Clockwerk - A 3D Printed, Three-Axis Tourbillon

by aaddaamm Jul 9, 2016
Download All Files

Thing Apps Enabled

Please Login to Comment

do i need to set horizontal compensation?or just print in 100% scale with 0 compensation should work?

Martin351 -

There is a list on the thing details tab - and he has all of the parts laid out in the beginning of the assembly video and as a picture above.
Parts List
4mm Bearing - Team Losi Mini T Bearings [1 or 2X]
2mm Bearing - Team Losi Micro T Bearings [6X]
M2.5x8 Machine Screws - [just get the 100 ct bag from amazon]
2.5 lb Barbell Plate 2X (or anything that weighs 5 lbs)

Like Adam says - this is not a beginner project - my biggest step toward a working model is getting better tolerance for the steel shafts. The ones available on Amazon have been too small and slip in the bearings. I found much better ones eBay;


Good luck - despite not having a working model yet - it has been a fun and engaging journey so far. Thanks to Adam for providing this.

Hey @aaddaamm, could you be so kind and post the amount of parts we need to buy to assemble this Clockwerk? That would be great. Thanx in advance.

Or someone else who already built this Clockwerk.
How many pcs of bearings are needed?

I can get it to tick in the single axis in my hand, then in the next axis its not so good. The impulse pin gets past the fork and jams up.

Hi, what is the approximate weight of the final printed parts?

What's your simplfy3d profile, i cant get mine to print reliably and accurately.(I also have a FFCP)

This is the file I use for printing on the FFCP. I put the Polymax PLA on the left nozzle, and the Polysupport on the right nozzle. I print with a buildtak platform. This will print with a one layer raft, which enables easy removal from the platform.

Sorry it took me so long to reply to this! I've been spending a lot of time working on fighting robot things and haven't been doing too much 3D printing.

Hi, After a lot of messing about I discovered at some point I must have changed the angle of the hairspring so I made an adjustable version so I could fine tune and also added another gear to slow the speed down, now runs really nicely with a good ticking sound. also changed the design of the bowl to use less filament and show the main gear - https://youtu.be/QKhcTPreu-w

Thought you might like to see my motorised version - https://youtu.be/Piv1TepBqBk made on my own design printer. I changed the design overall to include bearings on every moving part to enable a smooth a movement as possible. The DC motor is geared driven by a 2S LifE battery. I have a large intermediate gear driving the Tourbillon, speed isnt quite right matching the frequency of the fork and balance spring yet so you can see some snatching of the balance wheel, needs to be a little slower but I may experiment with a lighter balance spring yet

This is cool! You should upload the parts as a remix so others can use it :) I like the single bowl, and the other parts are super art deco looking.

It seems like you're using a thicker spring? That could be what's effecting the ticking. Can you get the escapement to tick by hand when it's disassembled? It seems to be slipping. This could be because it's too much torque from the motor, but a stiffer spring could do it too.

PS-LiFe batteries are cool... though I always seem to light them on fire. They're supposed to be super safe... but I've never lit a regular LiPo on fire and I've lit 3 LiFe on fire...

Thanks Adam, yes I tried to make it look a little art deco, its actually designed to be fastened to a wall the covers slide off to allow fixing in place.

I think the gearing is still a little too fast so its hard for the escapement to keep up, I will try a lighter spring see if that helps, I did quite a bit of experimenting with different weight springs to see the result but still not clear in my head how it actually works :)

Thanks again

I am working on this, though it will be a while before I have time to print all the parts. One problem I find is with the spring. Using Cura, it slices so that one of the screw holes is not fully surrounded by material, like this: https://photos.app.goo.gl/6fgGqbAo87DxpZMy2. This is true with various shell settings. I am using an ugly hack where I edited the STL in blender to add a little more material, but I wonder if you have a better solution.

Huh. That's weird. Doesn't really matter, it should still assemble. I don't think it was doing that in Simplify3D though :/

Nope, works with a weight.

The 4mm bearings that you have in your description linked me to these
and they were the complete wrong size bearing.What is the dimensions for the larger bearing?

That's odd, that's the correct part. It's a 4mm ID, 8mm OD, and 3mm width.

Hi ... The assembly files are combined as one part - how to un-combine so i can edit individual parts?

What software are you using to open it? It should open as a multi-body part so that you can isolate the part/body you'd like to modify.

I'm using Solidworks 2016 - When I open any of the files - I cannot separate them - they are fused together as one part. Do you have the full solidworks assembly with mates etc? That would be helpful. Or do you have any other suggestions?

This is when opening all of the step/x_t/sldprt files? This doesn't happen in my copy of 2016. Please note that it is not uploaded as an assembly, but rater as a multibody-part. You can isolate the body you want and put it into a new part on its own using this command:

Comments deleted.

NOTE!!! Not 100% 3d printable! (Despite the tag as being "3dprintable") Ya might want to change that Mr. Designer?

What do you mean? Many people have successfully printed this.

I guess he's just complaining about having to get metal parts.

He means he's a drooling idiot who can't read. It's fine.

While trying to get this to tick, I realized my probelm was that my 4mm rods came it at 3.98mm however the inner diameter of the bearings were 4.15mm. This is causing why wheel to slip along the rod bumping into the raised fork bearing support on the tourbillon cage. Any suggestions or is getting new bearings my best bet? I also thought about gluing the rod directly to the internal bearing..... I measured all of the bearings I have and they all are 4.15-4.16mm. The 4mm rods were right on spec.

Weird, were they the same bearings I recommended or a different brand? You can try loctite 680 but that's a lot of slop.


My solution was to print a small insert for the tourbillon cage. This kept it from moving along the rod and had just enough contact with the bearing to rotate freely. Everything is ticking well now.

Thanks for this - I had the same problem and used your solution.

Cool. Got any images? Sounds interesting.

I added a couple of photos to my make section.

I must have something wrong. After I install the hairspring, then flip the balance wheel over to install the impulse, it seems to be out, 90 degrees clockwise from what you show in the video. I double checked the file to make sure I didnt get a mirror somehow. I put a pic in the "I made one" section to illustrate. Any ideas?

It's hard to see but I would try flipping it it upside down. The spring should be inset into the wheel, while the impulse protrudes off the side of the wheel which was in contact with the bed of the printer. I think you've assembled it upside down.

Yes it fits correctly that way, just off orientation from what I thought I was seeing in the vid.
Oh yeah, now it tics!! Even tocs a bit too. Thank you!

Comments deleted.

OK so I'm in the process of making one of these and want to point out a few hurtles. If your printer is printing tighter than intended holes, slightly out of round circles, or otherwise imperfect prints, you'll need to make some adjustments. I discovered that my printer was making slightly oval circles due to a firmware issue so if you notice some kind of choppy movement from the bearings, they are likely being pinched along one direction causing a preferential orientation of the inner portion. While un-snugging the holes with drill bits i concluded that either Adam is very skilled at squaring off parts and hammering at them, or more likely as seen with my wire stapling abilities, i just am just terrible at it. It seems that just giving a whack at the bearings causes them to mash their way into an improperly oriented position within the groove. To circumvent this issue I made the holes ever so slightly larger so that I could nearly press the bearings in by hand, then placed the jig (squared to the surface) upon the bearing and GENTLY tapped it in with a much smaller hammer. The hammer i used is from a watch kit and weighs about 50g so basically you can use the handle of a screw driver if you don't have a watch kit hammer. The bearings have a tendency to cause a wobble for some reason and when/if this happens you'll see it once you get the dowel into the bearings of the balance wheel and give it a spin. if you do see this you need to find the side of the balance that wobbles closest to the O, orient that near the wheel as you match the two up and watch to make sure you dont end up tapping the balance against the O. If this rubs there will be too much friction for the movement to work. I'm not sure exactly why but if its not squaring up and youre getting a bunch of wobble, it got better for me when i took it apart and put it back together more gently. it helps to have a small punch and an anvil or metal plate with a hole in it to punch the rods and bearings back out. Oh and ive also noticed that the center portion of my pallet fork and the rounded portion on the opening of the impulse head-butt EVER SO SLIGHTLY on the returning motion. originally i thought this was a problem but it doesnt seem to be the cause of any issues now.

Great points. I tried to accommodate printing tolerances as much as possible but there are always issues. Someone else once mentioned using a vice instead of a hammer, and this should certainly help if you're having issues with alignment.

I had access to an arbor press and that was the best thing for getting things to fit together.

Comments deleted.

I love this but I am having trouble getting the escapement wheel to sit square to the cage. I have printed several with different settings in S3D but each one sits skewed on the shaft due to bearing misalignment. Can you offer some advice for printing this part? All the other parts fit beautifully. Thanks for giving us this mechanism, I am looking forward to completing my own.

hey Tom, I'm having similar problems and I think it may be due to a combination of imperfections in the way the printer makes the hole, plus not getting the bearings to drive in square. I have an idea to try and can let you know how it works out. I plan to first ream the hole that the bearing sits in slightly with a drill bit. Then I will place the 2 bearings onto a rod, slide a block (similar to the jig but with a hole clear through) over the rod and drive the bearings down while being braced/aligned

I look forward to seeing your results, I've tried drilling it out and it still skews over. I noticed that the shoulder at the top of the bore that the bearings fit into doesnt print very well even with supports. I think this is stopping the bearings from seating properly. I want to have a go at drawing up my own part for this but the laptop my 3D drawing software on has decided to stop working.

ok so after carefully reaming out the holes and ditching the full sized hammer for a watch hammer, i was able to mostly hand press the parts into place with some MINOR coercion from a small hammer. i think walloping on those tiny bearings with a claw hammer may have been a bad idea from the get. Im going to post more above. i also punched the 4mm rod into and out of the gear on the outside of the cage like 3-4 times forcing it at more of a proper/square angle

Comments deleted.

Can you upload a photo to Imgur? My first suggestion would just be to try and force it to be square, just bend it with your fingers until it fits. I could possibly give better advice with a photo. I think all I did was force it by hand until it was square-enough. Also make sure the bearings all fully-inserted into the wheel. vice/hammer them fully inside and they should be pretty square. The alignment between the parts of the escapement is the most annoying to get right.

This is one of the coolest things I have seen. :)

Hey adam

im trying to edit the fork and im having problem using the solidwork file you posted (the fork is part 23 in the assembly)

is there a way you could repost a better version or is there a special way to save the #23 out of the assembly?

Hmm. Is this happening in the solidworks, and the parasolid, and the step? What exactly is the issue you're running into?

i load the "Clockwerk-1_0.SLDPRT" into solidworks and i get a question if i want to do feature recognition like it was an STL file

after i press no, i get a list of like uneditable parts... i can right click and press "edit feature" but it asks for an aditional file...

Aaa, ok. Yes, I didn't upload the full history. It's really, really messy. If you tried to edit any of the features, it would just explode. One of those models where you didn't know what your making until you made it, and kept adding on changes, and it just gets ungainly.

I've included the parts as "dumb" files so if you need to change anything you can just build off of the provided parts using "move face" tools and such, or just extruding new bosses, cutting bigger holes, etc. Basically just add features onto the end. Definitely don't do feature recognition unless you want to freeze your computer. But you can't just right click and edit a feature unfortunately.

Congratulations, nice desing!!!

What is BPH (beats per hour) of the escapement?

I never timed it, since it was designed to be an exhibition piece/sculpture that shows off the mechanism, and not directly tell time. It should be fairly consistent, but each one will be a bit different depending on how it's assembled. Theoretically, you can adjust the number of screws and their depth in the balance wheel to get a nice even BPH of some sort. I'm not sure how well it will work over time though, as the entire sculpture is not perfectly balanced in every point of the 3 axis as it rotates, so it may slow down/speed up as it goes.

what is the relations between the rotation of the outer and the inner part?

I wonder if it would be possible to connect a motor to this and make it tell time accurately?

If the middle part makes 12 rotations for every rotation around the circumference it would be optimal.

The easiest way to make it tell time is to have marks on the wall where the counterweight falls for every minute, and then use it as a timer. It only runs for a few minutes (without a spring or motor modification) so anything more is a bit unnecessary I think. Feel free to do a remix though, I'd love to see it! All the gears were sized for easy of assembly and making sure it didn't bind up. Adding too much reduction between stages of the tourbillon would require too much counterweight.

To reduce the cost of the mechanism by about $13 Australian, why don't you make the centre axle 3mm in diameter (and the other 4mm short axle 3mm diameter) and use up the bearings from the Team Losi Micro pack that we are throwing away? that way you only need one pack of bearings and you can dispense with the need to buy the Mini bearing pack.

Nice design by the way :o) my copy of it is almost complete.

I considered doing that, but honestly by that point I had poured so much time into it, I just needed to be finished. Also, I was unsure if that might be too-small of an axle and bend during assembly. Probably it wouldn't, but I knew that the larger axle definitely worked.

Is there a tutorial on how to create each part, not just download them pre made. Id love to go through making each part to practice with my program, and have an actual thing to print at the end. Thanks!

Do you mean how to make them in CAD? No, I don't have a tutorial for that, that would be quite a long one.

Is it possible to print this in ABS or does it rely on the material properties of PLA?

I would imagine the shrinking ABS undergoes would make it very hard to get all the tolerances correct for the quite precise assembly. But maybe it would all shrink by the same amount? Not sure.

I've had pretty good success with getting good dimensional accuracy with ABS despite shrinking (I've done a lot of 3D printing rocket parts) so if that's the only issue I should be fine after a bit of filing.

I think ABS should be fine. The only question is the spring. I know my polymax spring lasts for a long time, I never had one fail. ABS should be as durable, and then it becomes a matter of stiffness. I'm not sure if it would tick faster or slower... but in the end that shouldn't matter either since if you're using it as a timer you could just time it with the different springrate and never know the difference.

They are nolonger selling the 2mm shaft on amazon, got any other places to get some?

Hi McCree. Any 2mm shaft should work. If you search "2mm shaft" or "2mm rod" you should be able to find alternatives on amazon, though I haven't tested any others I can't see why there would be any issue.

Make uploaded. Thanks for all your help Adam.

Hi again. Have the central part ticking really well - so nice :) My problem has actually come in the final gear on the radius of the bowl. It seems slightly too large and forces up the central part to get the two screws to mate. Its shank also seems slightly too long? About to reprint, just wondered if it was something anyone else had noted? PLA, 0.2mm.


Ultimately creates too much friction and puts great force on the two screws holding the outermost gear to the central timepiece.

That's quite odd. The gear is purposefully a bit large, so that it creates an interference fit which prevents it from slipping. It uses the "U" component as a spring to hold it in place, basically. So that's normal. Otherwise the teeth pop out and skip.

I'm not sure why it's popping out of the bushing like that though. Perhaps just too much interference on yours? Is it screwed in all the way? Did it pop out? It may be easier to assemble that gear before assembling the entire tourbillon cage into the bowl.

Thanks for the reply! I did notice it was a sort of a spring and acts to tension the U as you say. However mine starts to rotate and despite the screws being tight it starts to bow under the pressure, angles out and eventually pops off the bowl rail. Will try assembling prior to putting in the central cages. Possibly taking some material off the lug on the underside of the U opposite to this gear might allow more freedom?

Hmm. Weird. What material are you using? Perhaps it's too soft for the screws to get a good bite.

Sanding down the skid on the underside of the U will definitely lessen the spring force, give it a shot!

Just standard PLA. The screws tighten very well, it is just that the force is quite large as it rotates - seems beyond what is intended for the spring force on the U to my eye. Shall continue to potter :) Trying a gear reprint and will then try sanding the skid a bit.

Add: I personally love these challenge prints! I know some people who get frustrated if it does not work instantly.... nah! :) This print is topping the airsoft gun one I successfully made some time ago for difficulty!

http://goo.gl/yWyJW8 <- anyone who likes the challenge!

Hello! Great project !!
Would it possible to split the bottom piece so that we have 4 small parts ?
My printer prints slowly ;)

lr109tum created a version with 3 pieces here:


Remix: Clockwerk by Adam Wrigley

I have a simple clock that I made with a physics set (Thames and Kosmos? I might be wrong...), but it doesn't work no matter what angle I put the pendulum arm at to the gears. Maybe it's the shape of the gears? They're not shaped like the one in that swiss lever escapement video; They're just regular gears. The crown gear keeps slipping, and when that happens the counter-weight juts down and sometimes even comes off completely! Sometimes there's not enough force to move the pendulum back into position, so it just stops swinging. I've been trying to fix these problems for years, yet the only solution I've found is using a heavier counter-weight (which ironically causes more problems, like making the counter-weight fall off, for instance).
I assume that you are learned in pendulum clocks and how they work, so I hope you may be able to help.

P.S. Awesome, Amazing, Astounding... I need a thesaurus to describe this build!

I looked up that kit and found a youtube video. Very odd. Even in the video it barely works. The 'gear' you speak of is the escape wheel. It's extremely important, and as you note completely missing in that clock kit. My solution is unfortunate, I think you need a new clock kit.

Here is a nice pendulum escapement model:

This is the video I found:

Simple Graham escapement
by Dynasus

Awesome! That is the best mechanism model that I've eve seen before!
But I wonder, why my 3dprinted doesnt "soft" as posible ?
I using PLA fillament ( sợi nhựa in 3D ABS PLA Nhựa in 3D ABS PLA http://blogin3d.com/vat-lieu-cho-may-in-3d-soi-nhua-in-abs-pla-nhieu-mau-dep-gia-re

I'm not sure what you're asking?

Great design Adam, making now :) Just a note concerning the bearing requirements. The parts image and video only show and mention ONE of the smaller 5mm bearings but I believe two are actually required for each end of the 2mm shaft, correct? At least this is what the Solidworks shows. Just for those ordering the bearings individually locally, rather than as a set.

You can use 2 or 1, up to you. It's designed to use 2 but I only use 1 just because it's a bit easier to assemble/disassemble (which I had to do a lot in the process of designing it) and it worked fine without it.

Thanks for the response. It did feel like it would be a pig to disassemble as I put it together with 2 :) Also the video seems to show a circle part where the bearing housing is a through hole rather than just a recess in the parts files? Just curious, realise this probably had many many iterations :)

Just an earlier rev. :)

I´m using two of the small bearing in my clockwerk. And it works very well.

How long does it thick ?

This depends on the height that you mount it on your wall, but it runs for about 5-10 minutes usually. If you mounted it on the side of a building it could run a lot longer, ha.

cool thanks for reply

Bearing spec: 8 x 4 x 3 mm (ext, int, width) and 5 x 2 x 2,5 mm (ext, int, width).

I see you made this model a 3D printer to try and shed jobs.
Print the model, but you do not know the assembly sequence assembly.
Can you ever tell the order in which the Assembly?

Hello! May I have the bearings specs (ext/int diameter and width)?

Bearing spec: 8 x 4 x 3 mm (ext, int, width) and 5 x 2 x 2,5 mm (ext, int, width).

Hey, can you tell me the required diameters for the weights? I cant find 2,5lb here in germany :( Can I also use 2x 1,25kg?

Anything heavy you can attach to a string is fine. I used standard olympic weights designed for a 2" bar, but if you don't use my weight covers (which are purely aesthetic) then you can use whatever you want.


And to give it away for the benefit of the maker community, I tip my hat to
you Sir. May others learn from your brilliant example!

I can't print circles well enough in range for this too work. Totally bummed out right now can't stand my printer.

what printer do you have? cant you resize all?

I have the M3D. It's not the size of the prints but the circle , sun as the wheel1-1 literally prints like a warped square. I've been trying to get this backlash corrected and I'm close to trying again on making this work.

Hi, can't I use a high torque motor to replace the 5lbs barbells?

I was thinking of using motor if that's possible

Awesome stuff here! Cant wait to print it out

This looks surprisingly easy to print! Great job!

Thanks! I had a working version a month or so earlier that required a lot of support to print, but figured it would be much more useful to others if I redesigned it to eliminate that requirement.

I thought I was the only one who stared at watch mechanisms for long periods of time ;)

I would love to add some bearings, posts, and perhaps make another decorative shell. I noticed that a .part file made it into the archive. Are the 3d design/model/part files available because I would love to work with them.

Thanks again.

There are literally dozens of us!

I've uploaded a few different formats of CAD so anyone can make edits.

Thanks for the reply. and sorry to bother you.

I have tried using the CAD files to no avail (I am using SolidWorks2015 premium BTW). I can get the parasolid file to import, but since it is a large assembly, there are broken faces in the import diagnostics, and there are overlapping solids reported (improper mates perhaps, but dunno), any feature recognition, let alone editing, does not work. The files appear to be unusable (to me at least) for editing purposes as they all break, so if there is another way (individual part STEP or parasolid files perhaps?) it would be greatly appreciated.

I am eager to get started and congratulate you on a functional design.

Thanks again.

Hmm.Well the solidworks file I uploaded is a 2016 version. I'm not sure why the parasolid wouldn't be importing well though. Are the parts broken or just things like the screws? Have you used the auto-fix features in the import diagnostics? Those usually work well. The overlapping solids are most likely the screws, which overlap on purpose. I definitely wouldn't use the feature editing, that never works. I would just build on top of the existing solids by adding features, or potentially using things like move face, etc.

Oh well. I was hoping to use the Solidworks stuff so all of the controlling features would be there, but I may just have to wait until my software gets upgraded :/ I really want to keep/make the design parameter driven, such as number of teeth, so I can control/adapt timing as necessary through my modification process.

Thanks again for your time and your help.

I actually haven't uploaded any of the CAD with a feature tree in tact. It's just a complete mess and half broken. Most of the gear teeth are actually imported as sketches from an external gear drawing program anyway, so are not easily changed that way. I've never had much luck doing parameter driven patterns in solidworks like that, it just makes things explode.

Ahhh, I am not missing out then ;)

I will build the files up then. It should be fun either way.

Thanks again.

My PrintrBot unable to print largest parts, print bed max 150x150mm. If i'll scale everything on 0.8 it will work?
Sure bearings and shafts sizes also will be smaller.

If you could find 0.8 scaled bearings then I don't see why it wouldn't work, but I don't know where you could find such a thing.

I assume the biggest parts are the bowls and bowl supports right? If you print the bowl in quarters and supports in 3 parts instead of two, do you think that would reduce the overall size enough to support smaller printers?

Haha, Watch the Fork. Print version1.1 or it won't work.

Edit: When using Slic3r make sure to flip( it gets mirrored) the parts the right way or they will print reverse. To be sure just use Aaaddaamm's ver 1.1.

What was wrong with 1.0? I guess it doesn't matter since 1.1 works, just wondering.

On comparing, the middle prong of the Fork is on the wrong plane in V1.0. preventing the flywheel from oscillating. At least thats what Sclic3r is coming up with. Could also be a bug in slicer as I flipped the part around the fork part stayed put, very odd. My apologies if they're the same in your cad file.

BTW, awesome project !! Thanks.

Interesting. Solidworks is terrible at exporting STL files so perhaps that was the issue. Glad the 1.1 version I exported from simplify3d fixed it.

On further invesitgating, its gotta do with Slicer. when you flip a part in XorY it mirrors it (something I learnt). So in version 1.0 the fork is standing , I flipped it in Z and the in X to get it good for printing and of course, it had to be the wrong way. In version 1.1 you already have it the right way ready to print.

What is powering it? A pla spring?

And, what is it's purpose? Does it tell time somehow? Or is it more of a conversaional piece / something to be prideful of? ;D

It's powered by a 5 lb counterweight which hangs below it. It doesn't directly tell time, but if you put marks on the counterweight's string it can tell time intervals (basically a fancy egg timer). It's mostly a kinetic sculpture.

I nearly finished one but have a problem with assembling the ticking part. In the movie I can't see what is exactly happening during the very first assy stage. A detailed picture would be very welcome. And a text that describes the process part by part in the right order.

Have you watched the videos that I linked which explain swiss lever escapements?

Clockwerk? Do you mean, Clockwork? And btw, instead of making this, I am going to make one with a combination of my own 3d design skill and hand made parts, made by me. Mainly b/c yours looks simple and I want something more complex, no offence, this is really nice.

Look foward to your masterpiece...

Actually it's a portmanteau of "clockwork" and "werk." Werk is sometimes used to refer to things that are really awesome, and I was being a bit facetious. I also wanted something that would be easily remembered and searchable but it turns out there is a videogame character with the same spelling, oh well.

I look forward to seeing your remix!

Ah ok, but it wont be a remix as parts will have been hand made in real life, things like carved wood and stuff.

Incredible - genius

I added this to my ‘things to make’ collection, even though that's probaly wishful thinking.

some day i will make this, its amazing.

whoa... that's awesome!

Do you have to mirror the bowl when printing the other half? I ask because of the half tooth at one end.

Oh shoot. You're totally right. My bad. I thought they were the same when I was uploading parts.

edit: Added a new component "Bowl-Top_2" which is correct. If you mirrored the existing bowl it should be fine.

Hey Adam!

Great job! I want to make this clock and I am printing the parts right now. Can you tell me the parameters of the bearings? Inner, outer diameter, width, quantity? Probably I won't order it from amazon, I am planing to purchase it in a local store.

So exactly how would i make this i would 3D print it all then assembly it but isn't there a motor inside somewhere or something?!?!?!

Hi Adam,

A great design! I'm trying to print the parts. I had some problems printing the spring (Spring_1_0.stl). It appears as if the bottom layer is partially missing (I'm using Cura for slicing). Are there any special settings for printing the spring?
Thank you,

I just got around to re-uploaded the STLs from simplify3D. Maybe try again and see if this helps?

Thank you. I found the problem in my Cura settings. Both Spring_1_0.stl and Spring_1_1.stl print ok.

Thank you. I found the problem in my Cura settings. Both Spring_1_0.stl and Spring_1_1.stl print ok.


I'm not sure what would be wrong, there is nothing special about that part. I'm hoping to get some new parts uploaded tomorrow in the correct orientation/exported from simplify3d. Perhaps the reupload will fix it? Solidworks has a habit of making bad STLs.

nice work :)


Just going to throw a few things out there.

First of all, amazing job. I gave thought of making something like this a while back, but I didn't have the time to sit down and do the CAD work.

Second, What is the approximate diameter on the large bowl? I have a very quick little homebrew delta machine which only has a 180mm diameter working volume, so I might have to find someone with a bigger printer for those parts.

Lastly, the chamfer idea is a good suggestion. Not only does it help with the elephant foot problem, it also helps aesthetically.

Now, for my contribution. Provided I can build one, I'll do away with the weights and make it motorized. The idea is somewhat easy, you have a 3d-printed spring wheel with teeth on the outer edge which engage the teeth powered by a DC motor and the hub in the center attached where the counterweight would go. You place a contact switch somewhere along the spring so as it unwinds the DC motor re-winds it.

Thanks! and, oof... it's actually 181mm in diameter.

Yeah the motor/spring combo would be cool, I'm just not sure if a 3D printed main spring will have the necessary torque. It needs around 12.5 in-lbs.

the irony is sickening!

@base7 - feel for you, buddy.. :'(

Crap. :(

I'll just have to cut the bowl into thirds or quarters then, depending on the tooth count. I was going to remix your models anyway to add bevels and improve the holes. I never can get small holes to print properly unless I reduce them to triangular holes and then drill them out as needed.

I figured I would use a gear train to increase torque from the revised mainspring so the spring wouldn't have to be providing 12.5 in-lbs directly. ;)

Also, it might be worth using a block and tackle setup to get more distance out of the falling weight and increase run time.

If you shrink the pulley and use more weight you can get a lot more run time. I just didn't want to be hanging so much weight :/

Already remixing this. XD

I'm going to add time indication in the most ghetto way possible... i'll keep hours, minutes, and seconds on each of the axes by changing up the gear ratios. It looks OK in blender, no telling if it will work once I get it to print.

Hi Adam,

Amazing job!

Wanted to ask, my printer bed is not big enough to support some of these prints. Provided I found the right screws, will the mechanism still work if the size was scaled down, half for instance?


The big issue there is that this is built around specific bearings, so they won't fit if you scale it down. You'd also need a 0.2mm diameter nozzle and 0.1mm layers. Potentially the bowl could be cut into quarters instead of halves instead. But you might need to add more screw holes.

Wow, an awesome job. Well done.

You said you didn't know how to fix the orientation so I figured I'd let you know. It's pretty straight forward. First add a coordinate system via Reference Geometry -> Coordinate system. If the model is in the correct printing orientation when you're looking at it in the default isometric view (ctrl+7) you would click in the Z-axis box and then select the top layer as the reference and hit ok. If it's in some other orientation you'd have to change that accordingly.

Now as you go to save as STL you click the Options-button in the save dialog and change the Output coordinate system at the bottom to the one you just created.

And if I may give a small piece of advice for the design it is to give all the platform facing faces a chamfer of 0.5-1mm, it makes a huge difference in print quality IMHO. It helps get rid of the elephant's foot and makes parts fit together much better (especially nice with meshing gears for example). It can also help if you want a fillet towards the platform. See example here: http://support.3dverkstan.se/article/30-getting-better-prints#chamfers

I just got around to reorienting everything, but I did it via simplify3D. Solidworks is always awful at exporting STLs anyway, tons of errors.

Thanks for the tips! I'll take a look next week.

This is amazing! Looking forward to seeing more of your works!

Im just wondering the how that mechanism works. I would love to built that one but i have a 2 question for you.Is there any way to motorize that system.Im thinking about using dc motor instead of weight reel at the back. Also is there any way to speed that mechanism up ? Really nice build i love it

Thanks, glad you like it!

The movement should work with anything that can deliver the ~12.5 in-lbs of torque needed on the shaft. The torque should be consistent, or it could run oddly. Hence the counterweight being a good bet. However motors or barrel springs will work, if you find the right ones.

Speeding up the mechanism could be achieved in two ways. The ticking is a spring balance, so you can make it faster by lightening the balance or strengthening the spring. The balance wheel itself is a screw balance, and in the video I have all 6 screws in place. If you remove some it will go faster. The spring would need to be redesigned in CAD, or could be printed with a material stiffer than PLA (perhaps a carbon filament?).


Its good to know I will start thinkering about connecting with motor and speeding up soon thanks for reply

a good way to motorize it would be to have a barrel spring and have a dc geared motor to wind the spring at a fixed interval or a motor to rewind the counterweight. . Motorizing in any other way is just counter intuitive to the whole simplicity of the mechanism IMHO.

Great work though.. Has me thinking on a motorized version without killing the simplicity

Very cool, how long does it last after winding?

Thanks! The time it runs depends on how high you mount it. It runs with a counterweight, and my current setup is about 5 minutes per meter of height. You could slow it down with more weight on the balance wheel.