Kinematically-Coupled 3D Printer Bed Proof of Concept

Dr. Mark Rehorst's Kinematically-Coupled Bed Blog Post:
drmrehorst.blo...
E3D's Kinematic Coupling is now live on Github! Go check it out!
github.com/e3d...

[@cerberes]

I really like that. So much simpler than the Hevort style.

[@escknx]

This is true auto level. Looking forward someone start to implement it, I'd love to buy something like this in open source version so I can tinker with it further. Building this from scratch is not my level of competence. You rock.

[@SebastienChedalBornu]

i know you posted a long time ago but what about having the 3 coupler in a "T" shape instead of Y shape that have to point to the center of the bed ? i mean you're supposeds to heat the bed before, you jsut want the bed not to bend because it is constraint when expanding, but you don't really care if i'ts a little off centered don't you ? and it's simpler to build/design the 2 front just have an "X" coupler direction and the 3rd one an "Y" direction. What do you think ??? and it could be used with the same direction whatever the size of your bed.

[@340havoc]

Me again. I've committed to giving this a go on my printer. I wanted to upgrade to linear rails on my Z axis anyway, and there's no better time than now to convert to kinematic bed constraint. I was thinking, with how much you enjoy using kinematics a video explaining the geometry and how you implement the kinematic constraint system in your designs would be cool. For instance, every different printer with different sized frames will need different ball rail angles. I think I figured out how to set it up myself, however that could be a great topic for a video. Thanks again for the inspiration!

[@340havoc]

I was just planning on doing this! Your tool changer project actually introduced me to the kinematic alignment system. I have a diy 3 screw sprung flexible bed now and thought it would be easy to use that alignment system as a way to make things a little more solid.

[@PrintedStupid]

Me too, except I like the coupling better here than in my design.

[@klschofield71]

I can't believe you have less than 1k subs. Do you have another channel where you post more regularly? Seems like you're incorporating all the really good ideas in 3D printing into one platform. Look forward to any updates.

[@POOFjunior]

Aw, thanks, for the kind words. Nah, this is my main channel. Soon, I'll get these designs out and open-sourced, and maybe then I'll try to plug them into places for more views. At this point, these are just "mission logs" of sorts.

[@klschofield71]

I'm also curious if you're implementing any means of thermal isolation between the bed and the printed adaptors? I like the idea of having a tool-changing enclosed printer that would be capable of handling temperatures for materials like PEK, but anything in the chamber would also need to be resistant to the harsher environment. Those smaller parts would be easy to mill on a "cost effective" CNC.

[@POOFjunior]

@@klschofield71 This is a good point. The short answer is: Im just going to need to run it up to temperature and test it. While that seems a bit cavalier, dont forget that the coupling is really only touching the ball at two points, so thermal conduction is basically negligible. As far as radiation goes, Ill just have to find out. For pla, though, this should be fine. For other materials with higher bed temperatures, i may need to revisit this coupling and print in a material (ABS) that doesnt have the heat creep effects that PLA does.

[@VuLamDang]

@@klschofield71 I think it's understandable, JV representing the cutting edge of FDM 3D printing, so this is not for everyone

[@SebastienChedalBornu]

@@POOFjunior petg should be enough probably and is far easier to print than abs

[@seangassin958]

Could nesting force be applied using magnetic force? I'm using 1 inch magnetic disks with #10 countersunk holes on my delta. Each has approximately 25lbs of pull force and you can get polarily matched magnets from k&m (they are awesome) so you could have a 10mm steel sphere sandwiched between 2 25lbs pulling forces

[@POOFjunior]

Totally! I opted for springs in case I ever had a need to take them off and do some sort of "automatic" bed removal.

[@xxHansxx94]

Hello there Joshua, those three blue things called, are they 3D printed or where did you get them? Cool build by the way!

[@pyro1596]

They are 3D printed. I don't believe there are any manufacturers who make kinematic couplers for 3D printers

[@dutchworkx3d]

Hi, Do you have the STL files for the parts for the Kinematic Z-Axis? I really like your design.

[@POOFjunior]

Youre in luck! Everything is at jubilee3d.com ;)

[@smorrow]

Could you de-level (and on a macro scale) the bed on purpose to get a fifth axis of sorts? (And has this been done?)

[@pyro1596]

Thank you for posting this video. I was looking for a good design to base my own kinematic coupling on and yours takes the cake. It's so simple to make. Now I just need to figure out how I'm gonna attach the ball bearing to the bed. Might use a recessed hole on the underside of the bed

[@POOFjunior]

If it helps, Im using threaded 10mm stainless steel balls that are actually Kossel 3D printer replacement parts. These days, Jubilee has really matured, and youre welcome to have a peek at this bed pattern and the rest of the design at jubilee3d.com Have fun on your printer building adventure!

[@pyro1596]

@@POOFjunior I'm using 8mm ball bearing that I'm gonna have to drill and tap..... Not looking forward to that haha

[@POOFjunior]

@@pyro1596 oh maybe no need? These are 8mm and tapped already www.filastruder.com/products/tool-balls-for-jubilee-3-pack?_pos=3&_sid=b539fae50&_ss=r

[@Galvo3D]

Fantastic! I am kinda noob, once you setup this way is the leveling done via software, or adjusting the screw couplers? Thanks

[@POOFjunior]

No worries! It's done in software as a homing routine.

[@Galvo3D]

@@POOFjunior so you are using only one stepper motor and a belt, instead of 3 steppers? Is it better in your opinion?

[@POOFjunior]

@@Galvo3D This is still 3 steppers. Whoop, My video frame managed to crop out some important details. 3 steppers is necessary for auto-tramming the machine.

[@Galvo3D]

@@POOFjunior oh ok now I understand... so 3 steppers + 3 rails + 3 leadscrews, then follow this instructions? duet3d.dozuki.com/Wiki/Bed_levelling_using_multiple_independent_Z_motors?fbclid=IwAR0lAgZxaBGdxjHbcIe15En9_r340eSKapWouT_UXDbaGq1TbJH1GDqKss8

[@AcrimoniousMirth]

Just saw this vid linked on Thingiverse and its an insta-sub from me. Which is rare, there are channels I've been watching for years I'm not subbed to.

[@SebastienChedalBornu]

how do you setup your duet to move the Z axis with 3 motors instead of 2 and more important to make it perform bedlevelling correctly ???

[@SebastienChedalBornu]

nevermind found your github with the configuration. never though of considering the z axis like a delta for calibration. thanks, cant't wait for the full printer plans.

[@donz.519]

Where do I find this bed at? I like this design. Any updates to it?

[@POOFjunior]

Totally! This got fleshed out into a full-on toolchanging printer, Jubilee. Everything is open source, and you can even find vendors where you can pick up the bed too. More info at jubilee3d.com and feel free to join the crew on discord: discord.gg/jubilee

[@PRO3DESIGN]

Where did you get the ball joint screws?

[@ckafrouni]

How does Z-Hop work on a system like that since the bed is not screwed if I am understanding correctly what is happening here Thanks a lot for your work!!!

[@POOFjunior]

on the assembled version, there are springs holding the plate down in case you accelerate super fast or travel with the printer. (they're just preventing it from falling off.)

[@ckafrouni]

@@POOFjunior Hey, Thanks for the reply, I still had a question as for the Tool changing mechanism, Do you plan on adapting the firmware to Klipper? I am planning on modifying my Blv Mgn cube to your tool changing system, but I'm still saving up for a duet 2 and duex 5 for in a few month, in the mean time I wanted to try it in Klipper, watched a few videos and read about it, but I don't feel confident enough. What do you think.

[@ISFRedSandman]

I like it! The only thing to think of is that having multiple motors for one axis should cause desync issues, in theory that is.

[@VIDEOAC3D]

Sick design man! Great explanation too. 3yrs late, but... subscribed. :)

[@kevinaub]

Nice. Thanks for this! This is what RatRig have adopted now!

[@nirodper]

Planning to do the same thing on a concept printer, how well did it turn out after these couple of months?

[@POOFjunior]

It's holding up super well. Thanks for asking. Reference design is also now online too as the printer Jubilee. May it serve you well.

[@nirodper]

@@POOFjunior thanks!

[@Mogalize]

How were the 'slot' angles calculated for the thermal expansion of the bed plate? Have you had any issues of miss-alignment from the thermal expansion? ie: the ball not following the track during its expansion.

[@Ryan-dz3jo]

Great design and keep us updated with results.

[@LiquiDirt3D]

Great video I have a design similar to that for a 3d printer we are working on so great minds think alike. Great job there. nice looking printed parts as well on the machine.

[@aaronramly1705]

you were both twin in the past life maybe

[@v1Broadcaster]

this was before the hevort?

[@dariuscastor2331]

Sir you have a five axis 3D printer

[@Ed_Stoddard]

I had a CoreXY printer in the works, (well in my head anyway), and was going to use the three Z motor design I'd seen, but this idea puts a whole new layer of precision on that. Great idea. Look forward to updates and more ideas. Keep up the good work. Do you plan on going to ERRF in October? I hope to have a three head delta ready by then.

[@crazycreator8068]

hello @Joshua Vasquez, can you post the stl's from the beholder?

[@timschafer2536]

Does the coupling counteract z wobble?

[@POOFjunior]

It doesn't counteract any existing z wobble, but it wont introduce any z wobble by design. In other words, if something besides the bed (like a super bent leadscrew) is introducing wobble, it doesnt counteract that. But, as designed, it doesnt have any source of wobble. From the tall, skinny, prints Ive done, the bed has rock solid layer registration. Hope this helps!

[@timschafer2536]

@@POOFjunior thank you.

[@SebastienChedalBornu]

where did you bought your linear rails ???

[@POOFjunior]

I got mine from here: www.aliexpress.com/snapshot/0.html?spm=a2g0s.9042647.6.2.103f4c4d8GHuhy&orderId=104791534346103&productId=32906662047

[@solverbox]

@@POOFjunior Here's a correct link. The one you posted is to your order history! :) www.aliexpress.com/item/32906662047.html

[@SebastienChedalBornu]

Excellent idea, I really like it and will consider it for my future printer. But I will probably use aluminium for tool changer instead.

[@thomaslevendig7421]

Awesome concept!

[@BenSycha]

Love it. Cheers for sharing.

[@chloemcholoe3280]

I'm kind of a newb so sorry for asking but why would someone use this instead of just screwing the bed down to an extrusion/plate and using software comprehension I manually kind of get the bed close as I can with screws with springs and then use that. cause my bed is bowed in the middle. also 3 motors on my board wouldn't work it has only 5 steppers ;c but one day when I'm rich AF i'm gonna build one with a duet :D

[@POOFjunior]

No prob. This setup involves no manual setup. When the printer starts, it pokes the bed at three points with a limit switch. From there, the screws independently adjust themselves, autoleveling the bed. What's elegant about that is that no manual work is required. It's entirely done automatically. The other benefit has to do with heat. When the bed plate heats up, it expands. From setups that are rigidly connected, this expansion cannot freely grow, so the bed warps as it gets warmer (and bigger). With this setup, the bed can freely expand unconstrained. What that means is that the bed will still expand as it gets warmer, but it will not warp and it will not change its angle. Instead, it simply "grows" into the coupling. Overall, this is a bit tricky to explain in words. I may do a followup video with the autoleveling demonstrated. I hope this helps though!

[@chloemcholoe3280]

@@POOFjunior Thanks :) is there anything wrong with software comprehension? (like normal mesh leveling with a probe that makes the z axis go up and down!)

[@mattgraham4340]

@@chloemcholoe3280 Nothing wrong with it, per se. However, normal mesh leveling doesn't actually level the bed in a system that only has one individually addressable Z motor. Many printers have multiple Z motors, but many (most?) of them they are not individually controllable. On such systems the best the software can do it compensate for the non-level bed (which is usually good enough)

[@urizen878]

The weight of the bed is supported by the tiny balls of the motor bearings. Not the best idea.

[@sergeantseven4240]

I love this concept but i think it misses one of the major benefits to probed mesh leveling. You can accommodate for uneven bed surface or beds that rent totally flat. This wont do that unless you do two methods of leveling.

[@c-djinni]

But isn't probed mesh levelling needed because of the bed warping? If I understood correctly, this setup prevents/minimizes the warping by not constraining the bed, right?

[@ikbendusan]

@@c-djinni i know you commented this a year ago and probably figured this out for yourself lol but yes it will compensate for warp because of thermal expansion, but it won't compensate for bad initial conditions (a warped bed at room temperature)

[@c-djinni]

@@ikbendusan That won't be a problem if you buy high quality machined aluminum plate though.

[@ikbendusan]

@@c-djinni which many 3d printers don't come with from the factory, so the point stands

[@c-djinni]

@@ikbendusan many 3d printers don't come with kinematic coupling either...

[@syedmraza-ca]

i support u 100%

[@theGermanPrintingNerd]

Absolute cool

[@1FishinAddict]

Rat rig v core 3 copied this exact design…very cool.

[@zhaokate7685]

Nice machine

[@TheOrgonaut]

is this a rat rig v core

[@TrungLe-dw1mm]

MGN12H?

[@cigp]

subscribed

[@jakabgipsz4788]

This makes no sense. It's a waste of time and money. It is perfectly sufficient to connect the 3 axles with a ribbed belt and use a single Z motor. It is enough to adjust the height of the table once, by loosening the pulleys on the threaded shafts. Then the Z height never changes. One motor, one controller, simpler board. The Bambulab X1 and P1P are both famous for printing beautifully with a Z motor.