Threadless Ball Screw. No More Backlash

by MSollack, published

Threadless Ball Screw. No More Backlash by MSollack Jul 4, 2013

Featured Thing!

8 Share
Download All Files

Thing Apps Enabled

Order This Printed View All Apps


Use This Project

Give a Shout Out

If you print this Thing and display it in public proudly give attribution by printing and displaying this tag.

Print Thing Tag

Thing Statistics

76338Views 6361Downloads


This is a type of ball screw I read about at least 20 years ago in a technical trade journal. I have also seen it used in some linear actuators. Some advantages of being thread-less is that there is no backlash, no expensive threaded rods, and if you exceed the force threshold the unit simply slips without damage. A perfect combination for 3D printing.

7/6/13 - I have added a picture of a double unit.
7/7/13 - Added bearing mount with 10 deg angle. 4.14 mm per rev.
7/18/13 - Added bearing mount with 623 ZZ bearings and a way to adjust the tension. BOM added to instructions
7/18/13 - Added 623 bearing mount with 38.5 degree bearing angle.
7/23/13 - Added 624 bearing mount with 5 deg angle 8mm shaft
7/27/13 - Added base for 623 bearing mount



3) 625 2RS ball bearings. 5 x16 x 5 mm.
3) 5 mm machine screw 16 mm long.
3) 5 mm flat washer
3) 5 mm thin hex nut
1) 8 mm hardened shaft. McMaster Carr 6112K45 or similar
1) printed ball_screw_8mm.stl (bearing mount)

optionally print ball_screw_8mm_base.stl

Print the ball screw bearing mount at 100% with at least 60% fill. Press three 5 mm thin hex nuts into the pockets on the back side. Place a 625 2RS bearing then a 5 mm flat washer onto the 5 mm screw. Loosely screw all three assemblies into each top side pocket of the bearing mount. Do not tighten them yet.

Place the 8 mm hardened shaft through the center hole of the bearing mount and tighten each bearing assembly. The bearings should firmly clamp the hardened shaft so that the shaft will not slide with moderate force but rotate with silky smoothness.

If the rod is too tight or too loose reprint the bearing mount at slightly smaller or larger scale and re-assemble until the fit is perfect. If the screws protrude out the back of the bearing mount a washer may be added under the screw head as needed.


3) 623 ZZ ball bearings. 3 x10 x 4 mm.
3) 3 mm machine screw 12 mm long.
3) 3 mm flat washer
3) 3 mm thin hex nut
1) 8 mm hardened shaft. McMaster Carr 6112K45 or similar
1) printed adjustable_threadless_ball_screw_8mm.stl (bearing mount)

All Apps

3D Print your file with 3D Hubs, the world’s largest online marketplace for 3D printing services.

App Info Launch App

Auto-magically prepare your 3D models for 3D printing. A cloud based 3D models Preparing and Healing solution for 3D Printing, MakePrintable provides features for model repairing, wall thickness...

App Info Launch App

Kiri:Moto is an integrated cloud-based slicer and tool-path generator for 3D Printing, CAM / CNC and Laser cutting. *** 3D printing mode provides model slicing and GCode output using built-in...

App Info Launch App
KiriMoto Thing App

With 3D Slash, you can edit 3d models like a stonecutter. A unique interface: as fun as a building game! The perfect tool for non-designers and children to create in 3D.

App Info Launch App

Print through a distributed network of 3D printing enthusiasts from across the US, at a fraction of the cost of the competitors. We want to change the world for the better through technology, an...

App Info Launch App

Quickly Scale, Mirror or Cut your 3D Models

App Info Launch App

3D Print a wide range of designs with Treatstock. Easy to use tools to get the perfect result. The global 3D printing network that connects you with high-quality and fast working print services nea...

App Info Launch App

Can anyone please comment on the efficacy using this configuration vs. threaded rods or leadscrews (quality of prints at comparable speeds and jerk settings, etc).

Also, given the innate inaccuracies of at-home 3d printing, how stable are steps-per-unit and steps-per-revolutions measurements once made and entered into your config file?

Thank you!


What is the preferred bearing if I were to start from scratch? Thanks!

Jun 4, 2016 - Modified Jun 4, 2016

If anyone care, I calculated the formula to get the theorical travelling distance for one rev by knowing the angle.

PI x RodDiameter/Tan(90 - Angle) = Travel distance for one rev.
So, Pi x 8 / Tan(90 - 38.5) = 19.9914... mm/rev.

Playing a bit with the equation, you can found that: 90-ArcTan(Pi x RodDiameter/TravelDistance) = The angle of the bearings.

Just don't forget it's theorical. This kind of linear motion is simply not 100% precise. Still a very good idea.


Thanks for the calculations.


What is the force threshold? What is the acceleration limits. Im interested in this for cnc machining but I can see where it "slipping" and not having a positive motion at anytime during a machining cycle would be extremely detrimental to any part trying to be produced. Ive seen where others have used a threaded rod you still get the same type of motion but it also has a positive directional force that cant be "slipped". If there are formulas for this can you provide them for the threshold and acceleration forces itll work within?

Hi OrionCAM,

there are many factors involved in calculating the slip threshold, so I do not know a formula. I thought this would be good for 3D printing because of the low forces required. I do not think it would perform well in a cam situation. A Mister Park from Korea had an interesting twist where he used the bearings to ride in the valleys of an ACME thread. perhaps you can look him up on thingiverse. In any event It would probably work best for milling non metals using larger diameter hardened rods and not too much pitch.Tthey are cheap to produce. you can experiment if you like.

Thanks for the comment

Yes that is what I was referring to when referencing the threaded rod (acme or ortherwise) with the same drive pricipal and it having a higher threshold for slipping but the threadless ball screw is about the best idea for linaer motion and I do agree I think its smaller size and lighter weight components lends itself better to 3d printer and even in wood router and plasma tables where the force is more along the the weight of the tool rather than the work load of the tool. Still never the less a brilliant idea and concept with the right materials and angular contact this can be researched further to adapt it to a lot of motion controls where the primary force isnt trying to make hard material into some form or shape by milling or lathe work. It does bring to light that it could be used for precision location instead of scale glass or otherwise.

"Also I recommend using two units back to back as good alignment is important for accuracy. "

what do you think instead 2 unit, use 1 LM8UU and 1 unit in the top? Right now due to the holydays, are missing the 624 bearings :)

Nevermind.. i just found 10 624zz bearing :)

hi, i just printed the 624 bearing piece for the 8mm Rod but something is wrong because seem that the built part can accept 6mm Rods.
I used standard 624 bearing 3x13x5.

I did the 624 bearing version as a special request from someone. I didn't have access to the bearings so it was untested. I believe it worked for the person that I designed it for. Could it possibly be a scaling issue? I have made a number of the other bearing sized versions without a problem. perhaps you could try one of them. Also I recommend using two units back to back as good alignment is important for accuracy. Hope this helps.

thanks for the infos... is not a scaling issue :) thanks got from you (the design). I have now a working part designed for Prusa re-mixed from your. When working and ready i'll make the official remix for the Prusa Steel.

Thanks for the model :)

is there scad file available? can one be produce to use 625zz bearings?

Sorry, I do my work in Autocad. I did include some DXF files that you may be able to convert to SCAD.. The design will work with most bearings that can handle some side load.

Feb 2, 2015 - Modified Feb 2, 2015

This looks like a cool device, but I am puzzled by the Instructions. I see that the "original" flavor uses three of each fastener. (Machine screw, washer, thin nut, and bearing). The "adjustable" version also claims to need three of each fastener - but don't we need a FOURTH nut and machine screw, the ones that provides the adjustability? Are those "fourth" machine screws the same length?

And what about the machine screws that clamp the main part to the optional base? (I see these in the photo of the yellow variant.)

Comments deleted.

I have created a baseplate for K8200 / 3drag Z-axis. You find it here http://www.thingiverse.com/thing:399804http://www.thingiverse.com/thi...

Threadless Ball Screw Z-Axis K8200 3Drag

Test finished - works fine on double Z-Axis (and on single Z-Axis as well). I have added some documentation and pictures on http://www.thingiverse.com/thing:399804 The thing is not longer tagged "under construction"

Threadless Ball Screw Z-Axis K8200 3Drag

MSollack, I am working on my senior design project and we are making a large 3D printer. Do you have one of these designs made for a larger rod? something like 12mm or something like that? If not do you think you could point me in the right direction as to how I would make one for a larger rod?

Perfect!!! No more Z-wobble! Thank you.

I've tested by hand, Thanks Mark! This is really cool! http://youtu.be/TxOifzvnm70http://youtu.be/TxOifzvnm70

Your work in applying this concept is impressive as is the video. I am glad you are having success

having this as a variable pitch would be interesting as well

I bet you could use this device to push filament, too.

Hayoo, printed out the original ball screw for the 625 bearings. The rod spins really nicely on it but I have to get the rod cut before I can tell you how well it does. Going to try using it on the z-axis for the solidoodle, will report back. Two things though:
1: I had to print it at 102% the size, did anyone else have to scale it or is that a calibration issue on my end?
2: Is there some significance on why the center hole for the smooth rod is covered at one point?

Glad you had success. That was my original design and since it was not adjustable I purposely made it tight and would go through an iteration or two to get it right. also I added a feature to center it on the base which caused an overhang. I added a punch out to eliminate supports. I later determined that that feature is not necessary.

I highly recommend using two units back to back. It is much more stable and repeatable that way

Just wanted to say I installed it on the z-axis to fix some backlash/banding I was having, and it works fantastic so far. Cost me probably $25 for everything for the mod. Thanks for making this.

  1. That might be due to shrinkage if you print in ABS plastic.

So after 3 days of testing I am sad to say that the repeatability of positioning is not that accurate. I have these mounted on a machine that I am in the process of building. I had planned on using 8mm diameter 20mm lead screws on the printer to begin with. When I saw this featured I thought ooh my!

Well after getting them mounted and the steps correct. The tracking when going back to the start of the next layer wonders from layer to layer. I've gone slow 10mms and faster 80mms. No repeatability at all every big move is off from the last.


Would like to see your setup, any pictures or video?, do things move freely with no wobble in the drive shaft and a smooth action with constant load for stepper? My initial testing showed that it has potential but requires good straight parallel drive shaft and side smoothrods. Using a 20mm pitch with 8mm smoothrod could be too much angle to keep things accurate, try thicker rod and less pitch, am using 15mm pitch and 10mm smoothrods here. The shaft needs bearing on each end to help support it, need to add an extra bearing today, this is current test jig. http://www.youtube.com/watch?v=uzSgV8Ecu7Ahttp://www.youtube.com/watch?v... .

Nice video. My machine moves the same, smooth quiet....
I think I have found the issue. It would appear that the tension on the two halves is crucial for symmetric tracking.

The design that has no compression gap or the one that you have designed with both halves in the same assembly may get rid of the tracking error.

I am going to try one on the x axis of my machine. It will require a redesign to use that type on the y axis.

Great job on releasing this, LOVE IT! I am going to see if I might can use something like this to drive a small CNC mill... I did a major re-write of some existing code and made a fully customizable version: http://www.thingiverse.com/thing:125529http://www.thingiverse.com/thi...

Not sure yet if I have the angle correct, but it should at least work, just maybe not at the pitch intended... I'm printing one now to test out the functionality to see how strongly it "grips" the rod

Threadless Ballscrew (customizable, no backlash)
by texsc98

CNC requires considerable force. If you can't do it with a 5/16" rod try reworking it for a 1/2" rod. For durability when you have all the kinks worked out upgrade to a hardened shaft.

oh yea I was definitely going to use a much larger rod, the 5/16" was just what I had sitting around the house already to try it out. And yes for a CNC I would definitely use a hardened rod as well. Thanks for the tips :D

what happens if you make a version for say m8 thread? I mean, instead of the smooth rod make the bearings ride on the thread, keeping them in constant angle and working against.

or ride in valleys on a larger printed thread. this is a really nice idea though! and thought provoking!

I have seen something similar to what you ask about where a thread is ground into the bearing race and rides on a threaded shaft. Perhaps a threaded sleeve could be printed to fit over the bearings. Otherwise each bearing would have to be offset by one third and two thirds of the pitch to land in the threads properly.

i've tried printing remix of this project (customizable version allowing me to tilt the bearings) and it seems major issue is getting proper tilt vs minimal size of bearings. the smaller bearing is, the more problem the lenght of screw becomes.
eventually i've got rid of nuts as it seems forces are not so big and screws can be screwed directly into abs, and this allowed me to use 3x10x4 bearings with 8mm rod at 25 degree tilt. 20 deg tilt seems to be minimal what is required to prevent bearings slipping from the rod.

ad offset - i've managed to solve it by plainly adding variable amount of washers under each bearing. this makes the assembly bit assymetrical as with tilt each washer also moves the bearing off the track a bit, but it seems it's enough to reduce wobble to make possible screwing two symmetrical halves back to back.
i guess one should use spring washers as with threaded rod one does not really depend anymore on perfect bearing angles, and springs will compensate for rod wear over time.

to resume - it's doable and i hope to post results soon , definitelly worth trying :)

btw. i've used cheap m8 'construction grade' rod.

I would think a system like that would be more back-drivable than this system, but I could be wrong

This is a very nice "thing" I have a bot that is partially made so I can easily adapt my bot to use these. What cad source did you use to create this. Do you have any other options for exporting other than DXF and STL. I will prob need to add some features and adjustments to work with that bot.

Great design, works very well, thanks for sharing. One thing I noticed from the one I made (see http://www.thingiverse.com/make:42659http://www.thingiverse.com/mak... ) is that the bearings seem to run on their edges, not on the flat part of the bearing. Is this just an offset problem, or intentional?

Threadless Ball Screw. No More Backlash

This is intentional. If the bearings ran on center, friction would be the only force holding it in place. by running on edge the holding force is much greater due to microscopic deformation and rebound of the hardened surfaces. See comment by thebarron88 or my comment to bwhoward88 below.

I decided to test this. Using jspark's OpenSCAD version as a basis, I printed two different bearing mounts (623 bearings on 10mm hardened rod), one with the bearings running on their centres, one with them on their edge, as per the original. I had 400mm hardened chromed rods, which I ran each bearing back and forth on for 5 minutes, with a drill turning the rod at reasonable speed. Testing grip was a bit arbitrary; I did the preload up as tight as I dared, then pulled on the mount until moved. Findings: Grip on the rod seems proportional to the preload you can put on the bearing. With the bearing that ran on their edges, they ran rough if preload was too high. Grip was good, but gave way sharply. With the bearings that ran on their centres, much higher preload could be applied. Grip on the rod seemed equivalent (though preload had to be higher to achieve this), but seemed to lose grip more progressively. Wear on the rod, even a non-hardened drill rod, was negligable. Conclusion: depends on your application. I think running the bearings on their edges will have a shorter life (though this will depend on the quality of your bearings), and wear the rod they run on, but have slightly better grip. I'm going to set up a test to see how repeatable the accuracy is next.

I see a new rostock design idea.... too bad I don't have the time to invest in making it, maybe someone else can follow up on my idea.

currently, the Z-axis of a Mendel uses 2 smooth rods and 2 threaded rods. Would this allow us to eliminate the threaded rods alltogether? Or would we have to use 4 smooth rods and 6 (or 12) bearings? similarly, could this be used to eliminate the X or Y-axis belts?

It may be possible to do away with the smooth support rod on the Z axis if the ball screw rod has bearings on both ends and the motor is isolated with a zero backlash compliant coupler and assuming that the ball screws were well enough matched. It could also be used for one of the two support rods on the X or Y in a similar arrangement. good question

When I thought of this application i was really thinking of replacing the X and Y belts. I would think that the tandem Z axis would require a very well matched set to prevent the accumulation of error. I don't know that the current state of home 3D printers can achieve that kind of accuracy yet.

I don't know if this has already been mentioned or asked (tl;dr), but doesn't that mean that you use a high shear force to achieve a relatively low normal force? And doesn't this shear force eventually result in a "track" because the screw will kinda like work it's way into the rod? Even if the rod is hardened because hardening often goes alongside with lowering friction which makes you need more shear force for the same normal force. Sounds like an optimization task...

I would think that the possibility of forming a track in a hardened shaft would be no different than the hardened balls in the hardened ball bearing race forming a track. it probably does eventually happen but only after a long time. hardened steel is pretty hard.

Wow! Disquss has just told me that you replied to my comment... that's a hell of a latency. Well, however... We have this part working in a self designed 3D printer and it does a very good job so far. But for my own K8200/3Drag I'm gonna use a spindle with a TR8x1.5 trapezoid thread. It's a little less space consuming.

More info can be found by Googling "rolling ring linear drive", which seems similar for those who just want to read.

If the 38.5 degree works well, it might be interesting to try printing out with it. 5 degree, as DFarms said, will give a low pitch. Assuming you want to print at 80 mm/s, with 5 degree bearing angle, your stepper motor will have to turn at 36 rev/sec. Nema 17 won't go over 8 rev/sec without missing step (with 1/16 microstep and a strong value for acceleration/jerk).

With the 38.5 angle (20 mm/rev - 80 mm/s), it'll give a 4 rev/sec (more realistic). You could even thing about trying to print at 110-120 mm/s. I'm actually trying different Lead Screw pitches/sizes to get a good printing speed. I found some 8 mm pitch thread rod, they are working great and cost $40 (motor+lead screw). But it's quite impossible to print over 64 mm/s (8 rev/sec).

Belts' systems like Prusa have a "pitch" of 54 mm/rev, giving at 110 mm/s a revolutionary speed at 2 rev/sec. This is a really good speed for stepper but it gives a displacement precision of 16 microns (for 1/16 microsteps). A compromise would be a pitch between 20-25 mm/rev, giving a resolution between 6-8 microns and will still be possible to print up to 110 mm/sec.

Expensive threaded rods?

I was thinking about the ball screw type thread this would replace. They are expensive. A normal roll threaded screw is not very accurate.

Quality threaded rods (like square or ACME threads) can be quite expensive.

38.5 degree bearing angle is about 19.9915 mm per revolution.

5 degree is about 2.1988 mm per revolution

Tan(bearing angle)(shaft diameter)Pi

was just about to say that the formula should use atan and not asin. I have uploaded an semi parametric openscad double sided version with mount for 10mm pitch, very smooth. http://www.thingiverse.com/thing:124800http://www.thingiverse.com/thi...

Im pretty sure the formula is correct, It matches up pretty closely with the op's findings. the main issue is that depending on the tension i think the angle can actually change.

This is a really important and awesome thing. Good job linking the past and present.

How did you do the double unit? When I printed two and turned one upside down they went in opposite directions.

No need to mirror. All I have ever done is put two identical ones back to back. The second one flips upside down which then aligns the bearings in the correct direction

Interesting, when I did this and did not lock them together, they move apart. Are you saying because you lock them together they move correctly?

I just assembled a double unit to test and when I hold one and rotate the shaft it moves away from the other. When I hold the other one and rotate the shaft it moves toward the first one. When they are attached together they move in unison which is correct.
My other comment was in error concerning your question. That comment refers to accuracy

If they twist or tilt out of perpendicular to the rod it effects accuracy because each bearing will then track differently. Attaching them together keeps them perpendicular.
In a printer it is possible that whatever it is mounted to will keep it perpendicular

Print one mirrored

How do you print them mirrored? I have never seen a setting to allow that?

If you use Netfabb to clean up any of your models there is a mirror feature there to do it in the "Right Click" menu. Thats how I mirror things.

Thanks MSollack! I was impressed after assemble and test. The design was perfect. There's no difficulties during the printing and assembling on my UP printer. I've just downloaded and printed. There's really no backlash and it's hard to move without rotation of smooth bar. :)

Would this system be susceptible to wear due to dust and grit on the rod?

Yes, If the grit is harder than the steel parts. Although I would think less susceptible than the screw used on the Z axis or the linear bearings used on the other axis. The rollers would tend to push the grit away rather than pull it in. Plastic dust should have no effect. you should keep your machine clean and lubricated anyway.

Also regular steel rods are ok for demonstration but I strongly recommend a true hardened shaft for real use. see my BOM

Someone should design xends for prusa mendel that uses these, would be very easy to try them out then!

Is the mini double mirrored or inverted for the second side?

No, It is a second copy of the same part. They are just placed back to back.

This is a seriously wonderful idea to give to the 3D community. Thank you.

I like it.. do you suppose it will have a limited lifespan?

Doesn't this just shift the backlash over to whatever is going to be resisting the rotation?

Wouldn't that backlash be there anyway?

If you drive the shaft direct, it has as much backlash as the motor does. Drive it directly from a stepper, and the backlash is zero.

Wouldn't the backlash be there anyway?

Backlash is caused by play in the system when you change direction due to looseness or wear. In another words there is a dead zone where movement is not transferred.
The bearings and rod of this treadless system are always in constant contact regardless of a change indirection so there is no significant backlash

I feel really dense, but I am just not picturing how this works...friction drives the smooth rod at a set rate per revolution?

Please see my last comment to bwhoward88 below.

The three bearings center the shaft. The angle of the bearings simulate the thread pitch. As they roll over the shaft it follows the pitch of the bearings like a thread. Because of the point contact against the hardened surfaces the force is very high and keeps the shaft in line, not necessarily by friction alone.

It really does work. If it is sized properly the action is very smooth and considerable force is needed to cause it to slip. It may not be a panacea but I think it has at least some potential.

I would like to know how this effects the print quality. Does it stay in the correct position wen you print? Also I doubt I can print this part myself because I think the angle of the bearings need to be very, very accurate.

Anyway, nice upload :-)

That is the BIG question. Someone needs to try it in a real application. I am planning on rebuilding my Prusa into something like a MendelMax 2.0 using some of these. But that is probably months away from completion

I hope someone can try it and let us know the results even if it is just a single axis like an X or Y. I would hold off doing a Z first since it is a tandem system and more complicated. Or maybe a Rostock.

This is incredibly cool! Very nice work! I am actually playing around with making a custom version of this in openscad. I am playing around with different bearing's, different angles, and different shaft sizes. I am wondering if there is any math you use to figure out the X and Y translation based on the diameter of the shaft, the diameter of the bearing, and the angle of the bearing, or do you just eyeball it until you get it overhanging the center shaft's radius to your liking?

I am glad you are enjoying it.I have included some math in response to ChristianSi's comment for determining the travel (pitch) per revolution based on bearing angle and shaft diameter.

In my original design I pretty much kept modifying the bearing spacing until I got the result I wanted. in my later design I put a split in the holder so the spacing is not as critical. I have been adding the shaft diameter and the bearing diameter then dividing by two and subtract about a mm to provide some tension. That is the spacing of the bearing center line from the holder center. that get's me in the ballpark.

Hope that helps.

I saw that math, and it is very helpful! Makes sense what you are doing, I have been kind of doing the same, I thought you might be using some trig to get more exact, which is what I was trying to do, but my trig is a bit rusty. Thanks!

Very cool!

Any chance to get a version for 624 bearings, too? I have plenty of those from my old Mendel that is not used anymore (I have a MendelMax now) and I would like to reuse my old bearings instead of buying new ones (623's or 625's).


Same here, I just busted them out and a have an 8mm rod too that I would love to be able to use.

Belt-less Rostock printer anyone? ;)

I've been working on a rack and pinion system for Roatock. Maybe I'll use this instead....

Looks certainly like it would be a nice replacement for the threaded rods commonly used for the Z axis of Repraps, but I wonder whether it could be fast enough to replace belts? Assuming a good Nema17 motor can do 600 revs/min and still provide enough torque, the 10 deg angle version would allow up to 41mm/s. I currently do infill at 100 mm/s and travel moves at 150 so that feels a bit slow. Maybe it's possible to add a 15 deg version to make it faster?

You can calculate the theoretical bearing angle to pitch length (distance per rev.) as [ Pitch = Dia. pi tan(angle) ] or [Angle = arctan(pitch / (dia*pi))] you can calculate the arctangent in the windows scientific calculator by checking the INV box in the upper left corner.

Let's say you want to travel at 200 mm/s with a 600 rpm motor (most should do double or triple that) you will need a 20 mm pitch [mms / (600 / 60)]. Using the angle formula you would need an angle of 38.5 degrees with an 8mm shaft or 32.5 degrees for a 10mm shaft. I believe that the bearings should work up to about 45 degrees if a suitable holder can be made. Of course doubling the rpm would decrease the angle required by half. The possibilities are still there. I will try to make a holder with a 38.5 angle and see. If my math is wrong let me know

Very nice!

Is it necessary to use a double unit (with 6 bearings) to get precision movement in both directions?

It will move in both directions with a single three bearing unit, however any flexing of the unit will cause inaccuracies. Unless there is additional support to prevent flexing (tipping, wag) a six bearing double unit is much more reliable. I would recommend the double unit. Most commercial units have six bearings or more.

I will be posting a much smaller unit using 623 bearings on an 8 mm rod similar to Natko's with a tension adjustment feature soon.

Hi, love the concept, have built 623 version for 6mm rod.

Nocked out 2 using the same STL and found that the movement is not the same per rotation. Think slight differences in angle cause this, but it means at the moment I cant use this on a Z axis duel motor system :(

If you have two place them on the same shaft, holding both rotate shaft, after quite a lot of rotations you find one moves more than the other (get closure or further apart)

Love the idea tho, is smooth and seems firm solid enough to support printer heads

I am glad you had some success. It is a work in progress.

MSollack posted the files hope u don't mind .. after fideling found 2 things U need same tourque on screws an second dont put none axial torque (wag) on the bearing ... BRILLIANT otherwise

I never thought about doing this but now I want to build it. Is there any reason why you used 625 2RS ball bearings? Can other ball bearing be used?

The bearings are small to help keep the mass low. Also the 625 2RS can handle some side load because it has a fairly deep race. Plus they are cheap on Ebay. Other bearings can be used but the mount would have to be redesigned. These bearings should be good for a 10 mm shaft or possibly a 12 mm one with a proper base. Anything larger may need larger bearings
I am also working on one with a 10 degree angle and a design with adjustable tension.

I look forward to your designs :-) Correct me if I am wrong, but isn't the amount of force it can
hold up a function of friction and area. With a small friction force, due to
smooth aluminum on smooth aluminum, and a small area, due to point contact
between two rounded surfaces, the amount of force that your design can hold up
is significantly less than a threaded system? Also could your design not
greatly benefit from adding another three bearing on the other end of the
bracket and thus doubling the amount of force it can support? Lastly, would you
not be using long and thin bearings or rollers for industrial systems that would be used
to move large loads?

You can pair two of these to get double the holding force. Why complicate things with a separate design with 2x the bearings and 2x the mass?

Also, while there are better suited bearings for this particular application (tapered roller, perhaps?), the point was (at least if i got the OP correctly) to make it cheap and easy to obtain and build.

Very nice. I've seen something like this in telescope drives. You've got me thinking now...

Very good idea! Indeed, I've seen these in action and can confirm there's very low wear rate and predictable movement. Presumably can adjust the rate of movement per rev by simply varying the angle of the bearings.

That's correct. I set my angle to 5 degrees for a stroke of 1.8 mm per revolution. Other angles can be tried up to 10 or possibly 15 degrees. Two unit can be mounted back to back for higher loads or greater stability.

Actually, I was thinking about the reverse; Lower angles for increased resolution on a Z axis.

the pb is there is unpredictable slip, so you need a retro control for the real move for the electronic to know exactly where the carriage is. More you need a way to tune the fit which is going to change with the time.

you could do the almost same systeme with threaded rod and then no more slip (no more overload safety too )

but to move things with no accuracy or end switch it worth the try

There is no unpredictable slip.
The slip is very predictable - once you exceed the maximum holding force, it slips ;)

You may be surprised, The linear actuators I have seen using this technology are found in web guides used in the machinery that the company I work for builds. They move hundreds of pounds of mass easily with no slippage or loss of accuracy for years without adjustment. The feedback in that application is found on the servo motor not the actuator. As long as you don't exceed the load limit there appears to be no slippage.

That being said my concern would be in overloading the bearings and causing early failure. However the ABS plastic seems to have some flex that may reduce any overload. Time will tell and I am hoping that experimentation by myself and others will prove the superiority of this over belts and acme threaded screws.

I use a Starrett optical comparator at work that uses this kind of drive on the X axis. It has Heidenhain linear scales on it and the resolution is .0001". The accuracy is +-.0002" over 18". Most excellent contribution bud! Thank you!

I could make one but there is really not much to see. Each revolution of the rod moves it by 1.8 mm

Do you have a video of it working? this is awesome!

indeed this is awesome. I am looking forward for updates!

Yeh, this is really good. :)