Incredible! A huge thanks to you, Michael, for the tremendous work you've put into testing this system on your printer and addressing some integration challenges for specific situations! Your ability to explain the system, its background, and the installation process is truly remarkable - straightforward, concise, and genuine. I'm truly pleased that someone with your expertise can recognize the potential of a solution like this! Cheers!
And all this is based on your awesome work :D Also what do you think about this potential solution for the slipping problem with to little weight: Maybe increasing the distance between the Rods slightly could help keeping the ball on the tracks cauz it would sit lower inside of them? This would remove a little bit of the rotational axis becaus the wobblex parts are a bit closer to each other, but id guess theres isnt much rotational movement on an ender style z axis anyways
Thank you @MirageC for providing the parts for this review. I saw your original announcement of these and was considering purchasing a set for my Ender3. Not having any other printers I would not have spotted the issue @TeachingTech did. I have no problem with the weight solution but I look forward to seeing what fixes are found for this issue.
@@D4RKFiB3R I am unsure, my experience may not relate directly to the wobbleX, but I have for many years installed anti backlash springs/nuts in this fashion and it's always achieved better results.
Yeah this level of work is getting to the point that you can go belted Z. No grease, no wearing out, no backlash or weight issues. Only downside is if power is lost.
Did the same mod to my Ender 3 and hands down best mod I've done to it. Perfect Z layers now. On my Ender 5 though I installed another anti z wobble device, but didn't quite fix all the issues. Still prints great, but under the right light you can see the inconsistences.
would love a comparison between this and a more conventional, brass, Oldham coupler. It might solve the issues where you need extra downforce on the E3 as well
Oldham coupling does have the very same problem, although MUCH MUCH less!!. The clearance between the parts is still significant enough to cause the same thing! Although, since it would be MUCH less noticeable by eye'ing the print in progress, it would be difficult to recognize that the Z-banding comes from the this very problem. Another downside of the Oldham coupling compared to WobbleX is that Oldham can only mitigate movement on the X-plane, aka lateral. It can't compensate for rotational movement, which is exactly why the WobbleX was developed. (But yeah, it is MUCH cheaper indeed. The WobbleX seems to be very overpriced to be honest). But maybe this 'jump' problem, with both style of couplings, might be solved when you also add an anti-backlash nut? Dunno, can't test it out (my WobbleX are still in the mail and lately I don't have time anyways to experiment).
@@CookieTube the main issue with Oldham is that while it does decouple movement in the plane, it doesn't decouple the tilt of the axis. Whether that is good enough will need testing but I imagine that any rod non straightness will introduce some tilt.
@@JohnMeacham yep, indeed, hence why WobbleX was developed. You ALWAYS MUST compensate for tilt in the cases we use lead screws or balls screws for 3D printing. Hence the WobbleX. An Oldham coupling is simply not made for the purpose it is used in the 3D printing world!! Oldham couplings are flexible shaft couplings that are used to connect *TWO different shafts* : the *strait driving shaft* and *strait driven shaft* , which almost meet in the middle piece of the coupling. It is used to counter *parallel* misalignment that occurs between connected shafts. Important here is "parallel"... Aka, the two shaft are perfect parallel to each other (not crooked in any way), but their center to each other might be slightly misaligned. Many people don't get that (which you can also read in several comments about it).
Spring preload is definitely the way to go. My first thought due to the size restriction might be to use clothes-peg springs (aka a torsion spring). They are small enough with sufficient travel, a simple horizontal pressure point and have a handy hole in the spring section for additional mounting options.
These things are awesome. I used a set on my highly modified anycubic mega zero. Also saw the bounce issue and a couple of rubber bands 90° to each other, and that was enough to see huge improvements.
After seeing the initial video on the wobbleX I just looked at Oldham coupler models on Printables and remixed one for my ender. It helps my print surface be as clean as possible even if I don't have visible screw wobble.
I have a similar design that is a complete 3D printed solution for the frames, and found that a sizeable rubber band wrapped around the entire assembly works well on my CR-10
@@deathcube2006 check fit. sometimes they have small burrs (nicks) that keep em from moving smoothly. If that's the case wet sand 2000 grit on mirror/flat surface, clean then light machine oil.
Oldham coupling have the EXACT same problem! And it does not matter if you have the strait version or the slanted version; they both exhibit the same problems. But not as massive big as what you saw happening in the video with the WobbleX though. There can still be 'too much' tolerance in the Oldham couplings between the three plates. You can even feel it when you gently push or pull them apart. But is is not as much as with the WobbleX, which you can easily see happening during the print by the naked eye. As such, you should be aware that the very same thing happens with Oldham couplings. Using that style of coupling will not prevent it 100% because those tolerances need to be there for the main function of the coupling to work: sliding.
@@deathcube2006 Make sure you have the right version. Some models of the Ender3 must have the 2 hole, while others can use the 4. Get the proper screws and nylon lock nuts. Lube the coupler with a graphite or PTFE type lubricant. When you get the couplers, take them apart and inspect them very well for burs, machine defects, and debris. Any defects, or play in the dovetails will greatly hinder their usefulness. Assuming you got the brass version, if there is play, you can take some of the play out of it by tightening the coupler end piece in a vice with the jaws of the vice parallel to the travel of the dovetail. I'd also add a bearing to the top of the Z axis leadscrew to keep things vertical and allow the Oldham to just take up the side to side movement.
@@CookieTube The dovetails on an oldham can be much more accurate and reliable than this wobbler plate. And, if the dovetail has too much play in it, it can be tightened. Mine have 0.001" of side to side play in them. There is a reason some high end CNC equipment and other industrial equipment use Oldham couplers.
I think your assessment about needing to add tension to hold it all together is spot-on, when originally looking at this system that was one of my immediate concerns, that the gantry can lift up freely and thus looses an axis of constraint. Gravity is only good enough if your z axis moves freely enough. Also in general you may want to methodically work through your lifting z axis and see if there is excessive binding that's managing to hold the weight up, or a natural skew caused by where the X bar is tightened to the side plates. You need to lock the steppers to a known initial state then level the X bar from there usually, else it tends to jump to in-between positions etc. I'm personally a fan of the single-motor belted-z setups, they are rock solid once the constraints and parallelisms are worked through, but it requires some definite expertise to dial in.
@Roobotics : I totally agree with you. In addition, while adding extra weight to the upper component will help, it is adding weight to the bed assembly, with the potential extra wear on the Z axis lead screw/nut. A spring tensioning system would be preferable, I think.
Off topic here, but since you mentioned it: Have you ever found a place/video where they compare the two systems side by side (dual coupled Z-axis with 1 motor vs. dual Z-axis with 2 motors)? All I can ever find are 'superficial' videos where they talked about how to install them and how to set them up. I want to find an in depth tech video comparing the 2 systems *properly* with all the ins and outs, and not just the _"with 2 motors you can have automated gantry tramming"_ ... What about drift overtime in both systems, what about slack in the belt overtime, what about Z-banding in both cases, what about 2 motors working against eachother, etc etc And also: where to do mount the steppers best? On top (hanging Z-axis) or bottom (sitting Z-axis)... What style of coupling to use between lead screw and steppers (many styles exist), and their benefits or drawbacks... Do you 'fix' the screw on the bottom/top, or do you let if free standing/hanging, and if you fix it, do you do it in one plane or both planes or none at all but instead restrict the lateral movement it a bit... etc etc etc... Anyways, sorry for the wall of text....
Hi great video! I was already eying to buy the wobbleX for my soon to be rat rig, and now i am certain i will. However I do think that creality machines would benefit from a spring loaded double nut specially the ender3 whit the worn screw, i am running a set on my cr10v2 and they did a fantastic job eliminating backlash and elongating useful life of the screws.
I think you are probably spot on with the extension springs 3 or four would be best if possible they would also prevent some bed slinger wobble from the bed motion causing tilting front to back
or 1 wide tension spring in the middle, in which the lead screw can pass thru, but with flanges on top and bottom (thus needs to be custom made?) to hold the whole system together as one packet.... ..... or an anti-backlash nut....
I has been watching you and all the others for years, I went away for some years from 3d printing but now as I came back ofc we all aged but you look almost the same as I remember you years ago (I cannot say the same for the others) goo to see you still around) 👍
I have a modified Ender 3. I added a worm gearbox to my Z axis and I expanded the Z and Y axis to 1M (1000mm). The z wobble has been quite rough even before the expansion. So far, tests with the WobbleX have been impressive. I didn't know about the hole sizing on the screw and printed the wrong part but drilled it to fit and it's working fine. I would have printed it again but I have to take apart my worm gearbox to install the wobbleX and it was already apart when I discovered the size difference. The WobbleX for the ender3 doesn't have as much play as the others, and at about 500mm it does reach it's limit, but when I'm printing parts that large a little wobble is expected. One thing to watch out for is lifting the gantry. Those little magnets don't hold it together very strongly and you can easily lift the gantry and loose a bearing if you aren't careful.
Sounds like a nice idea, but with its challenges, and thank you for sharing. I thought to share my experiance in this matter. After a year of constant use of my ender3 I had started to get a slight Z wobble, so I knew it was time to replace the obvious parts. Well I thought at the same time I would upgrade the height and add an extra 150mm and so I bought a upgrade kit and it came with a perfectly straight ACME lead screw and I also bought the Z-axis T8 Anti-Backlash Spring Loaded Nut from Amazon After some tweaking in Jyres firmware it was all set for a full height test and the out come was perfect. So personally if you have a bent ACME Lead screw or a worn thread and a worn brass Nut I would just buy another set and also fit it with a Z-axis T8 Anti-Backlash Spring Loaded Nut. The Z-axis T8 Anti-Backlash Spring Loaded Nuts is good to help on the Z lifting as the bed or head moves up and when it is compensating the bed level or if you have the Z hopping setting on
LOL, This is the EXACT same I have: 150mm extension...... + dual, coupled on top, aftermarket dead straight lead screws + anti-backlash nuts on both. I've done the extension not so much to be able to print taller stuff, but far more because of convenience in having clearance on top for all the extra stuff without actually loosing the original Z-height (LED light, stiffer cables/chain so they don't flop on the bed, having clearance without the need to almost 'fold' the cables when on max height, the dual coupled Z-axis system, etc, etc). And also extra corner brace plates on the bottom + CR10-style diagonal braces. The whole thing is solid as a rock and even (very high) resonance vibrations are almost non existent.
Thanks for this review. I’ve built a custom 500x500x600 corexy triple Z axis printer and the 600mm lead screw’s definitely have some bend in them. I was going to go and buy some more expensive ball screws to remedy the problem, but for the price, I’ll give these a try instead. I should be able to adapt the ratrig bed mounts to my printer. ( machinist here, so I have access to machines to make changes on the fly as needed). Keep up the VERY informative reviews. It’s much appreciated!
My V-Core 3 was originaly 3.0 and one of the first kits ever. The Z-screws were not the greatest. The taller the print the more visible a slight 4mm-period Z-wobble became.I upgraded the Z-arms to a printed upgrade called wobble rings (sorry, forgot the name of the author) which does pretty much exactly hwat the WobbleX does. It completely fixed the issue. So much that I did not upgrade my printer to full 3.1, I just changed the already failing XY idlers to bearing stacks (slightly different size than 3.1 stock to accomodate 3.0 geometry). I did nothing new to Z axis because the wobble rings made it perfect already. Now I put together V-Core 3.1 for someone and the lead screws are way better. They were shipped inside the grooves of some of the 3030 extrusions so zero risk of transport damage and they looked perfectly straight. However I found that for a 500 size of the V-Core the bed itself is already on the border of what 500mm length of 8mm leadscrew can support without bowing sideways a bit and when you put a lot of plastic on top it only gets worse. Any tiny non-straightness of the leadscrews not even visible on them when they are on a table gets exagerated once they are loaded in compression by a heavy bed. The oldham couplers included in V-Core 3.1 that let the nut move in XY plane are good enough to fix this at least for the good lead screws I got on this machine. You actually see them moving as the bed travels if you look closely and yet a 500mm tall printed cylinder has absolutely perfectly smooth walls all the way. I think that for most printers the WobbleX with its 4-axis freedom of movement (translate in XY and tilt around X and Y) is overkill and just a sliding oldham coupler like the V-Core 3.1 would work too. But there might be some cases where the WobbleX fixes an imperfection other couplers would not.
I was in a similar situation as you with my v-core 3, but I waited and did the 3.1 upgrade and that also completely fixed my z issues. The 3.1 upgrade also incorporates a similar decoupling system that's more rigid than wobble-x, which I think is actually a better system.
@@AgentHubcap the 3.1 decouplers are for sure easier to use. You can lift the bed and put it back down without worrying about it and it just works. With the wobblex you need to make sure everything sits back into its place. Unlike the sliding brass Oldham decouplers the balls can move out of their grooves. However I get what the designers of WobbleX say about the 4 axis over 2 axis freedom. If the nut is badly made for example and it's top surface is not staying horizontal as the screw rotates (misaligned thread cutting tool makes the helix not coaxial with the outside of the nut) then the RatRig decouplers will tend to bind and not slide freely while the Wobblex will not even notice and just works. So for good components the RatRig decouplers are better because they are more robust and easier to use and don not require the enough weight on them like was shown in this video. But with worse components there can be situations the Wobblex will be better.
Great video. Excellent lighting on the prints that actually show the surface variation. Fortunately this was back in stock when I looked, so I bought one for my Monoprice Maker Ultimate III (Wanhao D6) printer. I am hopeful this will make my printer usable for prints that need smooth sides.
LOL I just ordered these a day ago. After MirageC's original video they sold out in no time. I saw they were back in stock and picked up 3, now Teaching Tech video is out they will probably sell out again. I am quite excited to try this out, it looks really cool and it should work fine on my Voron Trident with the heavy bed loading them up.
The one you printed for the Ender 3 looks like you could have a gantry wheels with flat spots on it. I've noticed this allot with my own Ender 3 and i have to replace them on a regular basis. either from it sitting to long from non use or normal wear and tear. so ive made a pratice of replacing my wheels once a year, and that equal length banding has gone from my prints! Just my thoughts and suggestions on replacing your wheels for the horizontal gantry.
This. I had the exact same problems with my ender 3 clone - coming with wobbly wheels fresh from the dropshipper - I replaced all of them and the wobble was gone.
@@amorton94 Those wheels are a dime a dozen now, that linear rails are the new fancy - and you gotta have a whole lotta more work to do, to retrofit linear rails onto an ender - ask me, how I know ;)
Yay! Thank you for trying out this potential solution with several different printers. Yes, compression is key to this since some machines enable Z-hop for travel across a print.
What about adding an Anti Backlash Nut with a extension spring between the top part and the te bottom part of the wobbleX? Then the assembly Is in compression. It will almost look like the anti backlash assembly on an Ender 3 but the spring is in extension instead of compression.
I printed similar setup without pins and magnets but with rubber bands. Magnets won't hold any position if printer tries to dig into bed, springs and bands give up in Z but keep nut from rotating
I can see this being an issue if you use Z hopping. Adding springs might help but the quick up and down movement of the Z might still cause problems. I think that if the screw is wobbling that bad then replacing it is a better option.
I've also used the same box-shaped POM clearance nuts that the Rat Rig V-core 3 originally used on my self-built printer with good results. In fact, I feel that on the combination of screw and nut, the appearance of layer lines comes more from the fit gap between the screw and the nut. It is well known that almost all screw nut combinations have very large fit clearances. Most of the screw nuts use trapezoidal threads. When the screw is shaken, the trapezoidal threads will cause height errors due to sliding along the radial slope. Therefore, in my humble opinion, the solution to the Z-axis error caused by the screw nut is mainly to solve the fit clearance of the screw nut. It is necessary to make them as tight as possible and avoid radial relative sliding. And the "Rat Rig V-core 3 original box-shaped POM clearance nuts" that we both use do a pretty good job of that. After the gap adjustment screw on the top is screwed to the proper position, the relative displacement of the screw rod and the nut will hardly occur in the radial direction. I think this is why your Rat Rig V-core 3 has a very good Z-axis performance at the beginning. . Its only disadvantage is that it is a bit bulky, and I don’t think the wear resistance is a problem. At least my machine has been printing for hundreds of hours, and there is no problem. I just readjusted the clearance screw halfway.
On the ball screw, the gap between the screw and the nut is very small, and most of the ball screws are relatively thick, and the straightness is relatively better. That's why it works so well in its raw state most of the time. If the ball screw bends, the bend will be amplified and faithfully mapped to Z-axis motion because there is no clearance issue with the screw nut. Looking at a section of the screw rod, due to the bending of the screw rod, the movement along the vertical direction has an angle offset, and the local distance relationship forms a right triangle, and unfortunately, the screw rod is at the oblique angle of the right triangle. On the side, the actual Z-axis travel distance is on a right-angled side of a right-angled triangle. According to the Pythagorean theorem, there is an error between them. As long as the ball screw bends, the actual movement distance of the Z axis will always be less than the distance output by the printer, which will cause the distance between layers to become smaller, which will lead to excessive extrusion and layering. Therefore, in my humble opinion, using the delamination device on the ball screw is mainly to increase the smoothness of the movement of the screw nut. Because when the two ends of the ball screw are fixed, if there is no delamination device to release the radial swing displacement, the bending of the ball screw will cause the pressure of the screw and the nut to increase, and the jamming phenomenon will occur. The coordination error of the ball screw and the nut will affect the axial motion accuracy. To a certain extent, this can explain why the solution to the layer pattern problem has almost become "metaphysics". The masters from all walks of life have exhausted all kinds of solutions, but they have not been able to solve the lamination problem caused by mechanical reasons in a 100% repeatable and stable manner, because it is a problem that cannot be completely solved, and the mechanical error is there. All we have to do is to make the screw as straight as possible and make the fit gap as small as possible.
@@jackgai1929 THIS! This is what I love so much!! Please make a youtube channel with stuff like that!!!!! Anyone??!! Not the _"it looks ok, so this is the best solution"_ , but the actual science, both theoretical and practical, and the very in depth look at things from a deep mechanical, and yes, even theoretical viewpoint (as in: you'd never ever would see a difference otherwise, but it is the thought that counts, and "many invisible errors make a large visible one").... Thank you for the very interesting reads!
Tension springs were my immediate thought, glad you mentioned that. Been following Hevort and love what Mirage and team are doing. I was also contemplating doing this on my vcore3since mirage posted it and glad you covered that as well.
@@shibbypoopalot Why not just 1 tension spring in the middle, fixed on the plate above and fixed on the plate below, in which the screw can pass thru? Or 4 small rubber bands around the corners? Or just a anti-backlash nut?
Great Review!!!! I have an idea for holding them together, use Lego's silicone elastic bands to clamp the parts together. They last for a very long time and will be easy to design for.
I would love for 3D printer manufacturers to start looking into this and implementing it as a feature instead of a DIY. I honestly would like for the option to choose what features I would like from a 3D printer at checkout, like the option for this fix over a tradition ball bearing, or a glass or rubber build plate. I think offering options like this on manufacturers websites when you purchase a product could be beneficial to people who dislike the whole DIY side of things. If they did this though, they would need to also offer a base line for people who don't want excess options when purchasing a product.
In the cnc industry we keep ball screws under stretch to eliminate this issue, among additional strategies obviously. I'd love to see that implemented somehow. In my experience, stretching a ballscrew can improve accuracy 90x. And it doesn't require that much stretch for a ball screw that heats up very little. Of course a glass scale or a linear transducer really would be infinitely better for this that relying on ball screw pitch error to be minimal.
a leaf spring would be a good idea here, take a look at whoppingpochard from the voron team's kinematic bed set, he uses leaf springs to keep the bed attached even with a kinematic mount
Oh great Michael... now I need to order these for my RatRig VCore 3.1/3.2 rebuild LOL... these look great. Thanks for the video and excellent demonstrations!
I use Oldham coupling on my original CR-10 it is working kinda the same principle as the Wobble X. Also worth watching HeVORT another video about the extruder gear causing inconsistent layers :D
Oldham coupling has more friction and only has translational degree of freedom, but usually rotational degree of freedom is also required, in particular with ender and cr machines, as z nut is mounted onto bended part, which usually is not perfect 90 degree bent, and may even deform under stress caused by feeder motor in bowden configuration.
What @Dima said + Oldham coupling does have the EXACT SAME PROBLEM, although much much less. You probably wont even notice it when you just eyeball the print in action, but it is there for sure. The tolerances on those Oldham couplings are not that great. You can even feel it coming apart when you push/pull the coupling. So, yeah, very same problem there. You might be able to solve it by also using a decent anti-backlash nut though.
@@CookieTube in Oldham coupling it is fixable, just make sliding parts not rectangular, but dovetail instead. i printed such a coupling, but it didn't turn out to be as good as just precisesly set z screw and nut.
@@AcTpaxaHeu No, a dovetail channels have just as much play in them up/down as with the rectangular channel. The shape doesn't matter. And If you think about it/visualize it, is logical that the shape of the sliding channel doesn't matter. I have several from several brands with several shapes, and all exhibit the same issue. It is just a matter of tolerances, not of the shape.
The guide rollers on your Cr10max might be what's holding one end of the z gantry up. You could try reseting them by loosening them, letting the gantry sit down hard on the lead screws and then tightening up. Also adding little springs to hold it all together will help too.
What do you mean by letting it sit down hard? I went from being frustrated with my Ender 3 to a v400 delta over Christmas, but I'd love to solve issues like this on the ender.
@@TheRealComradeOtis with Creality (and likely many other Cartesians with V-rollers) the X gantry extrusion is not locked into the bearing carriers so its possible for it to get askew from one side to the other, when this happens one of the two outer bearings on each side will have uneven force causing binding even if the tensioners on the inside are good, Good reason frame alignment trumps all in the factor of elimination print flaws, Linear rails have the same issue but to a much worse degree, they need to be perfectly aligned with zero misalignment or the bearing carriers will completely bind.
I would be interested to see the results of just adding weight to the z axis on the ender 3 for comparison sake and then maybe with other machines too.
Woah, this is sweet…this might have solved the issue I had with my OG Ender 3 Pro (Upgraded with a MicroSwiss Direct Drive extruder). Took it apart and re-assembled multiple times and couldn’t get rid of those inconsistencies.
Good review. A lot to change/modify. We will have to assess whether the z-wobble banding is sufficient to warrant the number of mods. Great idea, though. I'm not questioning your integrity as you are one of the few that do try and give an honest, unbiased opinion. However, I do believe in 100% unbiased, uninfluenced ( appearances - no room for doubt ), by citing the great Leo LaPorte "The Tech Guy" mantra which is anything he reviews he purchases with his own money so as to have skin in the game just as a normal consumer does. He has been providing tech tips, recommendations, reviews of anything electronics (computers, operating systems, phones, printers, scanners, audio equipment) since the early-1980s.
Another option instead of springs could be small o-rings, mounted/hooked in much the same way. Seems like they might couple the two halves rotationally with a lower amount of tension than required by springs.
The wobble alters the Z height it doesn't matter how much X or Y is compensated. It's at the leadscrew and brass nut contact location. A better solution, besides investing in a new straight leadscrew, would be to stabilize the leadscrew by passing it through a second brass or delrin nut before it contacts the first brass nut. Anti backlash nuts are also worth considering. If you have any better solutions or why you think my reasoning is wrong please share your findings! Have a good day!
Your over a beer reasoning is wrong. The Z height change due to wobbling is trivial compared to the swaying in the xy direction due to a bowed screw. The wobblex is a very good solution to allow a cheap screw to be used.
I changed my Ender 3 to a Dual Belted Z motion system and my banding improved massively along with hopping and fast Z motions, though this seems like a decent interim solution
Belt driven Z should not suffer from these lead screw artifacts. Right? From what I've seen SecKit is one of the few printers utilizing belt-driven-Z along with linear rails and 3-point bed kinematics. I don't think there's anyway to improve on that
I am grateful that my nearly-2-year-old printer still is not exhibiting Z-wobble; dual-Z steppers with a timing belt keeping them in sync + braces at the top of the threaded rods to ensure they remain stable has, apparently, been the best feature of this machine that I never even realized.
Regarding your Ender 3 problem. Have you looked at the extruder wheel wich is pushing the filament through the bowden tube? Sometimes it is not really centered causing filament beeing pushed irregularly, showing inconsistend extrusion.
Nice, even though I have a CR-10S with dual lead screws, I decided to buy a set because I want to contribute to the project by creating a design revision that works well for that setup
What extruder do you have on your Ender 3? MirageC actually has a really great video that looks at inconsistent extrusion with geared dual drive extruders like the Bondtech BMG. I am wondering if your Ender 3 prints would clean up with a non-geared extruder more similar to the original extruder the Ender 3 comes with. I have a BMG clone on mine and get some Z artifacts, and I am wondering if it is worth trying a non-geared extruder. I only switched because my original plastic one broke. That in combinations with another commentor's suggestion to change the wheels would be really interesting to see the results.
Oldham couplers are simpler, cheaper and don't rely on weight to keep things constrained. Great video, as always, but having to artificially weigh down a gantry seems counter productive
Mine will arrive in just a couple of hours! I have an ender 5 plus, so I think it will be sufficient weight to keep everything together. I also have some spare backlash nuts for some extra pressure, but I don't think those are needed.
9:51 This problem actually exist with the older MirageC wobbleX design (forgot the name of it). It rotates but the Z doesn't move unless its heavy top. Happen on my Ender 3.
I used a couple of Oldham couplers printed in abs and sanded to be as smooth as possible, with some grease in between and it's years I don't have banding... This is clever but even overcomplicated
I know you tried adding a thrust bearing in your previous video, but in this video it looks like your "rigid coupler" is hanging in the air. The problem with that is that the NEMA17 motor has a spring inside of the motor, and it is not built to hold steady vertical loads. If you want your rod to not move up-and-down because of the spring inside the motor, you must make it so that the rigid coupler is sitting pressed against the motor body with a thrust bearing. That way whenever the rod wants to push on the spring inside the motor, it doesn't have anywhere to go as the motor body and the rod are not allowed to get displaced on the vertical axis. I had some lines appearing at various heights on my cheapo biqu-b1, and after making the rigid couplers rest on thrust bearings that are pressed against the motor body those disappeared. I didn't even bother printing a part for them like you did in the previous video about this subject.
i have been using a fully printed version on my ender for about a year and love it. you know what are the hidden benefits of this that i love so much? you can lift the entire gantry up by hand to do maintenance any time you please since it is is no longer bound by the leadscrew. secondly, it is now harder to destroy the bed with unexpected or illegal negative z commands for the same reason. the Z leadscrew can no longer drive the nozzle below the bed as the nut will simply separate. sorry this only applies for bedslingers--you dont get this benefit for corexy. all printers should have this, and i suspect it greatly reduces the cost for the custom builders by allowing use of lower grade leadscrews and ballscrews in their builds. one piece of advice i will give is to get spare balls or inform mellow to include spare balls in the purchase as you will lose some.
I have 2 Ender 3 machines, one I bought about a month after launch and the second about a year later, both machines have perfect Z stacking of the layers. I
Simplest solution: loosen Z nut bolts to allow some lateral movement + never Z-hop. Free side rollers could be looser to allow smooth movement. Or, belted Z with as loose Z rollers as possible, and with counterweight on heavier (X motor) side to keep gantry from falling :)
Have you tried an anti-backlash spring loaded brass nut instead of just a plain brass nut? For the RatRig test it would have been interesting to see the difference between the 3.0 version and the 3.1 version and then the 3.1 version with the 3.1 version with wobbleX. By introducing two changes at the same time on the RatRig you don't know which one actually improved the print quality. You are just assuming that it's the wobbleX.
Same with even the lubrication of the lead screws. That should have been done before the base line test, not after it. Because if the lead screw wasn't so clearly bent, or whatever was the issue, the act of simply properly lubbing up the lead screw again might have solved it (a grindy, but otherwise strait lead screw can introduce similar artifacts on the print). Buit I guess that's the nitpicking boring scientist in me speaking, and not the excellent 3D-printing spokes man, and wonderful enthousiastic youtuber like he is... lol. Still, he should keep that in mind when doing AB tests. hehehe
Imo it's bit wrong to try to fix lead screw wobble by additions. If one allows it to flex too much cause twisted lead, it'll then result in changing pitch of the lead screw. Hence people should always check if their screws are actually straight. It's not hard to straighten them by force or downright replace them. After leaving or even checking that drive nut is straight and possibly leave it floating can increase quality quite a bit. Going for best possible results looking into different solutions is ok, but many seem to forget the very basics.
The low weight problem can be eliminated in a way that opening the gap between steel rods little bit more, then balls can't escape. I m using in my printer a similar design and I have been using with no problem.
The Wobble will only help with Z-banding if the printers lead screw is overly constrained. The imperfections in a lead screw will only translate to a print if it can't be taken up by tolerances elsewhere in the printer. This can be absorbed by a flexible coupler between the stepper and the screw and by removing any other bearings besides that coupler. I.E. a bearing at the top of the printer. Alex Kenis goes over this in depth if you want more info.
Just get some oldham couplers and if you need downward force perhaps try installing some anti backlash nuts on the top side or instead of farting around with a cheap fix, just get ball screws
All the time you were assembling, I was think there is no pre-loading being done. Seems that is a problem for the Ender 3. Hope you manage to fix it as the principle is good.
i think the tension springs would be a perfect solution to the weight issue, without adding said weight to the system and thereby mass to cause unwanted vibrations.
It is a clever solution and doesn't cost much, but I do wonder if simply going to the next size up diameter lead screw would be an easier and more straight forward solution. You'd need a new coupler for the motor, a larger nut and a support on the end.
The wobbly z-axis seems to pull the bed out of proper alignment and that seems to be the the mode that this fixes (by decupling the XY component of the wobble from the bed. I believe a better fix for this would be a z-axis guide that is rigid enough to be unaffected by screw wobble.
You can't really heal an overconstraint with misalignment by extra rigidity, to an extent it can help but not without adverse effects. If you have a more rigid linear motion system, either the leadscrew will start breaking its attachment point or the motor, or it will start skewing and breaking the frame, and indeed i think most of us have already experienced a situation sometimes where the Z-wobble was actually the frame going askew periodically, particularly if you started with a fully DIY i3 or anything of a kind. Or it'll find a way to break the linear motion system, the bearings etc. It's a little optimistic to suspect that the leadscrew will straighten itself out, but it's a possibility as well.
@@SianaGearz CNC machines use leadscrews, have way tighter tolerance requirements, and they generally dont discuss how to fix wobbly screws… In some cnc lathes you can see the feed screw wobble. So something CNC machines do, that printers dont. Rigidity is the main differentiator.
@@pyalot If you pay for a precision leadscrew, you get a precision leadscrew, which is not what we're paying for. They've still got to have an engineering solution for aligning it correctly when the machine is built. And then it happens that our leadscrews get bent in transit, machines have special shipping requirements, not just a cardboard box with a bit of foam and paper.
@@SianaGearz Sure they use better lead screws, but in terms of wobble this mainly means two things: it wobbles less, but since the lead screw is stiffer, any wobble it does have will pull on the frame much harder. On the bottom line, the linear motion system of a cnc machine must still be vastly more rigid than any wobble component the screw generates. It is not unreasonable to expect the linear motion system rigidity to exceed the lead screw bendiness by so much as to render the undesirable motion components inconsequential. All printers with lead screws are built on that assumption. The lead screw pulling the frame out of alignment, there is clearly an unhealthy rigidity balance. Have you tried just using more bendy screws?
@@pyalot There's actually a designed in lateral tolerance in the nut and leadscrew. As long as your wobbliness is below that, you are good, you're not putting undue force on the motion constraint system. You're also not supposed to use leadscrews like we do. You install it in two pillow blocks and give it a flexible connection to whatever drives it, be it a geartrain or a stepper or whatnot. You're also supposed to be able to zero out the position misalignment of those pillow blocks somehow. You're also expected to have some sort of thrust bearing assembly to prevent the screw from wandering alongside its axis. That's a positively unpleasant amount of parts. And then you're not supposed to have several constraints in the same direction, or you need to be able to align them precisely. So you have overconstraint built into the system in a typical 3D printer, because you have those two separate Z beams each with its own bearings. More bendy screws is what i'm actually doing today, i have m6, and m5 might be an upgrade, i also have made myself Z axis nuts which have lateral travel on the attachment points. M5 was what a lot of early reprap designers have settled on for Z. But my machine is kind of deliberately shitty, i'm just cycling cheap and shitty solutions in and out for the fun of it mostly to just observe failure and laugh at myself. Of course like this the Z hops are a little slow. People have been doing belted Z for a while, it's a very good solution for most types of 3D printers.
Would be well worth adding that 500g weight to a standard Ender 3 to see what effect that has without the WobbleX. Just to give a more direct comparison 😊
Exact same thought I had with when he lubricated the lead screw. Might have solved the issue (tbh, I think not, the screw was too bend, but still, same principle applies). He should not have that done for a fair comparison. Or, at least lube it BEFORE adding the WobbleX system and compare it with that. He added more variables in the mix... never a good thing when comparing stuff. (though, don't get me wrong, I'm very glad he reviewed it either way!! This is just a 'small' thing to be aware of when doing 'scientific' comparisons).
I'd like to see what you do with the compression springs. I have an old Prusa i3 clone with threaded rods that had the Z banding. I printed some mounts that allowed it to freely move but rested on the gantry and required some weight to keep it stable, it was fine for years. I changed to a lighter extruder and upgraded it to lead screws and have had nothing but problems since (z banding, jamming on the z axis, inconsistent z height on levelling) and all but abandoned the printer for a year. While it may not be a golden bullet I think this solution might resolve most of my problems if it can do it with a lighter gantry.
WobbleX is good, for sure, but is symptom-solving. The real issue is that most original designs (that the Crealities of this world have copied off of) are kinematically over-defined or poorly designed, and the mistakes get inherited mindlessly. Any design that has two parallel linear-rails under a print-bed ought to raise heckles of a mechatronic designer.
@Teaching Tech Michael, did you find an alternative way to get around Dual Z? I have an Ender 3 v2. I would love to use these on it. Thank you for all the great content.
Backlash nut on the top to provide tension? @TeachingTech I found that getting even load on the L/R wobble x was the hardest thing. The most annoying thing was the added components caused the nozzle to hit the bed first home firing the ballbearings everywhere, thankfully I found them. I got marginal gains, but gain all the same. (sidewinder x1).
Getring the springs in precisely the right alighnemt to avoid it tilting the pieces would be very difficult. The 1/2k weight isn't such a bad idea, really. It's on Z, so it's not like that has to move much.
Maybe increasing the distance between the Rods slightly could help keeping the ball on the tracks cauz it would sit lower inside of them? This would remove a little bit of the rotational axis becaus the wobblex parts are a bit closer to each other, but id guess theres isnt much rotational movement on an ender style z axis anyways
Or you could use bigger balls in that case (thus pins further apart and bigger balls)? What I also don't like (well... '"don't like" are big words for this tiny issue though) is that the pins do not rotate around, but are press fixed in place. At first, I thought the pins where able to rotate and thus create a real bearing system with the balls, but they are not, they are fixed in place. So, over (very long) time, either the balls or the pins will wear out and dent and/or show deformation (especially when low grade steel is used.... thinking about the copies and clones which WILL come)
@@CookieTube i dont think the pins need to rotate. The point of the pins is that the balls can roll over them to create a type of linear rail. In the case of the linear motion the ball is already rotating in the needed direction so no rotation of the pins is needed
@@KidCe. The problem is not with the lateral/linear motion. That gives a true bearing/rolling action. The whole major point of why and how the WobbleX was made, compared to other solutions, is so it can also tilt. That is why it was designed. So yes, the pins should in an ideal world rotate so that they form a true bearing with the ball, and not 'grind' on eachother. But of course, making them rotate properly also would require a whole other level of engineering and complexity for such a tiny tiny rod (if at all possible). I suggest to take a closer look at the cad drawings or videos to see how it actually works and what the true motion of freedoms are, and why they are needed (compared to an Oldham coupling). WobbleX it is a very simple, yet ingenious, solution (although it clearly also have some drawbacks).
Incredible! A huge thanks to you, Michael, for the tremendous work you've put into testing this system on your printer and addressing some integration challenges for specific situations! Your ability to explain the system, its background, and the installation process is truly remarkable - straightforward, concise, and genuine. I'm truly pleased that someone with your expertise can recognize the potential of a solution like this! Cheers!
And all this is based on your awesome work :D
Also what do you think about this potential solution for the slipping problem with to little weight:
Maybe increasing the distance between the Rods slightly could help keeping the ball on the tracks cauz it would sit lower inside of them?
This would remove a little bit of the rotational axis becaus the wobblex parts are a bit closer to each other, but id guess theres isnt much rotational movement on an ender style z axis anyways
A simple solution to fix the weight issue on the ender would be to mount the wobbleX & lead screw nut above the plate instead of below.
@@MyYewTubeAccount How would that add the needed weight/compreesion to the wobbleX? Just its own weight/gravity?
Thank you @MirageC for providing the parts for this review. I saw your original announcement of these and was considering purchasing a set for my Ender3. Not having any other printers I would not have spotted the issue @TeachingTech did. I have no problem with the weight solution but I look forward to seeing what fixes are found for this issue.
@@D4RKFiB3R I am unsure, my experience may not relate directly to the wobbleX, but I have for many years installed anti backlash springs/nuts in this fashion and it's always achieved better results.
Changed to a belt driven Z on both of my old Ender 3 Pro's using KevinAkaSam's belt mod. Brilliant system, with zero Z issues afterwards.
Yeah this level of work is getting to the point that you can go belted Z. No grease, no wearing out, no backlash or weight issues. Only downside is if power is lost.
Did the same mod to my Ender 3 and hands down best mod I've done to it. Perfect Z layers now. On my Ender 5 though I installed another anti z wobble device, but didn't quite fix all the issues. Still prints great, but under the right light you can see the inconsistences.
@@jaro6985 Just use a bigger stepper motor
would love a comparison between this and a more conventional, brass, Oldham coupler. It might solve the issues where you need extra downforce on the E3 as well
Particularly as the Oldham is much cheaper.
@@petercallison5765 and you can 3d print one that works fairly well
Oldham coupling does have the very same problem, although MUCH MUCH less!!. The clearance between the parts is still significant enough to cause the same thing! Although, since it would be MUCH less noticeable by eye'ing the print in progress, it would be difficult to recognize that the Z-banding comes from the this very problem.
Another downside of the Oldham coupling compared to WobbleX is that Oldham can only mitigate movement on the X-plane, aka lateral. It can't compensate for rotational movement, which is exactly why the WobbleX was developed.
(But yeah, it is MUCH cheaper indeed. The WobbleX seems to be very overpriced to be honest).
But maybe this 'jump' problem, with both style of couplings, might be solved when you also add an anti-backlash nut? Dunno, can't test it out (my WobbleX are still in the mail and lately I don't have time anyways to experiment).
@@CookieTube the main issue with Oldham is that while it does decouple movement in the plane, it doesn't decouple the tilt of the axis. Whether that is good enough will need testing but I imagine that any rod non straightness will introduce some tilt.
@@JohnMeacham yep, indeed, hence why WobbleX was developed. You ALWAYS MUST compensate for tilt in the cases we use lead screws or balls screws for 3D printing. Hence the WobbleX. An Oldham coupling is simply not made for the purpose it is used in the 3D printing world!!
Oldham couplings are flexible shaft couplings that are used to connect *TWO different shafts* : the *strait driving shaft* and *strait driven shaft* , which almost meet in the middle piece of the coupling. It is used to counter *parallel* misalignment that occurs between connected shafts. Important here is "parallel"... Aka, the two shaft are perfect parallel to each other (not crooked in any way), but their center to each other might be slightly misaligned. Many people don't get that (which you can also read in several comments about it).
Spring preload is definitely the way to go. My first thought due to the size restriction might be to use clothes-peg springs (aka a torsion spring). They are small enough with sufficient travel, a simple horizontal pressure point and have a handy hole in the spring section for additional mounting options.
If you use a smaller ball or spread the rods apart a bit, it should increase how much torq it can translate in movement.
These things are awesome. I used a set on my highly modified anycubic mega zero. Also saw the bounce issue and a couple of rubber bands 90° to each other, and that was enough to see huge improvements.
He could put O-rings vertically to tension the assembly. Like bands on teeth braces
After seeing the initial video on the wobbleX I just looked at Oldham coupler models on Printables and remixed one for my ender. It helps my print surface be as clean as possible even if I don't have visible screw wobble.
Genius, what an elegant design. And great catch for the missing link to the Ender's 3 issue. Fantastic video, thank you.
I have a similar design that is a complete 3D printed solution for the frames, and found that a sizeable rubber band wrapped around the entire assembly works well on my CR-10
You still need to try Oldham coupling. I put a set on my Ender3 and have super clean prints.
that's cool, waiting on mine, any advices on install?
@@deathcube2006 check fit. sometimes they have small burrs (nicks) that keep em from moving smoothly. If that's the case wet sand 2000 grit on mirror/flat surface, clean then light machine oil.
Oldham coupling have the EXACT same problem! And it does not matter if you have the strait version or the slanted version; they both exhibit the same problems.
But not as massive big as what you saw happening in the video with the WobbleX though.
There can still be 'too much' tolerance in the Oldham couplings between the three plates. You can even feel it when you gently push or pull them apart.
But is is not as much as with the WobbleX, which you can easily see happening during the print by the naked eye.
As such, you should be aware that the very same thing happens with Oldham couplings. Using that style of coupling will not prevent it 100% because those tolerances need to be there for the main function of the coupling to work: sliding.
@@deathcube2006 Make sure you have the right version. Some models of the Ender3 must have the 2 hole, while others can use the 4. Get the proper screws and nylon lock nuts. Lube the coupler with a graphite or PTFE type lubricant. When you get the couplers, take them apart and inspect them very well for burs, machine defects, and debris. Any defects, or play in the dovetails will greatly hinder their usefulness. Assuming you got the brass version, if there is play, you can take some of the play out of it by tightening the coupler end piece in a vice with the jaws of the vice parallel to the travel of the dovetail. I'd also add a bearing to the top of the Z axis leadscrew to keep things vertical and allow the Oldham to just take up the side to side movement.
@@CookieTube The dovetails on an oldham can be much more accurate and reliable than this wobbler plate. And, if the dovetail has too much play in it, it can be tightened. Mine have 0.001" of side to side play in them. There is a reason some high end CNC equipment and other industrial equipment use Oldham couplers.
You should use the "vase mode" to test the wobbling of the Z-axis, which is more accurate.
I think your assessment about needing to add tension to hold it all together is spot-on, when originally looking at this system that was one of my immediate concerns, that the gantry can lift up freely and thus looses an axis of constraint. Gravity is only good enough if your z axis moves freely enough. Also in general you may want to methodically work through your lifting z axis and see if there is excessive binding that's managing to hold the weight up, or a natural skew caused by where the X bar is tightened to the side plates. You need to lock the steppers to a known initial state then level the X bar from there usually, else it tends to jump to in-between positions etc. I'm personally a fan of the single-motor belted-z setups, they are rock solid once the constraints and parallelisms are worked through, but it requires some definite expertise to dial in.
@Roobotics : I totally agree with you. In addition, while adding extra weight to the upper component will help, it is adding weight to the bed assembly, with the potential extra wear on the Z axis lead screw/nut. A spring tensioning system would be preferable, I think.
Off topic here, but since you mentioned it: Have you ever found a place/video where they compare the two systems side by side (dual coupled Z-axis with 1 motor vs. dual Z-axis with 2 motors)? All I can ever find are 'superficial' videos where they talked about how to install them and how to set them up.
I want to find an in depth tech video comparing the 2 systems *properly* with all the ins and outs, and not just the _"with 2 motors you can have automated gantry tramming"_ ...
What about drift overtime in both systems, what about slack in the belt overtime, what about Z-banding in both cases, what about 2 motors working against eachother, etc etc And also: where to do mount the steppers best? On top (hanging Z-axis) or bottom (sitting Z-axis)... What style of coupling to use between lead screw and steppers (many styles exist), and their benefits or drawbacks... Do you 'fix' the screw on the bottom/top, or do you let if free standing/hanging, and if you fix it, do you do it in one plane or both planes or none at all but instead restrict the lateral movement it a bit... etc etc etc...
Anyways, sorry for the wall of text....
Hi great video! I was already eying to buy the wobbleX for my soon to be rat rig, and now i am certain i will. However I do think that creality machines would benefit from a spring loaded double nut specially the ender3 whit the worn screw, i am running a set on my cr10v2 and they did a fantastic job eliminating backlash and elongating useful life of the screws.
I think you are probably spot on with the extension springs 3 or four would be best if possible they would also prevent some bed slinger wobble from the bed motion causing tilting front to back
agree - I don't think adding weight is the best idea.
@@reasonablebeing5392 Extra weight would also wear the lead screw quicker and likely cause extra motor wear.
or 1 wide tension spring in the middle, in which the lead screw can pass thru, but with flanges on top and bottom (thus needs to be custom made?) to hold the whole system together as one packet....
..... or an anti-backlash nut....
Glad you made a review video, I ordered a few of these for 8mm lead screws and I can start thinking of a sprint clamp mod for my Cartesian printers
I has been watching you and all the others for years, I went away for some years from 3d printing but now as I came back ofc we all aged but you look almost the same as I remember you years ago (I cannot say the same for the others) goo to see you still around) 👍
Looking forward to the next step of this project. Have been looking at this part for my CR printer so was very glad to see the video
I have a modified Ender 3. I added a worm gearbox to my Z axis and I expanded the Z and Y axis to 1M (1000mm). The z wobble has been quite rough even before the expansion. So far, tests with the WobbleX have been impressive. I didn't know about the hole sizing on the screw and printed the wrong part but drilled it to fit and it's working fine. I would have printed it again but I have to take apart my worm gearbox to install the wobbleX and it was already apart when I discovered the size difference. The WobbleX for the ender3 doesn't have as much play as the others, and at about 500mm it does reach it's limit, but when I'm printing parts that large a little wobble is expected. One thing to watch out for is lifting the gantry. Those little magnets don't hold it together very strongly and you can easily lift the gantry and loose a bearing if you aren't careful.
Sounds like a nice idea, but with its challenges,
and thank you for sharing.
I thought to share my experiance in this matter.
After a year of constant use of my ender3 I had started to get a slight Z wobble,
so I knew it was time to replace the obvious parts.
Well I thought at the same time I would upgrade the height and add an extra 150mm
and so I bought a upgrade kit and it came with a perfectly straight ACME lead screw
and I also bought the Z-axis T8 Anti-Backlash Spring Loaded Nut from Amazon
After some tweaking in Jyres firmware it was all set for a full height test and the out come was perfect.
So personally if you have a bent ACME Lead screw or a worn thread and a worn brass Nut
I would just buy another set and also fit it with a Z-axis T8 Anti-Backlash Spring Loaded Nut.
The Z-axis T8 Anti-Backlash Spring Loaded Nuts is good to help on the Z lifting as the bed or head moves up and when it is compensating the bed level or if you have the Z hopping setting on
LOL, This is the EXACT same I have: 150mm extension...... + dual, coupled on top, aftermarket dead straight lead screws + anti-backlash nuts on both.
I've done the extension not so much to be able to print taller stuff, but far more because of convenience in having clearance on top for all the extra stuff without actually loosing the original Z-height (LED light, stiffer cables/chain so they don't flop on the bed, having clearance without the need to almost 'fold' the cables when on max height, the dual coupled Z-axis system, etc, etc).
And also extra corner brace plates on the bottom + CR10-style diagonal braces.
The whole thing is solid as a rock and even (very high) resonance vibrations are almost non existent.
Thanks for this review. I’ve built a custom 500x500x600 corexy triple Z axis printer and the 600mm lead screw’s definitely have some bend in them. I was going to go and buy some more expensive ball screws to remedy the problem, but for the price, I’ll give these a try instead. I should be able to adapt the ratrig bed mounts to my printer. ( machinist here, so I have access to machines to make changes on the fly as needed). Keep up the VERY informative reviews. It’s much appreciated!
Just ordered a v-core 3 and I was wondering whether I would benefit from these. Your video is spot on!
My V-Core 3 was originaly 3.0 and one of the first kits ever. The Z-screws were not the greatest. The taller the print the more visible a slight 4mm-period Z-wobble became.I upgraded the Z-arms to a printed upgrade called wobble rings (sorry, forgot the name of the author) which does pretty much exactly hwat the WobbleX does. It completely fixed the issue. So much that I did not upgrade my printer to full 3.1, I just changed the already failing XY idlers to bearing stacks (slightly different size than 3.1 stock to accomodate 3.0 geometry). I did nothing new to Z axis because the wobble rings made it perfect already.
Now I put together V-Core 3.1 for someone and the lead screws are way better. They were shipped inside the grooves of some of the 3030 extrusions so zero risk of transport damage and they looked perfectly straight. However I found that for a 500 size of the V-Core the bed itself is already on the border of what 500mm length of 8mm leadscrew can support without bowing sideways a bit and when you put a lot of plastic on top it only gets worse. Any tiny non-straightness of the leadscrews not even visible on them when they are on a table gets exagerated once they are loaded in compression by a heavy bed. The oldham couplers included in V-Core 3.1 that let the nut move in XY plane are good enough to fix this at least for the good lead screws I got on this machine. You actually see them moving as the bed travels if you look closely and yet a 500mm tall printed cylinder has absolutely perfectly smooth walls all the way.
I think that for most printers the WobbleX with its 4-axis freedom of movement (translate in XY and tilt around X and Y) is overkill and just a sliding oldham coupler like the V-Core 3.1 would work too. But there might be some cases where the WobbleX fixes an imperfection other couplers would not.
I was in a similar situation as you with my v-core 3, but I waited and did the 3.1 upgrade and that also completely fixed my z issues. The 3.1 upgrade also incorporates a similar decoupling system that's more rigid than wobble-x, which I think is actually a better system.
@@AgentHubcap the 3.1 decouplers are for sure easier to use. You can lift the bed and put it back down without worrying about it and it just works. With the wobblex you need to make sure everything sits back into its place. Unlike the sliding brass Oldham decouplers the balls can move out of their grooves.
However I get what the designers of WobbleX say about the 4 axis over 2 axis freedom. If the nut is badly made for example and it's top surface is not staying horizontal as the screw rotates (misaligned thread cutting tool makes the helix not coaxial with the outside of the nut) then the RatRig decouplers will tend to bind and not slide freely while the Wobblex will not even notice and just works.
So for good components the RatRig decouplers are better because they are more robust and easier to use and don not require the enough weight on them like was shown in this video. But with worse components there can be situations the Wobblex will be better.
Great video. Excellent lighting on the prints that actually show the surface variation. Fortunately this was back in stock when I looked, so I bought one for my Monoprice Maker Ultimate III (Wanhao D6) printer. I am hopeful this will make my printer usable for prints that need smooth sides.
LOL I just ordered these a day ago. After MirageC's original video they sold out in no time. I saw they were back in stock and picked up 3, now Teaching Tech video is out they will probably sell out again.
I am quite excited to try this out, it looks really cool and it should work fine on my Voron Trident with the heavy bed loading them up.
The one you printed for the Ender 3 looks like you could have a gantry wheels with flat spots on it. I've noticed this allot with my own Ender 3 and i have to replace them on a regular basis. either from it sitting to long from non use or normal wear and tear. so ive made a pratice of replacing my wheels once a year, and that equal length banding has gone from my prints! Just my thoughts and suggestions on replacing your wheels for the horizontal gantry.
This. I had the exact same problems with my ender 3 clone - coming with wobbly wheels fresh from the dropshipper - I replaced all of them and the wobble was gone.
At that point you might as well get rid of the wheels altogether and go with rods or linear bearings.
But he fixed the problem when he added weight and the print came out great. I mean he pretty clearly explained what the issue was and how he fixed it.
@@Hagop64 This was the (fairly similiar, but modded) dual-Z CR-10. The Ender only has one Z-motor.
@@amorton94 Those wheels are a dime a dozen now, that linear rails are the new fancy - and you gotta have a whole lotta more work to do, to retrofit linear rails onto an ender - ask me, how I know ;)
The observation at the end of the the inclusion of tension is a game changer I think! More investigation required 👍
Yay! Thank you for trying out this potential solution with several different printers. Yes, compression is key to this since some machines enable Z-hop for travel across a print.
What about adding an Anti Backlash Nut with a extension spring between the top part and the te bottom part of the wobbleX? Then the assembly Is in compression.
It will almost look like the anti backlash assembly on an Ender 3 but the spring is in extension instead of compression.
Fantastic testing and tweaking, Michael! Thanks for sharing! 😃
Stay safe there with your family! 🖖😊
In many cases I like rubber bands better than springs as they are easier to adjust for length and tension
I printed similar setup without pins and magnets but with rubber bands. Magnets won't hold any position if printer tries to dig into bed, springs and bands give up in Z but keep nut from rotating
I can see this being an issue if you use Z hopping. Adding springs might help but the quick up and down movement of the Z might still cause problems. I think that if the screw is wobbling that bad then replacing it is a better option.
I've also used the same box-shaped POM clearance nuts that the Rat Rig V-core 3 originally used on my self-built printer with good results.
In fact, I feel that on the combination of screw and nut, the appearance of layer lines comes more from the fit gap between the screw and the nut.
It is well known that almost all screw nut combinations have very large fit clearances. Most of the screw nuts use trapezoidal threads. When the screw is shaken, the trapezoidal threads will cause height errors due to sliding along the radial slope.
Therefore, in my humble opinion, the solution to the Z-axis error caused by the screw nut is mainly to solve the fit clearance of the screw nut. It is necessary to make them as tight as possible and avoid radial relative sliding.
And the "Rat Rig V-core 3 original box-shaped POM clearance nuts" that we both use do a pretty good job of that. After the gap adjustment screw on the top is screwed to the proper position, the relative displacement of the screw rod and the nut will hardly occur in the radial direction. I think this is why your Rat Rig V-core 3 has a very good Z-axis performance at the beginning. .
Its only disadvantage is that it is a bit bulky, and I don’t think the wear resistance is a problem. At least my machine has been printing for hundreds of hours, and there is no problem. I just readjusted the clearance screw halfway.
On the ball screw, the gap between the screw and the nut is very small, and most of the ball screws are relatively thick, and the straightness is relatively better. That's why it works so well in its raw state most of the time.
If the ball screw bends, the bend will be amplified and faithfully mapped to Z-axis motion because there is no clearance issue with the screw nut. Looking at a section of the screw rod, due to the bending of the screw rod, the movement along the vertical direction has an angle offset, and the local distance relationship forms a right triangle, and unfortunately, the screw rod is at the oblique angle of the right triangle. On the side, the actual Z-axis travel distance is on a right-angled side of a right-angled triangle. According to the Pythagorean theorem, there is an error between them. As long as the ball screw bends, the actual movement distance of the Z axis will always be less than the distance output by the printer, which will cause the distance between layers to become smaller, which will lead to excessive extrusion and layering.
Therefore, in my humble opinion, using the delamination device on the ball screw is mainly to increase the smoothness of the movement of the screw nut. Because when the two ends of the ball screw are fixed, if there is no delamination device to release the radial swing displacement, the bending of the ball screw will cause the pressure of the screw and the nut to increase, and the jamming phenomenon will occur. The coordination error of the ball screw and the nut will affect the axial motion accuracy.
To a certain extent, this can explain why the solution to the layer pattern problem has almost become "metaphysics". The masters from all walks of life have exhausted all kinds of solutions, but they have not been able to solve the lamination problem caused by mechanical reasons in a 100% repeatable and stable manner, because it is a problem that cannot be completely solved, and the mechanical error is there. All we have to do is to make the screw as straight as possible and make the fit gap as small as possible.
@@jackgai1929 THIS!
This is what I love so much!! Please make a youtube channel with stuff like that!!!!! Anyone??!!
Not the _"it looks ok, so this is the best solution"_ , but the actual science, both theoretical and practical, and the very in depth look at things from a deep mechanical, and yes, even theoretical viewpoint (as in: you'd never ever would see a difference otherwise, but it is the thought that counts, and "many invisible errors make a large visible one")....
Thank you for the very interesting reads!
Tension springs were my immediate thought, glad you mentioned that.
Been following Hevort and love what Mirage and team are doing.
I was also contemplating doing this on my vcore3since mirage posted it and glad you covered that as well.
I like the result and the idea of using springs for tensioning. I may just have to give this a try - good work.
Look into using keyboard switch springs. They're cheap and available everywhere. One on each corner I think would work wonders.
I dont think they will be nearly stiff enough if he needed 500g to remove the slipping. Keyboard spring are normally about 50g
@@KidCe. There would be 8 of them, so 50x8 is pretty close to 500g. If needed, there are 72+ gram springs. That's plenty of force.
@@shibbypoopalot Why not just 1 tension spring in the middle, fixed on the plate above and fixed on the plate below, in which the screw can pass thru?
Or 4 small rubber bands around the corners?
Or just a anti-backlash nut?
Great Review!!!!
I have an idea for holding them together, use Lego's silicone elastic bands to clamp the parts together. They last for a very long time and will be easy to design for.
Great video, and an impressive design of the MirageC team
Thanks for sharing your experiences with all of us 🙂
I would love for 3D printer manufacturers to start looking into this and implementing it as a feature instead of a DIY. I honestly would like for the option to choose what features I would like from a 3D printer at checkout, like the option for this fix over a tradition ball bearing, or a glass or rubber build plate. I think offering options like this on manufacturers websites when you purchase a product could be beneficial to people who dislike the whole DIY side of things. If they did this though, they would need to also offer a base line for people who don't want excess options when purchasing a product.
In the cnc industry we keep ball screws under stretch to eliminate this issue, among additional strategies obviously. I'd love to see that implemented somehow. In my experience, stretching a ballscrew can improve accuracy 90x. And it doesn't require that much stretch for a ball screw that heats up very little.
Of course a glass scale or a linear transducer really would be infinitely better for this that relying on ball screw pitch error to be minimal.
a leaf spring would be a good idea here, take a look at whoppingpochard from the voron team's kinematic bed set, he uses leaf springs to keep the bed attached even with a kinematic mount
Oh great Michael... now I need to order these for my RatRig VCore 3.1/3.2 rebuild LOL... these look great. Thanks for the video and excellent demonstrations!
I use Oldham coupling on my original CR-10 it is working kinda the same principle as the Wobble X. Also worth watching HeVORT another video about the extruder gear causing inconsistent layers :D
Oldham coupling has more friction and only has translational degree of freedom, but usually rotational degree of freedom is also required, in particular with ender and cr machines, as z nut is mounted onto bended part, which usually is not perfect 90 degree bent, and may even deform under stress caused by feeder motor in bowden configuration.
What @Dima said + Oldham coupling does have the EXACT SAME PROBLEM, although much much less. You probably wont even notice it when you just eyeball the print in action, but it is there for sure. The tolerances on those Oldham couplings are not that great. You can even feel it coming apart when you push/pull the coupling. So, yeah, very same problem there.
You might be able to solve it by also using a decent anti-backlash nut though.
@@CookieTube in Oldham coupling it is fixable, just make sliding parts not rectangular, but dovetail instead. i printed such a coupling, but it didn't turn out to be as good as just precisesly set z screw and nut.
@@AcTpaxaHeu No, a dovetail channels have just as much play in them up/down as with the rectangular channel. The shape doesn't matter. And If you think about it/visualize it, is logical that the shape of the sliding channel doesn't matter. I have several from several brands with several shapes, and all exhibit the same issue. It is just a matter of tolerances, not of the shape.
The guide rollers on your Cr10max might be what's holding one end of the z gantry up. You could try reseting them by loosening them, letting the gantry sit down hard on the lead screws and then tightening up. Also adding little springs to hold it all together will help too.
What do you mean by letting it sit down hard? I went from being frustrated with my Ender 3 to a v400 delta over Christmas, but I'd love to solve issues like this on the ender.
@@TheRealComradeOtis with Creality (and likely many other Cartesians with V-rollers) the X gantry extrusion is not locked into the bearing carriers so its possible for it to get askew from one side to the other, when this happens one of the two outer bearings on each side will have uneven force causing binding even if the tensioners on the inside are good, Good reason frame alignment trumps all in the factor of elimination print flaws, Linear rails have the same issue but to a much worse degree, they need to be perfectly aligned with zero misalignment or the bearing carriers will completely bind.
I would be interested to see the results of just adding weight to the z axis on the ender 3 for comparison sake and then maybe with other machines too.
Woah, this is sweet…this might have solved the issue I had with my OG Ender 3 Pro (Upgraded with a MicroSwiss Direct Drive extruder).
Took it apart and re-assembled multiple times and couldn’t get rid of those inconsistencies.
Good review. A lot to change/modify. We will have to assess whether the z-wobble banding is sufficient to warrant the number of mods. Great idea, though.
I'm not questioning your integrity as you are one of the few that do try and give an honest, unbiased opinion. However, I do believe in 100% unbiased, uninfluenced ( appearances - no room for doubt ), by citing the great Leo LaPorte "The Tech Guy" mantra which is anything he reviews he purchases with his own money so as to have skin in the game just as a normal consumer does. He has been providing tech tips, recommendations, reviews of anything electronics (computers, operating systems, phones, printers, scanners, audio equipment) since the early-1980s.
Another option instead of springs could be small o-rings, mounted/hooked in much the same way. Seems like they might couple the two halves rotationally with a lower amount of tension than required by springs.
The wobble alters the Z height it doesn't matter how much X or Y is compensated. It's at the leadscrew and brass nut contact location. A better solution, besides investing in a new straight leadscrew, would be to stabilize the leadscrew by passing it through a second brass or delrin nut before it contacts the first brass nut. Anti backlash nuts are also worth considering. If you have any better solutions or why you think my reasoning is wrong please share your findings! Have a good day!
Your over a beer reasoning is wrong. The Z height change due to wobbling is trivial compared to the swaying in the xy direction due to a bowed screw. The wobblex is a very good solution to allow a cheap screw to be used.
I changed my Ender 3 to a Dual Belted Z motion system and my banding improved massively along with hopping and fast Z motions, though this seems like a decent interim solution
I honestly think belted Z is in general the way to go. It's the right amount and type of constraint and isn't particularly expensive.
This is what I did on my CR-10s and it improved it massively.
Definitely seems like the most effective method you've tried so far
Belt driven Z should not suffer from these lead screw artifacts. Right?
From what I've seen SecKit is one of the few printers utilizing belt-driven-Z along with linear rails and 3-point bed kinematics.
I don't think there's anyway to improve on that
Michael, there is a variation on this, on the Tech 3D YT channel from a year ago that is using a compression spring pushing up in his design.
you could hols the part that screws into the x gantry and the part that holds the lead screw nut with strong rubber bands (ender 3 and cr10)
I am grateful that my nearly-2-year-old printer still is not exhibiting Z-wobble; dual-Z steppers with a timing belt keeping them in sync + braces at the top of the threaded rods to ensure they remain stable has, apparently, been the best feature of this machine that I never even realized.
Machine model?
@@stage6fan475 CR-6 SE, no mods
Thanks.
Regarding your Ender 3 problem. Have you looked at the extruder wheel wich is pushing the filament through the bowden tube? Sometimes it is not really centered causing filament beeing pushed irregularly, showing inconsistend extrusion.
Nice, even though I have a CR-10S with dual lead screws, I decided to buy a set because I want to contribute to the project by creating a design revision that works well for that setup
Michael, can you try oldham couplings next? Mellow has them in their store
What extruder do you have on your Ender 3? MirageC actually has a really great video that looks at inconsistent extrusion with geared dual drive extruders like the Bondtech BMG. I am wondering if your Ender 3 prints would clean up with a non-geared extruder more similar to the original extruder the Ender 3 comes with. I have a BMG clone on mine and get some Z artifacts, and I am wondering if it is worth trying a non-geared extruder. I only switched because my original plastic one broke. That in combinations with another commentor's suggestion to change the wheels would be really interesting to see the results.
THIS PRICE IS CRAZY
Spring load the coupler together.
Try a binder clip to add clamping force to it.
Oldham couplers are simpler, cheaper and don't rely on weight to keep things constrained. Great video, as always, but having to artificially weigh down a gantry seems counter productive
Mine will arrive in just a couple of hours! I have an ender 5 plus, so I think it will be sufficient weight to keep everything together. I also have some spare backlash nuts for some extra pressure, but I don't think those are needed.
How are they working out on the Ender-5 Plus Rachelle?
9:51
This problem actually exist with the older MirageC wobbleX design (forgot the name of it). It rotates but the Z doesn't move unless its heavy top. Happen on my Ender 3.
I used a couple of Oldham couplers printed in abs and sanded to be as smooth as possible, with some grease in between and it's years I don't have banding... This is clever but even overcomplicated
Nice comprehensive review with interesting results. Thanks a lot for your time :)
I know you tried adding a thrust bearing in your previous video, but in this video it looks like your "rigid coupler" is hanging in the air. The problem with that is that the NEMA17 motor has a spring inside of the motor, and it is not built to hold steady vertical loads. If you want your rod to not move up-and-down because of the spring inside the motor, you must make it so that the rigid coupler is sitting pressed against the motor body with a thrust bearing. That way whenever the rod wants to push on the spring inside the motor, it doesn't have anywhere to go as the motor body and the rod are not allowed to get displaced on the vertical axis.
I had some lines appearing at various heights on my cheapo biqu-b1, and after making the rigid couplers rest on thrust bearings that are pressed against the motor body those disappeared. I didn't even bother printing a part for them like you did in the previous video about this subject.
i have been using a fully printed version on my ender for about a year and love it. you know what are the hidden benefits of this that i love so much? you can lift the entire gantry up by hand to do maintenance any time you please since it is is no longer bound by the leadscrew. secondly, it is now harder to destroy the bed with unexpected or illegal negative z commands for the same reason. the Z leadscrew can no longer drive the nozzle below the bed as the nut will simply separate. sorry this only applies for bedslingers--you dont get this benefit for corexy.
all printers should have this, and i suspect it greatly reduces the cost for the custom builders by allowing use of lower grade leadscrews and ballscrews in their builds. one piece of advice i will give is to get spare balls or inform mellow to include spare balls in the purchase as you will lose some.
How weird, I was literally looking into this yesterday :D
I have 2 Ender 3 machines, one I bought about a month after launch and the second about a year later, both machines have perfect Z stacking of the layers.
I
Simplest solution: loosen Z nut bolts to allow some lateral movement + never Z-hop. Free side rollers could be looser to allow smooth movement.
Or, belted Z with as loose Z rollers as possible, and with counterweight on heavier (X motor) side to keep gantry from falling :)
Have you tried an anti-backlash spring loaded brass nut instead of just a plain brass nut?
For the RatRig test it would have been interesting to see the difference between the 3.0 version and the 3.1 version and then the 3.1 version with the 3.1 version with wobbleX. By introducing two changes at the same time on the RatRig you don't know which one actually improved the print quality. You are just assuming that it's the wobbleX.
Same with even the lubrication of the lead screws. That should have been done before the base line test, not after it. Because if the lead screw wasn't so clearly bent, or whatever was the issue, the act of simply properly lubbing up the lead screw again might have solved it (a grindy, but otherwise strait lead screw can introduce similar artifacts on the print).
Buit I guess that's the nitpicking boring scientist in me speaking, and not the excellent 3D-printing spokes man, and wonderful enthousiastic youtuber like he is... lol. Still, he should keep that in mind when doing AB tests. hehehe
Imo it's bit wrong to try to fix lead screw wobble by additions. If one allows it to flex too much cause twisted lead, it'll then result in changing pitch of the lead screw. Hence people should always check if their screws are actually straight. It's not hard to straighten them by force or downright replace them. After leaving or even checking that drive nut is straight and possibly leave it floating can increase quality quite a bit. Going for best possible results looking into different solutions is ok, but many seem to forget the very basics.
I thinkWobbleX would work if the Z axis had linear rails, i think linear rails would increase the weight on the WobbleX
The only thing is V-Core3 (v3.1) already has Oldham coupling on Z leadscrews so these ones are probably not really necessary there for a while.
The low weight problem can be eliminated in a way that opening the gap between steel rods little bit more, then balls can't escape. I m using in my printer a similar design and I have been using with no problem.
Tension springs make much more sense than a weight in that lateral forces perpendicular to Zed should affect the Wobble X less(?).
Just capture it with springs is a great idea Michael!
The Wobble will only help with Z-banding if the printers lead screw is overly constrained.
The imperfections in a lead screw will only translate to a print if it can't be taken up by tolerances elsewhere in the printer. This can be absorbed by a flexible coupler between the stepper and the screw and by removing any other bearings besides that coupler. I.E. a bearing at the top of the printer.
Alex Kenis goes over this in depth if you want more info.
Maybe try using a Keyback in the opposite way the Voron Switchwire uses one? To pull down on the gantry instead of pulling up.
Just get some oldham couplers and if you need downward force perhaps try installing some anti backlash nuts on the top side or instead of farting around with a cheap fix, just get ball screws
oldham couplers dont counteract tilt unlike wobblex
All the time you were assembling, I was think there is no pre-loading being done. Seems that is a problem for the Ender 3. Hope you manage to fix it as the principle is good.
i think the tension springs would be a perfect solution to the weight issue, without adding said weight to the system and thereby mass to cause unwanted vibrations.
It is a clever solution and doesn't cost much, but I do wonder if simply going to the next size up diameter lead screw would be an easier and more straight forward solution. You'd need a new coupler for the motor, a larger nut and a support on the end.
you could also use one of those anti backlash z nuts with a spring pehaps, maybe not enough force or hard to fit in tho
Try a TPU enclosure of somesort for the wobblex to retain the positions
10:05 add a rubber band ore spring steel clip that hold the top down, that can move from side to side but still clamp everything together...
The wobbly z-axis seems to pull the bed out of proper alignment and that seems to be the the mode that this fixes (by decupling the XY component of the wobble from the bed.
I believe a better fix for this would be a z-axis guide that is rigid enough to be unaffected by screw wobble.
You can't really heal an overconstraint with misalignment by extra rigidity, to an extent it can help but not without adverse effects. If you have a more rigid linear motion system, either the leadscrew will start breaking its attachment point or the motor, or it will start skewing and breaking the frame, and indeed i think most of us have already experienced a situation sometimes where the Z-wobble was actually the frame going askew periodically, particularly if you started with a fully DIY i3 or anything of a kind. Or it'll find a way to break the linear motion system, the bearings etc. It's a little optimistic to suspect that the leadscrew will straighten itself out, but it's a possibility as well.
@@SianaGearz CNC machines use leadscrews, have way tighter tolerance requirements, and they generally dont discuss how to fix wobbly screws… In some cnc lathes you can see the feed screw wobble. So something CNC machines do, that printers dont. Rigidity is the main differentiator.
@@pyalot If you pay for a precision leadscrew, you get a precision leadscrew, which is not what we're paying for. They've still got to have an engineering solution for aligning it correctly when the machine is built. And then it happens that our leadscrews get bent in transit, machines have special shipping requirements, not just a cardboard box with a bit of foam and paper.
@@SianaGearz Sure they use better lead screws, but in terms of wobble this mainly means two things: it wobbles less, but since the lead screw is stiffer, any wobble it does have will pull on the frame much harder. On the bottom line, the linear motion system of a cnc machine must still be vastly more rigid than any wobble component the screw generates.
It is not unreasonable to expect the linear motion system rigidity to exceed the lead screw bendiness by so much as to render the undesirable motion components inconsequential. All printers with lead screws are built on that assumption. The lead screw pulling the frame out of alignment, there is clearly an unhealthy rigidity balance. Have you tried just using more bendy screws?
@@pyalot There's actually a designed in lateral tolerance in the nut and leadscrew. As long as your wobbliness is below that, you are good, you're not putting undue force on the motion constraint system.
You're also not supposed to use leadscrews like we do. You install it in two pillow blocks and give it a flexible connection to whatever drives it, be it a geartrain or a stepper or whatnot. You're also supposed to be able to zero out the position misalignment of those pillow blocks somehow. You're also expected to have some sort of thrust bearing assembly to prevent the screw from wandering alongside its axis. That's a positively unpleasant amount of parts.
And then you're not supposed to have several constraints in the same direction, or you need to be able to align them precisely. So you have overconstraint built into the system in a typical 3D printer, because you have those two separate Z beams each with its own bearings.
More bendy screws is what i'm actually doing today, i have m6, and m5 might be an upgrade, i also have made myself Z axis nuts which have lateral travel on the attachment points. M5 was what a lot of early reprap designers have settled on for Z. But my machine is kind of deliberately shitty, i'm just cycling cheap and shitty solutions in and out for the fun of it mostly to just observe failure and laugh at myself. Of course like this the Z hops are a little slow.
People have been doing belted Z for a while, it's a very good solution for most types of 3D printers.
I should try this some time since I've been basically doing exactly this (but without the bearings) for quite some time now,
Would be well worth adding that 500g weight to a standard Ender 3 to see what effect that has without the WobbleX.
Just to give a more direct comparison 😊
I thought the same thing. Maybe the problem would be resolved with the weight alone?
Exact same thought I had with when he lubricated the lead screw. Might have solved the issue (tbh, I think not, the screw was too bend, but still, same principle applies).
He should not have that done for a fair comparison. Or, at least lube it BEFORE adding the WobbleX system and compare it with that.
He added more variables in the mix... never a good thing when comparing stuff.
(though, don't get me wrong, I'm very glad he reviewed it either way!! This is just a 'small' thing to be aware of when doing 'scientific' comparisons).
Your smile at the end.... priceless. 😁
I'd like to see what you do with the compression springs. I have an old Prusa i3 clone with threaded rods that had the Z banding. I printed some mounts that allowed it to freely move but rested on the gantry and required some weight to keep it stable, it was fine for years. I changed to a lighter extruder and upgraded it to lead screws and have had nothing but problems since (z banding, jamming on the z axis, inconsistent z height on levelling) and all but abandoned the printer for a year. While it may not be a golden bullet I think this solution might resolve most of my problems if it can do it with a lighter gantry.
WobbleX is good, for sure, but is symptom-solving. The real issue is that most original designs (that the Crealities of this world have copied off of) are kinematically over-defined or poorly designed, and the mistakes get inherited mindlessly.
Any design that has two parallel linear-rails under a print-bed ought to raise heckles of a mechatronic designer.
@Teaching Tech Michael, did you find an alternative way to get around Dual Z? I have an Ender 3 v2. I would love to use these on it. Thank you for all the great content.
awesome. now this i need to play around with on my ender 3 v2
Backlash nut on the top to provide tension? @TeachingTech I found that getting even load on the L/R wobble x was the hardest thing. The most annoying thing was the added components caused the nozzle to hit the bed first home firing the ballbearings everywhere, thankfully I found them. I got marginal gains, but gain all the same. (sidewinder x1).
I've tried everything on my Ender 6, converted to linear rails and a ball screw and it still has it bad. Will have to give this a shot I guess
Getring the springs in precisely the right alighnemt to avoid it tilting the pieces would be very difficult.
The 1/2k weight isn't such a bad idea, really. It's on Z, so it's not like that has to move much.
You can have the same result if you loose a little bit the screws of the brass nut, allowing some lateral movement
Maybe increasing the distance between the Rods slightly could help keeping the ball on the tracks cauz it would sit lower inside of them?
This would remove a little bit of the rotational axis becaus the wobblex parts are a bit closer to each other, but id guess theres isnt much rotational movement on an ender style z axis anyways
Or you could use bigger balls in that case (thus pins further apart and bigger balls)?
What I also don't like (well... '"don't like" are big words for this tiny issue though) is that the pins do not rotate around, but are press fixed in place. At first, I thought the pins where able to rotate and thus create a real bearing system with the balls, but they are not, they are fixed in place. So, over (very long) time, either the balls or the pins will wear out and dent and/or show deformation (especially when low grade steel is used.... thinking about the copies and clones which WILL come)
@@CookieTube i dont think the pins need to rotate. The point of the pins is that the balls can roll over them to create a type of linear rail. In the case of the linear motion the ball is already rotating in the needed direction so no rotation of the pins is needed
@@KidCe. The problem is not with the lateral/linear motion. That gives a true bearing/rolling action. The whole major point of why and how the WobbleX was made, compared to other solutions, is so it can also tilt. That is why it was designed. So yes, the pins should in an ideal world rotate so that they form a true bearing with the ball, and not 'grind' on eachother. But of course, making them rotate properly also would require a whole other level of engineering and complexity for such a tiny tiny rod (if at all possible).
I suggest to take a closer look at the cad drawings or videos to see how it actually works and what the true motion of freedoms are, and why they are needed (compared to an Oldham coupling). WobbleX it is a very simple, yet ingenious, solution (although it clearly also have some drawbacks).