August 2013

Around 2006, when ETA was taking its first steps to reduce movement deliveries, a small group of people came together, like conspirators. Among them was Andreas Felsl, a German entrepreneur specialised in patents who invented, among other things, Bionicon, a shock-absorber system used in the world’s best mountain bikes, and who created in Germany a bicycle company that was elected the best in its category on six consecutive occasions.
There was also Tzuyu Huang, a businesswoman based in Bienne and CEO of Momoplus AG, a blossoming company delivering watchmaking components, cases, dials, hands, straps… in short everything but movements. And it is these very movements that were set to become rarer and cause supply bottlenecks.
They therefore had the idea of starting out on a crazy adventure: what if we designed, constructed and produced our own base movement?! Competitively and produced on an industrial scale! At least 100,000, that’s surely a big challenge. A steep north face to climb… hence the name K1 given to this veritable expedition.

Andreas Felsl and Tzuyu Huang
Andreas Felsl and Tzuyu Huang
Achim Huber, Jonas Nydegger and Stephan Kussmaul
Achim Huber, Jonas Nydegger and Stephan Kussmaul

 ACCURAT SWISS - Operation K1

The company Accurat Swiss AG was thus established, two engineers and a strategy consultant soon joining the team, which worked with the utmost discretion. Among them was Achim Huber, an industrial strategy consultant, who was in charge of steering the development of the project. Stephan Kussmaul, a watchmaker and engineer, looked after the movement itself.
Before joining Accurat Swiss in 2009, Stephan Kussmaul was the head of R&D at Eterna, in charge of movement development at the brand that was, ironically, behind the creation of ETA. Before this, he had worked as a watchmaker at Patek Philippe. Working alongside him was Jonas Nydegger, another former employee of Eterna, where he was responsible for developing the processes for mass production of movements.


So the team’s objective is to develop and industrialise the production of a base movement, to which new functions could easily be added. A movement that is by no means a mere “clone” but whose encasing is nevertheless ETA-compatible. And, above all, a quality movement that can be sold at a competitive price. There are numerous sceptics and the search for venture capital is difficult. “You’re mad, you need 100 million” is the standard reply. So they decided to go it alone!

With the benefit of hindsight, Andreas Felsl and his team think that they were lucky not to have found a big investor: “we approached things more modestly, step by step, with full autonomy and self-sufficiency, without getting ahead of ourselves. Today, we are reaping the rewards.”
The slow ascent of this K1 has not been without its difficulties, however. While the small but powerful Armin Strom factory opened its doors to them so that they could machine the components for the first prototypes, the problem of the assortment was a lot more serious. After two years of fruitless research, Andreas Felsl, who does not come from the world of watchmaking and therefore does not have the same inhibitions as watchmakers, started to take a closer look at the patents registered for assortments in silicon. He noticed that, although everything had been done to prevent others from using this technology, there were a number of deficiencies. “We did a lot of research and found new solutions that did not encroach on existing patents. We did all we could to avoid being taken to court, we found a producer in Germany that pioneered silicon technology and we have managed to produce our own balance spring and escape wheel in silicon.”


A few years have passed and we are in mid-2013 and the K1 now starts its real existence with a first series of 40 prototype movements that are currently undergoing tests. But what does the K1 look like?
At a diameter of 25.4 mm – one of the most common diameters in use – and a little thicker (4.95 mm) than the ETA 2824, the K1 is a self-winding movement comprising 110 components for its basic configuration and up to 160 components for the most complicated version. It has a minimum power reserve of 45 hours and beats at the rather unusual frequency of 3.5 Hz, or 25,200 vibrations per hour.


The K1 movement by Accurat Swiss
The K1 is a platform that can be configured in 18 different versions, ranging from 110 components for the basic configuration up to 160 components for the most complicated version. It offers a power reserve of 45 hours and beats at a frequency of 3.5 Hz, or 25,200 vibrations per hour. The main benefits of the platform lie in its flexibility, with the majority of components being common to the different configurations. Modules for the different complications are pre-assembled and can be fitted to the base plate once it has already been assembled, adjusted and checked.

The main idea is to make it a platform that can easily be configured in 18 different versions, subdivided into three families: a big date family, a small date family and a family without a date. The small and big dates are placed at 3 o’clock, the power reserve display at 6 o’clock and the seconds are either central or small seconds at 9 o’clock.
The big advantage of this platform is the great flexibility that it offers. The great majority of components are common to the 18 different configurations possible and can therefore be produced in large quantities. The components specific to the different versions – small seconds, central seconds, power reserve, small or big date – are produced and pre-assembled as specific modules. Once checked, they are fitted directly on to the base plate, which has already been assembled, adjusted, checked and stocked.


This typically industrial-scale organisation does indeed offer numerous advantages. The fact that the basic design can be configured in 18 different versions allows customer demands to be taken into account using a “just in time” approach that is quick, reactive and flexible. Not to mention the economies of scale that make the finished product even more competitive.
And because the architecture of the movement is common to all versions, after-sales service work is also made much easier. This is no small advantage at the moment, whether for the end customer, the retailer, the brand or the producer.

As a next step, Accurat Swiss is on the verge of launching a first pre-series of 1,000. The feedback and experience from this pre-series will determine the future investment. One to watch very closely, therefore, in this period of impending dearth.


Europa Star: Why the chosen frequency of 3.5Hz (which is the co-axial movement frequency)?

Stephan Kussmaul: We think 3.5Hz is a modern frequency and are convinced that this frequency offers the right trade-off between dynamic performance and wear in our movement platform K1. Furthermore 3.5Hz gives us the opportunity to show that Accurat Swiss is different from others and we are following our own way.
Traditional frequencies are 2Hz, 2.5Hz, 3Hz, 4Hz and 5Hz. From a technical viewpoint, there is no reason for any limitation to these frequencies only. Movements can work perfectly with any frequency in between. We see a lot of 4Hz movements in the market, because in the past brands simply used ETA movements as the motor in their watches. Other companies use the same assortments as ETA from Nivarox for their regulating system in their proprietary movements because they have been widely available for a reasonable price up till today. But the situation is changing right now. Companies have to invest in their own technology and infrastructure to ensure their own supply of regulating systems. We are convinced that we will see more 3.5Hz movements in the future from other brands as well.

Stephan Kussmaul
Stephan Kussmaul

Is the self-winding mechanism uni- or bi-directional? Is it equipped with ceramic ball bearings?

SK: The self-winding mechanism in our movement is uni-directional because of reliability and simplicity. It uses less components than bi-directional self-winding mechanisms. Less parts means less risk of loose, lost or defective ones. At the moment we use steel balls in our ball-bearings for cost reasons. We are open to use carbide or ceramic-bearings in higher priced versions of our movements in the future.

Can you give us more details on the silicon hairspring and regulating organs? Is the balance-wheel your own or have you adapted an existing one?

SK: In the first two years of our movement project we worked with traditional regulating systems from well-known suppliers in the industry. At one point it became clear that our own supply would no longer be guaranteed in the future. At that time we decided to start researching other solutions. The hairspring has been developed from scratch together with our partner from the industry. It took us around 2 years of development time to recalculate the traditional metal hairspring we used before and build up our own know-how also by learning from our mistakes. Apart from the silicon hairspring, escapement and escapement-wheel, all the other components are made of traditional materials such as steel and copper-beryllium. The original balance wheel, balance staff and escapement staff have been modified to meet the changed requirements.

You speak about “new solutions”. Can you tell us more about them?

SK: Watch, listen and learn from the best in the watch industry and try to make it a tiny bit better.

Same with the gear train? Have you designed new teeth?

Today, in modern mechanical wristwatches we find mainly three different gear arrangements:

1. movement with direct minute hand in the centre and small second hand out of the centre (such as ETA Cal. 6497 & ETA Cal. 6498)

2. movement with direct minute hand in the centre and second hand in the centre (such as Rolex Cal. 3135)

3. movement with indirect minute hand and second in the centre (such as ETA Cal. 2824 & ETA Cal. 2892)

Some decades ago there was another gear arrangement (such as Felsa 756) where the designers of the movement put an additional wheel in a movement according to arrangement No.1. This additional wheel is fixed on the top of third wheel. The new wheel transfers the rotation to a small pinion in the centre, which holds the seconds hand on the dial side. This has been done to convert a traditional movement with a small second hand arrangement rapidly into a more modern movement with a second hand in the centre. To guarantee a steady rotational motion on the second hand without flickering it is necessary to use a friction generated by a spring. The gear train in the K1 movement has been positioned to meet our platform requirements. We have chosen an arrangement which is similar to Rolex Cal. 3135 in combination with some elements from the old Felsa Cal. 756 movement: We aligned our gear train in a way that allows us to put a small pinion at exactly 9 o’clock which holds the small seconds hand. This pinion takes its energy from the third wheel and is inserted from the dial side towards the end of the assembly process according to customer request. The main flow of force passes always by the third wheel in the centre to the escapement wheel. Because the third wheel is always present, the axle of the third wheel in the centre does not exceed the height of the cannon-pinion. In case the chosen movement configuration does not have a second hand from the centre, the minute hand will simply cover the axle in the centre. Therefore in our case the gear train arrangement in our movement is: a direct minute hand in the centre with a second hand wheel in the centre AND a small indirect second hand pinion out of the centre (when needed). With this configuration we guarantee a steady rotational motion on the large second hand without flickering in the centre and if needed a small second hand at 9 o’clock slightly flickering but hardly visible because of the short length of the hand (we do not want to add a friction to it).

To answer your second question: We were working together with Mr Michel Belot for the exact shape of the teeth on all our wheels. These geometries are industry proven like the ETA standards (from Heinrich Stamm, 1969) as well.

You say that you are “taking all the energy for additional functions directly from the barrel, without altering the movement works”. Can you say more about it?

SK: The force and additional module takes is always the same, independent of its position in the gear train. We try to take all the energy as early as possible, to keep the influence low at the end on the regulating system. The energy for the power reserve indication we take directly from the barrel (as ETA Cal. 2897). For the calendar mechanism we take it from the centre pinion (as Rolex Cal. 3135). Finally the module for the small seconds at 9 o’clock take the energy from the third wheel. The calendar and the power reserve module consume more energy than the module of the small seconds at 9 o’clock.

Which metal are the main movement parts made of?

SK: Apart from silicon parts we use traditional materials as steel, brass and copper-beryllium and bronze:

• Brass CuZn38Pb2 for the main plate and bridges

• Steel Sandvik 20AP for turning parts as pinions and axis

• Copper-beryllium for the balance wheel

• Bronze for some bearing bushes and bearing pins

You say that from the beginning, everything was done with the objective of an industrialization process. Can you give us more detailed examples of this approach?

SK: As an engineer not only working in the watch industry I have seen it several times: engineers and product designers develop a nice product which unfortunately does not fulfil manufacturing, assembly or maintenance requirements. Finally, the product is produced in an inefficient way till the company decides either to drop it completely or make serious modifications to meet at least minimum requirements. Industrialization is not rocket science, it is about listening and learning. In the beginning we started to engineer the movement in a straightforward way. As soon we had a rough 3D model with very rudimentary 2D blueprints, we started to get in contact with suppliers and machine manufactures etc. We presented our semi-finished engineering work in 3D and on paper and discussed the how we could produce each single part of the movement. We always left meetings with specialists with more knowledge than before and this knowledge led to design modifications on the parts. Doing all this was not a single step and is still an ongoing process. It is also not limited to production and manufacturing machinery. The movement has to be assembled in semi-automated assembly lines as well, so we also visited Lecureux and other companies over and over again. Maintenance is also a huge subject for mechanical watches, so we gave our first series of prototype to untrained watchmakers to identify possible difficulties during manual assembly and disassembly. Finally, all this leads to design modifications which make the product better, better and better, step by step. I can give you here just some rudimentary engineering rules (out of DFA Design For Assembly and DFM Design For Manufacturing-methodologies) in order to use automated assembly lines:

• A part is inserted from the top of the assembly: This motion can easily be executed by a hydraulic cylinder, the gravity helps to stabilise the part and the assembly personnel can generally see the assembly location.

• A part is self-aligning: Parts that require fine positioning in order to be assembled require slow, precise movements. The most common self-alignment feature is the chamfer on cylinder bolts and holes.

• A part is assembled in a single, linear motion: Pushing in a cylinder bolt needs less time than tightening a screw or riveting two parts together.

• A part is secured immediately after insertion: Until the part is secured, the assembly may be unstable, requiring special attention, fixtures, or slower assembly

• Prevent hidden oiling points: Hidden oiling points lead to more complex, slower and probably manual assembly procedures.

The manufacturing and assembly technology will change with increasing quantity. Where we can use wire-electro discharge machining for smaller quantities now on sheet metal components, we have to use stamping processes later on. This means high investments in tooling, but finally much lower costs per piece. As a result we already have to design the components and mechanisms under consideration of technologies we will use later on in higher quantities. When we do not do that we might be limited one time because we have no space or no possibility of such a serious design change. We are now at a point we have learnt a lot about processes. During the last four years we have identified all the necessary manufacturing processes involved to produce each single part in the movement. How many and what kind of different machinery is exactly involved in the processing of each specific part from raw material to the finished piece? What are the requirements on the machine and personnel to get the right quality? What are the limitations and requirements of the machine on our part?

Do you have some patents pending on the K1?

SK: No, we do not have filed any patents. The approach of modular design and platform-based development has been used in other industries for decades. We did not invent this methodology but consistently apply what we have seen, heard and learned outside the watch industry.

As a basis platform, could the K1, in future, receive other modules than the planned one? Could you, for example, add a chronograph module to it?

SK: The K1-platform consists of 18 different variants. Additional modules ultimately lead to more complex design of the movement itself and more complex components, which are then used in all other variants throughout the whole movement platform as well. As a negative result, it will make the very basic variants more expensive. We do not want to make everything more expensive! We want to become a movement manufacturer which offers reasonably priced movements to all our customers. We are convinced that our platform is well balanced regarding multi-functionality, flexibility and price. We will reuse proven mechanisms and components from the K1 platform in further developments of course.

Can you say more about the “multifunctional parts” (for example you spoke about two springs that can make one...)?

SK: Multi-functional parts are widely used in the watch industry where cost-efficiency and industrial thinking is an issue, because a single component not only costs money during the manufacturing process. It also generates costs:

• as a new item in the stock

• as a new item which has to be managed by the logistics department

• as a new part in a small box during the assembly

• as a new part in the construction manual

• as a new part which can get loose, lost or defective during assembly or everyday use

• as a new part which has to be stored for at least one decade for after-sales-service

When designing a new mechanism the objective should always be to use fewer parts and at the same time guarantee at least equal or higher functional reliability.

A few examples:

• A single spring can be aligned on the main plate by using two cylinder bolts force fitted into the main plate or alternatively by two studs milled into the main plate, which makes the cylinder bolts redundant.

• In our first prototype we used five different screws in our movement platform. Later on we were able to remove two of them.

• The spring that holds the date disc in place also limits the vertical clearance of the rapid corrector on the other side.

• Our combined setting lever jumper holds down the setting lever, the yoke and at the same time defines the position of the setting lever and pushes the yoke in the correct position with the spring.

What does the “semi-automated” assembly process involve?

SK: It all starts with fully manual assembly by skilled watchmakers (Mr Nydegger and myself). At a certain quantity there is a need for automation. There are tons of machines, assembly tables and other tools available on the market to raise assembly efficiency, for example:

• Oiling - This is a time-consuming task when done by a trained operator, always with a certain risk of human error: no oil, too little or too much oil. In a semi-automated process this task can be executed in a repeatable, stable quality with little supervision.

• Pre-assembly T0 - Several jewels and cylinder bolts have to be force fitted into the main plate and bridges. Automating this task leads to more stable quality. At the same time we can check the accuracy of the components against a read-out of the individual assembly force. All this helps us to give timely feedback to the manufacturing department or supplier before there is a shift in quality.

• Assembly chains T1 - Building up movement by movement is not an efficient way of assembling because we always need one tool for only one movement. When we assemble 20-50 movements at the same time, it is more efficient because I need the tool in my hand 20-50 times.

The huge advantage of our platform is in the assembly. We can produce 18 different models on one assembly line. Traditionally assembly lines are used just for one kind of movement. In the K1 platform, all 18 variants share the same architecture. This means all parts are on exactly the same position in every single movement, independent of the final configuration. This means no changing setup, no re-training of operators, no quality gap in beginning till everybody knows exactly what to do, etc. Furthermore, we only need one watch case, because all our movements have the same exterior dimensions.

You mention three different quality levels. What are the differences?

SK: We will offer two or three different qualities, similar to what you could get from ETA in the past. The differentiations are not yet defined. We are also open to offer a partly skeletonised version in the future, it depends what the volume-customers want to have...

Movements, read also:

Private Label by Vaucher Manufacture

Source: Europa Star August - September 2013 Magazine Issue