4. As outlined in the previous post, what customers want from Spring Drive today is different from 20 years ago. The desirable changes or improvements to the movement are:
a) No power reserve hand in the front (done)
b) Smaller size
c) Longer power reserve
d) No accuracy compromise vs. motor drive HAQ (partly done)

a) Rear power reserve (done)
There has been no new GS Spring Drive non-dive caliber with a front-mounted power reserve indicator since the 9R84 chronograph non-GMT in 2013. In 2016, the 9R01 debuted with the power reserve on the rear, followed by 9R02 and 9R31.

While 9RA5 provided this information on the dial for divers, 9RA2 and 9RB2 continue with power reserve on the rear. This is a direct response to changing customer needs.

b) and c) Smaller size and longer power reserve
These are discussed together because less space provided for the mainspring naturally means less energy storage, all else being equal. However, the worldwide market that GS is selling to now demands a range of case sizes that is difficult to serve with a single movement.

There are customers who today who enjoy a case of around 39-41 mm, while others desire a 36-38 mm watch. Today, the 9RA2 and 9RB2 are respectively positioned to serve these. The 9RA2 delivers 120 hours and 9RB2 72 hours of runtime.

Yet there is another dimension where even 9RA and 9RB are under pressure, which is thickness. Not just for the automatics, but even more importantly for svelte manual-wind calibers to follow the 9R31/9R02.

Post #6 will consider the limits of SD thinness and some Seiko patent applications which appear to address these.

However, if a thinner movement has too short of a power reserve as a result of also a narrower mainspring, customer convenience will be compromised. I don't think GS will accept a PR shorter than 72 hours for their mainstream SD movements, even resorting to a protruding barrel in the 9RB2.

Post #14 will examine some Seiko patent applications aimed at harvesting more runtime out of a mainspring's variable torque curve.

But before getting into the weeds, I want to emphasize the technical difficulty. The entire design of Spring Drive survives on a narrow intersection where acceptable values of torque, braking power, and electricity generation happen to overlap. Micro Artist Studio craftsman Yoshifusa Nakazawa alluded to this:

d) No-compromise accuracy (partially done)
Cal. 9RA and 9RB, nonwithstanding aging behavior of the quartz package, are engineered to be as accurate as 9F quartz - within 10 seconds a year. But being a new product, there are concerns about not meeting engineering objectives.

Five watchful years after 9RA5, the launch of Spring Drive UFA promising 20 seconds a year (20 spy) accuracy struck a compromise which captured the watch world's imagination and started a quiet conversation about autonomous accuracy as an aesthetic value even in the luxury market. (An ongoing conversation in specialized circles for decades.)

20 spy was a 'King Quartz', the lowest of the three tiers of Twin Quartz, slotting below Grand Quartz at 10 spy and Superior at 5 spy. The Suwa version of Twin Quartz, which is the closest precurosor to Seiko's modern high accuracy quartz (HAQ) was coincidentally also the brainchild of Yoshikazu Akahane.

The ultimate Spring Drive would therefore be one that deliveres 10 spy in mainstream calibers and 5 spy in limited editions, just like 9F quartz. That would unite two threads of Akahane's career: the pursuit of quartz accuracy and the pursuit of the ultimate mechanical watch.

But there is no commercial point of pursuing this in the present decade, if the other needs of the customer remain unmet, namely long(-enough) runtime within compact dimensions.

 

I'll never stop laughing at watch geeks acting as if getting battery changes is the greatest inconvenience they'll ever experience.

 

It depends on the situation. I have multiple spare batteries and tools at home, but I am not a normal watch owner.

In Seiko Epson's words (PDF attached):

Quartz watches had batteries that needed to be replaced, since their drive systems relied on battery power. This reliance on batteries was an issue from the start of quartz watch development, as there were concerns about issues such as the watches' practicability in regions where batteries would be difficult to obtain and the disposal of used batteries. Ever since the advent of the quartz watch, Epson had been laboring to develop a product that would overcome these obstacles.

Coin cell disposal can be haphazard in some regions. Despite official policy, most get dumped into the landfill. Being small and shiny, a few are then ingested by the children of scavengers from time to time. The rest leak out over a longer period of time.

 

Your content is always great to read as always.

For me, (just my worthless opinion) one wonders if they have new complications in mind.

They've done GMT, 3 hander, Chronograph / Chronograph GMT.

I would love to see a complete/annual calender, even a perpetual, or a moonphase.

A thinner chronograph with a more conventional subdial layout would be welcome too.

Of course just wishful thinking.

 

Apparently they have made a Spring Drive Moonphase. The reason we don't hear about it was it was never made for GS but Seiko SNR017. Learn something new everyday.

Seven Seiko Spring Drive watches you should know about

We take a look at seven Seiko Spring Drive watches, from both the past and present catalogue, we think you should know about.

timeandtidewatches.com

There was also apparently a Credor Spring Drive Big Date.

This Spring Drive Complication is Exclusive to Credor

Credor once again asserts its dominance over Grand Seiko

provenancewatches.substack.com

 

Twin coils are the thickest components of Spring Drive movements today, and appear to be largely unchanged from 1999. They are illustrated in a 2000 paper by Kunio Koike marked as 'coil block' on the right:

Photographed in the Eichi II's 7R14:

and, taken with the conductive cores, as large as the barrel of the 9RB2 and no smaller than the 20-year old 9R65's:

Koike's cross-section shows that the coils are almost as thick as the entire movement.

However, thanks to @Whiskey&Watch, a Seiko patent application came to this forum's attention. This document illustrated different arrangements and sizes of coil. The objective is to increase power generation efficiency so that "the electronically controlled mechanical watch 1 can be made smaller."

The 'first embodiment' from that document with elongated (i.e. thinner) coils.

In a key difference from classic SD, while a first coil still handles both braking and power generation, the second coil is always in power generation mode as the mainspring unwinds.

Further embodiments use a second glide wheel spinning at a faster rate to further increase usable power generation of the second coil.

A key problem that this overcomes is that, contrary to popular belief, braking per se does not generate power. It is necessary to 'chop' the braking in order to generate current. The proportion of braking time is referred to the duty cycle. Therefore, the higher braking force needed, such as near full wind, the higher the braking duty cycle and lower the electrical power generated.

I believe this exists at least as a prototype similar to the 1997 pre-7R88 shown by GS below:

However, there is a possible further related improvement, disclosed in a different patent application and the subject of my next post in a day or two.

 

I would think that the main focus for SD should be complications, power reserve and perhaps size.
There is no need to improve the accuracy since the SD always performs much better than the stated accuracy. I have owned 5R55, 5R56, 9R55, 9R56 and 9RA2 movements. And the non-thermocompensated ones all ran between +2 to +5 seconds a month. The 9RA2 ran +1 second over three months. I have never heard of anyone reporting an SD not having much better accuracy than the stated spec.

 

A more compact base movement with a healthy torque headroom would enable useful and/or cool complications, which I agree would help round out the product line. I should have mentioned complications earlier

 

Thank you so much for your excellent synopsis on the development of spring drive and the market forces driving it, and your insights into its future. Much appreciated 👏

 

I'd rather seiko delivers significantly better surface hardening and a completely toolless bracelet with micro adjust. Seiko bracelet is underwhelming when next to an iwc, GO and minase. Add a complimentary rubber strap and tang buckle will allow customers to change from bracelet to straps easily while adding very little to the cost.

 

Thanks for this, it's very insightful.

 

A little detour in a comparison of Spring Drive’s runtime efficiency. Below plots nominal volume of the movement in cc on the X axis and runtime in hours on the Y axis, for three main SD 3-hander calibers, their ostensible Rolex and Omega peers, and a selection of others

5 hz beat rate and SD are colored black
4 hz orange
3.5 hz red
3 hz blue

Data

The data shows that SD are comparable to peers, with 9RA2 a standout (almost 27 hours per cc) bested only by microrotors or 3 hz calibers.

Put another way, even the 21 year old 9R65 has been delivering decent runtime relative to its diameter and thickness.

 

The question is how long before we will finally see a smaller and thinner diver or gmt using the 9rbx? I hope next w&w, but unfortunately i wouldnt be surprised if they take 5 years.

 

Depends on sales more than technology, I think. SLGB003 seems to be selling well.

edited to add: I refer to 9RB2 in dive cases, or a ‘9RB6’ GMT variant, not to a whole new caliber. Any whole new caliber is still R&D dependent at this stage IMO

 

5. That mainsprings decline in torque during unwinding has been the bane of mechanical watchmaking from the very beginning.

This problem also affects Spring Drive, but in a different way:

a) At full wind and maximum torque, SD braking has to be oversized, otherwise the braking duty cycle would be too high to generate enough power for the IC.​

​

b) At low torque, SD has to mechanically stop the unwind*, otherwise the IC would be starved of power and the watch would begin to run very inaccurately.​

Putting this picture together, the usable runtime of a SD mainspring is truncated at both ends where torque becomes too high for braking and where it becomes too low for power generation.

Both restrictions are related to the IC's need for a constant supply of power. Therefore, as early as 1999, Seiko Epson filed a patent application (JP2000214271A) for a version of SD where an "additional power supply device" supplements the directly-generated power from the SD glide weel.

In the present invention as described above, when the generated voltage of the generator drops and the voltage of the electrical storage device drops to the set voltage or less, the power supply control device controls the additional power supply switch to supply electric power from the additional power supply device​

This was not referring to the capacitor that is present in all SD to date including the 9R65, but to something like a rechargeable battery charged by a solar cell.

This invention was not commercialized, a quarter-century passed, then Seiko Epson filed a new application** citing the first.

Now in English, Seiko Epson provides a diagram showing SD from left to center and then, at right, a "second power accumulation device" (#32) which is connected to the rest of the circuit by a "switch circuit" (#33):

The second power accumulation device is configured by a rechargeable secondary battery such as a lithium-ion battery and a solid-state buttery, and has a capacity larger than the first power accumulation device....​

​

Charging is selective: only conducted during high torque state, helpfully decreasing the 'chopped' braking requirement by introducing generation resistance:

​

the switch circuit is provided, and the second power accumulation device is configured to perform charging when there is an excessive amount of the output torque T of the mainspring. Thus, reduction of the voltage of the power supply circuit can be prevented, and an effect of increasing the duration [power reserve of the watch] can be exerted for a long time period.​

To reiterate that this is not 'classic' SD:

as compared to when only the first power accumulation device [a capacitor] is provided, the duration in which the hands can be subjected to correct speed regulation can be increased. ​

What it means for Spring Drive watches
First, this is not certain to be commercialized, but the effort to 'resurrect' the intellectual property is notable.

Rechargeable battery technology is now a lot more mature, with some chemistries having 20 year track records. Having a battery around could enable additional features, notably an elegant and foolproof perpetual calendar.

The 'smaller coils' invention of my previous main post includes a separate charging-only coil. What better synergy than to dedicate that coil to charging the battery, while taking it offline during low torque state?

These two inventions together could allow a 27 mm x 4.5 mm movement***, extracting a 72 hour power reserve out of a smaller barrel.

That would mean a 36 mm Spring Drive watch, or even as compact as 34 mm with effort. Thickness around 11 mm without compromising hand spacing.

With GMT complication, thickness could be around 12 to 12.5 mm.

To recap
Spring Drive is a means to a watchmaking end: smooth glide with the convenience of automatic winding and a long power reserve.

Market tastes during 2004-2024 allowed a temporary respite from the size constraint: Classic SD is bigger than the average bear.

But the time has come (at least outside the US market) for a smaller Spring Drive, with SLGB003 representing a good start. The technology to go even further seems to exist, but the question is if the reliability and manufacturability are good enough to proceed to mass production.


------
*This mechanical stop is why all SD movements have a PR indicator.

The definitive paper on torque in classic SD is SPRING DRIVE watch overwinding prevention and hand stop function denoted Hara (2001) containing this chart for the 7R mainspring:

Tsys is the minimum torque to run the IC. The watch is mechanically stopped by the PR works at some point higher than N2 winds, there the torque is a small margin higher than Tsys. With the new invention, supplementary battery power would allow unwinding below Tsys.

The paper calls this the 'lock mechanism':

**The improvement over the 1999 application is that charging is done from the mainspring: "A power generator for an additional power source is not required. As a result, the movement can be prevented from being increased in size, and a degree of freedom in design can be improved."


***The Omega cal. 8800 is 26.0 x 4.6 mm, while the new Rolex 7135 Dynapulse is 28.0 x 4.7 mm.

 

With the caveat that no single patent application indicates a future product, here are my prognostications on the future technological path of Spring Drive, with explanations based on both Seiko patent filings and my reading of the 2025 watch market.

View attachment 19307218 (from Akio Naito presentation during W&W 2025)

1. SD has always been a response to customer needs, as expressed by market forces
By the middle of the 80s, the broad strokes of motor-driven quartz had been sketched out. Accuracy has been tamed, first with analog thermocompensation (TC) to better than a minute a year, then with Twin Quartz, then in-IC TC.

Functionality was unlimited by the use of either digital displays (Seiko LC) or multiple motors (7A28) or both.

Miniaturization, which is the converse of functionality, was a matter of time (and Moore's law) as the 1.6 mm thin HAQ 8N41 proved.

The last frontier was customer convenience. Mid-80s quartz watches came with the major inconvenience of battery changes. It did not help that the fashion was thin and small, which meant the 16-fold efficiency improvement since the 70s was 'wasted' on having to fit a 1/4th sized battery.

(Another inconvenience was the non-perpetual calendar, but this was only 5 times a year with the benefit to the crown gasket of a little movement.)

To that generation of engineers, solar cells and rechargeable batteries had unknown longevity, while the mainspring was proven as energy storage, if it could be harnessed.

Spring Drive as envisioned by Yoshikazu Akahane and his colleagues was a solution to the inconvenience of the chemical battery. While of course being also chronometrically respectable in the context of the late 70s, i.e. matching ordinary quartz at +/-15 s/month.

But would the customer accept a large watch after being accustomed to slim models? No one knew, and the size of early SD prototypes didn't help.

Seiko therefore first tried another approach: the 1988 A.G.S. These movements fit expectations of the time at 27.6 mm diameter. Those first batteries were good for 75 hours runtime. When new.

Improved and marketed as 'Kinetic' in 1997, one can appreciate how long Seiko pursued this approach to the battery replacement problem. (Far longer than Swatch Group with its 'autoquartz'.)
View attachment 19307212

2. Market changes made SD not just technologically but also commercially feasible
By the mid-90s, marked by the Royal Oak Offshore 'Beast' in 1993, it was clear to any market researcher that thin wasn't as important at the upper reaches of the market as it was in the 80s. Watch enthusiasts wanted to wear a substantial-feeling and substantial-looking wristwatch.

The market portal had therefore opened to industrialize SD, first as the cal. 7R in 1999, then as the hefty automatic 9R65 in 2004.

Here was a quartz-accurate wristwatch, with automatic winding and power reserve sufficient for a weekend off the wrist.

The steel SBGA001/003 were 39 mm and quite thick for a quartz, but that was ok, or even a good thing.

Grand Seiko had relaunched with motor-drive quartz in 1993, then 9S automatic in 1998, and even a GMT in 2002. Finally, a synthesis of both.

9R6 would be the flagship for the next 16 years. 9R66 GMT followed close behind in 2006 and the mighty 9R86 chronograph GMT just a year later.

9R was naturally the engine inside the largest GS cases, debuting as 2016's SBGE037/039 and SBGC015/017. But that was near Peak Big Watch, viewed in hindsight, so those 46 mm models did not become the norm for high end GS sport watches.
View attachment 19307211

3. Market changes again after 2016.
At some point, the main selling point of Spring Drive became no longer the accuracy, but rather its perfectly smooth sweep of the hands. This included the chronograph, which is the only wrist chronograph with infinite resolution in the world.

The dial also gained prominence, and with it the negative reactions to the dial-side PR hand. Model nicknames like 'Snowflake' or 'Tatami' predominantly referred to the dial (not case or dial furniture) of the watch.

In all of these, customer needs were changing. The movement per se mattered less to the 2010's customer. The 'mechanical purists' were being outnumbered by those enjoyed the visual beauty of the watch in motion.

Not how the machine inside functioned to move the hands, but how the hands moved. That's why the old customers generally accepted the PR dial, but many new customers objected.

The new focus on watch appearance (vs movement geekery) is why unique dials have become such a GS selling point, an advantage skillfully used by GS to also sell many SBGH and SBGJ models to collectors.

The long-term implications of this change for Spring Drive will be discussed in post #2.

Fantastic write up. Though I would suggest that the Spring Drive chronograph is not the first with infinite resolution - I’d suggest the Omega Speedsonic with the accutron movement base (chronograph module being built on top) could be argued as the first.

 

Hopefully it is clear why my OP said "thin and hybrid." A new coil design, plus a tiny battery* which harvests the mainspring when it's too strong and returns the energy when it is too weak, could allow a smaller 3 day Spring Drive.

They've done GMT, 3 hander, Chronograph / Chronograph GMT.

I would love to see a complete/annual calender, even a perpetual, or a moonphase.

A thinner chronograph with a more conventional subdial layout would be welcome too.

Seiko has chrono modules, calendar modules, moonphase, you name it.

But will the complicated watch sell? Depends on market demand, which has been trending away from complications as people enjoy a pretty (and uncluttered) dial.

A more compact base caliber makes completions possible while keeping the overall size down.

I would be wary of any battery inside SD gen 3 being used for more than the bare minimum, but I must acknowledge that it is possible, and very space-saving, to use this battery to power motorized complications such as a perpetual calendar.

Maybe tastes will turn to no-fuss complications which don't clutter the dial.

a) There's no cleaner perpetual calendar than a quartz one.​

​

b) I'd also love a travel time hour hand that tucks under the main one. Seiko's had that in the early 70s Duotime:​

There is no need to improve the accuracy since the SD always performs much better than the stated accuracy.

9RA is observed to deliver UFA-level accuracy, but people would be happier with 'UFA' on the dial.

I'd rather seiko delivers significantly better surface hardening and a completely toolless bracelet with micro adjust

I'd like that, too.

I would suggest that the Spring Drive chronograph is not the first with infinite resolution - I’d suggest the Omega Speedsonic with the accutron movement base (chronograph module being built on top) could be argued as the first.

I'd say the accutron-based chronograph is visually stepless, but the underlying mechanics still have a very fine step. Spring Drive is fundamentally stepless which we could consider to be merely an aesthetic difference. @BrianBinFL has a detailed post with video comparison.

----
*Much smaller than a normal solar watch battery, because this one doesn't have to move any hands.

 

Fantastic write up. Though I would suggest that the Spring Drive chronograph is not the first with infinite resolution - I’d suggest the Omega Speedsonic with the accutron movement base (chronograph module being built on top) could be argued as the first.

I'd say the accutron-based chronograph is visually stepless, but the underlying mechanics still have a very fine step. Spring Drive is fundamentally stepless which we could consider to be merely an aesthetic difference. @BrianBinFL has a detailed post with video comparison.

The video that @Mandaue linked to is comparing Spring Drive to the 16 steps-per-second Precisionist (the same movement used in the Accutron II watches, but not the original Accutron watches). @IsaaH, the Accutron movement you refer to is one of the original tuning fork movements - though looking at photos of it I don't recognize the base movement that they modified.

In any event, all we need to know is that it is a tuning fork movement. All tuning fork movements have steps - albeit very tiny ones. All of the tuning fork movements use the movement of the forks to drive a little pawl that advances an index wheel with impossibly small teeth. The frequency rate of most (not all) of the tuning fork movements was 360 Hz - hence there were 360 steps per second. Not that those are REALLY tiny steps. So tiny in fact that I don't have a high speed camera fast enough to capture the steps. But if one had a 1000 FPS camera, and made a video of the movement of the seconds hand, presumably you'd be able to see the steps.

In any event, the tuning fork movement, though awesome (I have several), has a LOT of stopping and starting and bidirectional motion.

As Mandaue stated, the Spring Drive has literally zero steps. No stopping and starting. No bidirectional movement. Everything spins one direction and one direction only, continuously, smoothly, without ever stopping. At present there is no other such wrist-worn timekeeping device that I am aware of.

 

@Mandaue, they filed a new patent application on recovering eddy current loss-based energy from the two-inductor coils; this may push their PR beyond 120 hours in a significantly slimmer case design.

 

Very interesting. Do you have an image or application #?

 

OP, and everyone else that commented, thank you for an insight into what is happening at the cutting edge of watchmaking. Threads like these is why I joined WUS 14 years ago.

 

Fascinating reading. As a first-time spring drive customer in 2022 (you can probably guess the model) I had no idea that other companies e.g. Swatch had produced auto-quartz watches. What was also fascinating was I'd never considered that the electricity powering the motor could power a battery - wikipedia mentions a full charge lasting 6 months (though no citation). Can you say why SD wasn't coupled with a battery? Was that a step too far? Would having a battery really blur the lines of what is otherwise a mechanical watch? You're dead right though - as a contemporary customer I was looking at the dials and case finishing first and foremost, however at some point in my journey I decided that whatever GS I bought was going to be a SD. My only other watch is mechanical so it just felt so different.

 

I can’t say for sure why not, but I think the focus in 1999 was a Minimum Viable Product.

That first 7R88 ran for 48 hours, already matching or bettering many calibers of the time, such as the ETA 28xx, GP 3xxx, Rolex 30xx, to say nothing of the Peseux 7001.

A secondary battery was an additional thing that could go wrong. And battery chemistry was nowhere as good as it is now

 

We are going to be getting lots of SLGB models in the next couple of years, including a smaller "UFA Ushio" diver soon.

 

Interesting! Thanks for the trademark application link.

 

During the latest GS9 event that I went to (in Sept 2025), Joe Kirk talked about a thinner Springdrive called the 9RB2

Caliber 9RB2 | Grand Seiko

Grand Seiko official site. Made by hand for those who value perfection. Grand Seiko raises the pure essentials of watchmaking to the level of art.

www.grand-seiko.com

It looks impressive. I like Springdrive as it reminds me of the Accutrons of 1960-1977 except that they don't hum (although, to me the Accutron hum sounds more like the sound a metal crushing machine makes when it is tearing up metal)

 

That’s the movement in the SLGB001/003/005 models. Thinner than the 9R65 due to offset magic lever, and incorporating thermocompensation in the IC like the 9RA2 and 9RA5, but otherwise electrically and mechanically similar to 9R65. This thread is about calibers beyond that.

 

Unfortunately, I didn't keep a record this time as I kinda lost interest in the new SD after seeing the crazy price hike.
Will keep looking to find it again.