3D Printer’s Z-AXIS ULTIMATE Precision Upgrade - IS IT WORTH IT? (Linear Rails + Ballscrew)
ฝัง
- เผยแพร่เมื่อ 8 พ.ค. 2024
- Get professionally CNC'ed parts for your projects at PCBWay (5$ OFF for your first order) - bit.ly/PCBWay5off
STEP file of this project - bit.ly/KP3S_Z_LR_BS
👉🏻 In this video I am upgrading Kingroon KP3S 3.0 3D printer's Z-axis with two MGN9 H linear rails and the SFU1204 ballscrew. I will be showing you before-after print quality improvements and will be making an advantage list of this vs stock Z-axis solution.
🧡 Support my work at PATREON - / diyperspective
📋 RELATED ITEMS TO THE VIDEO (Affiliate)
- Kingroon KP3S amzn.to/3yDvHWe
- Digital Microscope AD246S-M amzn.to/3UnpM1L
- SUNLU Meta PLA amzn.to/3Vk9jvo
📷 VIDEO SHOT ON:
- Fuji X-S10 amzn.to/3Zar9BW
📢 OTHER MENTIONED THINGS:
GJB Test Cube www.printables.com/model/2931...
🕗 TIMESTAMPS:
00:00 - Intro
00:46 - The Design
01:10 - Mod/Test/Comparison
06:49 - My Conclusion
🔗 YOU CAN FOLLOW ME:
Twitter: / diyperspective
Instagram: / diyperspective
#kp3s #3dprinter #upgrade - วิทยาศาสตร์และเทคโนโลยี
The funny thing to me watching a bunch of your comparison videos is how many prints have errors in the exact same spots regardless of how many factors you change. That leads me to think that slicers still need more work to generate perfect consistency. It doesn't make sense that the same print with two different Z axis setups would have Z banding issues at the exact same layer unless something else was going on, ya know? Same with the Teflon coated screw and POM nut. Very interesting.
It's not the slicer buddy go look up the wobble-x adapter and see how a lead screw or ball screw with a tiny amount of unstraightness can cause this banding issue. In super Ridgid CNC mills where these ballscrews are normally used, a small wobble of the ballscrew will not affect a big milling machine with cast iron construction. With a machine with small aluminum extrusion frames, it will cause small side to side movement and banding. The wobble adapters are the best way around this problem and near perfect prints can be achieved on the basic lead screws and a pom nut with a wobble-x adapter.
@@clutchboi4038 I'm familiar with Z-banding as well as its causes and fixes like the Wobble-X and Oldham couplers. As I said "It doesn't make sense that the same print with two different Z axis setups would have Z banding issues at the exact same layer unless something else was going on" meaning that there is something other than the physical Z setup causing the issue. If the same Z band shows up when he has replaced both the screw and support method of the Z, how can it be Z banding caused by the screw or guide type? I may be missing something, but I'm just saying.
Thanks for the great video. Very informative! With that being said, when I built my Custom 235x235x275mm 3D Printer, I used Linear rails on all three axis. I also had custom hardened steel lead screws made. I use Oldham Couplers and a Backlash Nuts on my dual z axis. The difference between my Ender 3 Pro and my Custom Built machine is night and day. I get crazy accurate parts with my custom machine and much better surfaces with my custom built machine. I looked at ball screws, but with the Oldham Couplers and Backlash Nuts with springs, I get great accuracy and a great surface finish. I guess it is 6 of one and 1/2 dozen of the other. It is all a matter of what works best for you.
I never thought to just use my dial gauge to directly measure the Z axis accuracy. Thanks for the idea. I have wondered if the ball screw on my printer was causing my layer problems. Now I will just go measure it.
Great compairason!
I upgraded a few printers to quality single start lead screws and run the Z BOM wheels a wee bit tighter than usual and imo has been the best cost/performance upgrade as you no longer have to worry about the "magic number" .04mm z increments to whole step the z motor for most consistency.
With quality single start screws. Theres little to no wobbling to get transferred into the print and improves the Z step resolution to .01mm/step increments vs .04mm/step. And you can literally pick any layer height without worrying about layer inconsistencies.
Those same printers i swapped from 1.8deg steppers to 0.9deg steppers and even though they may be cheap bed slingers, their print quality is simply impeccable
Single start do you mean 1mm pitch/lead? I’ve on one of mine installed 2mm pitch/lead and sure movement is much more precise (over killing putting also a 0.9 degree stepper) problem I’ve found is that max speed and acceleration became way too slow… moving back bed or gantry after a full print height it’s a pain also on relative small printer…
Oh my, this hotend is simply awesome. Is it much better than the original one? does it use the default fans or different ones? I just bought a kp3s, still haven't arrived tho.
thanks. I've got s1 pro and v2 neo. and thinking of doing linear upgrade on all axis. but not sure if they're worth it? and which one should I be doing the upgrade on. if you can please advise, that'd be much appreciated.
Thank you for the truth. I had to subscribe for that.
I feel like the linear rail/ballscrew upgrade might work better for the x and y axis as they are the two axis that move the most out of the printer and would most likely be the most affected by any vibration. I would be interested in the results of such an upgrade.
I agree and, feel like linear rails, across all axis, is where the benefit will truly show. Given that, and ball screws, I believe you’ll see the highest improvement in accuracy.
@@TheJacklwilliams That printer already uses linear rails for X and Y, so he did end up with LR on all axes
And using a ball screw for x/y would make it SLOWWWWWW. There are also some other drawbacks. There's a reason a Voron 2.4 uses belts for everything.
Most printers are fine as they are. There is no need to modify as heavily as this. Very interesting video so thanks.
I suggest you to check out your extruder gears, that will make a real difference when you notice the problem.
What also would be Interesting is to see the difference between these to and a Belt driven Z-Axis.
I would also like to see this , I have a belted Z on my Pro 3D V-King 400 and I think they are much better than leadscrews but would like to see the difference in cheap ballscrews.
Even with the most advanced printers, slight imperfections can occur. However, with proper calibration, slicer settings, and high-quality filament, you can produce prints that are incredibly close to perfect.
For a single screw driven printer, it does not hurt your wallet. But if you have like a triple z axis bed, the cost really adds up. If budget is an issue, then using oldham couplings with standard leadscrews is quite effective.
Thank you for this interesting video! I also own a KP3S 3.0, and now I want to test this system myself. May I ask for the links of both the linear rails and ball screw you used? 😄
I love that you went full lunatic mode to prove your point. My KP3S came with some horrible Z wobble, but a different Z screw coupler and nut (I already had both) and a simple printed part to mount the nut and allow the Z screw to be 90 degrees solved it pretty well.
I'm still not able to get the quality on my KP3S that I see from may i3 Mega so now I'm exploring how vibrations from the part-cooling fan affect print quality. An ADXL345 sensor shows quite a lot of vibration and I can see that certain speeds (say 50%) seem to be much worse than others (30% and 70%). I saw in another video that you have a bunch of different fans to test with. Maybe a topic for a video?
Haha, thanks. Yeah, KP3S' can have some serious quality control issues. One of my KP3S had absolutely perfect Z-axis rod alignment and on another, I had to widen the holes to mount the brass nut in the correct spot. But for print quality wise I would say they perform extremely well.
I made many mods and calibrations on my i3 Mega back in the day, but it never reached the quality that I am currently having with my two modified KP3S printers (with V6 all-metal hotends). The stock hotend is good while it is new but with a PTFE tube inside it is just a question when you run into problems. And it also uses V5/V6/MK8 uncommon parts it is such a weird and annoying combo.
As for the fan vibrations, I doubt they affect the print quality in a noticeable way. Unless you get some really bad vibrations at certain speeds. Take a look at this video th-cam.com/video/TXw37ydld5g/w-d-xo.html This fan has way too much vibrations for my liking but only affected the nozzle at certain speeds. After I saw that I ditched the "fan vibration effects on print quality" idea. But maybe I should do it after all to show the worst-case scenario.
My kp3s had a bad coupler as well, seems like a common issue, though cheap and easy to fix. The Z nut bracket was also not very straight so modifed that and added an anti-backlash polymer nut as well.
I'm getting great quality prints, but I don't like how slow this printer is in it's default state, my Monoprice Select Mini V2 gave similar quality at significantly higher speeds, but it was also noisy af and had a tiny build plate.
@@ferrumignis Have you tried compiling Marlin or installing Klipper and increasing speed limits? I put Klipper on mine on day 1 and find it reasonably fast. I had to clean and relube the linear bearings. The mass of the x carriage is probably the limiting factor now.
@@beauregardslim1914 Klipper is on my todo list. I have a suitable Pi somewhere in a box in my loft...
Can you tell please what parts exactly did you used for this upgrade and what adjustment in pronterface, configuration file did you made? I want to try this mod on my kp3s.
keep in mind this could be inconsistent extrusion! adding a reverse bowden to my E3 S1 improved my layers a lot.
I wonder if there's much difference in structural integrity between the precision options.
Thank you for this review and perspective. I have one printer, Creality 10s Pro V1. I had to make some upgrades to improve performance. BL Touch, Timing Belt for dual Z axis screws and Direct Drive from Microswiss with their Hot End. I added Z axis Struts too. I am stiving for both detail prints of small parts and accuracy of larger Prints for my Model Train Hobby. Now i need to add another requirement SPEED. With one machine and selling train parts I occasionally need to print a lot of parts fast. I also do R&D and have to print a few parts in good detail. I was thinking about an Ender 7 as a second printer but saw Creality has a new Mother Board with 32 Bits, Now I am interested in that. Then I saw Linear rails for Z and X axis I like that. But I really don't have a handle in what is best and what works short term and long term. Would be able to offer any insight to this dilemma, One machine with all the upgrades thrown at it or second specialized machine for small parts printed fast. Thank You, Dennis
I don't know why you are so hooked up on one brand. If you enjoy modifying 3d printers you should have at least two of them and if you don't care about that and just want to print with speed and quality buy the P1P, the value of it is insane but it is also very loud. I think the days of bed slingers slowly coming to the end and hopefully we will see more and more CoreXY type machines in the future. Just do yourself a favor and don't buy one with POM wheels.
And for the 10s Pro I suggest upgrading to Klipper because it is almost unfair how much speed you can still push out of a 3d printer with it. Here is my old i3 Mega S running at 400+mm/s and 12K acceleration th-cam.com/video/n23zvv27flM/w-d-xo.html
Sounds to me as if by far the largest reason to make this upgrade is printing speeds. A quarter faster is a pretty big difference. Pretty interesting, thank you for testing.
It definitely helps to have perfect Z accuracy when printing thinner layers, but yes the biggest benefit of this modification is faster Z movement and acceleration. The print speed is so much more fluent and faster, especially if you have Combing mode OFF in CURA and Z-hop higher. Also manually leveling the bed is so much quicker.
@@DIYPERSPECTIVE Am I correct in thinking that backlash is also minimalised when using ballscrews against leadscrews?
Although the differences are extremely minimal I'm sure the ball screw will wear better over time. Also stiffness of the x gantry is a red herring.
Yeah, those things are great but the increased Z-axis speed and acceleration are something that I never thought I would appreciate so much. Now going to other 3D printers, Z-hop speeds seem like a joke. :D
With your single sided machine I was thinking if you changed the X axis print bed to a rotating circle bed that rotates with the bed width the original beds as it’s radius and instead of moving backwards and forwards it just rotates and the Y axis and Z axis works as normal you would increase the size by 4 times
I think 3d printer rigidity is not an issue. Open loop motion system is problem. Thats why old ink printers uses steppers than they changed to dc motors with microscopic lines with it. If all motion axis have this feature you can change the mass and precision in all 3d printers.
I remember the kp3s pro has only z and x linear rails and the original kp3s has y and x rails. I always wondered what would be the difference. Good to see the results
I have the KP3S Pro, you can also buy it with the dual Y axis linear rails upgrade kit. I am just waiting for it now as it is shipped from China. PEI sheet would also make sense as it is way lighter that the glass bed. I am just disappointed as they still have a lot of QC problems as with KP3S 3.0.
@@DIYPERSPECTIVE The rails shipped with the Y axis upgrade kit had horrible binding for me. My X rail also isn't super smooth, and the Z rail has even worse binding than the Y ones (KP3S Pro). Took them all apart completely and am letting them soak piece by piece in 99% IPA for around 24 hours, then will do a deep clean and lube them up. The QC on the rails is atrocious and they might as well be as bad as loose v slot wheels.
I only had Z binding rail. Y and X were smooth. One thing to take into account is that linear rails can come preloaded. Meaning that bearings will be just a little bit oversized depending on the preload level to eliminate any wobble between the rail and the carriage. You just lubricate it and move the carriage with changing pressure directions back and forth across all the rail until binding is gone. But if the rail binds in only at one side then it is the problem with the rail.
I wonder if the accuracy was increased by the ball screw or rail.
Shouldn't this benefit more on bigger printers like the nuptune 3 max ?
Cnc routers have to be really heavy and big to be really precise. I was wondering if was similar with 3d printing. Looks like you provided it doesn't.
I tried this with a single HGR15 rail and custom plastic belt drive and noticed a small improvement but I ended up going back to standard balls crew because I got tired of the Z falling when powered off. I even tried a geared kysan nema 17 and it would drop sometimes too driving the hotend nozzle into my print bed cooking the PEI.
You may want to re-test with layer heights of 0.1.mm and 0.2mm as 0.24mm is a bit tall of a layer when trying to determine quality changes.
On a cantilever 3d printer? Plus cheap ballscrews have more rotational resistance so funny enough falling Z is not a problem.
I doubt I will retest it because I try to aim at settings that are most universal and realistic for the majority. I did one test print with 0.12mm layers in the video + all the theory behind supports that the lower the layers the bigger the difference will be. So if you print thin layers, the perfect z-axis movement is definitely what you want to have.
@@DIYPERSPECTIVE I was talking about a belt driven Z that did not use a ball screw. I said I went back to the screw to stop the Z from dropping. I could have used a lighter leash/key leash or a counter weight but the quality difference didn't seem worth it.
I have spent years taking thousands of photos with and without macro lens under every lighting and angle combination I could think of. I spent just as much time looking online for photos of others print quality taken at angles and lighting showing the detail I was looking for. Since the move to 0.2mm layer height has become a standard even for purely aesthetic models I believe something has been lost. Maybe resin printers are to blame.
My bad. I mean for purely aesthetic purposes of the print the resin printers are on another level. Personally, I don't care about them as I am more of a functional print guy - quick modeling, printing, and solving or improving something. But yeah, when the "quality" norm becomes lower there is always something lost.
Don’t have the link available but I’ve seen someone has developed a board to be put between the stepper and the driver that allow to block the stepper when there is no energy sent from the driver… seems an interesting solution, belted are nice but bring all the issue of proper belt tensioning and other potential problem due to the belt…
You know I've been wondering for a while now if you could fit two mgn9h side by side on a 2040! Thanks for confirming that!
It's like they were made to be used here :D Just be aware that I had to sand down those V slot nuts, otherwise they won't hold the rail as they are too tall. I forgot to add this part to the video.
@@DIYPERSPECTIVE If they were that would make sense lol. I'll have to order some soon!
That sound is loud but kinda pleasant :)
Would you mind sharing which seller you bought your mgn rails, and sfu1204 ballscrews from?
I'm looking to buy them and yours look pretty decent.
You should just search for the sellers with a lot of sales and good reviews. I am not sharing from what seller I bought because I can't guarantee the quality and I don't like to disappoint viewers. I hope you understand.
I would love to buy this kit and build it as a personal project!!
Maybe in the future I will provide something like that.
on Ender-3 i regularly print on 0.05mm layer heights
just install pom t-nuts + z-sync mod for minimum wobble, fixed. z hardly moves compared to x/y. you may see unprecise movement in first layers, measure z-axis movement with dial gauge. you can write a slicer script to simply add extra movement on those layers. elefant foot gone. at the end this is fdm, precision of ~0.2mm in z and ~0.4mm+ in x/y so dont bother getting it perfect. for resin printers its 0.02 in z/x/y, btw those also use pom t-nuts. i installed linear rails on x+y, only gave me massive ringing (aligned to 0.00x precision) but fixed all other issues. it vibrates more and has more friction than wheels but stays in place (hiwin!). made my hero me wobble on a bowden, nozzle scratched layers - installed the klemco mount instead. motor vref bumped from 1.4v to 1.8v to allow 5000 accel, wasnt necessary for wheels. after alot of fixes its solid now. cr10s pro v2 turned into a reliable 5000 accel / 150mm/s machine ;) this is pretty much the max for bed slingers. printing a speed uncapped 20mm calib cube is violent...
How tight is your ballscrew nut tightened to the mount? I try to leave them loose and lubricated so they don’t force swaying on the arm/rails, so the nut only provide push/pull motion.
Mine didn't have wobble so it is tightened rigidly. You can check something like wobble x. MirageC have great video about it.
@@DIYPERSPECTIVE was yours graded (C0-C10)? or did you just get lucky?
It says C7, which is kinda crap a.aliexpress.com/_mrOuLni
But I used only 300mm long one so it's almost the best case scenario. There is also C5 for not that much more a.aliexpress.com/_mrcUYVE
@@DIYPERSPECTIVE C7 probably isn’t too bad considering C10 is the worst graded and there are a lot with no grade at all. I think i’d be pretty happy with C5 to C7. Not sure what grade the ones are that came with my printer. Its a trex3.0+
Great video, There is a solution to getting a wobbly ball screw you should really check out the 3d printed z-wobble wings from MirageC and his video titled, Fixing Z Wobble from 20$ Ball Screws with that simple device, I think they would be a great addition to the z mode in this video and they would help any one looking to do similar z mode to their printers without breaking the bank. I did a similar z mode but with one rail, one ball screw and his wobble wings to my printer and got similar precision as your mod.
Yeah I know, my only concern is that it won't withstand way higher acceleration, speed, and jerk settings of the Z-axis. And those are something special. Going back to 3D printers without this type of mod feels so sluggish now. It's like going from a 144Hz monitor back to 60Hz :D
@@DIYPERSPECTIVE WOW I did not think of the higher acceleration, speed and jerk of the z axis, when I did my mode I was aiming for the highest quality prints I could achieve. Thanks for responding.
I aim for that too. But I don't think I sacrifice print quality with faster Z axis. I am not pushing super crazy speeds or anything but still they are 4x faster than the stock settings, which feel so much faster. :)
@@DIYPERSPECTIVEnot really sure I grasp what you means, looking at miragec video the speed of the z axis movement seems quite fast and well in for non planar printing, wobble wing (expecially last revision) or Oldham brass coupler available also for ball bearing should allow for usage of non perfect ball screw without risk of binding the movement.
add wobble x should help
Did you test quality versus speed? Can the upgrade do better at higher speeds?
No, but higher speeds is the last thing you want to do on cantilever 3d printers. But sure it can print faster because it is more rigid than stock.
Do a test comparison for laser marking.
Have you tried turning off z-hop? I got way more consistent layers by turning it off on my Voron V0.1
Yes I did, but with this mod I already get perfect z axis movement. And with right coasting and wiping settings I can extremely fast z hop everywhere without downsides of it. Plus it is nice when nozzle doesn't act as a wrecking ball on curled parts of the print :)
Do you have an example of this part?
Interesting video. There is a 3d printed version of this mod on printables and thingiverse too
Thanks. Are you sure? I have only seen the one which uses two linear rails but not the ballscrew.
@@DIYPERSPECTIVE you're right I was focusing on the linear rail aspect and not a ball screw
@@MetalRhino42 Currently waiting on parts for the one you're refrencing. Quite excited to have rails on my Z axis and only just now finding this video and learning of ballscrews(probs passing). That said, I'm interested in the model you designed as the one currently shared on printables is multi-part, but can all be printed with a KP3s. Hoping the prints last as long as a metal one, but only time will tell and saves a ton of cost. Think the whole build is going to be around $35
Prints look like they have heavy wood patterns on them, my printer has the same. It's because of the inaccuracy of the gears in the BMG extruder. dual gear extruders run in direct drive suffer from this wood pattern, mirageC and Vez3D made good videos showcasing the wood pattern.
i wanna upgrade my Y axis to perfectly tight linear bearings. This seems like a very easy upgrade. i wouldnt even have to have any parts made. But the current Y axis on the longer lk5 pro (cr10) has 6 wheels on the center of the bed frame which squeeze into the sides of the one singular rail. the bearings arent made to have pressure in that axis and they move. ALLLL those style bearings move in the axis that they shouldnt have pressure applied. every axis. x,y, and z.. i already did a dual z motor upgrade and a few other upgrades but just now when i was leveling my bed i noticed when you press on the corners of the bed it will move what seems likea teeny tiny amount but is enough to throw your nozzle millimeters. the x and z axis dont matter as much because of the way the printer is buyilt, but the y axis holds the entire bed on it and the bed HAS TO be level or the prints will fail. ive been searching through mcmaster and amazon looking for linear rails and im stuck between 2 roads to go. the lk5 pro is a like 500 dollar printer. so it wouldnt make much sense to drop 2 grand on high precision rails made for industrial applications, but i also dont want to drop 73$ on amazon rails if they arent gonna be precise. if the cheap bearings have play in them then it defeats the purpose of.. i was hoping you had some insight.. or a review on the rails you bought..
Personally, I am not a fan of buying expensive parts for cheap printers. That is why I just order these from Ali - th-cam.com/video/O7DMtMtY22Q/w-d-xo.html I modded my LK5 Pro with these and it prints great.
@6:40 linear rail are show form the top, and... why on Earth I see a solid core with sheet metal over it to form tracks? AFAIK the whole point of a linear track is to have one piece, tight tolerance milled and honed rail to guide a slide. What was the price point for the rail that You have used?
BTW after spotting Your work on Instructables You have a new subscriber 😁
I have no idea what you are asking for :D The shot at that timestamp doesn't represent what I meant and it is my mistake to include it.
But the whole lubricating and moving back and forth and applying some forces in different directions on the carriage with one rail before assembly works well with preloaded linear rails if you have a sticky motion with your setup (usually not enough weight on the carriage). And the ones with clearance have extremely smooth motion (unless it has a defect) but can have a tiny play because balls are undersized by a tiny fraction. But when ordering linear rails that cost ~$10 per unit, usually you have no idea what version you will get. The most important thing to notice is if the binding/sticky motion is because of not enough mass on the rails because it is preloaded or if it is just a bad rail.
Wow. You're serious in this 3D printing game, using a precision dial indicator, using ball screws doing inspection using microscopes. Never seen anyone doing this on youtube. would like to point out, you should consider building CNC 3d printer from scratch using thick grade A granite surface plate with 10 micron precision hiwin linear rails. This will solve so many issues. However 3d printer nozzle also needs to be accurate when it extracts ABS, that alone is another rabbit hole to jump into. If you do this, you'll also have a super accurate CNC mill and PCB sensitizing liabilities all in one CNC platform.
Yeah, I go "all in" as I can. I would love to design a 3D printer someday, but I am a guy that is never satisfied with the results unless they are near perfection. Maybe in the future when I test more topics with definitive answers on what works better and what doesn't, but for now I am trying to stay away from that time sink, haha!
I tried on a kp3s with ball screw and it's not worth it 😉 but it will work with better stepper motors and drivers 👍
Has anyone tried replacing the sloppy brass lead screw nut with a plastic, anti-backlash nut? I used them when building fast, accurate pick and place machines. I haven't seen anyone use them on a 3D printer yet. They are readily available from a variety of sources, such as McMaster Carr.
Do you mean POM plastic nut? I don't think anti-backlash does anything with 2mm pitch and 8mm lead screws on 3D printers. But POM nut + PTFE coated lead screw could be the simplest way to get the Z axis with very little friction and no need for lubrication. I ordered the parts and will be making a video about them in the future.
The only part you nee upgrade in a printer is the Z axis lead screw to a ball screw to cancel backlash. Linear bearings add to much weight to the moving parts of a printer reducing acceleration times. Belts are the best.
How fast do you need Z to be? Lol I can see your point if talking about X &Y axis.
Off topic, but which parts cooling fan are you using? Looks interesting
It is the Mechatronics Fan Group B5015E24B-BSR www.digikey.com/en/products/detail/mechatronics-fan-group/B5015E24B-BSR/5209731
@@DIYPERSPECTIVE sorry, forgot a word, I meant fan duct :D
It is my designed mod for E3D V6 hotend on the KP3S. I only have shared it for my patrons (for regular 5015 fans, not the one I am mentioned).
I would think if you want to improve prints, improving x and y movement would be the thing to do.
The gap doesn't matter if it is always pushing the same direction
Hey, that pattern that you are seeing on the side of your print has nothing to do with z axis. It is issue 6. This is a pattern that results from dual gear extruders. Hevort was able to solve this problem and I suggest you check him out.
Hey, I follow him and I watched the video. But my question to you did you watched the whole video? Because you making a very bold claim and comparing completely different 3D printer designs without knowing all the variables, just saying you know.
sadly most companies just want a bigger margin, without regard to the final product
I can see why no one does it in this instance, but yes, it is still sad that instead of improvements we will see just another Ender clone.
Such a spindle is absolutely unnecessary, as two linear rails are already installed, which I also see as too many.
Then another part made of aluminum Vertigen is simply unnecessary.
Then you should invest the money in an industrial printer.
I myself installed the Z axes in the SX1 with a linear rail each, so that there is no longer any misalignment. It also turned out very well.
But as I said, you can also overdo it and do it if you have too much money.
But otherwise a good educational video
This is a Google Translation, sorry for any errors
As far as I have seen these kinds of inconsistencies are because of extruder issues. Direct extruders are too ' direct ' any small imperfection will show up in the print. One thing bowden setups have a leg up on direct drive is that it somehow dampens extruder stuttering or missalined extruder gears causing a fluctuation in flow.
This mod looks like AF tho.
Manufacturers don't use ball screws because good ones are too expensive. For an Ender 3 it would probably add US$300 to its price. And that's just for 2 Z axis screws. Ball screws are usually have grade levels ranging from 1 - 10, 10 being the worst. Those you show for $40 each are probably around grade 14.
Add some wear to the wheels and lead screw nut and test again
Install a wobble-x adapter and try again.
Why was printing in plastic impossible? Genuine question.
With that shape, I could appreciate flex near the base, but given the clearance a different shape could have provided significantly more rigidity.
It will introduce a wobble to the X-axis from the heatbed movement, which will make printing results worse than a stock printer.
@@DIYPERSPECTIVE I'll take your word for it, though I remain skeptical. There is a printer called "The 100" that was designed to be an ultra-high speed 3d printer (eg. 10min benchy), that uses a completely plastic, 3d printed frame (from a series of bolted together parts) and does not seem to suffer from wobble given the rigidity (lack of flex) due to the part thickness.
@SeanLumly okay, but I am talking about using aluminum part vs plastic in this specific scenario, where it holds the whole arm and the printhead. There is no room to increase the thickness of it to be as stiff as the metal one with a reasonable size. Of course, you can make a plastic part that can be more rigid than the metal one, but it must be way bigger and better engineered. Nothing is an apples-to-apples comparison when it comes to this topic, you know.
@@DIYPERSPECTIVE I'm looking at the design, and yes, room would have to be made for the screw (and possibly couple), but I believe the the screw can be threaded through the structure allowing the mount to extend out in the y-direction, up in the z-direction.
I'm not suggesting you do anything in particular or that you did anything wrong. There's no need to be so defensive -- I'm suggesting a potential path for an alternate solution.
Bro, there's no way you asked a question and I am trying to give you my best answer and you are saying I am somehow "so defensive" lol. In each comment you introduce new variables to the original question and somehow wrongly interpret my answers. Sometimes I ask myself why I bother responding.
Your result was quite expectable. You should solve the stiffness issues first before you'll move on the better screw and/or better linear movement elements.
There are no rigidity problems with the metal bracket upgrade. This is probably as good as you can get on a cantilever 3D printer.
Cheap ball screws aren’t precise, they’re rolled to shape and not machined like good quality ones.
You should try that upgrade but with a normal Z screw!
if you want to test a z-axis improvement shouldn’t you test over a wide z range? the small z-axis range means you are mostly testing near the bed which minimizes a lot of z- errors
Not really. If the alignment is good and rotation is smooth all the way I don't think that is needed. Since the video, I made a few taller test prints, and the layers look exactly the same no matter the height.
I'm a belted z fan boy
YES! This is on my list to test one day :) On what printer? Did you make before and after comparisons?
@@DIYPERSPECTIVE I swapped a voron V0.1 to belted z and now I have a V2.4. Never had issues with lead screws and didn't really notice a difference after the switch but I love the speed and it's one less thing I need to grease every month.
That's true, even the greasing alone worth it if the print quality remains the same.
Thanks for doing this so i didnt have to lol
the only way to benefit form this setup is to improve extrusion and the motors as well
This was with Trianglelab DDB v2.1. I printed cylinders in vase mode and I didn't notice any repeating extrusion inconsistencies. And I really doubt "better" motors will make any difference. Maybe I just should make a video about it to prove it or eat my own words lol.
thx @@DIYPERSPECTIVE ,you could increase layer height to minimize the spread a bit just to highlight extruder related inconsistencies. both filament diameter and extruder precision (I mean the gears) contribute. I wonder if you can find a test setup to isolate extrusion, maybe use a stylus to mark a hair thick line on the bed? but then pressure will twist the gantry, I saw that it's optimized for speed but not for rigidity. this setup should be considered heavy and strong, so more balanced between speed and precision/torque. maybe use a laser and a fax paper. anyway at one point I'm sure controller and motor type will be the limit (try high end servos?) I notice the backlash is still there but smaller, I wonder where is comes from.
@@DIYPERSPECTIVE about the repeating patterns : some frequencies will be mangled if the vase mode object has himself a layer frequency that don't create visible harmonics. depending on the type of extruder (gear sizes) you can have multiple freqs. to test well you could try to sync more the layer volume to one gear turn.
This is a good point. But if I would properly test the extruder consistency I would just assemble a jig and with a dial indicator measure the consistency of full rotation. This would completely eliminate other variables. I feel like I reached a point where the upgrades give only diminishing returns and chasing every last bit of quality improvement is too crazy for my taste.
@@DIYPERSPECTIVE true, it depends on the base platform that you work on, this one shouldn't go too far. But as a viewer I dream of a perfect 3d printer. I feel that the extruder and filament is the next focus for the community due to their contribution to the overall quality. The servos are a possibility but limited to high end large printers for repeatability and speed/torque. The software and motor controller have some room too, like non linear and better handling of flow and advanced quality improving pathing, but people can't easily improve it themselves with the limited offers on the market (f-in hard to develop). Generally I feel that people invest too little in the printer compared to how much filament they go through. If most are enthusiasts, we should see more high end, next gen offers. Look at the mountain bike amateurs for instance, they often invest like 4000$ for the bike only and they can't even use it every day living in a city..., then throw in that we are still in the prototype phase of this industry, barely after the ford model Ts (prusa/creality), where most of the "modern" models are hand crafted, tailor made unique art pieces extremely expensive. today we can get the best of the performance of the 30's car in the cheapest one but with even better usability. we need more people like you pushing the boundaries with great analysis.
Rigidity, good point! Lesson learned!
Yup! More rigidity = fewer artifacts. Quite a few people even suggested using ball screws on X and Y axis to eliminate the ghosting artifacts, basically what Pantheon 3d printer is. But it needs high-quality ball screws and they cost a ton! :D
like the 6th video i've watched were results didn't make a difference yet people still install lin rails ...
No measurable difference, so no difference.
It would be more interesting to see those lead screws on X & Y axis.
linear rails and belts on z are better
should you do this? as a hobby yes, as a business no.
belt will be better
or buy bambulab a1 mini direct and have ams multicolor XD
nice video, but whats the sense of using cheap chinese rails and a cheap chinese ball screw(tou cant call that thing a "ball screw")... vibration come from the screw that is made in a plastic lathe. you should try to do the same with a good quality rails and screw
Thanks, but what sense does it makes to use a $100-$300 ballscrew on a $200 3d printer? I am jus giving a sensible scenario on what manufacturers could do out of the box and what results we would get. The ballscrew I received didn't have anything that would cause problems. I checked the z movement accuracy and rotation smoothness. No problems whatsoever. I don't understand why you hate it so much. :D Just because it is cheap and Chinese? Kinda weird, kinda weird.
Direct drive on a cantilever printer... *sigh*
Okay the shouod-be Elephant in the room is…. WHAT QUALITY / Acuuracy rating ballscrew are you using!? As well as Grade rating of the former lead screw??
Believe me I’m all in on ballscrew usage I have some awaiting going on my FT-6 Large bed printer with tons of weight and custom aluminum bed plate and platform! That’s 700x330mm and weighs a ton! That I needed Z-hopping to even remotely be a possibility Vs a total joke via the friction on lead screws (30-70% transfer efficiency) vs ballscrews and their (85-95% power transfer efficiency). That in this scenario with a tiny weight on a short lever arm and hardly any mass to lift up and down … I’d personally not expect to see much improvement and more likely without a decoupling device for the ballscrew have the majority of people doing a mod like this to see some
Z banding and issues from the fact ballscrews are so rigid and their radial deviation being forced to move elements of the chassis that usually lose the tug of war.. I’d say generally speaking most people will get more consistently decent results from a lead screw Vs ballscrews lest in the ballscrews is very short and very constrained with many (4 linear rails carriages in this case) please can you tell me where you bought that particular ballscrew and it’s rating C0-C5 ??
To have no issues I’d assume you’d need to fork out a lot of cost for a C3 or so! Without a decoupler!
Lest YIURE very lucky or bought a ground ballscrew again also high prices and usually the only way to get C0-C3 accuracies as the cold rolling process imparts too much axial deviation into the steel .. and sadly in the smaller 12mm SFU1204 (types I use on all things) vs. SFU1605 or another common 16mm type those internal stresses in the steel are less likely to manifest in inaccuracies / wobble around center of axis because there’s more mass to resist the curvature after the grooves are pressed into it for the bearings to run into..
Anyway off topic now I’m going into TMI.
And I only have watched 2/3rds of your video maybe in the end you share the ballscrew quality rating!
I’d be very eager to know that and source of it! As … it just seems to me you’ve gotten very lucky to be able to harm mount the ballscrew nut onto that machined aluminum
Piece (which btw does that PCB shop have good pricing on small CNC parts? Is that some secret I should know about as it’s usually very expensive)
But that made me worry you’d for sure have unwanted artifacts in the print just due to the quality of these often unrated or off the charts in amount of inaccuracy type ballscrews.
Any info on that would be very appreciated as anytime I see someone without issues bolting it directly on I’m
Eager to hear the source of the ballscrews! ??
Please? Ty 🙏
I am impressed that you wrote all this but only watched 2/3 of the video :D But being serious I really appreciate the detailed comment. I used a cheap rolled C7 SFU1204 300mm ballscrew from Aliexpress for ~$20. I don't buy ballscrews from there on a regular basis, so I can't tell if I just got lucky or if the quality is decent for my specific use-case scenario (I also bought an Oldham and other types of coupling for the worst-case scenario, but I didn't need it as I had no binding issues). Buying a more precise ballscrew that was ground and not cold rolled, those have some insane prices.
For the pricing on CNC parts, you should just upload the STEP file and see the price for yourself. This aluminum bracket's price was ~$75 ($144 with all the taxes, import fees, and DHL shipping, no discounts).
A couple of days ago I made this test print ( i.imgur.com/fmL8xr2.jpg ) and it looked way better than even on a dual motor z-axis 3D printer with regular lead screws. I couldn't see any repeating irregular lines and the photo on the left is the worst-case scenario with lighting from the top (the best scenario on the right). At this point, I think I am running into filament consistency limitations. That was printed with +/-0.03mm filament.
RIGIDITY is key .
Has to do with patented parts and shit. Cant make everything standard.
No
Dude, just work little bit on your pronunciation!😊
Yeah, when I record the audio it seems fine, but later when I forget what I am supposed to say and watch the video, I clearly noticed that. Thanks for pointing that out.
Overengeniring
short answer. HELL NO.
it's not worth it.
not only is it extra hard for the steppers to rotate it.
this looks like a chinese piece of shit that just like a leadscrew. will be bent or off centre.
i litterally have the same ones. and yes they were bent
I like your content but your to liste to you speaking English is really exhausting. Can you please pronounce words more clearly.
These upgrades are for LONGEVITY, not precision. For someone who isn't printing professionally, completely unnecessary. For a pro, less maintenance and upkeep.
What's the point of getting a cheap printer and then spending time and money to upgrade it? Get better one to begin with.
The reason is: nema 17 have not enought torque to '''wobble''' around the printer. And therefore make the prints bad.
This upgrade is not about quality You misdirect people We use linear rails for speed and rigidity for a long time
Dude, what the hell you are smoking? ;D You just contradicted yourself with this comment. Lol.