Hacker News new | past | comments | ask | show | jobs | submit login
DIY makers who build their own watches (bbc.com)
217 points by raleighm on March 24, 2018 | hide | past | favorite | 71 comments

As a watch enthusiast, I've wanted to build my own watch in the past. However, as I got deeper into watches, I realized I could never really make the watch my own creation unless I developed and manufactured the movement myself. Otherwise, I'll be dressing up a generic ETA movement [0] with a personal dial, hands, bezel, etc. This is what most of these DIY watches are. With that said, ETA movements are some of the most solid movements out there with decades of support. They're also capabale of COSC [1] chronometer timekeeping if adjusted properly. For me, the DIY magic dies with the movement because I'm limited to the dimensions of the ETA movement. It's the heart of the watch.

[0]: https://en.wikipedia.org/wiki/List_of_ETA_Movements

[1]: https://en.wikipedia.org/wiki/COSC

I've been making my own movements for fun for a while. Most of the parts you can print/order, but for accuracy purposes I tend to use an existing balance spring (generally sourced from any old Seiko movement).

I recommend 2 books if you want to make your own movements

1) Watchmaking by George Daniels 2) Practical Watch Repairing by Donald de Carle

The first is a textbook and second deals with repairing, but actually taught me the more practical side of how to assemble & regulate a watch, and what makes a movement accurate.

I've made my own hand-winding movement. It is within COSC specs but only when it's more than 75% wound, perfectly flat, and in a cool place. Oh did I mention it's the size of a ping pong bat. It makes me appreciate ETAs and their reliability even more now. It's time consuming and drives me nuts at times but totally worth it.

That's super cool. Can you please, pretty please post some pictures?

I've been wanting to attempt to make a coaxial movement, roughly that size as well. As there's no chance I could make one that would fit in a watch.

I've also pulled a few Unitas (pre-ETA acquisition) movements out of stopwatches and assembled wristwatches with them. It's not "watchmaking" but it's fun nonetheless.

This is cool, do you have any recommended blog posts on the end-to-end process? I’d love to build a movement but would like to gauge the effort/cost/time involved

It's a never ending thing!

You could start by sort of reverse engineering movements (disassemble and reassemble).

Start to build your own tooling.

Then make a copy of the bridges and baseplate.

Try to change the geometry of the bridges and baseplate, keeping the same moving parts.

Then build your own moving parts.

Start from massive clocks to "table clocks" to marine clocks to wristwatches movements.

And it could be very expensive, wood clocks are cheaper, less time consuming. There is great stuff and ideas online. For example, see the timer from Ugears there is mind-blowing tricks hidden in this model.

super interested. i do machining on the centimeter scale and have no idea how to work on those parts.

do you buy gears? if not what is your gear cutting rig look like.

it seems pretty clear you need a jewelers lathe and some decent files. what else is indispensable?

It's not possible to buy gears due to the "module" (I don't know how to translate it), it's a factor of scale that influence the spacing between gears.

The hardest parts are the springs (balance spring and mainspring) and the screws (there are so small, it's a specific norm, see NIHS for Normes de l'Industrie Horlogère Suisse). For the gear cutting, there is an example here [pdf link]: http://www.cowells.com/docs/cutter.pdf

There is a lot of custom machine tools, many of them are needed to resolve problems that an individual doesn't care. But there is always alternative ways and it's the fun and watchmaker's task to find them!

So there is not really indispensable tools as it depends a lot of the project. Expect if you want to make specific watch decorations, like perlage, guillochage, soleillage...

I can't find links but there is an operation named roulage where the guiding parts of the pinions are hardened, it's not indispensable if you don't want an industrial grade, kind of.

The simplest design is the roskopf movement.

> "module" (I don't know how to translate it)

fyi, it is called module in English too.


This is more clock making rather than watch making, but check out the clickspring channel on youtube. He builds a mantel sized clock from scratch in his home machine shop. Amazing stuff and beautifully done videos.

I went to an exhibition at the Science Museum in London on George Daniels.


Fascinating story - he spent years making a watch from scratch in every spare moment. For no other reason than that he wanted to. And look where it led for him!

I've had the first book (Watchmaking by George Daniels) on an Amazon wishlist for some time but the price never seems to drop below £50 (and used copies go for far more). Always a sign of a worthwhile book :-)

Oh did I mention it's the size of a ping pong bat.

That sounds more like a clock movement, but you quickly realise that the miniaturisation is the truly difficult part; making a mechanical clock is not nearly that hard in comparison.

> I realized I could never really make the watch my own creation unless I developed and manufactured the movement myself.

Exactly. But clocks are definitely doable and once you have mastered clocks who knows, you just might be able to use your experience and graduate to a watch small enough to be worn. The good news is that there isn't anything that 18th century clockmakers knew that hasn't been documented somewhere.

This guy has the right idea and has some interesting movement workarounds for tools he could not afford:


It's unconventional but it is an actual watch movement. Far more impressed with that than the generic movements in some case.

Before diving into wristwatch movements, a step by marine clocks could be interesting and fun too!

It's funny how when you start tracing clock making you end up reliving pretty much all of the industrial revolution.

Those people were really artists improving their products and their tools at every step and their heritage is still alive today.

The watchmaking/clockmaking industry trained a whole generation of precision machinists which led to a revolution in mechanics, toys, calculating machines and ultimately set the stage for worldwide navigation and the industrial revolution.

True and clocks and watches are still the most precise mechanical tool/instrument.

These are amazing too:


I've never been able to find one that I could afford though.

What do you base that on? I thought some industrial applications actually required much finer tolerances. Things like telescope mirrors. It probably depends on what you consider to be a mechanical tool.

Sure, most watches have a tolerance of +4/-5 seconds by day (86400 seconds). It's about 58 ppm.

I don't know which precision a not electronically assisted telescope have, but it's sure it's impressive.

-That would be pretty good watches - the requirement for being labelled a chronometer by the Swiss is +6/-4/day (with some further requirements - positions and temperatures)

Seiko tune their top-of-the-line Grand Seikos a bit tighter - +5/-3/24h.

They also have some limited edition VFA (if memory serves - Very Fine Adjusted) watches capable of +/-2s/24h.

The engineer in me finds this kind of accuracy in a mechanical device just a hair short of black magic.

>The engineer in me finds this kind of accuracy in a mechanical device just a hair short of black magic.

I'm curious where you get this intuition. As an engineer with a background fully in electronics/software my intuition is that multiple-seconds per day of error is horrible. I never got into mechanical watches because of this. There are so many great quartz based watches that keep time so well that I've never seen the allure of the mechanical. I'm now curious from this discussion how big of a feat it is anyway but have no intuition for it.

https://www.amazon.de/Bauen-erleben-begreifen-Technikgeschic... has a chapter where you‘re building several types of real, working clock movements. Very imprecise, of course, but still.

You could probably build a movement yourself these days with one of those $500 SLA 3D printers, mostly-plastic movements have definitely been done before.

Though it sounds like these people are using CNC mills for the cases, that is definitely not cheap equipment.

SLA printing services also provide a pretty cheap option (Shapeways or dirtySLA). You can also print (and then fire) ceramic parts printed with SLA printers (Vitrolite), I'm not sure very small parts would work well however.

Milling services are also available (and not super expensive in China) but the parts are very small, and there are a lot of them so it might not be a easy option.

The idea of printing the entire movement is certainly interesting though...

Not really.

$500 SLAs gives you LCD SLA printing with medium resolution.

You will need at least $1300 for a Moai kit(with Laser Galvoes):


Or at least a $1500 DLP kit with a good resolution for creating small parts.

Fo really small parts you need the strength of steel or titanium. But you can made then in plastic, metalize the plastic and use something like EDM(electrodischarge) to make your parts.

What do you mean 'medium' resolution? The Moai has 70 micron laser spot size. Form2 is 140 micron. The $500 SLAs have 47 micron pixels. They have major disadvantages vs the expensive laser SLA machines but resolution is not one of them.

While I don't doubt that clockworks can be 3d-printed, doing a watch-scale movement with even a $5k printer sounds pretty incredulous.

Have you seen what the new cheap SLA machines can do? Several ~$500 machines out now with 47 micron X-Y resolution, using phone LCDs for the UV mask. Wanhao D7, Anycubic Photon, etc. Not sure if it's quite good enough for watch movements but they're certainly capable of much finer details than any FDM machine including the very expensive ones. In principle they should even be capable of finer details than the $3500 SLA Form2 which has 140 micron laser spot size.

> Have you seen what the new cheap SLA machines can do?

I'll have to admit, I haven't seen what the state of the art is like. I did some googling of the printers you mentioned and the level of detail is pretty impressive (although most reviews were printing pretty figurines instead of engineering applications), so maybe 3D printing watch movement is not that far fetched after-all. That being said I'd still be quite impressed seeing a 3D printed watch movement. Even more so (and this is moving the goalposts) if the movement ran whole day on a single winding, demonstrating good meshing and low friction.

https://www.thingiverse.com/thing:1249221 is very cool - '3D-printed Watch with Tourbillon' (it does look pretty big though ;)

I hear a lot of you talking about ETA movement.

Great movements, but buy them up while you can.

Swatch Group will cease selling their movements to just anyone.

There's enough movements in the secondary market to keep most home hobbiests happy though.

The Swatch Group, and The Wilsdorf Foundation (Rolex) will not sell parts/movements to pretty much everyone.

These practices are killing the watch repair industry, and making fools out of anyone who buys a quality watch.

Certain supply houses (Cousins UK) are fighting the monopolistic behavior, but the lawsuits are going nowhere.

So that custom watch, with the ETA movement is really not yours if you can't buy parts for the movement.

Our government (USA) knows about the problem, but has bigger fish to fry?

Mechanical watch movements are really cool! But I'm not sure you can call it monopolistic behavior if, first, you are able to name many major competitors, and when the substitute good (a quartz movement) of equivalent accuracy is $5... It seems a bit silly to complain overmuch. Especially when other posters are talking about printing the parts off on a $1500 machine!

Nobody cares about ETA at this point.

The patents and registered designs of ETA movements have expired. You can buy exact clones of ETA movements legally.

ETA clones by Sellita and Soprod are said to be better than the original. Sea-Gull (and other Chinese) aren't bad either. Especially for the price...

"I realized I could never really make the watch my own creation unless I developed and manufactured the movement myself"

Aren't you equally disappointed that you didn't mine the metal from the earth, and raise the cows from which the leather was taken? ;)

The movement is he heart of the watch. I'm not sure you could call yourself a watch maker if you are not designing and building the movement. You work on your specialty and outsource everything else.

Hey, that's my hobby! Here's my most recent watch:


This is one of the coolest things I’ve seen in a long time. Do you work in a machine shop to have access to that type of tooling?

I work at Google, we have a hobby machine shop for engineers to play around in :D

Ughh I’m totally not incredibly jealous.

Really cool to see it being used for interesting projects, kudos!

Extremely awesome walk through. :)

Two questions, if that's ok?

* What was the bad experience that caused you to stop using a parting tool?

* What sort of mill are you using? (kind of thinking a 4th/5th axis being added might help too) :D


I've had a few parting tools explode on me before, haha, which had left me afraid of it.

I'm using a bridgeport for most manual stuff and a CNC converted ACRA mill for the CNC work. I don't do that much CNC, since manual is a lot faster in my experience, so three axes is plenty for me.

That's got as much to do with watchmaking as putting decals on your car has to do with engine fabrication.

If you want to make watches for real you should start by making clocks and work your way down in size as you get more experienced. It will likely take you a lifetime to master the skills unless you can find a watchmaker to train with (and not a movement packager).

Modern software engineering has a much to do with software engineering as putting decals on you car has to to do with engine fabrication.

If you want to engineer software you should start by making transistors ands putting them together into AND gates and work your way up. It will likely take you a lifetime to master the skills unless you can find a programmer to train with (and not a JavaScript hacker).

Let’s not “no true Scotsman” these folks or anyone for that matter. Everyone starts not knowing everything. Everyone ends not knowing anything.

These people are making cool watches where none existed before. Therefore they are watchmakers. The aesthetic design is an equally important part of the package: who cares if that’s were they decided to focus their first efforts. I imagine a large number of us so-called software professionals started the same way: changing an HTML tags color, hacking a new sprite into a favorite game or changing the color of our DOS prompt.

> Modern software engineering has a much to do with software engineering as putting decals on you car has to to do with engine fabrication.


> If you want to engineer software you should start by making transistors ands putting them together into AND gates and work your way up.

No, but if you want you can:


> Let’s not “no true Scotsman” these folks or anyone for that matter. Everyone starts not knowing everything. Everyone ends not knowing anything.

Indeed, let's not belittle the people who - you know - actually make watches. This is a very hard to master skill, a sub-branch of clock making and it's just plain insulting to pretend that watch making is the same as watch packaging.

> These people are making cool watches where none existed before.

If you put DIY and 'build their own watches' in a title then the least I expect is someone who knows how to cut a gear to be present on the other side of the link. Otherwise it is just another repackaged movement.

> Therefore they are watchmakers.

Watchmakers would beg to differ.

Source: have worked with and for actual watchmakers in several capacities, once where I wrote the software that controlled specialty machinery to make parts and once where I looked in lots of detail at a company that repaired watches.

> I imagine a large number of us so-called software professionals started the same way: changing an HTML tags color, hacking a new sprite into a favorite game or changing the color of our DOS prompt.

I'd hope they know the difference between software engineering and changing HTML tags color.

I'm sorry, I'm not of the 'participation prize' generation.

If you want to call yourself a watchmaker: show me a watch you've actually made.

Note that many of the major watchbrands are also no longer proper watchmakers, they couldn't make a movement that lasts economically if their lives depended on it. The unit costs are so low that making watch movements economically is hard for even the largest players.

But that doesn't mean you couldn't make your own watch if you wanted to, it would be a labor of love.

Some nice links elsewhere in this thread that are much closer to actual watchmaking than TFA.

I think the post you're responding to is sarcastic ;)

The root of the problem is flooding out subgroup discussions, which eliminates those subgroups from the hobby.

There used to be a usenet group a mere 30 years ago for homebuilt computers which had a FAQ explaining they mean home built as in boards full of TTL and using a commercial LSI CPU was considered kinda cheaty. This worked for awhile but eventually crossposting and eternal September utterly destroyed that group as a discussion area or subgenre of the hobby. You can't talk about designing a single bit serial ALU using nothing but 7400 NAND gates, some flip-flops, and wire jumpers if for every on topic post, there are over 100 cries for help from people looking for windows video card drivers for their homebuilt computer, where by homebuilt they mean they purchased a COTS motherboard and a COTS case and screwed the together. And that's how a hobby dies.

Likewise these "watchmakers" are perfectly valid legit jewelers who happen to mount COTS watch movements in their jewelry instead of photographs or gemstones or whatever else other jewelers do for fun. There's nothing wrong with being a jeweler other than misidentifying as watchmaker thereby annihilating all social discussion of actual watchmaking by vertical flooding discussions of watchmaking.

A good analogy of the situation would be I have passed adult and pediatric CPR certifications and several first aid classes but it would be wrong to portray myself as "an emergency room doctor", and get a couple million people to think that way and you'll destroy the medical schools eliminating the possibility of having actual real ER doctors in the marketplace of ideas, nothing but guys who passed a CPR class a couple years back.

An even better analogy is there's about two places on the internet you can get nice plans for wooden clock movements, but they're almost impossible to find because there's an infinite number of websites for wooden clocks on the subject of "build a wooden box, put a COTS movement in it, call it a wooden clock". Its virtually impossible to discuss actual real wooden clocks because of the infinite flood of box-makers who install a movement in their wooden box. There's absolutely nothing wrong with building wooden boxes unless its destroying an online community or commercial marketplace of wooden clock makers.



I've been working on a project from the latter site, on and off, mostly off, for some years. Nice plans and nice designs. In comparison google reports 41,700,000 results found for wooden clock, of which the top result, at least for me, is a plastic wood trimmed clock on Amazon using a quartz movement that atomically regulates itself to the radio station WWVB, pretty much the opposite of what I'm looking for. I did not check the other 41,699,999 other sites but they're probably very similar, fake wood box for sale that happens to contain a quartz clock movement or make your own wood box as an exercise in wood box making that can have a COTS quartz movement installed.

A similar analogy is that gray market low quality counterfeit products ruin the marketplace and social discussion of the actual non-counterfeit subject.

Or in summary, the folks in the linked article are doing something cool; it has nothing to do with watchmaking, but it is cool to make jewelry.

Lets not discourage people from being interested in things by pedantry. This would be like yelling at kids for reading harry potter because it's not Proust.

If someone gets into understanding more about a thing by doing some DIY then they're getting into the thing. They don't have to go from scratch on every single part.

People don't have to learn to cook a burger starting with planting seeds and raising a steer. Nor do you have to be capable of working on the Epic graphics team to be able to build a good video game. No you just need to use unity and make something fun.

Encourage people to get into the field and let them self select on what interests them, not what interests you.

More people having even basic assembly knowledge the better.

I don't know anything about watch making, I'm just wondering are the gears etc. these days mainly milled with a CNC machine?

I assume there's a lot of skill in assembling as well.

I'm very curious how they made gears etc. say a 100 years ago too considering they're so tiny.

There are many skills that come to play in watchmaking.

Gears are typically made with a gearcutter, a precision indexing instrument that revolves a 'blank' through all the gear spur positions to cut the gear halves one-by-one from the blank until the whole gear is done. If you're going to use a high quality hardened steel bit for the cutter and a much more mild material (such as brass) for the gears then you should get a lot of stand time out of a cutting bit.

Another technique you might try is etching, which can be used to make quite high precision parts. And if you want to do some minor automation it is fairly easy to see how a bit of knowledge about stepper drivers and basic machining might help you to create a reasonably accurate and quick automated gear cutter.

An off-the-shelf CNC machine that will cut gears for you is possible but one that is precise enough for watchmaking will set you back a small fortune. Making your own tools and toolbits is a good part of the fun of a project like this.

Cheers I'd not heard of a gear cutter before.

Also etching sounds an interesting approach too.

IIRC tools for making time pieces were the most precise of their time; custom ordered and very expensive. The history of the persuit of time keeping accuracy, and its important to much of modern civilization, is very interesting IMHO.

It depends of the quantity you need, but there is some that use stamping to make gears.

But you can also saw the profile manually.

Assembling is hard but doable, the hardest part is the adjustment!

> But you can also saw the profile manually.

That is something I've wondered; until what scale is that really feasible? Because fancy indexing lathes, gear cutters and CNCs bring in kinda high barrier to entry, but a jewelers saw (presumably?) and a vise are far more approachable for a hobbyist.

Interesting! I'd not thought you could make them by stamping.

100 years ago they were made on a hand crank hobbing tool, or wheel engine.

I don't know about these days.

You'll note that the article never uses "watchmaker" or "watchmaking" to describe what they do, probably exactly because "watchmaking" has a different meaning than merely "making watches".

I looked into this a year or so ago with the idea of making my own mechanical watch movement, not just a kit watch. I could only find one source of a hobbyist doing this themselves: https://ashtonwatch.com/2012/01/04/its-alive-and-a-bit-of-a-...

It's still a work in progress but there is an initiative at openmovement:


I'm a watchmaking engineer and I know two people on the project, I'll be pleased to tell them any feedback.

By the way the watch you linked is a clone of Unitas 6497, the work is awesome nonetheless.

This is the English link, as it seems to default to German:


That aside, this looks pretty nifty. :)

Reminds me of Dan Spitz (Anthrax lead guitarist) who left back in the 90s to become one of the watchmakers in the world: https://vimeo.com/161489498

After the sad end of Pebble I've dreaming of, someday, making my own smartwatch. All the other existing smartwatches don't compare to Pebble's simplicity and robustness.

The nRF52840[0] seems to be a very interesting chip for this use, with lots of resources and low power consumption. An e-paper display[1] seems also quite interesting since its power requirements are so low.

Do you know any interesting project on DIY smartwatches?

[0] https://www.nordicsemi.com/eng/Products/nRF52840

[1] https://eink.com/product.html?type=productdetail&id=14

This may not be your idea of a smart watch, but is a lot of fun and has cool hardware: https://goodwatch.org/posts/introducing-the-goodwatch/

ePaper usually dies around freezing temperatures, and you won't be able to safely insulate it unless you mount it on the inside close to your wrist. Transflective LCDs are pretty good though.

I think a low cost EDM would be perfect for fabricating small watch parts. Stupid accurate, no forces on the cutting head, could easily be desktop sized (at least for a work area of a few cm) and very quiet. Can only cut metal, but that's a pretty decent compromise I think.

I just found https://www.geartechnology.com/issues/0610x/edm.pdf which seems to confirm that EDM has been used for watch gears - "Wire EDM can be used for micro- machining applications, including the gears found in watches"

Do you know of any low cost EDM machines out of interest?

I've actually been playing around with designing one, but lack the free time to do so. I have a couple of Fusion360 models set up with commercially available parts, but I don't have a physical prototype yet.

I'm planning on setting one up this summer.

I know of nothing available commercially at a reasonable price and size.

This reminded me of the book "Longitude", a tale of truly epic clock-making: https://en.wikipedia.org/wiki/Longitude_(book)

It’s a good book. The story is more nuanced than people often realize. Most definitely Harrison was treated poorly, however, while he created chronometers that met the latter of the challenge, he didn’t really help with this spirit of the problem. A true solution needed a mass produced chronometer.

It’s also often misrepresented that he was the person who solved the “how do you measure your longitude“ problem. It was well understood that you needed to be able to accurately measure the difference between local time and time at a known meridian. The questions where about how to accurately and practically measure that.

Clocks and watches, like old/homebrew computers, are very cool. They are hard but not too hard to the point that an amateur can't do it. I always envy the amount of ingenuity someone has that went into making a watch.

There is an exceptionally good podcast about a vintage clock restorer's life I listened to last year called Shit Town. It describes how the life of someone so brilliant, so smart could be so bizarrely and tragically dysfunctional. I think the story rung a bell to myself, and I couldn't be more sympathetic after listening to it. It has to be my favorite podcast of all time.

The article and this thread is simply amazing to me! I had no idea that were so many who are doing this. I've always had a fascination for clocks of all kinds. Building your own mechanical watch from scratch seems like it would be one of the ultimate achievements for a hobbyist.

And here I am trying to decide whether I should build shelves in the garage myself, pay someone else to build them, or buy shelves at Home Depot. Downright embarrassing to mention in a thread like this.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact