I've just now merged ZV Input Shaping into Marlin Firmware, so it will be included in 2.1.2. It's labeled as "experimental" because it's so new and we haven't tested it in every scenario. Thanks for taking notice of our work-in-progress and -as usual- for your thorough coverage of the topic. We're really excited that we can gain both speed and quality on all our old 3D printers without adding any new hardware or requiring an additional board. Our implementation requires a fair amount of SRAM, but it is so computationally efficient that we can use it even with AVR boards like the common RAMPS. More complex algorithms like your two favorites MZV and ZVD will require more work, and likely a 32-bit processor. But even the basic ZV shaping does a remarkable job and allows for much higher print speeds with no loss of quality. We hope everyone finds it useful. I'm certainly going to be upgrading all my Marlin-based printers this weekend!
I find this so valuable and amazing. This means that a lot of old printers will probably be used for way longer because they can now keep up with the speeds of more modern machines. This will probably result in less E-waste. A great upgrade to have if you don't want to build your own LIS3DH cable harness, etc.
depends on the old printer. not all have replaceable driver chips. many - especially creality made - have soldered on chipsets that technically "can" be removed, but much more difficult, and not a 'drop in solution'. The E-waste is going to be just as bad really due to the cheap manufacturing of boards and other components. I mean if you really worry about e-waste, start talking to Apple and other Phone makers who make a few new models every single year - only to have most of them sit on shelves/warehouses for over a year.
To get input shaping on any printer, you don't need to throw anything out, no ewaste. All you need is a pi (or any other computer) and different firmware. You can add an accelerometer if you like for ease of use, but that's it. no garbage.
You can still use an old printer. Most people print a few parts per week, so a slow print speed does not matter. Input shaping is like car tuning. It's fun, it's a nice hobby and they find a lot of excuses why they need it. Be honest: you want to "waste" time and that's absolutely ok. I will stick to my old Geetech A10 with low print speeds.
I'll like to say that GH-Enterprise support is just great. Thanks a lot to Andrea who replied quickly, clearly and with detailed answers to all my questions. Now I've made some progress and I achieve better results. My printer is clearly not able to print at 300mm/s with 25000mm/s2 acc., but I'm now having good results at 80mm/s with 3000mm/s2. I still have plenty of tests to perform, but ringing is clearly reduced.
It is always nice to see that companies actually innovate and provide a different solution to the problem. I am just not sure about the price. Raspberry Pi Zero 2, which can run Klipper with Input Shaping perfectly, comes at ~2x lower cost and it is a complete overhaul to a 3D printer.
@@ssm445 Fortunately, there *IS* another solution: a second-hand thin client. This is how easy it is to set one up: th-cam.com/video/yHv9QOwaaS8/w-d-xo.html
@@KeithOlson the thin client in unknown condition (used) costs the same as a new rasp, consumes 10 times the power and is just too bulky to fit below the bed
That looks amazing, probably the next great evolution in extrusion 3D printing. Since I already have a Pi running Octoprint and a community firmware build of Marlin on my CR6, I reckon I'll wait for Marlin support, but now I'm going to be actively monitoring the progress!
I took the plunge on Klipper some months ago and one thing I do have to say is that for bedslinger printers input shaping does also cause some rounding/smoothing of fine details, which can cause issues in the case of high detail models and sometimes cause clearance issues with functional parts. Of course, if you run a little more conservative acceleration, these effects diminish significantly.
@@daliasprints9798 What can I say, I like to push the limits. Part of the test model that Klipper uses is designed to show the smoothing effects of resonance compensation/input shaping so that you can choose a more appropriate max accel. to use. The smoothing does occur at all accelerations, it just happens that the effects only start to become noticeable once you start pushing the limits.
@@poonholder5643 I was not aware that there was an implementation of pressure advance that is synchronized with input shaping. Pressure advance is designed to reduce blobbing/bulging in hard corners, so are you saying that rounding occurs due to too high a pressure advance value?
Live your videos mate and I assure you I took the plunge to buy my 1st printer about 18 months ago probably because you made it so accessible with your tutorials, but I have to say that it is hard to keep up with some of your videos. This is one of them 🤯
If yours isn't modded with more steppers, a CoreXY mod or such, mechanically it's OK, it's a cartesian machine even tho it looks like a CoreXY one. You might need a new motherboard tho if you haven't upgraded yours, in this case tho a Pi might be about the same price (or less with a Zero 2 one). If you've already upgraded your motherboard to one with replaceable stepper drivers then it's good to go already.
Thank you for that technical tip. I am rather new to 3D printing and i have been a fan of your videos for awhile now. So again thank you for providing this great content. Keep up the great content.
Your ringing is formed from bed movements on both directions (X and Y). I've got similar bed swinger and had almost identical resonances on X/Y. I concluded that it must come from bed only since X-parts are super light compared to bed. I put the shaping on only for Y-axis and it worked like a charm -> the bed induces ringing on both directions.. With bed swingers, you can use only 1 stepper driver to compensate the bed induced ringing 💪
@@bigdaddy0690 Tornado bed is on top of screws/springs and major mass is fastened on top. When you check the printable pattern, you'll notice that bed moves in both cases (X and Y pattern). Y pattern accelerates bed to greater speed than X pattern but the acceleration happens at same value. As the resonance is on low freq and the spring/screw setup resonates easily, you'll get the same resonance pattern on both directions.. As I wrote, I have very similar setup and it works like that. If you have a bed swinger and have similar (low) resonances on both axis, try to remove X compensation and test the results.
Really impressive way to upgrade an older printer for $40. I'm only running Klipper CoreXY printers now, but this is a great way to get shaping on an old Ender 3 type unit!
Unless your old Ender 3 Pro is now an IDEX, dual side lift belt drive with 10 micon resolution (with microstepping off), rails, Octopus board and klipper. The only thing Ender on mine is part of the frame lol I blame the boredom during the covid lockdowns o.0
Love you encouragement that although it may seem challenging it isn't overly hard. Thank you! The detail and effort you put into all your videos is greatly appreciated! Could you possibly throw a link in with your vids to the models you print?
I too have an old Tevo Tornado, an older model than the one that you have, but still using the old MKS board. This should be a really easy drop in upgrade I'll have to try! Thanks for the helpful video!
Interesting IMHO most of us that have been in 3d printing for a while have at least one machine that would benefit from this tech. Thanks for the video.Just need to remember where my firmware file is stored LOL.
Nice video, as always. Some people apparently took the "without changing firmware" bit too literally - and it's hard to blame them ;) Also, am I the only one slightly triggered by the incorrect "0.5 Hz - 0.5 second on, 0.5 second off" remark? Because that's 1 Hz ;)
Hi! After having watched this video I've ordered a kit to test these drivers and LA on my Anycubic Mega S. My printer was already quite customized (TMC2209, Marlin 2, 0.9 motors, etc..). After having installed these new drivers, updated FW accordingly, set voltage, all my prints was half the size, took me some time to figure out that the default microsteps value was 32 and not 16, so I've to cut MS1 pin (not quite well explained). I've also had difficulties to enable the LA, just because I didn't used upper case M92 X80 command (for strange reason lower case works from Octoprint terminal, but not in a gcode file). Command I've to use as my M92 X and Y are 160. I've printed the calibration file with the new X and Y code, the print was successful (without LA enabled I wasn't able to prine beyond 40% without layer shifting) and looks great. Acceleration was supposed to go up to 18000, so I was very confident. Then, I've tried to print a 3D Benchy boat, a usual at 60mm/s and twice my default acceleration (1500 -> 3000), and the result is catastrophic, lots of layer shifting. I don't know what's wrong, what I'm doing wrong, I'm sure that I can obtain better results, but at the moment I'm not really satisfied as the printer is also quite noiser than with the 2209.
Hi @Alkerion73 I'm Andrea of GH Enterprise regarding noise you could try to switch to silent mode (Voltage instead of current). Regarding layer shifting could you try to reduce travel speed?
Fantastic idea, the only problem that I personally have with it, is the same problem that I have with the Crality Sonic Pad, or company's coming out with Pi replacements/alternatives... Theyve all come out right after I bought a new PI4, and set up Klipper. If I'dd waited only a cpoupl;e of weeks, I probably could have saved a fortune lol
@@BeefIngot Yes. Plus I'm able to configure it with 2 accelerometers for the bed slinger so I don't constantly have to move it, I'm able to use the Pi to also run open VPN for remote access, plus whatever else I wish to do with it (Makes for a handy media PC for the kids when my niece and nephew visit)
Thank you for this review. It wasn't until the last minute that my hopes for putting these on my CoreXY machine were destroyed :D. I still keep my hopes up for the Creality Sonic Pad as it runs Klipper as an external device similar to a Raspi.
Terrific video as usual. This will help a lot of people. This would be great for my older cr10s pro if the drivers weren't integrated into place. Tooo lazy to change the board.
For all people having no jumpers on the mainboard, simply cut the pin MS1 on the driver so it doesn't connect to the Board. This will put the steps to 1/16 instead of 1/32, if your Board has this by default. There is no loss in quality, the opposite is the case because of the interpolation
I think it looks even way better than the 500mm/s² print... The little towers dont look round at all in the base print, and very smooth and round with input shaping.
3D printing is such a tinkering thing. I love 3D printing, and I use my modified Ender 3 Pro weekly. It's never just a set and forget thing though. For example, my machine developed a weird issue a while back out of no where, it tries to print the 1st and 2nd layer at the same height. It doesn't rise up for the second layer like it should, however this only happens when going from layer 1 to layer 2. I haven't been able to figure out why, so I've just been babysitting the machine during starts and manually baby stepping the Z up by the layer height when it starts layer 2. Just today it started another weird issue where it stops extruding during the outside wall, but only during the outside wall.
Those both sound like slicer issues to me. I'd install a different slicer and do some test prints with that. It sounds like some settings are preventing it from outputting certain lines of code
@Teaching Tech : I've tried to reproduce the castle prints, with my default settings (0.2mm, 60mm/s and 1500mm/s2) Cura say it should take 2h07, with acceleration set to 3000 it take..... 2h07. If I rise the print speed to 150mm/s it take 1h47, and I don't think my printer is able to print at that speed. So perhaps Cura is wrong (can't print now with IS as there is something wrong with my setup), but what are your settings to achieve 1h34 with such quality ???
Amazing video, I have no idea how you can know so much about so many things. So to check if I got this input shaping concept...lets say the machine moves naturaly in X between -1 and 1 during one second, and ideally we want it to always be at zero (zero being the coordinate it should be at any moment of the path it should follow to print the part).......then with input shaping we make it move between +1 to 1 in a second, so we neutralize that -1 to 1 produced by vibrations. Something like that?
Ok then make it happen. You can bulk order them and sell to all the low IQ masses that are totally smart enough to program their own manufacturing equipment.
I think you should stress a lot more that input shaping will cause smoothing of sharp corners when pushed too far, so a test print specifically to evaluate the maximum acceleration WITH inpu shaping. Otherwise nice, I didn't know about those chips!
this feels so complicated vs using an accelerometer and especially vs Bambulab which just does it invisibly for you. Bambulab really set a standard for what auto calibration should be. Hopefully similar tech comes to klipper, especially the auto pressure advance tuning.
@@itsnotthat_ I think you are in a rush to belittle Bambulab because it's closed source and missed a lot of nuance in your comment. Bambulabs solution doesn't need you to take 15 minutes or evaluate a png graph on your raspberry pi Linux computer. It also doesn't need you to print calibration prints you'll then need to measure by eye for pressure advance. This means, especially with the latter, that you have much less friction in tuning. You don't have to tune pressure advance for different filaments. You don't have to get out calipers for input shaper or mount an accelerometer, or review graphs. Many times hobbyists do this annoying thing where they discount how much hassle something is because they don't have something better. I feel that's what you are doing. I have a printer with klipper and an accelerometer. I made the profile myself, I made mounts for the accelerometers, I have tuned pressure advance and input shaper. It's a significantly worse solution than Bambulab which does both for you, and rather than downplay it, you should be trying to get similar features into open source printer firmware.
Mike - a little OT... I am using your calibration page/tools. Is there a way to select on there an equivalent to the option in Prusa Slicer for "Verbose G-Code"? I find I have to have that enabled for my sliced objects to be picked up by Octoprint and work on my MK3s+. Sorry to be OT on your video remarks, I was uncertain how you engage for feedback or support on the work you produce :).
Thanks for helping us all. Now for the help I hope you can show me please please with a cherry on top. I’ve been using Onshape for the past two months and trying out all the tools, but I’m at a point where I’m stuck. I’ve been using the helix tool and that’s great, so here’s my problem I want to make a helix pattern on a curve or on a torus but it keeps going wrong it either go straight up in the air or off to the side and never follows the curve or the torus, I’ve looked online and there is a command that’s called Helix ^2 but I still can’t get it to work. Please can you help me or do a video for us new people that have no idea lol. Again thank you for your help.
Thank you for your super vidéo. I have an ender 7 so that technology dont.work for.me but him waiting.to see what the sonic pad from creality will do to my printer. What do you think of the ender 7 printer??
Michael can I ask you a question, I am currently tuning a printer that has Marlin 2.1.2 I have been following your guide for as far as I can remember LOL, and just have 1 question. should I tune input shaper with the ringing test before or after I tune Linear Advance. I know in Klipper they recommend tuning input shaper before you tune Pressure advance, and a lot of what I have read linear advance in Marlin is the equivelant to pressure advance in Klipper. Thank You for Your Time Rick
Looks great though I would have to get a new controller board and I had just upgraded from an 8-bit to a 32-bit SKR Mini E3 V2.0 recently. Is there a board with replaceable stepper drivers that is a drop in replacement for the Ender 3 without having to replace the connectors like on some other SKR boards. I might be better off using a raspberry pi and klipper
Introduces the video by saying flashing to Klipper is beyond a lot of people, and this is simpler... then proceeds to explain how you must either directly edit your Marlin firmware somehow, or start measuring current going to your old motors, and converting them to new values, etc. Seriously, buying a Pi and flashing Klipper on my Mega-X was so easy I can't imagine anyone who have to go through the above steps to not be able to do this
Hi. Did you have problem with error: Input Step Rate to high? I got this with speeds around 100mm/s when connected to Prusa MK2.5 via external board (needs some soldering). Below speeds 100mm/s it works ok.
I think this is a major step forward however unsure how this will be used in the future. If marlin is in the works for firmware and klipper has it also then it could be limited. Perhaps though with some of the newer marlin builds not fitting on 8bit boards these will help but either way it’s great to see technology moving forward.
Marlin has had some advances with 8 bit recently. I run very comfortably on 8 bit. It's especially good if you use ArcWelder and enable arcs in Marlin. You are more likely to need 32 bit if you have a delta, though. Input shaping for Marlin is also now in the bugfix-2.1.x code.
I am just curious if you know if these stepper drivers will ever be offered here in the US the company says they can be ordered on Amazon but have not been able to find them here. I am very interested in the input shaper capability in Marlin. I only have 1 printer on Klipper and it is an excellent feature.
I'd be interested if you cover marlin implementation of IS because it is basically free and requires just building new firmware. Also you didn't mention that IS does cause quality loss, especially on high acceleration and it is noticeable if you print functional parts and not things like castle because those suddenly might refuse to fit due to smoothed corners.
@@ThantiK In the video, he said Marlin IS is in the [development] pipeline, so I assume it's available as a snapshot (pre-release) build, but I haven't actually looked.
@@pnt1035 So an hour AFTER I posted this comment, you're all "correcting" me because it was merged... when I posted this comment, it hadn't been. And it is STILL only ZV, and EXPERIMENTAL. Parent comment asked TT to cover the marlin implementation...which did NOT EXIST at the time.
I don't understand why making all those movements at the beginning of the print reduces ringing. I think I'm more interested in how this process will be streamlined for the masses in the future.
The specific sequence of movements is just a coded message to the stepper driver module to configure it. The actual reduction in ringing is done by a microprocessor that is packaged in the module. It uses a well known method of smoothing motion in a clever way to counter resonance in the printer.
I've been meaning to ask about something. What is it with slicers and some layers being thinner than others? It doesn't seem to matter what model I slice, if I turn off adaptive layers in order to get better detail, I notice that 9 times out of 10, there's always this group of layers about half a centimeter thick that are just more compact than the rest of the print. You can even see it in the preview of the slice before exporting it. I don't really care that it looks bad, but I have noticed that even though it's denser than the rest of the print, it still becomes a more likely fail point where layers start separating after the print is finished.
Sounds like there would be a risk of interpreting normal print moves as programming. Once it's been programmed once, is it locked against interpreting programming again until power cycle?
Maybe it's because English isn't my native language, but I found the title a little misleading. Yes your firmware stays marlin, but you still have to recompile to work with the new stepper drivers. And that's the biggest hurdle for most I think. If it wasn't for the shortage I find a pi zero 2W more bang for your buck and the Klipper config way easier to read than the one from Marlin.
I've just now merged ZV Input Shaping into Marlin Firmware, so it will be included in 2.1.2. It's labeled as "experimental" because it's so new and we haven't tested it in every scenario. Thanks for taking notice of our work-in-progress and -as usual- for your thorough coverage of the topic. We're really excited that we can gain both speed and quality on all our old 3D printers without adding any new hardware or requiring an additional board. Our implementation requires a fair amount of SRAM, but it is so computationally efficient that we can use it even with AVR boards like the common RAMPS. More complex algorithms like your two favorites MZV and ZVD will require more work, and likely a 32-bit processor. But even the basic ZV shaping does a remarkable job and allows for much higher print speeds with no loss of quality. We hope everyone finds it useful. I'm certainly going to be upgrading all my Marlin-based printers this weekend!
Nice work. Is there an ETA on the release of 2.1.2 so we can all play.
I see it's only 2.1.1 right now. Or will it be in the 2.1.x nightly?
Amazing progress!
@@AAEmohawk I see it merged into bugfix-2.1.x
@@runningfromtheheat oh cool thanks I will check it out.
The GH smart shaper now works on corexy printers too!
I find this so valuable and amazing. This means that a lot of old printers will probably be used for way longer because they can now keep up with the speeds of more modern machines. This will probably result in less E-waste. A great upgrade to have if you don't want to build your own LIS3DH cable harness, etc.
depends on the old printer. not all have replaceable driver chips. many - especially creality made - have soldered on chipsets that technically "can" be removed, but much more difficult, and not a 'drop in solution'. The E-waste is going to be just as bad really due to the cheap manufacturing of boards and other components. I mean if you really worry about e-waste, start talking to Apple and other Phone makers who make a few new models every single year - only to have most of them sit on shelves/warehouses for over a year.
To get input shaping on any printer, you don't need to throw anything out, no ewaste. All you need is a pi (or any other computer) and different firmware. You can add an accelerometer if you like for ease of use, but that's it. no garbage.
I'm using an old lg g4 phone. Installed ubuntu. (chroot) with linux deploy and klipper + octoprint on top of it. It is not an easy install by the way.
You can still use an old printer. Most people print a few parts per week, so a slow print speed does not matter.
Input shaping is like car tuning. It's fun, it's a nice hobby and they find a lot of excuses why they need it. Be honest: you want to "waste" time and that's absolutely ok.
I will stick to my old Geetech A10 with low print speeds.
@@1234fishnet you do you
I'll like to say that GH-Enterprise support is just great.
Thanks a lot to Andrea who replied quickly, clearly and with detailed answers to all my questions.
Now I've made some progress and I achieve better results.
My printer is clearly not able to print at 300mm/s with 25000mm/s2 acc., but I'm now having good results at 80mm/s with 3000mm/s2.
I still have plenty of tests to perform, but ringing is clearly reduced.
Not so much anymore apparently. I havent received any assistance since I sent an email a month ago.
Thanks for keeping us update with ways and technology that makes our 3D prints better, faster, easier, and/or cheaper.
It is always nice to see that companies actually innovate and provide a different solution to the problem. I am just not sure about the price. Raspberry Pi Zero 2, which can run Klipper with Input Shaping perfectly, comes at ~2x lower cost and it is a complete overhaul to a 3D printer.
@@renevile It’s easier than u think, and gives u a lot of cool features
The Pi Zero 2 is easily double the price now though.
@@ssm445 Fortunately, there *IS* another solution: a second-hand thin client. This is how easy it is to set one up: th-cam.com/video/yHv9QOwaaS8/w-d-xo.html
@@KeithOlson the thin client in unknown condition (used) costs the same as a new rasp, consumes 10 times the power and is just too bulky to fit below the bed
@@ssm445 True in general, but if you can't *FIND* a Pi, a low-cost thin client or even an old laptop motherboard *are* viable options.
That looks amazing, probably the next great evolution in extrusion 3D printing. Since I already have a Pi running Octoprint and a community firmware build of Marlin on my CR6, I reckon I'll wait for Marlin support, but now I'm going to be actively monitoring the progress!
Finally! My prayers have been answered. Thanks so much for featuring this solution.
I took the plunge on Klipper some months ago and one thing I do have to say is that for bedslinger printers input shaping does also cause some rounding/smoothing of fine details, which can cause issues in the case of high detail models and sometimes cause clearance issues with functional parts. Of course, if you run a little more conservative acceleration, these effects diminish significantly.
Rounding should only happen if you're using acceleration or corner velocity well above what IS can cleanly suppress the vibrations for.
@@daliasprints9798 What can I say, I like to push the limits. Part of the test model that Klipper uses is designed to show the smoothing effects of resonance compensation/input shaping so that you can choose a more appropriate max accel. to use. The smoothing does occur at all accelerations, it just happens that the effects only start to become noticeable once you start pushing the limits.
@@quinnobi42 you mixing up 2 separate things here. If you tune p_A correctly you won't have rounding.
@@poonholder5643 I was not aware that there was an implementation of pressure advance that is synchronized with input shaping. Pressure advance is designed to reduce blobbing/bulging in hard corners, so are you saying that rounding occurs due to too high a pressure advance value?
@@quinnobi42 If You have rounded corners You might be using too agressive input shaper, but anyway You should calibrate PA AFTER input shaper
Live your videos mate and I assure you I took the plunge to buy my 1st printer about 18 months ago probably because you made it so accessible with your tutorials, but I have to say that it is hard to keep up with some of your videos. This is one of them 🤯
love the oscillation demonstration chook 🙂. I'd be keen to see more on this topic. I'll have to investigate if my Ender 5 pro can use this upgrade.
If yours isn't modded with more steppers, a CoreXY mod or such, mechanically it's OK, it's a cartesian machine even tho it looks like a CoreXY one.
You might need a new motherboard tho if you haven't upgraded yours, in this case tho a Pi might be about the same price (or less with a Zero 2 one). If you've already upgraded your motherboard to one with replaceable stepper drivers then it's good to go already.
Thank you for that technical tip. I am rather new to 3D printing and i have been a fan of your videos for awhile now. So again thank you for providing this great content. Keep up the great content.
I like the responsiveness of GH to your advice and requests :).
These look like a neat little change.
Good Ole MKS Gen L. Still using one after 4yrs.
Your ringing is formed from bed movements on both directions (X and Y). I've got similar bed swinger and had almost identical resonances on X/Y. I concluded that it must come from bed only since X-parts are super light compared to bed. I put the shaping on only for Y-axis and it worked like a charm -> the bed induces ringing on both directions.. With bed swingers, you can use only 1 stepper driver to compensate the bed induced ringing 💪
This is illogical unless you didn’t have your print oriented perfectly or you have some other mechanical issue like loose belt or wheels.
@@bigdaddy0690 Tornado bed is on top of screws/springs and major mass is fastened on top. When you check the printable pattern, you'll notice that bed moves in both cases (X and Y pattern). Y pattern accelerates bed to greater speed than X pattern but the acceleration happens at same value. As the resonance is on low freq and the spring/screw setup resonates easily, you'll get the same resonance pattern on both directions.. As I wrote, I have very similar setup and it works like that. If you have a bed swinger and have similar (low) resonances on both axis, try to remove X compensation and test the results.
this is awesome i left them a comment some weeks ago to send you and cnc kitchen some for review
Thank you for sharing this with us. I hope new machines will come with these drivers built in and a slicer profile ready to go.
Really impressive way to upgrade an older printer for $40. I'm only running Klipper CoreXY printers now, but this is a great way to get shaping on an old Ender 3 type unit!
Unless your old Ender 3 Pro is now an IDEX, dual side lift belt drive with 10 micon resolution (with microstepping off), rails, Octopus board and klipper. The only thing Ender on mine is part of the frame lol I blame the boredom during the covid lockdowns o.0
Love you encouragement that although it may seem challenging it isn't overly hard. Thank you! The detail and effort you put into all your videos is greatly appreciated! Could you possibly throw a link in with your vids to the models you print?
Nice results. Feels like there is a second frequency present. You eliminated lower, but higher persists.
Love the Senna t-shirt
I too have an old Tevo Tornado, an older model than the one that you have, but still using the old MKS board. This should be a really easy drop in upgrade I'll have to try! Thanks for the helpful video!
Really, really interesting! 😃
Thanks a bunch, Michael!!!
Stay safe there with your family! 🖖😊
I think I'll wait till it lands in Marlin. Interesting video, though, I was not aware of this trick. Thanks.
Awesome shirt mate and great video 👌
Great find!
Thanks for sharing your expirences with All of us 👍😃
Interesting IMHO most of us that have been in 3d printing for a while have at least one machine that would benefit from this tech.
Thanks for the video.Just need to remember where my firmware file is stored LOL.
Nice video, as always. Some people apparently took the "without changing firmware" bit too literally - and it's hard to blame them ;)
Also, am I the only one slightly triggered by the incorrect "0.5 Hz - 0.5 second on, 0.5 second off" remark? Because that's 1 Hz ;)
This will definitely go on my must try and do list
This is simply amazing. What an easy and effective mod, wow. Cant wait to try them out.
Hi! After having watched this video I've ordered a kit to test these drivers and LA on my Anycubic Mega S. My printer was already quite customized (TMC2209, Marlin 2, 0.9 motors, etc..).
After having installed these new drivers, updated FW accordingly, set voltage, all my prints was half the size, took me some time to figure out that the default microsteps value was 32 and not 16, so I've to cut MS1 pin (not quite well explained).
I've also had difficulties to enable the LA, just because I didn't used upper case M92 X80 command (for strange reason lower case works from Octoprint terminal, but not in a gcode file). Command I've to use as my M92 X and Y are 160.
I've printed the calibration file with the new X and Y code, the print was successful (without LA enabled I wasn't able to prine beyond 40% without layer shifting) and looks great. Acceleration was supposed to go up to 18000, so I was very confident.
Then, I've tried to print a 3D Benchy boat, a usual at 60mm/s and twice my default acceleration (1500 -> 3000), and the result is catastrophic, lots of layer shifting.
I don't know what's wrong, what I'm doing wrong, I'm sure that I can obtain better results, but at the moment I'm not really satisfied as the printer is also quite noiser than with the 2209.
Hi @Alkerion73 sorry for the problems you are experiencing. Can you send us via mail and send us your G-Code?
Hi @Alkerion73 I'm Andrea of GH Enterprise regarding noise you could try to switch to silent mode (Voltage instead of current). Regarding layer shifting could you try to reduce travel speed?
Can’t wait for this to be a feature in Marlin.
Fantastic idea, the only problem that I personally have with it, is the same problem that I have with the Crality Sonic Pad, or company's coming out with Pi replacements/alternatives... Theyve all come out right after I bought a new PI4, and set up Klipper. If I'dd waited only a cpoupl;e of weeks, I probably could have saved a fortune lol
You say that but look at how painful this is to configure vs an accelerometer
@@BeefIngot Yes. Plus I'm able to configure it with 2 accelerometers for the bed slinger so I don't constantly have to move it, I'm able to use the Pi to also run open VPN for remote access, plus whatever else I wish to do with it (Makes for a handy media PC for the kids when my niece and nephew visit)
Just realized I need a lot more technical concepts explained using chicken mechanics 😂
most underrated comment, lol
So thankful for this
Thank you for this review. It wasn't until the last minute that my hopes for putting these on my CoreXY machine were destroyed :D. I still keep my hopes up for the Creality Sonic Pad as it runs Klipper as an external device similar to a Raspi.
Wellp on core xy you should have way less ringing ect
the smart shaper now also works with corexy printers
Terrific video as usual. This will help a lot of people. This would be great for my older cr10s pro if the drivers weren't integrated into place. Tooo lazy to change the board.
For all people having no jumpers on the mainboard, simply cut the pin MS1 on the driver so it doesn't connect to the Board. This will put the steps to 1/16 instead of 1/32, if your Board has this by default. There is no loss in quality, the opposite is the case because of the interpolation
I think it looks even way better than the 500mm/s² print...
The little towers dont look round at all in the base print, and very smooth and round with input shaping.
3D printing is such a tinkering thing. I love 3D printing, and I use my modified Ender 3 Pro weekly. It's never just a set and forget thing though. For example, my machine developed a weird issue a while back out of no where, it tries to print the 1st and 2nd layer at the same height. It doesn't rise up for the second layer like it should, however this only happens when going from layer 1 to layer 2. I haven't been able to figure out why, so I've just been babysitting the machine during starts and manually baby stepping the Z up by the layer height when it starts layer 2. Just today it started another weird issue where it stops extruding during the outside wall, but only during the outside wall.
Those both sound like slicer issues to me. I'd install a different slicer and do some test prints with that.
It sounds like some settings are preventing it from outputting certain lines of code
The legend F1 Racer.
"👍👍💯💯🙂t-shirt "
I found out about them 2 weeks ago and I want to try them at some point!
Haha that chicken !! Nice
5:16 is that a spark?
Amazing video Michael.
Looks awesome new to all of this working on getting the prints to look ok then I will need to learn how to use g code alter it and whatnot
Simply spectacular!
These might be a nice upgrade for my old Printrbot Simple Metal.
Wow, that is awesome. I might need to see about upgrading my FF Creator X that runs Sailfish
Both Amazon and eBay links are sold out.
Available again from 10th November
Wild. Just as Marlin dropped their own implementation of input shaper too
As with so many advances, it is an idea whose time has come.
@Teaching Tech : I've tried to reproduce the castle prints, with my default settings (0.2mm, 60mm/s and 1500mm/s2) Cura say it should take 2h07, with acceleration set to 3000 it take..... 2h07. If I rise the print speed to 150mm/s it take 1h47, and I don't think my printer is able to print at that speed. So perhaps Cura is wrong (can't print now with IS as there is something wrong with my setup), but what are your settings to achieve 1h34 with such quality ???
Amazing video, I have no idea how you can know so much about so many things. So to check if I got this input shaping concept...lets say the machine moves naturaly in X between -1 and 1 during one second, and ideally we want it to always be at zero (zero being the coordinate it should be at any moment of the path it should follow to print the part).......then with input shaping we make it move between +1 to 1 in a second, so we neutralize that -1 to 1 produced by vibrations. Something like that?
May I know the brand of the printing head depicted at 1:46? 🤔
the smart shaper now also works with corexy printers
Very interesting! Marlin currently has input shaping integrated in the current release... in beta status...
You can do UART on the gen L that came with the Tornado. It just requires some jumper wires
I'd love to see these available locally Instead of shipping overseas). Seem pretty interesting!
Ok then make it happen. You can bulk order them and sell to all the low IQ masses that are totally smart enough to program their own manufacturing equipment.
Thank you so much, i might get some for my printer.
I think you should stress a lot more that input shaping will cause smoothing of sharp corners when pushed too far, so a test print specifically to evaluate the maximum acceleration WITH inpu shaping.
Otherwise nice, I didn't know about those chips!
Can you do it on a Prusa MK3s?
Nice!
🐾🐾🐈🥰 1:45 resonance frequency = 🐔wiggle 👍👍
Can this be a drop-in replacement for the TMC2209 on a BTT SKR v1.4 Turbo ?
Avaiable on EBay (GH smart shaper)
Amazon too
this feels so complicated vs using an accelerometer and especially vs Bambulab which just does it invisibly for you.
Bambulab really set a standard for what auto calibration should be. Hopefully similar tech comes to klipper, especially the auto pressure advance tuning.
@@itsnotthat_ I think you are in a rush to belittle Bambulab because it's closed source and missed a lot of nuance in your comment.
Bambulabs solution doesn't need you to take 15 minutes or evaluate a png graph on your raspberry pi Linux computer. It also doesn't need you to print calibration prints you'll then need to measure by eye for pressure advance.
This means, especially with the latter, that you have much less friction in tuning.
You don't have to tune pressure advance for different filaments. You don't have to get out calipers for input shaper or mount an accelerometer, or review graphs.
Many times hobbyists do this annoying thing where they discount how much hassle something is because they don't have something better. I feel that's what you are doing.
I have a printer with klipper and an accelerometer. I made the profile myself, I made mounts for the accelerometers, I have tuned pressure advance and input shaper. It's a significantly worse solution than Bambulab which does both for you, and rather than downplay it, you should be trying to get similar features into open source printer firmware.
Mike - a little OT... I am using your calibration page/tools. Is there a way to select on there an equivalent to the option in Prusa Slicer for "Verbose G-Code"? I find I have to have that enabled for my sliced objects to be picked up by Octoprint and work on my MK3s+.
Sorry to be OT on your video remarks, I was uncertain how you engage for feedback or support on the work you produce :).
Hmm. Can I switch the two x& y drivers for these but still leave current drivers (TMC 2209) installed on my SKR E3V3??
Has anyone here tried the new Marlin input shaper in 2.1.2? I can't seem to get it to make a difference.
Thanks for helping us all. Now for the help I hope you can show me please please with a cherry on top. I’ve been using Onshape for the past two months and trying out all the tools, but I’m at a point where I’m stuck. I’ve been using the helix tool and that’s great, so here’s my problem I want to make a helix pattern on a curve or on a torus but it keeps going wrong it either go straight up in the air or off to the side and never follows the curve or the torus, I’ve looked online and there is a command that’s called Helix ^2 but I still can’t get it to work. Please can you help me or do a video for us new people that have no idea lol. Again thank you for your help.
Thank you for your super vidéo. I have an ender 7 so that technology dont.work for.me but him waiting.to see what the sonic pad from creality will do to my printer. What do you think of the ender 7 printer??
Michael can I ask you a question, I am currently tuning a printer that has Marlin 2.1.2 I have been following your guide for as far as I can remember LOL, and just have 1 question. should I tune input shaper with the ringing test before or after I tune Linear Advance. I know in Klipper they recommend tuning input shaper before you tune Pressure advance, and a lot of what I have read linear advance in Marlin is the equivelant to pressure advance in Klipper.
Thank You for Your Time
Rick
doesn't appear available on amazon US...
Correct me if I'm wrong but doesn't Marlin already have backlash compensation as a feature? Could that also help with ringing at higher accelerations?
Backlash? What 3D printers have a leadscrew on the X and Y axes? o.O
@@alejandroperez5368 The Taz Pro has backlash compensation and gt2 belts on all axes.
@@alejandroperez5368 Backlash happens on belts too, not just lead screws.
This is completely different to backlash
2:27 if it is the name brand, there seems to be compeditors that are cheap and do the same thing.
I wonder if those could be combined with closed loop steppers ...
Looks great though I would have to get a new controller board and I had just upgraded from an 8-bit to a 32-bit SKR Mini E3 V2.0 recently. Is there a board with replaceable stepper drivers that is a drop in replacement for the Ender 3 without having to replace the connectors like on some other SKR boards. I might be better off using a raspberry pi and klipper
Thinking of using something like these with a BTT SKR 1.4 Turbo on my Prusa MK3S to get some fast and relatively quite printing
Will it work if I have 2 Z axis motors? I run Marlin.
Yes, this is for X and Y axes... not Z.
How would i make these work on a flashforge creator pro?
Can the "Input Shaping CPU" be emulated on modern 32bit boards?
Will these work on Prusa i3 mk3 printers?
out of curiosity, what happens when you have Klipper's Input shaper trying to work in conjunction with these?
Loved it
Pretty clever. Maybe I'll try it on my modded A10...
Introduces the video by saying flashing to Klipper is beyond a lot of people, and this is simpler... then proceeds to explain how you must either directly edit your Marlin firmware somehow, or start measuring current going to your old motors, and converting them to new values, etc.
Seriously, buying a Pi and flashing Klipper on my Mega-X was so easy I can't imagine anyone who have to go through the above steps to not be able to do this
Hi. Did you have problem with error: Input Step Rate to high? I got this with speeds around 100mm/s when connected to Prusa MK2.5 via external board (needs some soldering). Below speeds 100mm/s it works ok.
Would this help my quality if I have the crealoty 4.2.7 with the tmc2208 drivers already?
How can I upgrade Ender 3 with this?
I think this is a major step forward however unsure how this will be used in the future. If marlin is in the works for firmware and klipper has it also then it could be limited. Perhaps though with some of the newer marlin builds not fitting on 8bit boards these will help but either way it’s great to see technology moving forward.
Marlin has had some advances with 8 bit recently. I run very comfortably on 8 bit. It's especially good if you use ArcWelder and enable arcs in Marlin. You are more likely to need 32 bit if you have a delta, though. Input shaping for Marlin is also now in the bugfix-2.1.x code.
@@tombrazier6172 It’s already in the bugfix? Oh I thought it was only in the initial stages yet. Time for me to get a wee looksie
I am just curious if you know if these stepper drivers will ever be offered here in the US the company says they can be ordered on Amazon but have not been able to find them here. I am very interested in the input shaper capability in Marlin. I only have 1 printer on Klipper and it is an excellent feature.
Hello, if you want you can buy it on Ebay, or make a request directly on our site.
Would this work on a stock ender 3 and a cr6?
May I ask what tables you use? :)
Talks about the hassel of changing firmware just to go in and edit firmware.
I'd be interested if you cover marlin implementation of IS because it is basically free and requires just building new firmware.
Also you didn't mention that IS does cause quality loss, especially on high acceleration and it is noticeable if you print functional parts and not things like castle because those suddenly might refuse to fit due to smoothed corners.
What marlin implementation? Marlin doesn't have an implementation that's comparable. Even RRF; which just got its implementation - is lagging behind.
@@ThantiK In the video, he said Marlin IS is in the [development] pipeline, so I assume it's available as a snapshot (pre-release) build, but I haven't actually looked.
@@ThantiK It's in recent bugfix versions.
@@pnt1035 So an hour AFTER I posted this comment, you're all "correcting" me because it was merged... when I posted this comment, it hadn't been. And it is STILL only ZV, and EXPERIMENTAL. Parent comment asked TT to cover the marlin implementation...which did NOT EXIST at the time.
I don't understand why making all those movements at the beginning of the print reduces ringing.
I think I'm more interested in how this process will be streamlined for the masses in the future.
The specific sequence of movements is just a coded message to the stepper driver module to configure it. The actual reduction in ringing is done by a microprocessor that is packaged in the module. It uses a well known method of smoothing motion in a clever way to counter resonance in the printer.
TLSmoother Compatible??
I've been meaning to ask about something. What is it with slicers and some layers being thinner than others?
It doesn't seem to matter what model I slice, if I turn off adaptive layers in order to get better detail, I notice that 9 times out of 10, there's always this group of layers about half a centimeter thick that are just more compact than the rest of the print. You can even see it in the preview of the slice before exporting it.
I don't really care that it looks bad, but I have noticed that even though it's denser than the rest of the print, it still becomes a more likely fail point where layers start separating after the print is finished.
sound slike you have the variable layer height setting in your slicer turned on
Sounds like there would be a risk of interpreting normal print moves as programming. Once it's been programmed once, is it locked against interpreting programming again until power cycle?
Maybe it's because English isn't my native language, but I found the title a little misleading. Yes your firmware stays marlin, but you still have to recompile to work with the new stepper drivers. And that's the biggest hurdle for most I think. If it wasn't for the shortage I find a pi zero 2W more bang for your buck and the Klipper config way easier to read than the one from Marlin.
Could have make it stay with a eeprom so that you won't need to run the code to program it every time u power up
Looking forward to the v2 with this