw that holds the oldham in
add on the top constraint on top of that and it turns into a wobbly mess
it only takes 1.5deg to push the leadscrew this far forward like what I'm seeing in person. this angle also explains why I don't see the oldham actually compensate for any lateral motion: there is none.
now
how do I fix this.
I've no fucking clue.
If this is true, then when you assemble it all on the printer, the coupling is allowing the axis of the stepper and the axis of the lead screw to be angled to each other. That angle in the connection would cause the wobble.
I suppose you could try loosening the stepper mount part, attaching the coupling & lead screw, and looking at the stepper mount to see if it appears tilted in relation to the extrusion or linear rail.
I don't think it's the coupler, I think it's the motor mount/how it is installed
ah, yeah
By design there is slop in the motor coupling, so when it clamps down it could be "securing" an angular misalignment between the motor and lead screw, which would be a wobble
I give up
Just gonna live with it I guess
Switching filament and enabling performance mode is no change, though the wobble is much less noticeable on darker materials
Yes, I'd say that sounds expected... It's a mechanical problem that isn't going to go away until it's mechanically eliminated
Yeah just ruling out the things I don't expect just in the off chance I might be missing something
I was really kinda hoping something about the stepper configurations was playing weird games since these are not LDOs and that's what the config is set up for. The performance profile would've at least changed the symptoms if not eliminate them if that were the case tho
Might just stick with dark colors and .3 layers. Hardly noticeable.
Honestly probably just switch to tr8*8. Even if they wobble it isn't as noticeable.
My understanding is that TR8x8 has too much pressure angle... the bed is able to drop when the steppers are not powered.
I'll say it again... you have something wrong in that specific Z drive.
If you've taken that arm off (left the stepper, coupler, & upper constraint), and it still wobbles... the most likely issue is that either your stepper is bent or your coupling's two holes are drilled off-axis.
Swapped couplings and it still does it? It pretty much has to be the stepper or your lead screw being bent.
I've totally unmounted and remounted the motor mounts, ensured the frame was square, totally replaced all couplings (known good split couplings with the red separator from another machine) and rotated the leadscrews to different z positions. literally nothing has made a difference.
one lead screw does look like it has a slight wobble to it, but it's absolutely nothing the oldham shouldn't be able to compensate for.
on a granite countertop it's near imperceptible, but I can force move Z up and down super fast or super slow and notice it once installed. still though, it's so constrained and the oldhams are there so... I really doubt that's the cause.
Sure, as long as the lead screw is in the center of the oldham/arm on average, but you showed a picture where it was hitting the side of the opening in the arm, right? Either too much wobble, or too far off center to allow the oldham to work
the problem with that is that it wasn't wobbling towards the screw, it was aligned that way. it was simply sitting too far forward. there was no lateral motion within the oldham.
The wobble would be perpendicular to the screw... it wobbles the bed arm side-to-side when it pushes on it with the wobble
I've mostly corrected this simply with taking off and remounting the back z motor mount. it sits more in the center, still no noticeable lateral movement in the oldham though.
The video you had taken before had visible wobble in the lead screw... was that the oldham that had no visible motion? If so, it must not be free-sliding enough to allow it... worst case scenario is if there is SO much wobble that it twists the lead screw nut, and that forces extra friction into the oldham causing it to bind. Not too likely since lead screw nuts have slop built in, but not impossible
yes that was all the same back z. that same leadscrew is totally flat on granite, when reinstalled it wobbles again.
that's when I decided to take apart the back motor mount thinking that it was angled forward or misaligned in some way. in this same step I rotated lead screw positions. no more wobble in the back z. no more wobble on the same leadscrew. slight wobble on the other leadscrew now that they are all fully installed but it is super slight. all lead screws sitting near center enough to the oldhams.
I can manually 'wobble' the leadscrews when installed and see the oldham is compensating for the movement, but that is very exaggerated. I still don't see very much lateral motion in them during printing.
Wait, so the wobble moved to where you moved the lead screw? Did you move it alone, or with the coupling and/or motor?
wobble moved to the other other leadscrew.
as in the same leadscrew did not wobble in a new position, and the leadscrew that went into the old position did not wobble. the other leadscrew had the wobble.
00-11-22 - wobble on 11
02-10-21 - wobble on 02
the original wobble on 11 was pretty bad yeah, shown in the video. this wobble is nothing
Have you considered that you possibly had terrible countertops?
yes, I also checked on tempered glass
Well, there are only two possible sources of wobble... a bent lead screw, or something causing it to bend
my money is leading towards the latter - but that would lead to either some sort of major part issue or installation issue and... I don't have enough info to really sus it out so I'm stuck hitting things one by one till something changes unless someone who's had this before chimes in.
wobble on a v-core is uh... not exactly well documented.
I've had multiple people tell me know "well it shouldn't happen"
like
okay, but it is. and?
that's kinda why I went into looking at other filaments and print settings, see if I can hide it enough for my purposes.
Not sure how having had wobble issues makes someone an expert on resolving it, but ok
I just mean it would be helpful to even know where to start looking when it comes to the unobvious things to check
sometimes that is where the answer lies, and if it isn't documented anywhere then having found it previously is the only way to know about it.
Well, it's basic mechanical debug to deduce what the cause is... not sure anyone could point you to a miracle solution, but good luck with that
don't take me the wrong way, I just mean that I have tried to debug the obvious mechanical solutions and we seem to be running out of things to try.
I'm looking at part replacements or someone with a 'oh yeah, I had this, try...'
my money is on part replacements
sadly
that's why I mentioned the tr8*8
I mean 4 would probably work just as fine
and maybe grab the mandala oldhams and ditch the cheapo 'upgrade kit' oldhams since they seem to be doing nothing
TR8x4 is what the kit normally comes with (they got rid of the 8's for the above reasons). I have run 2's for 2.5 years on VC3.0 arms without an issue.
In that video it is obviously wobbling, but I can't tell if the lower coupling is wobbling. If not, then it 100% has to be a bent shaft in the stepper or the lead screw isn't mounted into it straight.
Yes, that is exactly why it is so frustrating to troubleshoot, we are right back to where we started this conversation.
Every symptom points to a part being bent, but every test/replacement I am able to do shows that there is nothing wrong with the individual parts. The wobble is only detectable when all parts are fully installed - meaning it is more likely an issue with how the parts are installed versus an issue with the parts themselves.
And again, even with the above said, the wobble is so slight that oldhams should absolutely compensate for it instead of transferring the motion to the bed/part.
The only thing I have yet to fully replace are the lead screws and oldhams simply because I don't have more on hand.
I can run the risk of getting tr8*2s again but if I'm correct and nothing is wrong with the current screws nothing will change. at least if I get 4s or 8s the wobble will be spread out and less detectable. I don't give a shit if my bed drops when the power is off, I can't even guess why that would really be a problem.
Progress!
Progress is progress, at least
Honestly if I just slowed it down it probably be loads better
Clean the plate perhaps.
Issue is with DDEX you need some sort of wiping, like a tower or such.
this bed sheet is trashed, totally sacrificial for now
still need to figure out some way to get both nozzles to primeblob at the same time
also susi wasn't playing nice with setting up prime towers because I had a .6 and .4 nozzles defined, just switched them to both .6
I was really hoping to get the best of both worlds but there's no good way to do that unless youre only using one of the nozzles for certain features like supports.