What I've learned about your 3D printers (and mine)

3D printer extrusion issues have been bugging me for months. As I'm designing my own extruder, I've gotten a bit obsessed about figuring out what's going on. And I'm not talking about the common issues we already know how to fix. It's the subtle ones that only show up in special circumstances. Many of you helped me by printing test patterns and submitting your findings. Turns out it's more widespread than I thought. In this video I'm showing what I've learned so far, and in the next one I'll dive deeper into one of the issues and how I managed to patch it for my printers. Still got to film that one, so it'll take a little while.
Part 2 (measurement and fixes) • What I've learned abou...
Part 3 (intertwined waves) • The intertwined waves ...
Extruder test models mihaid.com/extrusion-test
Spreadsheet with the gathered data docs.google.com/spreadsheets/...
Patreon / mihaidesigns
Buy me a coffee mihaid.com/coffee
Instagram: / ​
mihaidesigns.com/​
#mihaidesigns

[@3DPrintingNerd]

This is AMAZING to see! I'm so thankful I'm subscribed and get to see the content you make. You're methods of investigation are great, and you're testing every permutation. I cannot wait to see more.

[@MihaiDesigns]

Thanks Joel, much appreciated. My other video goes deeper into the issue and there's even a potential fix.

[@josephbodhorn6511]

Isn't this the exact issue teaching tech was correcting when he did the vid with his voron 2, accelerometer, and input shaper on klipper? It's variations in ringing right?

[@hologos_]

@@josephbodhorn6511 nope it's not. It's inconsistent extruding.

[@MO-ProMaxDesign]

you nailed it, its based on a printer software problem and also the slicer software. there was a vid made by @GeekDetour he also came up with inconsistent extruding @@hologos_

[@FilamentStories]

What incredible research you've done. Outstanding. I can’t wait to see the next video. Totally awesome.

[@MihaiDesigns]

Glad you enjoyed it!

[@iXmerof]

that's a valuable content. fear not making more detailed ones on the case

[@marsgizmo]

your work is incredible 👏😌 I’m happy that you’re digging into solving this issue, especially that you focus on engineering facts, not wishful thinking, as it’s often the case on the internet 😅

[@BikZom]

I can see where mihai draws inspiration from

[@MihaiDesigns]

Glad you enjoy it!

[@kerndama]

I watched this video and was 100% sure it was marsgizmo :0 The voices are so similar!!

[@ClickItYT]

@@kerndama Yeah, like they were brothers. ;)

[@kerndama]

@@ClickItYT Are they?lol

[@marsgizmo]

Thanks!

[@GeekDetour]

Mihai could you fix it in the end? This design of dual-gear might have a non-linear transmission of torque, due to the design of the teeth, so it generates a "strong/weak/strong/weak/strong..." push of the filament and when you print a regular shape like cones/rectangles/squares/prisms... it creates a variety of Moiré Patterns. By the way, SUPERB RESEARCH!!!! Exhaustion of tests!

[@MihaiDesigns]

Thank you. Yes, that's exactly how I believe it happens too. I do have a patch for my printers, although I'm not sure it'll work with very flexible filament. A different design of the gears would be best, but also more costly.

[@alexanderrobinson3219]

I hadn't looked into the gears closely, what did you see on them to give you this idea?

[@cebrailaktas6427]

@@alexanderrobinson3219 Might have measured the distance between the gears' teeth or between both gears while running I guess ?

[@MihaiDesigns]

@@alexanderrobinson3219 I'll explain in detail in the next video (yet to be filmed)

[@julienboucaron9780]

May be more teeth in contact help: I wonder if the LGX Bondtech is less impacted

[@mlgisnotdead7752]

I have a suggestion, so hear me out. Maybe the gear that is fixed to the extruder motor is not concentric with the shaft, thus the filament is not pushed evenly through the hot end even if the motor spins at a steady rate. As the splined gear that grabs the filament may not have the inside diameter identical to the outer diameter of the shaft, and is secured to the shaft only from one side with a grub screw, this could be the source of the problem. Even a tiny deviation from concentricity is enough to create that effect. Let us know if you find out anything on this matter. Love your vids btw!

[@MihaiDesigns]

That would be a good subject for a test. Is not the case here, I imagine it could lead to layers seemingly not aligning well in specific situations.

[@simonquvang6073]

@@MihaiDesigns The bear project/community have come to that conclusion though.

[@RFi731]

If that is really the case, then i got an idea to fix this. How about we add something like extruder concentricity compensation value into the firmware. Something similar to how linear advance compensates the filament flow, and how input shaping compensates for frame flexibility.

[@expression3639]

Look up MirageC, he did a video on this issue.

[@cebrailaktas6427]

Investigating, then analyzing, then finding a solution and THEN only publishing results : this is the right methodology to adopt 👌🏼 Can't wait to see a lot more of your videos in the future, thanks for your content Mihai !

[@MihaiDesigns]

My pleasure!

[@cebrailaktas6427]

@@MihaiDesigns And the longer format is a delight, just saying 😏

[@mikeneron]

But did you try kicking the printer? Seriously though, that's taking investigation to a whole other level. Hope you are able to get it figured out. You will become a 3D printing legend.

[@alexanderrobinson3219]

So refreshing to see an engineer's take on 3D printing rather than only printing for fun or even walking blindly into projects. Your content is so much more satisfying because of your approach please don't stop🙏

[@joemoment-o1275]

I'm glad he could satisfy your needs princess. Please let us all know next time when you need something my sweet.

[@Antrim3d]

You Sir, are a wealth of knowledge and a gentle speaking teacher. Thank you for all you're accomplishing in 3d printing environment and sharing your findings with all of us! Keep up your passion!

[@mungbean345]

Wow. I've never done any 3D printing and may or may not ever own one, but I subscribed because I liked your cheerful short videos, and I'm glad I did after seeing how much effort (and all those careful labels!) you exerted attempting to solve this mystery. Bravo!

[@leonhardwolf4626]

I've been following your work for a couple months now... I'm happy to see your subscriber numbers explode 😍 love how thoroughly you worked on this! Stay awesome!!

[@MihaiDesigns]

Thank you so much!! You stay awesome as well! 😃

[@xmodalloy]

I've got a very mild case of this pattern when printing with a totally custom build that does not have a dual gear extruder. It's running a drive wheel straight off the motor shaft and a simple ball bearing pressing the filament into that. Direct drive (not bowden) with a very well constrained and short filament path. H bot gantry system with standard 1.8 deg/step motors everywhere. Slicing with Repetier Host using Slic3r.

[@MihaiDesigns]

Can also be other reasons to get a somewhat similar shape. See my other videos on how to measure and determine where it could come from.

[@riyadhgany]

You’re phenomenal. I can’t believe how fast you’ve grown your following. No surprise there though. I’ve been following you from Reddit.

[@richardr3178]

Very nice in-depth investigation. Keep up the great work!

[@Loneman_OG]

All this research you're doing on behalf of the community is very much appreciated. :thumbsup:

[@MihaiDesigns]

Glad it's helpful.

[@iiianydayiii]

Thank you for saving others so much time, effort, and frustration. It's very kind of you.

[@FloRian-xr3ug]

I am really looking forward! I have the exact same issue and almost replaced everything on my v2, but still got those lines / patterns.

[@chrisdixonstudios]

Vibrations come from printhead, cooling fans, any moving part that is not dampened, loose or generates frequencies like sound..each can contribute to fine detailed extrusions of hot plastic "woodgrains". Match the vibes and see where to fix. I wonder if sound absorptioncan be tuned to temperature and type of filiment? NOT impossible to tune a fork..however you can't tuna fish, hehe. There is indeed a limit to print details depending on each machine's absorption capabilities.

[@TheBekker_]

That is a lot of tests! awesome work! looking forward to the next video, even though this is bit of a tease :D

[@andrewroy9263]

Jesus christ, now THAT'S experimentation!

[@StJohnny22]

I just got VERY EXCITED as I could see my prints having this same problem on some models and smaller layer heights. Larger layer height seems to mask it more. Either way, I am subscribed now! Thanks for digging into this!

[@rustynutsnboltz]

Your investigative efforts are mind blowing. Thank you.

[@tuegeneumann9961]

Wow! Just wow! That is an unbelievable amount of work you have put into this. Even more than CnC kitchen. Seriously keep this up just don't over work yourself! Absolutely amazing!

[@MihaiDesigns]

Thank you! Will do!

[@ing.resenas7345]

I I have a home made printer with a board that I made and my own firmware. I will test this prints! I spend a lot of time solving similar problems. Very good job!

[@thehighwayman78]

Fantastic troubleshooting!!!!! Hats off to you sir!

[@likevvii]

Thank you for actually testing and sharing. Extreme appreciation. I have switched to 5-phase steppers from oriental motor. Their motors can reach 1000steps/rev. I never have grains/stripes/waves. Clearpath servos also work great, but they have a little latency (2ms). Super smooth walls. The only thing I have now is inconsistent extrusion from sudden change in flow depending on the part geometry as I run near my extruder max flow rate.

[@MihaiDesigns]

Flow change does affect both extrusion % and print texture, especially if you go above 1/3 of the rated max flow rate. This is a real problem for me as well. My only solution is to go with bigger heating blocks. Even my super volcano is not enough in my experience. So I try to keep the volumetric flow constant where possible. Of course, there's always slow down at corners so there will be some variation.

[@WildIppa]

Not sure how I got here, I don't even own a 3D printer, but good work my dude.

[@redline6871]

Wow going to say this is my fav video you have done yet good work sir!!!

[@jamespray]

I ran the prints and submitted to your survey tonight. Looking forward to your findings!

[@ytivarg5371]

Well done. Im a newbie and thought thenwood patterns were normal. Im excited to see your findings and trying to adapt them to my ender 3 pro

[@pauldecastro1523]

Good job! i admire the Quality you are pursuing!

[@BillBaran]

Reminds me of an article I saw about 10 years ago dissing 3d printing when it was first gaining popularity. "The secret about 3d printing they don't wan't you to know - no two prints are exactly the same."

[@siwiskate]

Me likey! Cannot wait for the outcome of all this extensive and excellent testing

[@Creative_Electronics]

Really good and interesting video!

[@JRPW]

Wow so much work. Thanks all the testing.

[@MihaiDesigns]

You are welcome!

[@medyk3D]

Wow you made a gargantuan work on this one. Superb!

[@david-wtf]

This video was so great!

[@sincity337]

I am anxiously awaiting your next video!

[@MihaiDesigns]

Coming in several hours.

[@siwiskate]

Me likey! Cannot wait for the video with the outcome of all this extensive and excellent testing

[@MihaiDesigns]

Coming soon!

[@siwiskate]

@@MihaiDesigns watched it already, kudos to you sir!

[@Adam-my8qx]

You sir just earned yourself another higher tier patreon, keep up the good work.

[@MihaiDesigns]

Wow, thanks! 💗

[@mah6786]

I too tried to solve this issue by following the advice if someone online and bought a Bondtech extruder for my MK3S. Didn’t help at all, waste of money and time. Thanks for solving the mystery.

[@markmacherey4744]

incredibly detailed diagnosis, and incredibly frustrating... nice work..

[@slypork6178]

I had a similar issue and did dozens of tests and replaced a bunch of parts, updated firmware, and tuned a lot of stuff. Nothing improved the quality of my prints until I stopped using my Cura profile and restarted with a default Cura profile. It improved the prints dramatically, despite having the same settings as the Cura profile I used for years

[@akshaylalla4901]

Your content is amazing! Great stuff!

[@MihaiDesigns]

I appreciate that!

[@billyjoe3309]

We all love you Mihai, keep it coming brother!

[@MihaiDesigns]

💗

[@taoistflyer]

Incredible reaserch, the fact that it's a consistent pattern and not random noise should be a big clue, and seeing how it's so consistent across printers and setups says it's probably related to something very basic in 3D printer designs. With some math you can measure the distance between the peaks and troughs compared to the speed of the print head and determine the frequency at which the inconsistent extrusion is happening. The cooling fan is too high of a frequency to be causing that but something else moves spins or vibrates at that specific frequency and an imbalance in that is most likely what is going to be causing the problem. The fact that the extrusion is so consistently inconsistent in a way that causes a pattern is a big clue. Something is causing the printhead to move or wobble out of alignment at a very specific and consistent frequency.

[@1MoreTurn]

Lots of work you've done. whew.

[@makes3dprint468]

Very informative! Great research which tells me that the better is the enemy of the good ;)

[@mariokuntzag1237]

Genial, so viel mühe und Energie!

[@FakcioR]

It was also happening on my Artillery Sidewinder with single drive Titan clone, but more elongated. It was caused by eccentricity of small motor gear. It was pronounced when printing with 0.8mm nozzle, 0.4 layer heigh and 120% flow. I reduced it by putting adhesive tape on the shaft, now there is no play of the gear, so after fastening grub screws it does not move to one side.

[@billyjoe3309]

Mihai, I need your swapable hotend. My printer just broke, If I had a swapable solution, I could just take off the broken one and replace with another. I can't wait for your swapable to be available! :(

[@miiklaa]

Awesome work you did. 🧠❤️

[@RaphaelBeguin]

I have no words, you're awesome 😅 👏

[@xPixelationNationx]

You’ve got no quit in you. I would stop trying ages ago

[@culturecanvas777]

Great channel, super useful. New sub. Love it. 👍👍

[@DanielCaaz]

Totally meant to be a part of this research since i was having this issue and thought it was related to other issues I was having (apparently not) Moire is insanity so i look forward to seeing the conclusion to this

[@truegret7778]

Unfortunately, I interpreted "What I've learned about yout 3D printers (and mine)" as meaning you had a solution to the wood grain anomaly as well. My Creality-CR10S does not experience this issue. Did you investigate the relationship of the infill (pattern and % overlap) to the walls?

[@MihaiDesigns]

Yes, my other video is now out. Also explains why Creality printers don't have this.

[@labemolon5739]

This is typically caused by vibrational resonance that becomes amplified over time. Printer components and even the sturdiness of the desk/shelf the printer is mounted on can affect resonance. Another youtuber (Teaching Tech) covered this well recently and used an accelerometer on a Rat Rig, with Klipper Input that contains specific values in the firmware to address this and came out with amazing results.

[@mouseFPV]

I had this issue when switching to DD on my e3pro with the 4.2.2 board (loud). TL smoother on E fixed the problem. Will likely just get a better board soon.

[@pr0xZen]

Wood grain/rings often come from extruder gear issues. Commonly hobbed gears that are non-centric, either by manufacturing defect, bad mount to the axle, or the axle (and thus the hobbed gear) is at an off-axis angle vs the filament path and idler. Obviously the same goes for the other gear on dual gear extruders. The stepper axle itself can be bent/off-axis, but the far more common culprit is the extruder gear(s) and/or their mount. It can be difficult with with many direct drive constructions as they're often enclosed, but a good troubleshooting step is to snip 2 sensibly long pieces of filament off the spool, then using a hair dryer or _very_ gently a heat gun, straighten them out from their curve off the spool. Something slightly (!) less rigid than PLA can be favorable here. Next, drive the stepper to run the 1st filament piece through the extruder, without the exiting filament rubbing/bending on anything _NOT_ going into the hotend). Run it really slowly. Observe the extruder gear(s) and idler at work, really up close. Is there _any_ kind of non centric movement? Any movement on the tensioning lever? Does the distance between the center of the screw/pin/axle of the gear(s)/idler, and the outer edge towards the filament, change during rotation? Is there any movement in where between the gears/idler the filament rides? Lastly, on the filament itself. Is there any curve to the exiting filament? Are the grooves from the hobbed gear(s) biting into the filament, precisely perpendicular to the length of the filament? Is their depth and spacing consistent throughout the length of the he filament? If dual gear, is the depth, angle and spacing along the length of filament consistent, identical on both sides? Next, do the same with the other piece of filament, but at higher speed. The speed when printing these parts, ideally. Watch and inspect the exiting filament for the same as above, but this time also look for signs of slipping, grinding or missing steps.

[@MihaiDesigns]

Excellent tips @pr0xZen. I did check for some of these aspects even though some things I can only measure with limited precision, but I find your test procedures interesting. I haven't gotten a test for non-centric gears. It would require a much larger print, but is something that can be designed specifically to detect this. I've noted it down for a future experiment.

[@pr0xZen]

@@MihaiDesigns In terms of stuff like the distance from center to edge of the gear(s) and idler, you don't neccessarily have to measure it precisely. If you watch the gear(s)/idler up close when running filament through them at a constant speed - especially when moving at speed - if it's there it should be visible as an oscillating motion. You can sometimes also feel that oscillation through the tensioning lever if your extruder has one, if lightly holding a finger on it. For the hobbed gear "bites" along the filament, _inconcistency_ in the "bite marks" is usually not too difficult to observe when inspecting it by eye afterwards. If inconcistency is caused by non-centric or improperly cut/unevenly worn hobbed gear surfaces, the inconcistency will be a repeating pattern spaced roughly every 10-20 bite marks; there's not that many "teeth" around the sircumference of those gears. With dual gear, look for the same _inconcistency_ and repeating pattern on both sides. And, you don't have to count every mark to compare to the marks on the other side. Just see how they line up with the opposite one. Then spot-check by eye every 5-10 mark. See if there are any sore thumb spots popping out at you, or if the spacing vs the other side, changes significantly over distance. See if the grouping tightens and expands, vs the opposite side if dual gear. A tiny bit is not a significant issue, but if it varies by 5-10%, then that 5-10+% repeating deviation is gonna be attenuated *a lot* at the plastic squiriting end, unless you're printing with a 1.25+mm nozzle. Heck, even if you ran a 1.7mm nozzle, non attenuated 5-10% repeating deviation will still cause visible artifacts. And, it's gonna induce repeating artifacts in the print too, although in more visually recognizable patterns in prints of simpler geometry, and on direct drive, because the extruded distances are very repetitive and there's not a tonne of retraction to offset or "randomize" the effect on the squirted plastic. If you've ever changed your extrusion factor or adjusted e-steps on thre extruder, then you know how much of a difference a few percent change at the extruder end, can mean at the nozzle and print part end.

[@tvathome562]

WoW Seriously impressed by the methodical hard work and detail to identify this problem, I would love to help but I cant see any inconsistent extrusions from printer, probable as currently in pieces in the spare room :)

[@redtidetogo]

Problems like this drive me crazy. Thanks for your dedication to trying to solve these issues. Have you made any head way on this issue? Perhaps it's caused by the elasticity in the belts? I wonder if going to a chain driven or another more pricise x axis drive could solve it. Another idea i had could be the consistency of the nozzel temp, i think slight fluctuations in nozz temp could cause a color change. Best of luck!

[@MO-ProMaxDesign]

Perfect you did a great job.. this was also the reason i created my own monster based on a ender 5 pro. i try all but almost no problems with my printers. voron 2.4 a little wood texture but my heavy modded ender 5 pro has no problems at all. the ender 5 has core xy and a 4x linear rail z mod, btt1.1 PI TFT50 Touchscreen + 8 TMC2209 Drivers, i used my own made sofware. 10 mm bed with heatmod up to 200c 330x330 buildvolume because off the z mod i got more space to make a bigger bed. hotend is a micro swiss M2601 with a ProMax design cool head (becomes available FOR FREE (thingiverse) after 3 month test is done. all of this is my own creation put together on an ender 5 pro base

[@jkersting]

Great work got my sub.

[@atombell991]

KZbin just recommended this gem of a video to me on the home page (probably the other guys in the 3d print space I subscribe to leaving nice comments). Immediately subscribed 👍

[@john-lenin]

It’s a characteristic - almost never a problem. It’s like complaining that wood has a woodgrain problem.

[@lunchbox1341]

its not a characteristic, layer lines are a characteristic, the printer not feeding the plastic at a constant rate is not.

[@saltysteel3996]

Not getting any of these patterns with the Micro Swiss direct drive and hotend on my Ender 3 Pro, running the Creality 32bit V4.2.7 board with Marlin firmware, using BLtouch, glass bed with solid bed mounts, dual Z motors, and gantry brace rods to stiffen the frame. Been mostly printing CF nylon, PC, and CF PC.

[@seanmostert4213]

Interesting, thank you for your thorough research! I have some observations on the subject that may be of interest, this is of course based on my on meandering experience combined with a few sound physics principles. I suggest that the pattern may be simply created by the gravitational force of earth passing through a fluid (molten plastic), which then sets as it cools and this locks the image into the material. If you want to confirm if the pattern is the same as that found in gravity, simply observe water as it’s poured from a jug into a cup, it has the same pattern, so does timber grain, or water at the beach as it recedes back to the ocean etc. there are many examples of this pattern throughout nature due to the common force that creates the pattern. Not sure what you can do to overcome the issue if the cause is as I suspect, however, maybe printing at a lower temperature will improve the finish as the material will not be so pliable? There’s always a trade off to any decision I guess. Anyway I hope this offers an alternative perspective.

[@Max_Marz]

As soon as you said "dual gear extruder" I knew the issue, as a machinist we see this on the lathe and even in the "lobed" nature of bearings creating diameters that are not truly round. Bless you man, so much respect for your analytical approach.

[@MihaiDesigns]

Thank you! Sounds like we need more people like you in this 3D printing hobby community. I know next to nothing about machining and am sure experienced people could figure some things out in no time. It being a hobby market for the most part, I can imagine it wouldn't pay much, so skilled engineers look for other domains to use their skills.

[@doritosicecream2820]

dam..... the amount of work that went to this video event that i am just a begginer from 3d printing but you deserve a sub and a like 😅

[@3DJapan]

One of my printers, the EcubMaker ToyDIY, always has lines that look like ripples in water, on the +X side of the model. It has always had that problem and I never figured it out. What makes this more challenging is that the machine was used at a convention before I got it and at least one part was broken so it may have been mishandled by a previous user.

[@dinosoarskill17]

fantastic stuff! SUBSCRIBED

[@MrTartans]

Newbie’s aren’t as critical as we are with quality of prints. What we see, they would not. Not sure what kinda message this gives to to 3D community. Personally I’m really interested in your findings. Nice work!!!!!!

[@MihaiDesigns]

Indeed I imagine many people wouldn't care about these details. I'm more hard core so I nitpick.

[@justinantonius2823]

Very interesting I want to test this on my ender 5 pro I never notice any patterns but I use a Bowden tube so retraction is set high due to the springing of the tube

[@super_slo]

Wow. And that advice at the end about not printing the test samples seems like excellent advice 🤣 Can't wait for the next one!

[@Twin-jl1bi]

Could you do a video on your ideal 3D printer setup

[@MihaiDesigns]

I might try. There's still a lot I don't truly understand about 3D printers, but am learning more every day.

[@Dbm_-oe7zv]

I have a PRUSA mk3s from 2019 and have experienced this issue from day 1. The patterns displayed on the side of prints had me concerned for the longest time that I had set up my Prusa incorrectly. It was not till I looked at the plastic parts printed by PRUSA themselves did I realize that their print farm printers are having the exact same issues.

[@MihaiDesigns]

Good observation! I replaced my stock printed parts so long ago that I didn't think of this. Check out my other video where I have a potential fix for the MK3S extruder. Someone with a print farm tested it with great success for all their MK3S printers.

[@noway8233]

Well , they printed in a Prusa , by the way, the looks pretty bad for 1000$ printer...

[@digibluh]

even prusa denies this problem.... its so odd. yet here we are. the og extruder on the mk2 was perfect.

[@AppliedMathematician]

Do you work in quality assurance of 3D printers? That is impressively thorough.

[@MihaiDesigns]

No, I'm a software engineer, but I'm obsessed enough with details to dive into this stuff and got into 3D printing over the last 2.5 years.

[@moldoveanu8]

Esti boss, buna observatie Mihai.

[@tomaszkorytkowski1399]

I think you need to talk with CNC Kitchen (and others) to solve those issues together

[@nickbuhling]

That would be super interesting to see indeed!

[@MihaiDesigns]

I'm not sure he's having this issue, but would be interesting.

[@tomaszkorytkowski1399]

@@MihaiDesigns I think he would have to look close at the prints like you.

[@chrismorrison9140]

I would stay the course, go solo. This is your thing, no need to get anyone else involved. I appreciate seeing a different view of 3d printer concerns and you provide that. Thank you!

[@tomaszkorytkowski1399]

@@chrismorrison9140 English is hard for me sometimes when i want to say something more complicated and it doesn't have the same meaning as in my mother tongue.

[@ReverseDeadshot]

Have you tried loosening the bolts. Squaring it up and re-tightening them. Them sitting the whole printer on a concrete flag sat on a towel to reduce vibration. Thats how i fixed mine. I still get layer lines if i push the speeds but no longer get the wood look.

[@_v_e_c_t_o_r_]

Simply awesome 😲

[@tiestofalljays]

I missed this, I’ll print up the models too :-) I’m curious.

[@philippeleca8239]

Did you check the good centering of your pulleys ? Did you check the cleaness of the teeth of your pulleys, and of the nuts+threaded shafts of the vertical axis ?

[@shaker2321]

Did you tested changing the whole extruder ? even to a bowden type? and also changing the extruder gears tension (tension applied to the filament), regards.

[@Rustaceann]

6:06 did you check for Z-Axis Wobble? Run your gantry from lowest to highest point and check to see if it moves side to side or if the rods deviate from center while it's moving.

[@DaveCharnock]

Could the patterns be down to some kind of harmonic between between the length of filament put down per layer and gear circumference? I'd be interested to see if random seams make it disappear.

[@alengusic4259]

How long did this changes and variations testing took, I am as an engineer impressed, we are working in our daily jobs 😂, have 3 Vypers the 4th is for testing, after a long day at work have no motivation to research artifacts 😅, but I am speechless about your amazing work, working on a screw extruder, my hypothesis is the backlash in the gears, maybe in the Bowden is enough filament which compensate for the backlash, as you say stay awesome 😊

[@bluedominator650]

Wow, you went through a lot !!!

[@bluedominator650]

So, what’s the best set up, configuration to have the best results ?

[@PaxDrakonis]

Thanks for putting so much effort into this Mihai, if I could add an idea... if converting the single gear helps the problem (presumably because it alters the pattern of torque transmission) what do you think would happen if we designed an extruder with multiple sets of dual drive gears? Imagine one with 2 or 3 matched sets of dual drive gears, at offset angles of the filament path, with each set of gears having different number of teeth for both drive and filament contact but balanced for the same rate of feed. Could we somehow achieve 99% smooth torque transmission?

[@MihaiDesigns]

Great point! Multiple gears with offset teeth would reduce the issue most likely (although I have no way of testing). If you take the number of such gears up to infinity, what you end up with is very similar to double helical gears which as I understand it are better in every way from these spur gears, except they cost more to manufacture.

[@PaxDrakonis]

@@MihaiDesigns I think we need to crowd fund a modified Bondtech LGX for you to play with!

[@z3r0f84]

I used a TL Smoother board on the extruder motor for my ender 3pro fixed this issue for me

[@zacbavdek2818]

If only I'd found this sooner, I have a cr10-v3 and a flashforge 3 printer I would have happily tested with

[@Chrom35kull]

I've wondered whether using a low pressure air pump to gently blow into the bowden tube would help with friction and also reduce thermal absorbtion when the extruder retracts. It would only need a t-piece near the motor end to prevent extra mass or movement restriction on the extruder. Unfortunately my direct drive printer died and have given up repairing it (time for a replacement).

[@Vez3D]

Have you tested the LGX extruder? Might be worth a check. Different gears ..bigger with more grip. :)

[@MihaiDesigns]

I have not, but I intend to. I'm really curious to see how those big gears change the game.

[@air21511]

have you tried Power Filter? This issue can come from your power. You could in theory power your printer with Lab grade 12/24v power supply instead of power supply in your printer. Even better would be to connect power supply to the battery and than run printer from the battery - it would further ensure perfectly clean input power for the test purposes. (Sidewinder heated bad needs replacement - look at Fula Filament sidewinder bed upgrade - I have Artillery as well, but Genius with LGX + ACE copper head ultra short filament passage + CHT nozzle 0.4mm + Klipper with FluiddPI. I just finishing setting those upgrades, so can't comment on results, but have high expectations - ALSO, Klipper PID control of the bed should work a lot better. Also have you considered if those issues could be due to nozzle temp variation - poor PID tune of the heater similar to the bed issue on sidewinder. We really need to move our SW into Sinewave modulation instead of basic on off squarewave control logic for PID output.

[@johnhansen1684]

I have wondered about the Slicer post-processing results if the codes themself sometimes can be to play for poor performance. Would be very interesting to check to codes in an NC code verification program and see if there is anything of interest. I'm sure most cam programs convert a cad model into a mesh even if it doesn't show it when programing so I think it shouldn't be a problem for a slicer to make printing paths on an stl/mesh however the tolerance might be a lot smoother in a real cam program but might not be anof to make any difference.

[@3dPrintingMillennial]

I didn't do testing as widely as yours but I spent an entire day doing something similar. In the end, the only solution that worked was printing EXTREMELY slowly on the outer wall. I tweaked practically every other setting and still ended up with some type of pattern. After testing in vain, the research I found is that the cause is generated from the stepper/signal and resonance. There was a 10+ year old video that I came across that made sense. I'll have to find it again and repost it. If anyone can find a solution, I'm sure you will!

[@Karavusk]

My pattern looks a bit like the one at 1:09. Flsun QQ-s pro delta printer. I heavily modified it (duet 2 and many other things) and I tried a fake titan extruder, an original titan extruder and a fake bondtech extruder (all with a bowden setup). All of them have this pattern. I think relocating my extruder to get a shorter bowden tube helped a bit but the extrusion is still far from "perfect". To be fair you only notice this on flat surfaces with a strong light. Did you try out a 0.9 degree motor for the extruder? That is definitely on my to do list. At least I think this should reduce the effect unless it is caused entirely by gearing. Do you have any data from an extruder like the zesty nimble? With a 30? to 1 ratio mayhe it is smoother?

[@MihaiDesigns]

Wow this is super interesting to learn about because I only had one data point from one clone Titan so I did not include the original Titan. On one hand we need to consider that it might not be the extruder that's at fault. On another hand, if it is the extruder, then that's really interesting. Yes, flat faces are required, so it is in a way an occasional issue, but as we print more and more mechanical parts it shows up more often. I did try 0.9 degree motor and a 3:1 geared motor as well. I saw zero change. No data on the nimble.