GHOST BUSTING! A systematic approach to stop ghosting. also, should printers have shock absorbers?

I got called out in my last video about fuzzy skin for having a bad printer, or at least in desperate need of tune up. Serious ghosting issues in my prints.
So in this video I go through several settings and ways of tuning your printer to get rid of ghosting or ringing with a mix of hardware and software tuning.
This includes belt tightening, speed changes, wall order, printer mounting surface, alternating wall directions, flow changes, and trying to dampen the motion system.
Why is no one dampening the motion system of the printers with physical dampers?
This was all done with my very old Lulzbot Kittaz I built in Oct 2014 with very minimal changes. If you have an older printer without a lot of the modern features, this could really be a great option to look into.
__ Some of my favorite 3D Printed Accesories __
Kasa Smart Power Strip: amzn.to/49Uq4Uc
Blink Pan and Tilt Camera: amzn.to/3y2BFDt
Nozzle Cleaning Needles: amzn.to/3Ults4S
NoClogger Extruder cleaner: amzn.to/3xZeIkB
Fulament 3.0 Print bed: amzn.to/3WoGrp6
Slice Plastic Repellent: amzn.to/3JI4W96
Slice Torque Wrench: amzn.to/3UEQXqI
Slice Silicone boot: amzn.to/3UI0TA1
Deburring Tool: amzn.to/44klc9L
Brass Brush: amzn.to/3UpomEI
Plastic Razor: amzn.to/3Wm3SiM
* Above are affiliate links *

[@willofthemaker]

Hi everyone, if you are interested in the model used in this video, I just posted it to Thangs.com Check it out: thangs.com/designer/WilloftheMaker/3d-model/1033464

[@3dexperiments]

Dampened and non-dampened belts with and without input shaper would be a very interesting test. Great video.

[@willofthemaker]

Thanks. I think dampening belts would help reduce the need for input shaping and processing power while still getting good results. Especially if there is proper dampening values. I had a whole talk about it but cut it bc the video was already getting too long. Additionally having the print head be dampened from the carriage would make a bit impact.

[@fredhamilton1701]

@@willofthemakerDamping definitely decreases the *need* for input shaping (and thanks so much for this video - I think you really did discover an aspect of printer tuning that’s been largely ignored), but if you want the best possible print quality, you should still use input shaping to take care of the rest of the ringing. It won’t have to work nearly as hard and the results should be even better. I’m not aware of any downside to using input shaping.

[@willofthemaker]

@@fredhamilton1701 thanks for the feedback. I think the only real downside to input shaping is it adds cost through extra accelerometers and higher requirements on processing. Adding a simple damper can help in that regard, especially on older printers with 8 bit boards. That being said, input shaping was definitely a big advancement in the industry for quality and speed

[@charlesstaton8104]

​@@willofthemaker input shaping, processing power, >8 bit boards, these things are all cheap now. The amount of time you spent on this is several times what it would have taken to install a 32 bit or 64 bit system with input shaping and been done with it. Something like a SKR Pico is cheap ($35 on amazon, $19 on AliX) and 32bits. Add to that a BTT PI ($34 amazon or $10 on AliX) and a ADXL345 ($20 amazon, $8 AliX) and now you have massively improved stepper drivers, input shaping, you'll be able to run way faster, smoother, quieter, better quality prints, and a better printer overall. You'll be able to access your printer remotely and monitor progress. Add a USB webcam and you'll be able to actually *_see_* the print progress from the comfort of (anywhere) and be able to shut it down in the event of calamity instead of being tied to the room where the printer is. With >=32 bits and >100MHz you will find yourself no longer even giving a thought to "processor budget." You will thank yourself for doing the upgrade. If you think of that as a "sell-out" ("just throw money at the problem") solution, I get that, but I don't agree. It's not like getting rid of your Camaro and getting a tesla because you can't get the carburetor tuned right. It's more like getting EFI and a supercharger for your Camaro. It's typical tinkerer stuff, fair game, no judgment. Nobody will think less of you, and if they do, they're the 3D printing equivalent of the "wannabe artisan" woodworker type that drills holes with a sharpened stick and bowstring because even a hand auger is "cheating."

[@GlennBrockett]

Rather off topic, I know, but Klipper has a way to compensate for the resonance in your printer. However that should be the last thing you do, after your basic tune-up. Tight belts, mass, and lower speed can reduce your ghosting by a lot by themselves. I don't use any foam or other decoupling device between my printer and the table it rests on, but the table is VERY solid. Only by cranking up the speeds and accelerations will it even move. (3500 mm/sec^2, and I cannot remember the print speed.) Placing it on a paver will lower the frequency and amplitude of the ghosting.

[@willofthemaker]

A lot of people have recommended adding Klipper to the machine. I may need to look into it at some point. Part of me likes keeping it stock purely for historical purposes and it is very reliable. Print quality is pretty good (relatively) now with these changes too. I was always a bit skeptical to how much decoupling the printer from the table top would make but it does have some effect. I guess the different vibrations will effect different parts of the printer differently.

[@GlennBrockett]

@@willofthemaker It all has to do with the resonance and moment inertia. If you have a heavier total mass, the frequency goes down. Adding reinforcements to the frame corners will also reduce the amplitude of the vibrations. If you are running printers concurrently, decoupling the printer from a common plate will remove the outside vibrations. Resonance is an interesting topic.

[@igiannakas]

That is the way to go. These settings are great but the solution is to dampen the vibrations at source.

[@GlennBrockett]

I like the idea of printing a ruler on the test piece. It can make measuring the ghosts easier.

[@willofthemaker]

Thanks, I'll upload the model soon

[@yolo6741]

This is a great video, I especially like your systematic approach with the tables to compare and the special test pieces. I wonder a little about some comments like" you have a poor printer" or "with clipper you nowadays have not such problems". First of all, it's a good idea to get the best out of the hardware you have. Also, from my point of view, it makes sense to get to the bottom of these problems like ghosting and understand them. Klipper does a similar thing in the end to avoid these effects. You shouldn't let know-it-alls get you down, I think both the approach and the execution of these videos are very good. And even if the fuzzy skin video had the idea of compensating for a few shortcomings, in the end it is a great video with the result of using the technology to get a different kind of printing surface. Please keep going creating such great content.

[@willofthemaker]

Thanks so much! Very kind words. But ya there is a lot of great new tech that almost makes this moot. But people still have older machines or cheaper machines that might not be so feature rich so still good to know some of this stuff. I've had to focus on a couple other things (starting a business selling 3d prints I create) but should try and get back to video production soon!

[@end_of_6]

It looks like one of the issues that your printer has is too low of a current on the x and y motors. It looks like you have layer shifting that could be resolved by increasing the current to your motors. Since you increased the tension of the belts, I thought that would have resolved some of the issue, but you might want to try adjusting the motor current. I hope this helps.

[@mongini1]

Next up: Put an ADXL on your Toolhead and Bed and fine tune resonances, cause your mechanical parts are at their peak already. If you dont have klipper yet: get it and set up input shaping... this killed all my ringing on my Ender 5 which got a CoreXY mod. Printing at 200mm/s now without artifacts. Only some cooling issues on smaller bits ^^

[@willofthemaker]

So many people are suggesting this i might just have to look into it. I don't think my hot end could keep up with those speeds but even getting reliable quality prints at above 100 would be a win for sure

[@mongini1]

@@willofthemaker even if you just go at 100mm/s that will be a big difference. I switched to Klipper about 5 years ago and never looked back. The fact that you just exit a text file instead of compiling a whole firmware every time was reason enough for me ^^

[@willofthemaker]

@@mongini1 ya definitely a lot of advantages. Web interface for sending and monitoring prints would also be really nice and convenient.

[@mongini1]

@@willofthemaker so what are you waiting for? Order a Pi already 😁

[@willofthemaker]

@@mongini1 Truthfully, just life gets in the way. too much going on some times 🤕But Im going to try and find a way to squeeze it in. Theres also a few videos I want to do before that happens. One being a comparison of this old printer and my new Wuxn WXR could be interesting to see how things have changed.

[@four_0_four]

Next video - Klipper on the one of the oldest printers on KZbin? Go for it!

[@willofthemaker]

So many have suggested it, I think im going to have to haha. Not sure about next but sometime in the near future.

[@josephpk4878]

Jerk Setting = Your buddy is walking by and as he does, you grab him by the arm and yank him toward you. It's more of how quickly the head changes speed after it prints a corner. Another helpful video.

[@willofthemaker]

What a jerk! haha thanks

[@dieterbulcke]

Hi. I watched this and the previous video (about fuzzy skin). I think you have a really poor printer there. In my experience even an old stock Ender 3 without much tuning will give you better results. While your video's are informative (they do point out causes and possible solutions for print quality problems), I don"t understand why you put this much effort in getting this poor printer better. And "better" is still not good at all. If I were you, I'd ditch that printer. Use it for parts on other builds, it will serve you a lot better. I tore down an old and really bad Anet A8 and built myself a Prusa bear clone from scratch. Learned a lot doing so, redesigned and modded it heavily, also Klipper-ized it. This now is by far the best bedslinger I have and it's quite fast too - for a cheap bedslinger, that is ;-) How does your printer compare to the other one you have (timestamp 12:15)? I can't identify it, but my guess is print quality is superior. Sorry for being critical, but I'm sure you'll make better videos with better machines.

[@GlennBrockett]

Nothing really wrong with the printer, it just needs some tuning. It won't be as fast as newer ones, but it's the one he has. The printer of Theseus, upgrade as you go. stiffen the gantry, upgrade the board, maybe upgrade the servos...

[@willofthemaker]

It's an old printer, lulzbot kittaz which is essentially a Taz 4 in kit form. I built it in October 2014 and it's honestly been incredibly reliable. It's a workhorse for sure. I've done really less maintenance than I should have over the years. Most of my prints don't have super sharp corners and while still some ghosting, this test piece really pushes it to be the worst possible it can be. the other printer I just got about 2 months ago. It's a Wuxn WXR. Definitely can get better quality and speed from it. Been pretty happy with it over all and expect great things out of it as well. It's probably on par with a Prusa mk4 for half the price, built in America. (Although motion system is LDO). No need to apologize for being critical. It's an old printer and it shows it especially compared to what is out now. Has no bells and whistles. Bed leveling is with me screws and the bed is supported by springs. Which means there will always be some vibrations from y axis movement unless I rigidly mount it. This video made me pay a lot more attention to it than I ever have haha

[@willofthemaker]

Also, thanks for watching both videos. There will definitely be more videos to come with the new printer. One I think would be interesting is to do a side by side comparison of nearly 10 years development in 3d printers. At least with bed slingers. The Wuxn WXR doesn't have everything available nowadays but definitely a lot of improvements. One thing I love about the lulzbot is it is so easy to work on.

[@willofthemaker]

​​@@GlennBrockettya, it's honestly normally not this bad. This filament was good awful. I use it for selling things on Etsy that are circular and it is more than adequate. Slow yes, but reliable and has a big build plate I can fill up. And with those parts the quality is still pretty decent. And better now with the tune ups from this video. A lot of people have suggested adding kilpper and accelerometers which I might explore at some point

[@Dr.Ratio69]

I’ve finished this video. But have you heard there’s a thing called “clipper”?

[@willofthemaker]

haha yes. Ive heard of it and know it does amazing things. A LOT of people have suggested it. So it will probably be added to at some point in the near future.

[@joescalon541]

Interesting concept, but I’m confused why there is any ghosting at those speeds and acceleration. Before Klipper my SV06plus had zero ghosting at 80mm/s and 3k acceleration and 8 jerk. Only needed input shaping after upgrading Klipper for 6k accel and 150mm/s print speed.

[@willofthemaker]

This was all done on my nearly 10 year old printer with mostly all original parts, including belts (only replaced items are a couple nozzles and adding PEI bed). I built it in october 2014 and printers were nowhere near what they are now. A lot of the printers were still made out of laser cut wood. If you want to get an idea, watch "print the legend" on netflix. A lot since then and this one is showing its age. Both in tech and possibly starting to wear out. Its still very reliable, but not the greatest quality. Also, i never updated the firmware. So it is still running 2014 software which doesnt have a lot of the features that are available now.

[@MikeKasprzak]

Interesting idea adding a belt absorber.🤔 It might be worth investigating the belt tightness differences with and without the absorber. I haven't tested this myself, but one random piece of advice I read was that if a belt was tight enough to play a musical note, its too tight. That seemed logical to me if you think of how guitar strings work, but as you showed your quality did improve as you tightened.

[@willofthemaker]

I think the belt likely got too tight. The tightening method on this printer is not the greatest to say the least so there was probably room for me to get it better initially. Could be a contributing factor to the improvement seen. A better design would have been to maybe have a clip on damper that isnt wedged in with the belt.

[@Hi0beczek]

we are missing the last test dampening of belts + gcode improvements! :)

[@willofthemaker]

Haha I know. I'll have to reprint with a different filament. Maybe do a short with it.😅😅

[@davidconner-shover51]

I did the TPU feet act, but I made them like coasters, probably much stiffer than yours, 10% infill gyroid, 95a about an inch and a half thick to the bottom but 2 inches wide where the standard feet set on. I did an input shaping afterwards. The biggest issue was external vibration from the other printer on the same table, it got the same treatment. as well as the nearby washing machine. it worked out the worst of it. Now my speed demon can thunder along, while my slower printer isn't interfered with while doing slow delicate work

[@willofthemaker]

Awesome! Great to hear you had a good result from the TPU feet. Mine was Ninjaflex that is 85A so definitely softer. Maybe some more tinkering could help on my side.

[@DiThi]

I migrated my super old Prusa i3 (classic one build from pieces) to Klipper and I have no regrets. It's a replacement for the firmware so it's not as easy as setting up octoprint for example, but it's worth it.

[@willofthemaker]

Ya I think it's going to happen for me some point soon. I haven't had a chance to look into what I'd need and how long it takes. How was it for you?

[@logicalfundy]

11:40 - to apply a car analogy, acceleration is how far down the brake / gas pedal is, and jerk is how quickly you move your foot to get to that position.

[@willofthemaker]

Thanks, I completely understand that. I think what throws me off is the unit just doesnt make sense for that IMO. i think it should be mm/s^3 in that case.

[@SianaGearz]

Do you have smooth idler or tooth idler? I had to shop for idlers a couple times... the cheapest tooth ones you can buy basically don't work, so they're a little pricey if you get usable ones. But i found it to be good. No i mean you won't spend a fortune on them, but they won't be cent items either.* The idea is that as the belt travel over idler, there is different number of teeth contacting the smooth surface, depending on phase different practical tension, so the belt repeatedly relaxes and tensions, causing belt vibration. Good tooth idler stops that, since every tooth has its dedicated contact patch that keeps the same distance without tooth flattening. However the quality of smooth idler is unimportant, it might seize and you won't even notice. For tooth idler, the tooth profile (deep enough) and bearing quality is absolutely critical, or it's trouble.

[@willofthemaker]

Definitely good info. The pitch of the belt and any toothed part in the system need to be aligned and equal for sure. Mine are smooth idlers though so that's not an issue for me

[@SianaGearz]

@@willofthemaker I switched from smooth to toothed specifically for tension improvement. Depending on belt phase, there is different number of teeth touching down on the smooth idler and different distortion on the teeth, causing the whole belt to oscillate. Just jogging the belt under higher tension feels much smoother on quality toothed idlers than on smooth ones.

[@MiG82au]

Those TPU feet are still extremely stiff with such thick walls and minimal height. I designed some for mine for noise reduction that are about 60 mm high to the supporting face and only have 1 wall. I have one stiffer one under the corner that's weighed down by the PSU. I agree with you that the feet are unlikely to affect the ringing.

[@willofthemaker]

The walls are stiff for sure which i somewhat wanted to make sure it stays on the printer nicely,. Inside on the surface where the printer actually rests is fairly soft. IDK there could likely be more tuning to get a better result but probably not worth the effort tbh

[@MiG82au]

@@willofthemaker yeah the gyroid infill is soft but the shear stiffness of stubby walls is high for the purpose of isolating the fairly low frequency vibration the printer outputs. I haven't calculated the response though, that's just gut feel. I hoped to get some damping of frame vibration too (the motor excitation resonance in particular) but unsurprisingly I couldn't tell the difference.

[@willofthemaker]

@@MiG82au hmm I wonder if having a TPU gasket between the motor and it's mount could help. Could be hard to ensure a good residual torque on the mounting screws though.

[@torrosua]

You've started tuning in the wrong direction. You already have an elephant foot, missing layers and layers of movement. (based on the video). It would help if you went back to the temp/fun tower, checked retraction, and checked the hot-end and extrusion rate. You definitely have some issues with mechanical parts. Vibrations shouldn't have much effect on this speed and acceleration(they're creating most of the vibrations). You have pretty bad results with low speed.

[@willofthemaker]

The issues you mention with missing layers and movement were largely due to that filament. Same gcode and their gone with other rolls. That roll had multiple tangles and was generally inconsistent diameters. Kinda why I used it for this to not waste good filament.

[@stevencullen6261]

Try a block of cement instead of tpu you have low weight in comparison to a car so just trying to stop it outright should be a help and for the 80's car that bounces its the stuts that are bad not the springs if it was the sping your ride hight would drop due to not being able to support the weight of the car anymore

[@willofthemaker]

I think in a cement block will be more helpful with multiple machines next to each other (like I have now). Definitely not the spring in the 80s car example. I mentioned the shock absorbers being shot causing that as there is nothing to dampen the energy. struts are a combo spring and shock

[@cgcgcg]

What plugin do you use to categorize - with icons - your Cure control panel(s)?

[@willofthemaker]

Its called "tabbed settings"

[@MisterkeTube]

Is this printer running Marlin or Klipper? As I see no mention of the obvious one for Klipper (input shaping), I guess Marlin, right? And if so: did you consider going to Klipper (and changing the slicer to match)?

[@willofthemaker]

It's marlin cerca 2014 lol. Never upgraded the firmware. A lot of people have suggested klipper. So much so I will likely look into it

[@MikeKasprzak]

FWIW newer Marlin versions are supposed to support Klipper-esc compensations. Upgrading Marlin will likely be far easier than switching full-hog to Klipper

[@garbonzoe]

I think it would have been helpful to calibrate for your material first. Flow rate/temp/linear advance, that type of thing. Then started playing with general settings. I think comparing different slicers would be interesting too. I recently switched to Orca from Cura and have found that my prints look better. I think Orca has more classes of line types that allows it to be smarter. Curious what to others think though.

[@willofthemaker]

I've heard nothing but really good things about orca. Honestly my days with cura are also likely numbered. It's really beginning to feel like a bit of a has been when it comes to slicers

[@riba2233]

​@@willofthemakeryeah you should, cura has some wacky stuff going on, orca is great

[@willofthemaker]

@@riba2233 Will definitely look into it. Think ill do another couple videos with trying to get the most out of cura and then switch over.

[@riba2233]

@@willofthemaker cool, sounds good :)

[@henrymach]

Printing outside wall first will ruin your overhangs (when you have them)

[@willofthemaker]

Can have an effect for sure although I haven't seen it be that bad. Although I generally design my parts to minimize overhangs in the first place

[@bwloawesome]

You should throw a link to your test piece stl in your description.

[@willofthemaker]

I will! Probably today after work

[@willofthemaker]

Finally got it up to thangs. thangs.com/designer/WilloftheMaker/3d-model/1033464

[@doublegdawg]

For anyone who may not know: Since velocity is your change in displacement over time, and acceleration is your change in velocity over time, then jerk is your change in acceleration over time. My physics courses have definitely helped me in a few areas of my life, and 3D printing is one of those places, lol

[@willofthemaker]

Thanks for these post. This is one of the things that throws me off about jerk. If jerk is the derivative of acceleration, shouldn't it be in mm/s^3 in that case? I haven't been happy with some of the descriptions I've seen about it. One where I saw it was more about braking to a speed into a sharp direction change. Mm/s makes sense for a speed. I'd like to believe it is basically a change in acceleration and it is how quickly it gets up to the 1000mm/s^2 but if jerk is in the range of 10-20 mm/s there is a disconnect somewhere for me.. Def not trying to argue if it's coming off that way. Interested in a better understanding

[@riba2233]

It isn't, it is basically the speed you keep in the corners instead of slowing down to zero, using the inertia of the motion system.

[@willofthemaker]

@@riba2233 This was one of the descriptions I saw and made the most sense to me given its unit which is why I mentioned this, although to some extend i mentioned it as braking into the turn and not necessarily at the turn.

[@christianmontagx8461]

Sorry, not enough material for a subscribtion. The shown is 2018. Nowadays we have Klipper and Inputshaper. There is no reason for extrem ghosting anymore since a few years. Even with bad hardware like yours it is possible to print fast and in good quality.

[@willofthemaker]

Sorry to disappoint. Not everyone has newer machines with klipper (or older with klipper added on). Better luck next time. Ill probably do a video converting this old printer to klipper at some point soon

[@johnbeima6413]

Your conclusion is not supported by your tests… You are wanting to draw a conclusion that putting the TPU on the belts reduced everything. This finding is incorrect, since for all you know, actually tightening the belts more did this. The biggest challenge you have is your belt clamping system is so poorly designed that you can’t actually properly tighten your belts. The last thing we need is bad conclusions like this one floating around confusing people.

[@willofthemaker]

Thanks for the feedback. You are right that tightening the belts could be a contributing factor to the end result. I think that is part of it but not all. The sound being a lower thud on the lower belt path says dampened vs high pitched "ting" of tight belt. A better thing would be to have a way of just sliding it on the belt more akin to the tennis racket example. But this was a quick and dirty . Maybe something worth exploring.

[@johnbeima6413]

@@willofthemaker The thud does not say it was dampened. Since tightening it more does the same thing. What you truly need to do is tension the belt as per the sec for your belt. Which is easy enough to do. Your findings do not conclude there is more than one contributing factor. Now don't get me wrong, I do believe there is in fact more than one factor, however your test of TPU is completely unsubstantiated until the belts are set to the proper tension. All of the tests you have run, really need to be re-done once the belts are properly tensions and input shaper is properly implimented. Otherwise this is simple misinformation.

[@johnbeima6413]

@@willofthemaker One added note. Your test on TPU feet was also pretty conclusively done by Thomas Sanladerer in a way more complete fashion. This "I did it so you don't have to" stuff... Really... If you are not going to put real effort in to-do things correctly, and not "quick and dirty" all you do is confuse people who do not know better.

[@MikeKasprzak]

@johnbeima6413 geez dude, try being constructive without being rude next time.

[@leeburrowdotcom]

@@willofthemaker the TPU dampening is interesting as an idea. I would say be very careful suggesting the way you did it as a solution for other people to use. Basically you have changed the intended geometry of your belts - your print head will move slightly more or less depending on where it is on the rails. I think isolating out obvious printer setup issues first might make for a better starting point - like belt tension.