Hacker News new | past | comments | ask | show | jobs | submit login
New oscillator system in Zenith mechanical watches (hodinkee.com)
167 points by nichodges on Sept 16, 2017 | hide | past | web | favorite | 63 comments

I am a long-time watch nut, and I loved this write-up - it concisely explained what makes this mechanism unique and was most educational.

Fun to see a departure from the norm - until this, the probably largest innovation horology over the past couple of hundred years was George Daniels' coaxial escapement (which basically eliminated friction from the equation, ensuring all transfer of energy in the escapement took place between components tangential to each other, allowing for excellent long-term precision)

The plural of anecdote is not data, but my ten year old Omega with coaxial escapement (in a c.2500C) is still accurate to within a second a day without ever having been serviced. That,in my book, is remarkable.

(It is due for a service soon; I do not intend to run it into the ground...)

> the probably largest innovation horology over the past couple of hundred years

There was this little thing called the quartz revolution though :) Just because it turned out you can make really cheap watches with it doesn't make it less of an innovation.

But limiting the field down to mechanical watches, there is also Seiko's springdrive which manages to combine the durability of mechanical watches whilst approaching the accuracy of quartz watches by using a complete new method of escapement.

Co-ax movements are surely a worthy innovation, but I wouldn't want to claim it was the most significant horological innovation of the last few centuries.

Fair point. Spring drive is a very fascinating concept and one could debate for weeks on end in the right (or, rather, wrong) circles whether it or co-ax holds the top spot. (Or whether the spring drive is even a fully mechanical movement or a hybrid!)

(And I should have been more precise - I meant in mechanical horology; my apologies)

Not that there's anything wrong with quartz (except its inherent inaccuracies; I say this tongue-in-cheek as my interest in all things horological extends to my having a rubidium standard in the man cave - a hand-me-down HP 5065A. Sadly, my wife does not appreciate the beautiful engineering, and suggested a Seiko wall clock would have to do for our everyday time-keeping needs...)

> one could debate for weeks on end

That is certainly and certifiably true :)

Although I'm very fond of Springdrive (and Seiko in general), it's just an example that I thought would be more recognisable than, say PP's use of ceramic components. I wouldn't want to claim any of these advancements as more significant innovations than Co-Ax escapements in what is increasingly an art-for-the-sake-of-art (or tech-because-we-can) endeavour anyway.

It's best just to marvel and to start thinking of how to explain the expense to the significant other.

How can we debate whether the spring drive's fully mechanical or not? It's got a quartz oscillator and an integrated circuit.

Well, for one, the escapement doesn't depend on the quartz crystal for timekeeping, but on magnetic induction.

Spring drive would theoretically work without the quartz components. Seeing how they would generate an electrical current because of the principle, they build a tiny generator to drive a circuit that adjusts the escapement, but doesn't run it.

Also, Spring Drive depends on a fully mechanical wheel train and barrel spring to function; it could not be made to work with a battery.

So what exactly constitutes mechanical and what does quartz here? Is it the wheel train? Does electro-mechanical count as mechanical? Is the mere presence of a quartz crystal in a secondary circuit enough to call it a quartz?

I found it easier to see it a as a new different system for timekeeping instead of trying to place it on an arbitrary scale between "fully mechanical" and "quartz."

> Well, for one, the escapement doesn't depend on the quartz crystal for timekeeping

No, it uses the quartz crystal for timekeeping. It keeps time like any other digital watch, only it outputs that time by controlling the rotor's speed, and gets its power from the rotor.

It could keep time mechanically instead and still use electronics to maintain the rotor's speed. But in fact it uses a quartz crystal.

How does the spring drive actually work? What is the mechanism which keeps it moving at a consistent speed? A quick google only turned up rather vague descriptions.

At the heart of every mechanical watch lies a coiled spring. Within that spring energy is stored that wants to be released.

This energy is released through the wheel train, which is a set of cogs with reductions (some of them attached to the hands).

Left to its own device this system would just spin as quickly as possible until the spring is unwound.

To prevent this, somewhere in the train there is something called a lever escapement. Simple put, this puts a stick in a cog at regular intervals to prevent it from spinning and unwinding freely. This stick is timed by a weight that spins from left to right. This stick literally is what makes watches tick, since that's where the ticking sound comes from.

Spring Drive doesn't poke a stick into a wheel to prevent the system from freely unwinding. What it does do is add magnets to the wheel and stationary coils around it. The wheel can't spin freely since it has to overcome the magnetic induction as the magnets pass the coils. That's the reason SpringDrive doesn't tick and the hands glide instead of stopping and starting.

If it's still unclear, try looking up escapement levers first. There should be a lot of clear animated explanations online.

I understand (roughly!) the basic idea of a watch mechanism, and how mechanical escapements work.

What i don't understand is exactly how the spring drive keeps the wheel moving at a constant speed.

I appreciate that this involves magnets, and eddy currents induced by those magnets, but that's a description, not an explanation. What is the mechanism, physically and mathemetically, that keeps the speed constant?

From a physics angle we're mostly dealing with electromagnetic induction. Faraday's law of electromagnetic induction states:

"that a voltage is induced in a circuit whenever relative motion exists between a conductor and a magnetic field and that the magnitude of this voltage is proportional to the rate of change of the flux".

The magnetic field here is attached to the glide wheel (rotor). The conductor is some stationary spools placed near the glide wheel. This means the voltage output depends on how fast the glide wheel spins (rate of change of the flux).

The other part of Springdrive comes from Lenz’s Law of Electromagnetic Induction:

"the direction of an induced electro magnetic field is such that it will always opposes the change that is causing it"

That means that the current creates a electromagnetic field of its own that opposes the changes in the magnetic field that caused it.

So that opposing field depends on the rate of flux according to Faraday. That rate of flux directly correlates to the passage of time (the spinning of the glide wheel).

By adjusting the placement and configuration of the circuit (the spools) we can influence the flux, thereby influence the induced voltage, thereby influence the induced magnetic field that opposes the change that is causing it, which is the spinning of the rotor that correlates to the passage of time, to a degree that hopefully is around 1 second per second (or 1 rotation of the glidewheel per 1/8 second).

Note that if the glide wheel accelerates for some reason, the change of flux changes, the voltage changes, and the opposing magnetic field increases, forcing the speed of the glide wheel down. On the other hand, if the glide wheel slows down, the opposing magnetic field decreases and allows the glide wheel to spin faster.

So technically the "smooth" second hand isn't smooth at all, but slows down or speeds up about 8 times a second, so it's not exactly "at a constant speed," but since it's not a start-stop motion, it sure looks smooth to an observer.

Inertia. The torque applied by electromagnetic force on the rotor plus that of friction and air resistance equals the torque applied (indirectly through a gear train) by the mainspring. The net torque is zero and the rotor keeps going at the same rate of rotation.

The integrated circuit counts how many times the rotor has spun compared to how many quartz crystal oscillations have happened and adjusts the electromagnetic force somehow. My guess is by increasing/decreasing the power passing through some resistor (creating heat).

There's one basic question I couldn't figure out from the article. Where is the spring that you wind up and that stores the "power reserve of about 60 hours"? Is it part of the silicon? Or is it external? I figured out that the three crescent-shaped parts are springy ("flex laterally to provide a restoring force as the oscillator vibrates through six degrees of amplitude"), but it doesn't look like there's anything else that could stretch/relax over 60 hours. My sense of intuition is useless, of course, at this scale and with these materials.

This video may help, as you get to see the movement in action: https://youtu.be/xWh0p9Irznw

It doesn’t clearly show the mainspring, however at around 10:45 you can see the small gear spinning. That is being driven by the mainspring, and if you removed the escapement it would spin super quick until the mainspring was unwound.

It’s the constant rocking back and forth of the regulator that limits the speed that small gear (called the escape wheel) can rotate, therefore limiting how quickly the mainspring releases it’s energy.

Indeed it helped. Thanks.

That spring will be connected to drive the escapement. The resonator is the equivalent of a pendulum, even though it contains a spring. The escapement wheel is driven by a non-constant force from the drive spring. The resonator regulates its motion to tick at a regular rate, and in so doing draws power to keep resonating. The rest of the watch is driven from the same source.

The energy is stored in a big spiral spring, the mainspring, as in a conventional mechanical watch. The oscillator does away with another spring, the balance spring.

Not only is the content incredibly interesting, but that website layout is perfect. So clean.

The ads page is interesting:

> The typical HODINKEE reader holds a master’s degree or higher, browses from his professional office, earns more than $200,000 per year, owns seven watches, and purchases, on average, three watches per year with an average value of $7,000 or higher per watch.

An attrition rate of three $7,000 watches a year is a lot!

Hodinkee is a great site! Some of the videos of people discussing their collections and stories are fantastic and well worth checking out.

Do bear in mind though, even very wealthy collectors also regularly sell pieces and 'trade up' over time. You can only wear one watch at time after all... unless you decide to start 'Schwarzkopfing'

A nit: the article states "A guitar string is a harmonic oscillator; no matter how soft or hard you pluck it, a G string is going to play a G note."

It's not that simple. I have a steel string guitar which, when a string is stuck loudly, will play sharp, then asymptotically approach the Hz it has when sounding its quietest. I hear it, and my digital tuner displays the effect as well. I speculate that when the the string is more perturbed (arced further away from straight), the tension increases more than proportionately to the displacement from straight.

Much more practical but much less accurate than an atomic wristwatch[1]. I wonder how small a cesium frequency standard could be made. Miniaturization isn't typically a priority in their design, there's surely room for improvement.

[1] http://www.leapsecond.com/pages/atomic-bill/ (Yes, that is a guy with an HP-5071A strapped to his arm.)

High-precision quartz wristwatches with well-designed temperature compensation can get accuracy within single-digit seconds of drift per year, and even that level of accuracy is enough to make them a niche product. Also, you need pretty good battery life in order to ensure the battery doesn't run out before the drift becomes relevant - some of the higher-end quartz movements already struggle with this a little, and chip-scale atomic clocks are much worse and likely to remain that way. Having accuracy of a second in 5,000 years doesn't give you much benefit over quartz because the battery life is measured in days.

You can also get a radio controlled watch

Of course, or a GPS-synchronized smart watch. But where's the fun in that?

It's likely that this is a paid post. And if that's the case, then I think it's money well-spent, as this is a beautiful post and an effective advertisement.

Only 10 have been made and all have been sold for some exorbant amount, so mostly it’s genuinely motivated coverage of some cool tech.

Zenith is also debuting a new material for the case: a composite called Aeronith. Aeronith is essentially an aluminum foam; melted aluminum is poured into a mold "where a procedure initially developed by Hublot transforms it into an open pore metal foam," according to Zenith. The voids in the foam are then filled with an "extremely light polymer" and the result is 2.7 times lighter than titanium, 1.7 times lighter than solid aluminum, and even 10% lighter than carbon fiber..

F1 engine tech, in a watch case, and that's not even the innovative part. Sooooo cool.

You've obviously never heard of the concept of a halo product.

Ford doesn't make GT's because they're profitable. They make them to sell Mustangs.

All this technology is amazing, but does it have any applications beyond mechanical timepieces?

Realistically, no. The design is very clever, but nothing is truly new here. You could have made this out of wood 200 years ago, but it would have been much bigger and less precise. Using silicon for a mechanical device is neat, but not new. All the technology in terms of the movements- springs, levers, escapements, wheels, gears- has been known for a while.

New designs are pretty much all digital (for better or for worse). If a spaceship needs a backup timepiece, it will just carry a second digital clock. If a spaceship needs a mechanical lockout, it will just do it with microswitches, solenoids, and digital logic.

Is silicon-material manufacturing now within reach of industries other than IC manufacturers and high-end watch manufacturers? If so, you could imagine all sorts of mechanical devices that will experience a renaissance now that they can be precisely designed and manufactured in tiny sizes and in huge numbers.

I suspect I just reinvented nanotechnology. For some reason, this watch article makes it feel much more real.

Have you heard of MEMS before?

There seems lots of cool applications for that, apparently also Timex is bringing out a watch which uses the technology in some way ( http://www.ablogtowatch.com/timex-silmach-watch-movements-me... ).

Also MEMS is used in DLP projectors, where lots of tiny little mirrors can be moved which seems rather amazing.

Ah, of course. Thanks. Yet another set of dots I hadn't connected.

From the Wikipedia article, one might even say this is a MEMS watch. MEMS describes devices up to 1mm having components up to 100 microns. This watch has parts as thin as 20 microns, so although it is overall larger than MEMS, its components encroach on MEMS territory.

No. Expensive watches are jewelry for men.

Accuracy of this watch has no real value for the owner. He just wants to own technical mobile.

Do analog watches have any applications to begin with, these days?

EMP resistance. This isn't exactly a practical justification for most uses, of course.

Would they be resistant to EMP, surely you've got lots of possibility for induced currents that will warp/melt fine components sufficiently to damage the watch?

A sundial is what you need.

I feel like most EMP damage on small electronics comes from blowing out small in-chip wires and transistors. I wonder if an EMP could induce enough current to overheat those little (but large by chip scale) silicon wires.

I wonder what are the mechanical properties of the case material (aluminium foam filled with plastic). This seems like entirely new class of composites.

This one is super coarse, there are much finer ones:


I've played around a bit with such super light metals and aerogels as well, they are very interesting materials with surprising properties.

So, these watches are all pre-sold and we can't buy one? When would be able to buy a watch with this innovative mechanism?

Eventually you will, hopefully it won't be as expensive as the gyrotourbillon that one is the most amazing mechanism I've seen.



Wow, that is truly amazing, are there large versions of the gyrotourbillon out of interest, as in for a clock.

The thing about clocks is, they sit in only one position, so there's no justification for a tourbillon, which averages out the errors caused by change of position.

Wasn't one of the first tourbillon mechanisms in a carriage clock though?

I thought this was pretty cool regarding tourbillons: http://www.laimer.ch/laimer_tourbillon

I'm not sure what that would be classed as though.

Edit: Oh, doh, I misunderstood you, yeah I see what you mean

Yeah, that would make sense. And (googles) you can get a tourbillon clock in your carriage today -- if your carriage is a Bentley Bentayga.

Haha, awesome. Since the tourbillon spring mechanism seems to be constantly rotated with the gyro version, I wonder if there is another way to mechanically keep the tourbillon always level?

Wouldn’t that render it useless though? The purpose of tourbillon is specifically to cancel the effect gravity has on watches or clocks that are stationary. This was a problem with pocket watches that had a mechanism always in the same direction. Tourbillons in wrist watches are largely only for decorative purpose as the hands move and thus the gravity effect averages out.

Some number of tiny swinging weights could be used to establish the direction of gravity and the escapement could be slowly rotated until it's perpendicular to that direction. Edit: Or let the escapement hang. There's probably reasons why that's difficult.

The usual answer to these types of questions is “In 20 years, when the patent expires.

"Patent? What's that?" - an engineer in Shenzhen

They only made 10 of them.

I'm guessing they will make other watches with this movement eventually.

If you want a watch with a neat movement that is widely available, get a Swatch Sistem 51. Granted it has 5x as many moving parts as the Zenith Defy Lab, but it is actually affordable and pretty neat.

Although Watch Guy [0] was not amused...

[0] http://watchguy.co.uk/review-a-trip-inside-the-swatch-sistem...

Wow, that really is some serious cost-cutting. As he says none of the Chinese manufacturers have gone that far in cost-cutting; even their cheapest mechanical watches have a fully metal movement that is theoretically serviceable as normal (and some of them are really cheap).

This movement is the output of a combined R&D lab for LVMH watch brands (Tag, Zenith, Hublot, Bulgari). So the movement will likely to be rolled out to watches by all of those brands.

If the whole thing is mostly based on etched silicon shouldn't they be as cheap and common as CPUs? /s

What interests me about this movement is that it's mechanically simpler than a conventional movement and should eventually be cheaper to produce. We might actually see these in relatively affordable watches rather than sold-out unobtanium stuff. This is a lot more exciting than yet another watch-maker piling on ridiculous numbers of complications to make a watch nobody will ever want to actually wear.

How durable would Silicon be in this application? Those teeth look awfully delicate.

Love the irony of how silicon is used.

Awesome write-up.

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