MakerBot Print is our newest print-prepration software, which supports native CAD files and STL assemblies,
allows you to interact with all your printers via the Cloud, and many other exciting new features.

Download Now


Watch Escapement Desk Toy

by LarkysPrints Jan 18, 2019
Download All Files

Thing Apps Enabled

Please Login to Comment

I can't tell from drawing 1/4 if the Ratchet Spring pin goes under the Barrel Ratchet or on top of it.

I have mine working, but when I wind the mainspring tighter, it runs a lot faster than it should. I must be doing something wrong. Any help is appreciated.

Just curious where are you guys are getting your hardware? I spent like $15 buying each piece individually at Ace hardware.

Works great with the 6mm hair spring remix at 100% with PETG. I had no luck at all with the standard spring.

Having trouble getting the pallet fork to move with the rest of the machine. When I take it out, all the gears spin just fine. I have tried to adjust the pallet fork but it won't move when it is touching the escape wheel, causing the whole machine to lock up. It seems like the fork is too far in to move back and let the gear advance, but I can't seem to adjust the screws to fix that. Any suggestions on how I should adjust? Is there a video somewhere of on how to adjust the pallet fork?

EDIT: So it seems like I got the pallet fork dialed in, but in order for the whole thing to work I have to rock the balance wheel back and forth. I have only printed the files that was originally given, but I will be printing the remixes soon. Is the issue a lack of weight from the balance wheel, or do I have to tension the hairspring to get it going?

All brackets holding the wheels must mount in a planar fashion from my experience. If the pallet fork mount causes the teeth to be ever so slightly too close to the escape wheel, it may not work. At least one cause for non planarity is due to elephants foot. I have posted about this here. I have built two of these and both worked without modification

No elephants foot, but the 3 sections do not lay completely flat. By loosening the screws, it flattens out but then the parts aren't nearly as stable. When I flatten it, it adds a bit more time when the spring is tightened up but as the spring unwinds, the pallet fork eventually gets stuck on a tooth. I have printed the modifications and the larger spring which did give more time to the balance wheel. I plan on reprinting the balance wheel, how heavy should it be? (I am using a filament similar to NylonX, current weight is at 33g) Any suggestions?

Try this. Wind it up. Then place it on a flat surface, and holding both ends, force it to lie flat. See if it will work.

So I reprinted the balance to be heavier, still no fix. When I wind up the spring, it runs just fine for about 10 seconds then the balance wheel needs to be turned. It stops because the pallet fork gets caught on the teeth of the gear, and doesn't have enough energy to move. If I spin the driver gear with my hand at a consistent force, the machine works just fine. This leads me to believe that it is something wrong with the driver spring because everything works consistently without using the spring as the driving force. This is all happening while the parts are flat, but the screws from the brackets to the individual parts are all loose in order for them to lay flat or else they bow upwards. Any recommendations? I can send you a video of the issues so that you can get a better idea of the issue.

so I was thinking about making a gramophone would you mind if I used your main spring design?

He works great.

Still working on keeping it ticking more than a few seconds..... Didn't know it could be such a delicate mechanism ... Very fun! Thanks to @LarkysPrints

Nice print. PEPE SMYTH :D How many hours di it take?
Adding a motor would be a good solution I guess, unless you find a proper metal spiral spring like those on pendulums.

does not work i have done al i can and it will not go on like in the video have to help it a little and then it stop again

You have a weak spring.
Print a new one with PETG with 70% infil.

Wow! what a complete post! Drawings, etc are a knockout. I cannot believe you posted this at no charge, your time is worth quite a bit. Please keep up this excellent work! It really is appreciated.

Hello. I made the impression but the mechanism did not work. is there an update? seems to me to be problems in the model. the fork pallet seems to be misaligned. I await help. thank you

https://www.thingiverse.com/thing:3464466 I printed this remix get about 11 minutes run time. Look at my make it has a couple issues but the run time is great!

Watch Escapement Movement - Steel Spring
by phoney1

I printed this on my Monoprice Mini V2. Everything fit at 100%, except the spring barrel, which I had to print a 99%, but it works fine. I did need to change my bed size to 124mm x 124mm to get it to work though. I also had to turn off skirts for a few of the larger pieces because I didn't have the room.

Mine went together fine. I couldn't find any M3x25mm screws locally, so I used a #4-40 x 1" screw and nut that I got from Lowes. Lowes also didn't have any M4x16mm screws for the base, so I used 20mm screws instead. Lowes does stock those screws, but both stores I went to were out. Other than that, Lowes had all of the fasteners I needed. If you want to save a few bucks, just use longer screws where you can. I used a 40mm screw instead of a 25mm screw to go through the hairspring assembly and it works fine, just sticks out a little bit on the end.

Mine worked fine almost right away with very little tweaking. The only issue I had was that the brackets that connect the bases were not level for some reason, so the one end where the hairspring assembly is was lifted up a mm or so. When I manually held it down to the table and made it perfectly flat, it ticked perfectly. I think I had a slight elephant's foot on each of the brackets that was causing the issue. I just filed down the bracket where each base piece sits and it seemed to fix the problem. I get about 2 minutes run time, which is plenty, IMO.

I did print the following remixed piece though, which I think helped:

6mm hairspring:

V2 Train Frame Front:

Here's a video:


6mm tall HairSpring for Watch Escapement Desk Toy
Watch Escapement TrainFrameFront V2
by quartit

Do you think it can work with 3D printed fasteners?

Comments deleted.

Best thing I've printed yet! Great detailed design. Well done! Printed and fitted perfectly. Getting around 3 mins run time, no modifications. I used PETG which avoided any warping issues that may be responsible for the problems some are having. Like others, I removed the adjusting bolts on "balance frame front" as I don't see the need. The pallet fork's travel is limited by the escape wheel. It looks cleaner without them also.

Absolutely amazing design! well done. I work in an engineering workshop and we want to make this out of aluminium entirely (excluding springs etc). would it be possibly for you to provide me with the solid parts so that i can process them into my cnc machines for machining?
i would love to be able to make this in this way. Thanks.

You should make a gear system powered by weights that'll run for a few years

Great design!!! Supper detail on assembly. Once printed, everything went together so clean! Awesome !!!
Nice modular design. Any plans on adding other parts to make a complete clock?

Cant get mine to work. Every gear is loose and spinning freely including the balance wheel. The hair spring was printed at 100% infill with the remixed file (6mm). It doesnt even start.. is there any trick to tune the impact pin screws?

Thanks for the help guys. I got it working by adjusting the impact pin a little better =)

Print looks nice! It's gotta work. First thing I reckon is to remove the adjusting bolts from "balance frame front". The pallet fork's travel is limited by the escape wheel. The impact pin housing adjustment is out by the looks. Mine was close to centre of adjustment, and the travel of the impact pin in the pallet fork on mine is far from symmetrical.
Also note that the locknut on geartrain is very close to the escape wheel. i had to remove the washer to leave clearance.

Try adding a bit of thin machine oil to all contact surfaces. Though plastic seems quite slippery, in fact there's quite a lot of friction drag, and a thin coat of oil can overcome it and bring your make to life. :) https://youtu.be/IYJtfdTx-T8

That is tricky. I had some issues as well. You have to tune the impact pin housing. I had to put a little tension on the hair spring and adjust that housing just right until it had enough spring back to lift the fork and return. A lot of trial and error. Deffinatly have to have the lock nuts loose enough to have free movement of the balance wheel.

Thanks for the tip guys. I will try it out =)

Excellent project. I used the stronger balance spring, assured the base was flat, and printed all major parts with 100% filled PLA. Works well. I'm confident more tuning could yield longer run times. Thank you!

I have added a remixed 6mm tall HairSpring: https://www.thingiverse.com/thing:3469862

Same 10mm overall height but the spring part is now 6mm instead of 2mm.

6mm tall HairSpring for Watch Escapement Desk Toy
Comments deleted.

What’s the purpose of the barrel ratchet/spring?

It locks the Barrel when applying winding force.

I've made some modifications to this Karl's excellent design in order to support a steel recoil spring from a chainsaw (<$10 on Amazon).
Run time is approx 20 minutes. A remix will follow shortly if there is interest.
Sorry about the long winding but I wanted to show the number of coils.


Wow. That's a lot of winding. I'm interested.

I beefed up the hairpin spring like many have suggested, but I still cant get it to work. Anyone have any suggestions?

A thin layer of machine oil on all contact surfaces.

Make sure you have the hairspring on in the correct orientation, many have made this mistake.
Remove any elephant foot from the drive train components.

I should say that its also much faster than the one shown in the video, and the spring doesn't flex anywhere near as much. The fork only barely catches enough because of this, and often slips.

The spring is in the correct orientation. I can get mine to work for a few seconds, but after that, the hairspring just sort of loses its power until I jump it off again by giving the balance wheel a quick nudge.

Is the Hairspring tightly securely fastened at the base with the nut so it can't slip?

Printed it and it doesn’t work. Very Is disconcerting considering the time spent printing. I’m printing the Hair Spring again but larger as per comments. One major flaw is that all the gears should have support from both side. The Gear Train gear exhibits a lot of torque from the main spring.

I believe there is a remix that provides the support to the gear train (https://www.thingiverse.com/thing:3399736) but I can tell you that is not the source of your issue. Post some pics and maybe we can offer some assistance.

Watch Escapement TrainFrameFront V2
by quartit

My n00b tips for getting it running - I now get 30s of runtime where before I got none:

Watch escapement desk toy, minor fixes
Watch Escapement TrainFrameFront V2
by quartit
6mm tall HairSpring for Watch Escapement Desk Toy

Why is the Spring Barrel file so large while the others are around 1-4 mb? I am guessing this is what causing 3 failed print attempts. It states it is at 20 percent within a few minutes which is not right and then it stops about 45 min. In mid print. Is anyone else running into issues printing the Spring Barrel?

So, i designed some rings that i printed at 100%, and put around the main balance wheel. Even got it to 6 minutes now.
I can remove the rings without dismantling anything.

And this is ring nr3

Did you print your mainspring at 100% or just the hairspring?

Only the hairspring

I've added the rings and got some improvement, although I couldn't quite fit the 3rd ring on without rubbing up against the bracket. I'm still only at 1 minute, though. What bearings did you use for the balance?

Cool Design - As others noted, I also had to add some thickness/height to the hair spring coil (~ 6 mm) to make the escapement work properly, but other than that all worked well. Neat project.

How are you increasing the thickness of the spring without increasing the thickness of the socket in the center?

I loaded the original STL file into 123D Design (could have used Tinkercad) and used Smart Scale to increase the height of the hair spring in the z-direction by 6 mm. I then inserted a block on top of the hub, moved it down 6 mm, and used the combine feature to subtract the cube from the hair spring hub to delete the unwanted 6 mm section of the hub. You can do the same thing in Tinkercad.

Can you share your design with the hairspring coil

Yes, see attached file. I printed it at 100% infill.

Comments deleted.

Thanks for the design. Worked on first try!
Very small adjustment needed with aligment, but it tick really nice.


Having "Bowing Up" problems. I tried to eliminate elephant feet by trimming the brackets so I don't think that's it. Any other advice to prevent bowing up? The center brackets do NOT lay flat, they are about 3mm off the table. If I push them down it runs great, but when I let go it bows up and stops.

I'm considering heating up the joints on the brackets and bending it a bit then letting it cool.

Awsoome! i am printing currently and i don't get the barrel rachet part, is it above the rachet spring and putting it in the gear? mayb a picture of it? Thank you

Look at your files...there are four assembly drawings. Review #1. It is all there in detail

Great design printed well, big issue is the balance just can't get it to rock no matter what even after adjusting screws etc etc did at one stage have it going for around 30 seconds - thanks

I printed it about a week ago and had problems keeping it running more than 4 or 5 clicks.
I beefed up the hair spring by a little bit (about 1 mm thicker) and it runs great now.

Thanks for the great design.

After printing everything, waiting on my hardware and looking at how to assembly this. On assembly sheet 3 of 4 it looks like screw 35 going though the impact pin housing doesn't go anywhere. Could you please tell what I'm not seeing in the drawing. Thanks.

It engages the pallet fork. Just install it as shown. You will be fine.

Thanks!! Printed the whole thing in PET-G, but had dto change some things to make it really work 100%
-Spring printed 100% infill in tough-pla
-Printed the modifications found in the remixes, also tough-pla
-Put a bearing in the balance rotator.
-Added weight to the balance rotator.
My personal record now is.... 4 minutes 59 seconds. When it stops, there are maybe 10 more clicks in the spring.
Thank for the design!!!

Amazing!! Having spent literally hours in adjustments, I still have only a maximum 2 min run time! I won't be able to sleep until I find out how you got 5 minutes! Could you tell us specifically how you weighted the balance wheel? Thanks! Dennis

can you please link the modication files in a reply or email to pharmnfranck@msn.com

I'm still curious if this can be scaled down 5% without effecting screws.

I just cant get mine to work. The Balance Wheel always stops after some spins. Does it have to be heavier? I printed it at 30% like in the instructions but it seems really light. I dont really get the concept of how it should just keep spinning when nothing is really driving it. Also it somehow doesnt align with the escape wheel that well although every print came out near perfect

Did you resolve this? If not, post pictures.

I kinda did by just sanding the edges of the pallet fork down to a curve, that way its a little faster than intended but acually gives my hair spring enough inertia to make a full round

will this work at 90% - 95% scale? I'm limited by 120mm build space and one part is a little over 124mm. can I print everything at 95% without issue? will i still be able to use the included list of screws at 95%?

I can't seem to get it to run for more than about 3 ticks. Can anyone help?

I've started testing from the left:

  • Mainspring assembly works well
  • Gear train assembly works well. Low friction, no binding.
  • But I can't seem to keep the balance moving. Just 2 or 3 times and then it dies out.

My understanding of the physics of it is that: as a tick happens, the escape wheel needs to impart some energy to the pallet fork, which will transfer it to the balance by hitting the impact pin. That makes the balance swing out. Then because it's on the hairspring, the balance will come back with (hopefully) enough energy left to trigger the next tick.

So somewhere in there, I'm losing energy. Either:

  • Escape wheel isn't giving enough to the pallet fork
  • Pallet fork isn't giving enough to the balance
  • Balance is losing too much during its swing out and back in
  • Triggering the pallet fork / escape wheel to make the next tick is requiring more energy than it should

I've tried printing the balance with high infill, and I've tried printing the hairspring with 100% infill. Any more ideas?

Well, I figured it out. Basically the entire problem was the hairspring not being strong enough. To fix it, in Simplify3D I scaled the Z axis to 200% to make the spring portion twice as thick. Then I had it stop printing 50% of the way through, so that the overall part wouldn't be any bigger. Once I did that, it stopped being finicky and just worked.

I'm going to give this a try as I think I'm having the same exact issue. My balance wheel just doesn't swing the amount of motion as the original video.

Here is my order from Bolt Depot (boltdepot.com) for all the hardware needed. It seems possible to get all this at a hardware store but they didn't have all the exact sizes and it was a pain, so ~$10 for the screws + $5 shipping gets you the exact sizes with plenty of extras.

Confirm Your Order
Quantity Description Unit Price Subtotal
5 pieces 17888 - Metric machine screws, Phillips pan head, Zinc plated class 4.8 steel, 4mm x 0.7mm x 20mm. Each $0.07 $0.35
5 pieces 17892 - Metric machine screws, Phillips pan head, Zinc plated class 4.8 steel, 4mm x 0.7mm x 40mm. Each $0.10 $0.50
5 pieces 17894 - Metric machine screws, Phillips pan head, Zinc plated class 4.8 steel, 4mm x 0.7mm x 50mm. Each $0.12 $0.60
5 pieces 17895 - Metric machine screws, Phillips pan head, Zinc plated class 4.8 steel, 4mm x 0.7mm x 60mm. Each $0.14 $0.70
5 pieces 17873 - Metric machine screws, Phillips pan head, Zinc plated class 4.8 steel, 3mm x 0.5mm x 25mm. Each $0.06 $0.30
5 pieces 17869 - Metric machine screws, Phillips pan head, Zinc plated class 4.8 steel, 3mm x 0.5mm x 12mm. Each $0.05 $0.25
20 pieces 17887 - Metric machine screws, Phillips pan head, Zinc plated class 4.8 steel, 4mm x 0.7mm x 16mm. Each $0.06 $1.20
1 bag of 100 4784 - Metric hex nuts, Zinc plated class 8.8 steel, 4mm x 0.7mm. Bag(100) $1.86 $1.86
1 bag of 100 4783 - Metric hex nuts, Zinc plated class 8.8 steel, 3mm x 0.5mm. Bag(100) $1.69 $1.69
5 pieces 4802 - Metric hex lock nuts nylon insert, Zinc plated class 8.8 steel, 4mm x 0.7mm. Each $0.06 $0.30
1 bag of 100 4525 - Metric flat washers, Zinc plated steel, 4mm. Bag(100) $1.29 $1.29
5 pieces 4524 - Metric flat washers, Zinc plated steel, 3mm. Each $0.05 $0.25
Edit Cart
Subtotal: $9.29
Tax: $0.00
Shipping: $4.95
Total: $14.24

My order from Bolt Depot is in and this worked out great. Everything is in sturdy little bags with great labels, this made assembly really easy.

Thanks a lot ! I just place my order !!! Thanks for all the part numbers ! It is super easy to order from them !!

You are a genius! Thank you!

Ich schaff es nicht das meine läuft, es scheint als würde die Hairspring nicht genug kraft haben (Gedruckt am Anycubic i3 Mega)
werde sie mal neu drucken, hat vl wer Settings für den druck wo es funktioniert?

Printed mit ECO-PLA from 3DJake, doesnt run.... i dont know why... :(
The Hairspring only does 2 spins and stops

I was unable to get mine running at first. I reprinted the hair spring with 100% infill after up scaling the thickness by 4% and now it runs perfectly for over a minute. I think the original hair spring (at least with the PLA I'm using) is a little too weak to consistently keep it running.

ball bearings upgrade, that is really super nice and easy to perform

I'm having trouble downloading the SpringBarrel.STL because it's such a large file. Any advice? And what makes that particular file so large?

would it be possible to share the cad files 3d please ?

I just calculated how much filament is required for this project: ~430 grams, 44 m. 33 hours to print all the parts.

Hi, I'm having difficulty getting the device to cycle automatically. If I provide minimal downforce on the hairspring or push slightly on the downstroke, it will make full cycles. With no help, it only makes half a cycle. It looks to be a hairpsring tension issue. How can I get more stored torque on that hairspring to power through a cycle? without the fork, the mainspring cycles through just fine, and the balance will bounce on its own a few times before resting. I have the balance pin at 9:00 when the balance is at rest. Fork max and minimum are slightly past where the fork needs to go to complete a cycle

My balance pin sits at about 8:45 (ish) when at rest. Check for friction at every interface.

Comments deleted.


was ich festgestellt habe, wenn ich die 12 Schrauben der 4 Füße locker mache, läuft die Uhr einwandfrei bis zum Schluss

As I was troubleshooting my build, to eliminate the possibility that the feet were the problem, I mounted all parts on a pair of 2020 aluminum extrusions and created spacers to keep the extrusions at the proper distance apart. I'll post a pic later. I'm attaching an STL of the spacer. Use either M4x8 or M3x8 bolts and T-nuts.

I discovered the same thing on my build. That is consistent with the information I gave you in my previous response.

thank you for your work i have printed it at 25% scale for testing and it's fine except for the axes

Really nice mechanism! I sadly can't get mine to run :(. Can anybody help me with this? See pics attached. Might it be because my gears dont run smooth enough or the balance wheel is too light? Any help would be appreciated.

Feel free to PM me...would be happy to help you troubleshoot it.

I looked at your pictures and it all looks fine. First test would be to remove the pallet fork. Then wind up the spring a little while holding the escape wheel. Then release the escape wheel. It should spin out rather rapidly. If it does not, then your gear train has a problem. Second, with the pallet fork removed, test the balance wheel. Does it oscillate back and forth freely when you give it a start? One thing I did notice is that the screws are set pretty tighly on the balance wheel frame front. To start, you can remove those screws...they are not critical at first. With the pallet fork replaced, does it rub against the impact pin housing? The closer you are to the balance wheel, the more sensitive you will be to friction. Last...when you set the thing on a planar surface, to all of the brackets lie flat on the surface? Does it bow upward? I found on my first printing that, due to elephants foot, the brackets where causing the whole structure to have a bow upward which significantly increased friction/binding in the gear train. It will work...it is a brilliantly done design.


Und was kann man machen wenn die Struktur einen Bogen macht bzw. Kippt?

In my case, the bend was caused by imperfect printing. I show with the two files, 1) how I printed the brackets and 2) side view drawing of bracket showing where elephant foot will cause a slight angle when attaching a frame element.

A marvelous working device, easy to make (except for finding the screws), runs beautifully for about 2-1/2 minutes with one winding. Congrats and thanks!

In Zeitlupe bestimmt, sonst glaube ich das nicht!!!

Ja, es ist toll zu sehen. Hypnotisch wirklich!

i printed mine with low infill, this means my balance wheel is very light weight, and it does not really run, feels like it need momentum to work, did any one add weight to the balance wheel ? how ? how much ?

I did not add weight, but I did have to print it twice. One of my printers produced a bit rougher finish on the gears, they didn't run as smoothly as it should. Printed the gearset on a better printer, that was the magic. I used 40% fill throughout, PLA. I did have to fiddle with the settings a bit to get the action centered and running free.

thanks, all my parts are printed 15% infill, 8 layer walls and top bot, so it is very light, all my parts are perfect and nice in surface, all parts fits and run smooth, just cant get the balance wheel to go tick tock as it is supposed to, think it is too low in weight, could be nice to know a wanted perfect weight in grams ?? how heavy is your balance wheel alone ?

My balance wheel weighs 34 grams. Printed @ 100% infill with Inland PLA. Impact pin,housing and all fasteners weigh an additional 8.55 grams. Assembled mass would be 42.55 grams.

A few things come to mind:

  1. Is your hairspring not "springy" enough. This is a critical element that works in conjunction with the balance wheel to bob back and forth, thus allowing the escapement to function.
  2. Are you sure there's no binding through the gears, causing them to not forward the mainspring force to the escape wheel. On mine the gears all looked good and they rotated freely individually. However they were a bit oversized, causing binding once assembled. I had to reprint them on a known-good 3d printer, that did the trick.
  3. According to Simplify3d 4.1.1 my balance wheel is about 36 grams. If I lower the fill from 50 to 30 percent it drops to about 31 grams. I use two solid outer layers, 2 bottom and 4 top.
  4. Any chance you put the hairspring on backwards or not centered properly? This could certainly have an effect.

I suggest you wind the mainspring then manually move the balance wheel back and forth. I'm confident that careful observation will elicit the problem. For example, if you give the balance wheel a push, does it start to work and then just peter out?

Good luck, you will get there.

mine is running super nice now, only a few minor things to mention see attached file, has to cut away 2mm marked with black on this red part, to make it run, it my make I explain more details and minor bugs found, my balance wheel alone weight 34grams, dont know what is designers best wish for that, and also no idea or methode to measure the springyness of the hairspring, also i have a problem with the main spring, the more i turn it, the less the clock want to run, it works best if I only turn it a little bit, if the main spring force is too big, it will not work.

it is bigger than i tought :p

I also made it. It worked perfectly. I was impressed.
That's excellent.

GREAT design, GREAT idea thanks so much for sharing, I am having a blast printing this thing.
Just one little salty side note : if you use chamfers (45 deg edges) instead of Fillet (rounds) for the edges
it will be SO MUCH easier to print with a good results, specially on less than pro grade printers..
so i had to spend some time on repair with hot tools after printing, we dont need to do that, if the designs where more uptimized for manufacturing on cheaper tools :-)

Excellent design and instructions, Thank you so much. cannot wait to see whats next :)

How did you get the first picture animated?

I love the engineering drawing style assembly instructions. Very cool idea!

I'm having all kinda of issues with mine. When I first put it together it worked beautifully, but I tried it out today and it clicks like 4 times and gets stuck. Do you have any good troubleshooting steps? Thanks!

Is it ok to use M4x20 instead of M4x16? Maybe I can use it with a few extra washers?

I printed all the parts for page 1 (the Mainspring Assembly), but can't get them to fit together. It seems like either the square part of the Ratchet is just a little too big, or the matching holes on the other parts are too small. I was able to get the spring on there by forcing it with a hammer, but there's no way I'll be able to get the winding stem on there.

Any advice? Is there an easy way to print that lug a little smaller without affecting the rest of the piece?

Most likely a printing problem on your end. I have printed and built the whole thing. Because I had a little runout due to "elephants foot" I had to trim some parts. In fact, I think I had to trim the mainspring part you referenced. This was not a problem with the design, but with my printer settings. You are probably having the same. I have since adjusted z-axis offset to fix the elephant!

I found a setting in Slic3r that was doubling the amount of filament extruded on the first layer. Turned that off, also turned down the extrusion multiplier to 0.6 and reprinted. Now things fit!


Great job on this project! I've printed everything and am now waiting for the metric fasteners to be delivered so I can assemble it.

Would you be willing to publish the solid model files so we could make our own modifications? If not, even DXF files of the 2D profiles would help build some of the parts.

Thanks for your efforts and I'm looking forward to getting it running soon!

I got everything working yesterday and it's now running smoothly - and for a surprisingly long time!

A couple of observations:

  • I really don't understand the purpose of the Pallet Fork Limit Screws. As long as the Impact Pin is within the fork the pallet fork is retained. If the pin were to come out of the fork, the limit screws don't adjust tight enough to constrain the fork, and the escapement teeth prevent any motion outside the normal working bounds. Scratching my head on this since someone else wrote they had to adjust them to get it to work properly.
  • The motion was very inconsistent at first. I'd get a few ticks then it would stop, or it would refuse to start altogether. I seemed to have solved this by filing the contact points of the Pallet Fork and Escape Wheel smooth. I also made sure the corners were very sharp. Initially my printer (Prusa i3 Mk3) printed slightly bulged corners and after watching the motion it was clear that this was part of the problem.
  • The adjustment of the Impact Pin definitely made a difference, but the mechanism is quite sensitive to its radial location, so expect to fiddle with it until the motion becomes smooth
  • I managed to "strip" several of the M3 nut pockets - in particular on the Impact Pin Housing. Not a big deal, but maybe a tighter fit or use of a square nut would be better here
  • I get quite a long run time with only a small (90-120 degree) wind of the spring. Anything more and the friction between the pallet fork and escape wheel (and I'm sure in multiple other areas) rises and diminishes run time. I think that if the Mainspring Assembly was stiffened up a bit (perhaps with a bridge between the two frame pieces) and some finely-sanded escapement surfaces it would run for a lot longer. I think I will try brushing some CA onto the contact surfaces and then sanding/polishing it smooth. That should create a lower friction coefficient, but we will see :-)

I bought a set of M4 fasteners from Amazon, and the long ones I bought from McMaster-Carr

Where did you get the metric fasteners from? Im having a hard time finding the longer sizes

For other projects, I've bought metric hardware from either Amazon (if I need it sooner) or AliExpress (if otherwise). I think I have most of the hardware on hand, though I'll need to buy the longer M4 fasteners.

I combined the BOMs on the assembly drawings to come up with totals for prints and hardware...it's attached. (For prints, you need 4 "Bracket"s and 1 each of everything else.)

I do not know where you live, but I found 40, 50, and 60mm at Lowes.

can i use pla for the spring or does it have to be a flexible filament

I used PLA. Works fine.

I'd like to Rube Goldberg a winding method that uses a small motor or a pendulum. Something that keeps it running without manual winding for over an hour. Has anyone looked around for metal springs?

Prior to the advent of stepper-motor clocks, electric clocks that still used balance wheels worked as follows. There was a spring that, when wound, kept the clock running for a short period of time...perhaps ten minutes or so (I cannot recall for sure). Once the spring was close to losing energy, contact was made that completed a circuit through a solenoid. The solenoid then wound the spring for the next cycle. This was repeated until the battery was dead. I don't think a direct application of this idea will work because it takes a considerable amount of energy to wind this thing. However, I discovered (while assembling) that the drive train does not need a lot of power to operate. Not sure how to take advantage of that...but it is a very interesting problem to ponder.

This was especially common for clocks in cars; I had one in a '77 Cutlass Supreme that was of this type. Arcing of the contacts from making and breaking 12VDC would often damage them to the point that the motor would no longer get power and the clock would run down and stop.

I am trying to make it and have a small diference in the M4 Nut I get 18 needed.

M4 Washer 24
M4 X 16 13
M4 X 20 2
M4 X 40 2
M4 X 50 1
M4 X 60 3
M4 Nut 18
M4 Lock Nut 4

M3 Washer 3
M3 X 12 2
M3 X 25 1
M3 Nut 4

Will post pics once I am more advanced.

Thanks for the files

Comments deleted.

TIPS TO GET THIS RUNNINg (for me anyway)

Pallet fork; loose enough to freely move was not enough, another 1/4 turn to give it some wobble helped immensely. Possibly because of the impact pin surface issue mentioned below.

Impact pin to pallet fork; Make sure that you're not getting binding between the face of the impact pin housing that is parallel/touching the pallet fork. I was getting binding here and loosening the pallet fork more seemed to help.

"gear train"; Check your "gear train" axle. I assembled mine so it freely moved but this is a counter rotating part which means it can self tighten, mine did even with the nyloc.

Assemble each section and test from left to right and make sure everything spins freely.

Adjustments; The only one I found that really mattered was the impact pin housing (round part) too far one way or the other and it swings too far in one direction. I adjusted this so the ticks and and tocks were roughly even timing, and as such the swing was centered. The two screws on the bracket which limit the pallet fork travel seem to only matter in speeding up or slowing down the tick tock.

Nyloc nuts are not intended to be used for parts subject to rotation. You should use either a drilled nut with lock wire, a castellated nut and cotter pin, or a straight pin with keeper if required.

Those are not proper for this application either and you're attempting to correct the wrong person. I simply used what the creator called for in this case as it's just a tchotchke. If I were actually engineering this there would be thrust washer involved.

I wasn't trying to "correct" you, I was just providing additional information regarding why a nyloc nut is not correct for this particular use case for whomever cared to read it.

BTW: Use of thrust washers has nothing to do with the improper use of the nylocs. Now I am correcting you.

No, now you're just not reading properly, because I didn't say the two were related in any way.

Then use separate paragraphs.

That's not what paragraphs are for, it was written correctly. You have a nice day now, goodbye.

Also, I have almost exactly a 3 minute run time.

I am just astounded... and I am so going to print this... ASTOUNDED SIR! ASTOUNDED!

I wish I could get mine to work. I spent many days printing. I'm wondering if PLA will even work. It seems like it should and perhaps I just need more time to tinker with it.

PLA works. I just completed mine this morning. You have to make adjustments. You have to make sure that the balance wheel screw is not too tight. make sure that all of the brackets lie flat (planar) on the surface.

Great work designing this! I just finished assembling one that I printed at 75% scale. It barely functions due to the hair spring being too thin. Do you mind if I post a remix after re-engineering the spring tension and sizing so that others can do smaller versions?

Hello! Did you use different hardware after scaling down to 75%?

Sorry for the late reply, I missed the notification until cleaning out my emails.

I ended up using M3 screws for just about everything, maybe one or two M2 screws where necessary. I did drill some holes a little larger to fit the screws since everything was now undersized.

Would it be possible to take a photo from behind the balance wheel, so I can see how the spring is attached. (If you have pre tentioned). I am currently having issues with the ballance either having to be manually moved due to not enough momentum or the part spinning out. Spent half an hour moving from one side to the other and think I have it nearly there but the spring resets to middle.

Yes, i have put my spring on the wrong way around :) quick 180.

I don't think it matters. I've seen the hairspring mounted both ways in videos on youtube and the thing still runs. Check this out:


As opposed to LarkysPrints video.

Let is know if you got it working!

Having a issue that until it is nealy fully wound up, it will not turn. then it goes like billy goat. This is on both sides, was trying to get an even stroke but had the issue that even at fulll top it would just bounce at the bottom releasing.

This thing is amazing, thank you!

Check the central gears.I had problems with the center Gears the did not move freely all the way around. Mine also stopped after 2 to 4 Ticks.

I used a Hairdryer to soften them a littlebit until the move freely.

Too bad that there is no adjustment help, my part stops after 3-4 bars. And for what are the screws 21 + 22 the pallet Fork comes to any of the screws.
After 3 hours I give up, too bad for the 3d printing .. Or still have to be reworked on the parts?

Kind regards


Maybe if you post some details, picture, someone could help you. I have not completed my build (maybe today), but I have the gear train and balance wheel built and working just fine.
The structure must sit firmly on a flat (planar) surface. I noticed that if not, the pallet fork engages too closely with the escape wheel causing problems.

The model is great, but why the metal screws and the nuts? I am waiting for a completely printed plastic model.

I tried that jonnie, printer is not able to print nuts that small. I printed the bolts and nuts and they dont screw in. best to just buy the m4 nuts and bolts.

It seems to be dependent on a printer. I am successfully printing even M3 screws and nuts on Prusa MK2 and MK3 using ultra-detail profile. But if a printer is not precise enough to print standard ISO threads, there are always more error-tolerant versions with higher pitch values and deeper teeth, not to mention other alternatives to screws, such as snap-ons etc. :)

I'm having trouble getting mine to work properly. It only works if I move the balance wheel by myself but it does not get enough power to move on its own. Any advice on how to properly configure it so it works?

Did you find a fix for this? Mine is the same, as if the spring for the balance wheel isn't providing enough force

Yeah I just got it working pretty nicely. It really was several things that I improved. The main thing was taking my time fine tuning the impact pin position and most importantly for me, reprinting the hairspring. I think on the first print I accidentally printed it on 30% infill, now with 100% infill it works much better.
For the fine tuning I made sure that there was no pressure on the main spring and positioned the resting position of the impact pin like it is in the beginning of the creators youtube video.

I had the same problem. It wouldn't start on its own and only ran for a few seconds. Then I removed the balance and the hair spring. I rotated the hair spring 90 degrees, and pushed it back on the balance. Reassembled it all and it worked great. Getting the tension on the hair spring, along with the position of the impact pin housing and the limit screws of the pallet fork was the key to get it running well. Also look at some of the others' pictures. I used that as a starting point for the limit screws and the position of the impact pin housing.

I dont think the hair spring is supposed to be pre tensioned.

I think my comment wasn't clear. The hair spring should not be attached to the frame when you rotate it 90 degrees on the balance. All rotating does is position the two screws that hold the impact housing 90 degrees from where they currently sit, which makes a difference in the adjustable limits of the housing as it sits on the balance. I initially assembled it without regard to the position of the screw holes on the balance for the impact housing. Therefore I didn't have the correct adjustable range for the housing until I rotated it.

I NEED this!!
Can I have it printed for me, as I don’t have a 3d printer!

I offer 3d printing service. I'm about to print one myself and could sell you one if you contact me.

YES! I’ll take you up on that! Contact me and let’s DO it!

Thank you. It works. It was fun to make it.

I wonder for how long does it run when you start it

I wonder if printing the springs in petg would be more effective, as it's a little more flexible, and less brittle.

This is a superbly well designed project and the drawings are spot on. Thank you! I share your interest in all things mechanical and this is an ideal project for me and to share with my grand-son. The only improvement I can think of is the use of bearings but the fact that all the cogs and brackets are printed is exquisite!

Build mine. Wow works nice.
Took abit of tweaking, but all fits together pretty good.
:) :) :)

Comments deleted.

Super cool!!! It's interesting, I just happened to watch that old "how a watch works" video about a month ago.

What software did you create all these parts in?

Your design is the first project we are printing out on our new Robo R2. My students love it - well done! One question (so far...): we have tried to print out the Spring Frame Front out twice and both times it failed. It prints the 'knob' first and then has nothing but air to print onto for the 'A-frame'. It seems upside down. Am I missing a setting? Should it have supports? - Thanks!

you need to rotate the part 180 degrees so it lays flat

Thanks. It was such a 'newbie' question. I just downloaded the Robo software and rotated it. Now I'll try to print it again.

Turn the object around 180° in your slicer so the 'knob' is on top. Several parts need to be turned around or on there side ( bracket ) in order to print them without overhang.

Thanks. The software was pretty intuitive so I was able to rotate the part. Maybe one day 3D printers will have a way of printing 'in mid-air' without supports...

They can but they're incredibly expensive, as in "Price on Application" levels of expensive. Selective Laser Sintering 3D printers work by depositing layers of powder and sintering (melting) the sections that are meant to be solid. The layers of powder provide support for overhangs and you just pull your printed part out of the bed of powder when the job is finished. The remaining powder gets recycled for future print jobs.

This morning I started printing the last part, the main spring and I did not notice I had to turn around this part also, had to start over. You really have to turn every part in your slicer to check wether it is printed in the right way without overhang.Succes!

How long does it go for?

Great design I am looking forward to printing it out. What length is the bolt you use for the impact pin it is not listed?

It looks wonderful, how much time does it takes to print (approximatively)?

hello, this is a cool 3d printed design. i am interested in how you made the main spring and hair spring for the watch escapement desk toy, i saw a little but if you had more info for me that would be awesome. i am a horologist and i know how crucial the spring length and strength of the is for the overall beat and timing. how did you get the beat and timing right? and what is the exact beats per hour? and what is your gear ratio? thanks.

Have you considered submitting this project over on Hackaday.io? They are currently running a 3d printer gears pulleys and cams contest and I think this project would do really well in it.


Hello, When I search for the m4 bolts, it seems they are all full thread. Is that a problem ? I'm just about ready to start assembly, so these are the last pieces to the puzzle. Thanks for such a wonderful design ! I hope to finish and post pictures soon !

Shouldered screws are not required, fully threaded bolts will work just fine.

Excellent ! Thanks !

Hi Everyone,
what a beautiful design.


Has anyone already tried to reduce the print size?
Let's say 50%?
Is that possible ?
do the m4 bolts become m2?

I want to know if this can be scaled down 95% without effecting screw list or operation.

Hey when I put the main spring into the creality slicer (I have an ender 3) it said it is going to take 9 hrs to print. Is this right or do I have a setting wrong. All the other times look right but the main springs takes way longer than everything else. Even the spring barrel says it will take 3 hrs.

I used cura to slice the main spring. It was quoting me almost 16 hours. I started the print and noticed something pretty quickly. It would start doing the infill as normal but would only complete several rows before jumping to a new section of the spring. Basically instead of infilling everything on one spiraling motion it would do a section, jump, do a section, jump until it was done filling in that layer. It created a lot of unnecessary motion which really added up to the overall print time.

I stopped it after 3 layers and downloaded slic3r. Sliced it in this new software with same settings. It ended up printing in about 3 hours because it wasn't performing tons of unnecessary motions. Never realized how big of a difference slicing software makes. I will not be going back to cura.

On my settings the spring took around 5-6 hours to print with a speed of 45mm/s and 0.2mm height

Ok thanks, I also changed the fill to 100% for the springs like you said in the comment below this one. It pumped it up to 10hrs but I'm still unsure. Guess I might have to do multiple prints if it doesn't work. Thanks for the quick response!

100% fill at 0.1 mm layer height took me 11 h at the speeds I print at.

How much infill for the springs?

For mine I used 100 percent

with some adjustments, could this work as a Benjamin Franklin Clock? like this one here https://www.thingiverse.com/thing:3149176
awesome design btw

Benjamin Franklin Clock
by Zarlor

My other clock which is a remix of A26's Tourbillion 1.5 is more like this watch mechanism than my Ben Franklin clock which is just a simple graham escapement. https://www.thingiverse.com/thing:3313820

EB Tourbillion Clock
by Zarlor

My current model will require heavy modifications and tuning to get the balance to oscillate to a known frequency, then a suitable gear ratio could be applied. My model just shows the basic working principles of a mechanical watch.

Comments deleted.

As I wrote before, I edited winding stem and impact pin housing to print without any supports (they should print well on 0.2 and lower layer height), and also made a little upgrade to winding key to make it flat on one side to lay better on print platform.
@LarkysPrints If you like these upgrades, feel free to add them to thing files, because I see no sense in posting them as remix
Edit: I also added slightly changed file for ratchet


nice design, thanks for sharing!

three questions/suggestions:

  • could you please copy "TheCze"s part list to the main description?
  • you forgot to mention which software did you use for the design schematics?
  • why on earth is this thing tagged "NSFW" ??

I have updated the main description with the part list. As for the software I mainly use CATIA for any CAD work, and the NSFW tag must have been a mistake, I have removed the tag since I have no idea what it means.

Not Safe For Work

Thank you for this thing! Thumbs up to you!

  • Posted with 3D Geeks Thingiverse Browser App

On the Gear Train (Misspelt "Teain") Assembly page (2/4), only one of the two Brackets are called out (#12).

Thank you for pointing out the mistake. I have updated the assembly drawing, please feel free to message me if you have any queries or spot any errors.

I'm printing this right now :D I've upgraded few parts, but I will post them after I assemble and test everything - I've edited those parts which require supports to eliminate need of supports, and I've also split balance wheel to two parts - outer and inner, so they can be "assembled" in slicer and then it's possible to use per object settings to give outer part 100% infill to make it heavier, but to be honest I have no idea if that will improve anything :D

Required parts as I counted them:

M4 Nut: 19
M4 Locknut: 4
M4 Washer: 24
M4 x 16: 13
M4 x 20 : 2
M4 x 40: 2
M4 x 50: 1
M4 x 60: 3

M3 Nut: 3
M3 Washer: 3
M3 x 12: 2
M3 x 25: 1

The old instructions have one M3 taking a M4 Nut, the designer confirmed and fixed the mistake now, the correct amount of M3 nuts is three.

Also should probably note it looks like the bolts need to be fully threaded, partial threading on those longer ones is going to be super common

Thanks for compiling this! I counted through all the pieces too and come to the same results; so I'm glad I didn't overlook anything. This is useful info!

Thank you for pointing out the mistake, I have already updated the exploded view.

What would be required to turn this into a functional clock, and how long does a wind last?

In order to turn this model into a functional timekeeper, the oscillating frequency of the balance will need to be fine tuned. Then the gear ratios could be calculated to achieve hours and minutes.

Printing this right now, cool design!

Thanks for sharing love mechanical things just starting on 3d printing and will have to print it

nice design !! thanks for sharing !!. thums up

So... I would bet that the makes are all awaiting delivery of the correct bolts!

I guess that I must be a little slow, but, I must print out 4 (four), of the brackets. Is that correct? Also, this is the only part that requires more than 1 copy, correct?

Yes, you will need to print out 4 of the brackets and all the other stl files will only need to be printed out once.

Thanks. Just making sure.

That's fascinating. Thank you for sharing, and congrats for the nice design !

amazing design! Can't wait to print one of these

very cool, thank you!

Super! What is the maximum diameter of the parts?

The largest part is the Spring barrel which has a diameter of around 120mm

Thank you for sharing this very interesting idea and the great model.
I have to print one :D

Just brings bad-assery to a whole new level! cheers

Really Awesome Design

What documents/sources did you use to come up with the geometry for the swiss lever escapement?

When I was coming up with the geometry of the model I didnt have any proper documentation except for pictrures to gudie me through the process. But by understanding the concept of how it works helped a lot, you could refer to the video link i have posted to gain a better understanding of how it works. This is merely a basic model, I am planning on designing a more sophisticated and calibrated version in the future.

I like the exploded assembly documents you've included. What program did you use to create those?

Everything is done in CATIA

Looks like CATIA

This looks rad, can't wait to build one.

I was thinking of something very similar just the other day. Oh well, onto the next one :)

Great Works!

could you make a list with all the screws you need?
that would make it easier to search

thank you

All the parts are listed, not hard to count.

Very cool! What is the total running time?

One my model the running time is around 2 minutes.

This is one of the coolest things ive seen on here. I knew i needed to print it as soon as i saw it. Ive got half of it printed already. Thanks for your time and effort to model and post it.

Awesome design!
Everything is printed in PLA?
And PLA is flexible enough for the coil springs to actually flex and bend?

PLA can definitely handle the flex on a design like this. I would just make sure your PLA isn't old. I'm not familiar with Sunlu +, but I print with eSun PLA + all the time and it is not very brittle. I will be attempting this build with eSun+

Great Design! Would there be any chance you could add a STEP file of the assembly?

Great design! Thanks for sharing.

I was also hoping for some solid models, if you could manage to export to another solid format (IGES, parasolid, etc...) I would be extremely appreciative. Even 2D DXF would be amazing. I'd really like to add some fun tweaks and simulate the results.

Hello Guy
Thank you for the interest in my design. Unfortunately my licence does not allow me to save files in STP formats. However if you have any queries or problems with the assembly,please feel free to message me for support.

Awesome post! Definitely going to be a Thing of the week!! Thank you for posting!

Cool design. Can this be scaled?

I really wanted to scale it down as well. I was going to until I though about the hardware. You would need smaller everything and I'm not sure which scaling factor to use to be able to use a common sized fastener.

I'm currently making one scaled down to 80 percent to fit my build area. Some of the gear parts I've tested seem to work fine but I won't fully know until I finish the build. I'll try and reply with an update once I get everything together.

Hi PandaCycle, any update on printing this at 80%? Did you have to change the hardware?

I’m not sure about that but scaling down may change how tight the gears mesh together.

Really Awesome Design
Thank-you for your time and effort to share