(Guy's note. Often, the behind the scenes conception, design, successes, and failures of the products we hold dear are unknown to us. We appreciate Dagfinn's giving us the full story on the birth of such an interesting watch - the Sinclair Radio Watch.)
The story behind this watch has never been published, although the story of the Sinclair computers etc has been fairly well documented. The project was from the outset shrouded in secrecy, standard Sinclair practice, especially as the watch was intended to link into the development of the mobile phone.
Sinclair Research LTD (SRL) Telecommunications Division had been established in 1982 to develop and exploit technologies that would enable SRL to become a major player in telecommunications, especially in developing mobile telephone products.
I graduated from the Royal College of Art in London 1979 with a postgraduate Masters Degree in Design and set up Pankhurst Design and Development (PDD) with my friend and colleague Paul Pankhurst. The work in PDD was interesting an varied but I wanted to gain more experience, particularly in New Product Development (NPD) and manufacturing.
I joined the SRL Telecommunications Division in Winchester in 1982 as Senior Product Designer and went on to become the Product Design Engineering Manager, specifically to develop the FM Radio Watch and micro mechanical, electrical and electronic technology for mobile telephony. The job was advertised in the Times newspaper and there were many hundred applicants. The selection process was very rigorous and included several interviews and a days worth of psychometric testing and assessment by Cambridge Recruitment Consultants. At interview I produced two conceptual models of telecomms products in order to visualise my design vision and skills, one was a pager and one was a watch. This initiative was well received and I got the job.
There was no written design brief, Clive simply said to me: Design an FM Radio Watch as small as possible. The watch must have an integrated speaker and aerial. The aim is to design the first/smallest integrated FM Radio Watch in the world. That is exactly what was achieved.
The project was a serious challenge and afforded me a thorough apprenticeship in New Product Development from concept through to production and I loved the blending of innovation, design and technology which the experience allowed me to master. Although there was a managing director in Winchester, on product design and aesthetic matters I reported directly to Clive. I usually travelled to London by first class train on Friday afternoons to meet with Clive and discuss progress on design. Working with Clive was a great experience and I enjoyed responding to his challenges and discussing design issues with him. He was a very demanding boss but once he had approved the design he threw his full support and financial muscle behind the project. SRL had an executive jet aircraft and on more than one occasion I had approval from Clive to use the jet to transport watch related components from England to Dundee airport in order to speed up the manufacturing process at TIMEX. Clive also provided me with a great working environment and the very best of equipment that money could buy. I shall always be grateful to Clive and SRL for giving me this one in a million opportunity to develop as a designer and manager.
The technical challenges in designing the worlds first and smallest FM Radio Watch were numerous and unforgiving. The wavelength of the FM signal is about 1.8m, yet the antenna only 180 mm long further reduced to 60mm as sitting around the wrist. This is a matter of the laws of physics and effectively meant that the FM Reception would be great if the receiver was in direct line of sight of the transmitter. If the transmitter was not visible then reception would be poor or non existent depending on distance and signal strength.
I gave my father Sverre Aksnes two of the FM Radio Watches and he enjoyed excellent FM reception and use over a 20 year period in Bergen, Norway. I think he must have been Clive’s best satisfied customer and longest serving unofficial product tester!
One of the first conceptualisations of the FM Radio Watch was a solid model machined from acrylic, painted black and assembled with standard miniature piano hinges and a rubber watch strap (picture). This concept model allowed us to have a clear vision and image of the future product and in fact the final design did not differ greatly visually except that the form was more refined and the final execution of the design utilised different materials and finishes.
A significant feature of the watch design was a strategic decision to split the watch into 4 separate units: Watch, speaker, FM tuner and battery compartment which was placed in the clasp in the middle of the strap.
The watch, speaker and FM tuner were made as die castings, connected by stainless steel case backs hinged together (Photo). This design concept allowed the watch to take up the shape and size of the wearer’s wrist so that it would adjust comfortably to individual requirements.
This concept also effectively distributed the watch around the wrist, thereby reducing the apparent size of the watch and balancing the components. This allowed me some latitude to adjust the critical dimensions of the individual units during the design process.
All the electric connections between the four units were achieved by using a flexible printed circuit board based on Kapton Polyiamide polymer. This produced all the interconnects and power lines as well as the antenna circuit. Part of the antenna circuit which was adjustable was produced from Delcan copper alloy buried in the strap. A separate flexicircuit connected the speaker to the amplifier. Whilst the connections could have been achieved through micro cables in a harness, this would have been a manufacturing and reliability nightmare and I quickly won the argument about the flexicircuits.
Another advantage of the flexicircuits was that it allowed the connections to run in the neutral plane of the hinge, thereby totally avoiding metal fatigue in the conductors which could have become a problem.
The gaps between the three hinged units are covered by neoprene bellows (0.3mm thick) which protect the flexicircuits and provides a visual continuity between the three diecast cases
The design process I deployed has a great deal in common with the processes now promoted by IDEO and the International Design Management Institute; simple, straight forward and above all manageable.
Altogether the development time was 2 years. I worked mainly alone developing the concept and designing the watch during the first year, then Gary Shaw joined me as a very bright and design conscious draughtsman for the second year helping with the production start up. Gary’s contribution was invaluable and we worked extremely well together as a team. The atmosphere was one of great creativity and technical expertise blended with a fantastic sense of fun and achievement. The office humour was certainly second to none other I have experienced before or since. How lucky we were to have such a positive experience!
As this was the era prior to CAD becoming available to designers in a user friendly and affordable format, Gary and I worked in pencil on tracing film, a fast and effective medium which proved to be ideal for the purpose and with Gary’s skills it was also fast and visually beautiful, especially his stylised writing. Because of the small dimensions of the watch, some of the drawings were done at 10:1 scale in order to describe the geometries accurately.
All the production critical components were prototyped and tested thoroughly,
including sets of cases machined from acrylic and brass, hinged casebacks pressed from brass sheet with acrylic tools, prototype mouldings machined from ABS, snap fits, and various prototype components which I made.
There were a few gold plated watches which I made for Sir Clive, and some versions which had coloured titanium bezels. Some early prototypes had different styles of stainless steel bezels. There is also a one off black cased version which I made as I was curious to see what it looked like.
I used a Swiss ACIERA F1 miniature precision watchmakers milling machine for much of the prototyping.
• A 3D exploded study of all the real production components which went into the watch
• A 3D study of innovation, prototyping and experimentation documenting the design and development process
• A 3D study of some of the materials and manufacturing technologies deployed
There are two patents covering the innovations I worked on in connection with the watch. A third patent was drawn up to full application, however this was not proceeded with due to circumstances unconnected with the watch project.
FM Watch strap assembly 1984 Patent number EP 0125930
FM Watch strap attachment 1984 Patent number EP 0126629
FM Watch tuning mechanism 1984 Patent application prepared
We prepared tooling and production equipment based on an annual production of 500,000 units. This allowed us to amortise some tooling over time with the tool cost built into component cost and some of the risk was through this arrangement shared between SRL and the component manufacturers.
I adopted a right first time approach to production start up based firmly on careful risk assessment and functional testing of all critical components prior to tool making.
Where I considered that there was an element of risk or where I could foresee potential difficulties arising from wear of the tool over time, I specified that the tools be manufactured with replaceable inserts. This allowed for components to be easily and inexpensively modified in case of any functional difficulty. Similarly tool wear could be rectified by replacing an insert from spare stock rapidly (minimal downtime) and without having to manufacture a new tool. This approach proved to be extremely robust and produced 100% results throughout the development of the watch. In fact all our tools worked right first time, unlike previous Sinclair products and many others! These tangible benefits were in my view more than justified against the small additional costs incurred in the toolmaking.
In the functional testing I had great help from a very talented and skilled freelance prototyping engineer Peter Nutting who modelled some of the miniature components in realistic materials using vacuum casting and other technologies to achieve some fantastic accurate and testable components. Sadly he passed away a few years ago.
Manufacturing and Production
The output of all the design and prototyping activity was a detailed specification and a set of design engineering blue prints fully toleranced for manufacture. I subcontracted all the component manufacturing mainly in the UK where we had XX suppliers (list) and in Hong Kong where I had the watch circuit manufactured.
The product was in fact tooled up and mass produced at TIMEX in Dundee, Scotland. I was responsible for establishing the production line with John Logan, general manager of watch production. This war a fairly modest operation to begin with. I remember training six highly skilled assembly workers to build the watches. They immediately produced excellent results and we never had any difficulties in production of complete watches. The Sinclair flat screen CRT and television production line was also running at Timex at the same time, what fun we had! We assembled somewhere in the region of 11,000 watches and they were nearly all shipped to the US for distribution. However a warehouse fire destroyed them all, at least that was what I was told at the time.
FM Radio Watch Innovations
FM Tuning Mechanism
This invention required innovation in design, materials application and manufacturing processes. The resulting mechanism is unique and a good example of a miniature electro mechanical mechanism. Although a patent specification was drawn up, SRL never proceeded with the patent for unknown reasons however there can be no doubt that the invention was patentable at the time.
A radio receiver must be able to tune into a range of radio stations and discriminate between them. The FM frequency spectrum covers 88-108 MHz and the tuner must be able to adjust the frequency smoothly and repeatable within this band.
The conventional tuning circuit consists of a parallel connection between a capacitor and a coil, also known as an LC circuit with the L being inductance and the C being capacitance. This circuit is connected in series between the aerial/antenna and earth. Tuning is in principle a variation of the resonance frequency of the circuit and this is conventionally achieved through the varying of the capacitors capacitance value whilst the coil inductance remains constant. Variable capacitors for this purpose are manufactured in a wide range of types. The tuned signal is then taken from the two circuit junctions and rectified through a diode, then amplified.
At the time of development of the FM Radio Watch there existed no variable capacitors small enough to be built into the available space volume of the watch. This was a very serious problem as the watch had to have a tuner capable of tuning the FM frequency spectrum. The invention of a tuner mechanism was a critical path item.
As a boy I had built many simple radio circuits and I had always wondered if the tuning could be achieved through other means than those traditionally used. Would it be possible to vary the inductance of the coil instead of the capacitance of the capacitor?
The solution which emerged was to develop a novel and unique tuning circuit in which the inductive element or coil inductance could be tuned instead of the capacitance of the capacitor, the overall variation in resonance being equivalent. This was theoretically sound, however the practical execution represented a major inventive step as nothing like this had ever been achieved on the miniature scale required for the FM Radio Watch.
The Design challenges which faced me in this task can be summarized as follows:
• Design a tuning mechanism for the FM frequency spectrum of 88-108 MHz where the inductive element is varied instead of the capacitor.
• The whole tuning mechanism must be compatible with the electronic tuning and FM Radio circuits and physically small enough to fit inside the FM Watch This meant it had to be smaller than existing variable capacitors – a tough requirement.
• Design and manufacture of a suitable coil, which had to be compatible with thick film hybrid technology manufacturing processes. This meant that the coil needed to be assembled onto the ceramic substrate using normal pick and place robotics technology and be compatible with the vapour phase re-flow soldering process running at 205 degrees Celcius
• Design a method of varying the inductance of the coil and design a mechanism which would allow a human being to carry out the tuning process.
• Design all components of this mechanism in such a way that they can be successfully manufactured through conventional manufacturing processes and assembled into the volume/space constraints of the FM Radio Watch
Electrical Principle Design
A test circuit was designed and built using a fixed capacitor and a fixed wire wound coil. The circuit was connected to a frequency generator and a frequency analyser. By experimentation it was found that the resonance frequency of the circuit could be varied by presenting a triangular shaped piece of aluminum at the open end of the coil and moving the triangle slowly across the coil in such a manner as to vary the amount of aluminum covering the end of the coil between minimum and maximum to achieve full coverage of the 88-108 MHz FM band.
Through calculation and further experimentation it was established what the values of L and C needed to be and the shape of triangle required to achieve the tuning function.
Initially a conventional wire wound coil was considered and this would have been possible. However due to the small size of the coil and its fragility in manufacture and operation, it quickly became apparent that a coil produced by flexible printed circuit board technology would be better, more repeatable and lend itself much better to handling and processing during manufacture. The flexicircuit is double sided with the two halves of the coil circuit connected via a plated through hole in the middle. (Picture)
Coil Pillar Design, Coil Assembly, Pick&Place and Vapor Phase Reflow Soldering
The coil needed to be presented at a certain height above the thick film substrate in order for the tuning band to avoid clashing with any of the chip components on the substrate. To achieve this and a rigid positioning of the coil, a plastic pillar was designed. The pillar had to be carefully dimensioned and asymmetrically formed to miss neighbouring chip components and their relatively wide tolerance bands.
The coil was wrapped round the pillar and secured on the underside by four headed brass pins. The pins acted as mechanical and electrical connectors. This produced a robust assembly which could be handled by the pick and place robot and once soldered in position present the coil in a position suitable for the tuning mechanism. In order to allow for the assembly to pass through the Vapour Phase Reflow Soldering Process running at a constant 205 degrees Centigrade, it was necessary to specify a plastic material which could withstand this. Poly Phenylene Sulphide (PPS) can handle 225-235 degrees C and is inexpensive and was chosen for the pillar. The whole assembly performed extremely well during manufacturing. I worked with Scorpion Plastics in Essex, Dowty Circuits and CorinTech Circuits (www.corintech.co.uk) in Fordingbridge to implement this.
The tuning mechanism relied on a metal taper shape moving linearly across the coil to effect the change in tuning required. To achieve this it was necessary to produce an endless band containing the metal triangle. In practise I worked with Timex Feltham to design a triple lamination of polyester film/aluminum/polyester film. The polyester film pieces were offset in relation to each other to allow for a good overlap for bonding when formed into an endless band and this arrangement proved satisfactory.
With the functional principles established and an outline design for the mechanism, a draft design for the tuning frame was sketched out and immediately modelled in 3D. This component was machined from ABS on my Aciera F1 watch maker's milling machine. On assembly it worked right first time and a second prototype was drawn with some more detail and sent for prototype machining. This prototype proved entirely satisfactory both mechanically and electronically and was subsequently drawn up as a final design and sent for toolmaking and manufacture.
These consisted of two precision turned brass shafts, one with a secondary op milled spade end to allow for driving and two PVC extruded sleeves for the tuning band to engage with. The drive depended on the characteristic of polyester adhering closely to PVC by the exclusion of air and this effect was utilised very satisfactorily to drive the tuning band.
Control Shafts and Control Knobs
These components were made from 6% glass filled nylon and turned brass. There were 3 different shafts with a standard control knob fitting all 3. The shafts were designed to snap fit into the watch case and the control knob was secured radially by by 2 prongs and axially by a self tapping screw into the end of the control shaft.
One of the control shafts had 7 different functions, all of which were individually prototyped and tested at scale 1:1
The control shafts were moulded by Dynacast’s New York subsidiary and the first samples arrived back 6 weeks after drawings were produced and everything worked right first time. There was no need to adjust any of the tools.
The potentiometer was a small trim pot 6mm square designed for adjustment with a screwdriver. The connection with the trimpot and drive shaft could not be achieved in plastic due to high torque and sharp edges of the pot digging into the plastic materials. It was also difficult to attach anything to the pot as the hole through the rivet was only 0.8mm diameter. I had made an effort to measure and describe the geometries involved in order to establish the parameters which a solution had to engage with.
The solution to this problem was a drive tag, shaped like a T and manufactured from etched beryllium copper. The etching process was over etched so as to produce three thin knife edges with sufficient resilience to be force fitted into the hollow rivet by press and remain permanently. This idea came to me – the Eureka moment – on a business trip to South West Wales as I was driving and listening to music. I immediately stopped at the nearest telephone box and telephoned my draughtsman in Winchester and described the idea to him. He then produced a drawing which was faxed to London for manufacturing. This was done on the nightshift and when I arrived a work the following day, a motorcycle courier was waiting for me outside the office with a little bag of etched drive tabs. The first one went into the rivet without a problem and this assembly never gave any trouble. Rapid Prototyping has been around for a very long time!
The watch cases were made by Dynacast’s zinc die casting process.
This was extremely accurate, fast and relatively inexpensive. However the finish required for the plating process was difficult to achieve with consistency. We were stretching the parameters of the Dynacast process! This led to some intensive Dynacast process development. The cavity gate design was evolved to achieve required finish and vibratory bowl finishing by Walter Trowal was introduced. In the end I had to switch the electro plating contractor in order to achieve the consistent satin chrome finish required. In retrospect we should probably have gone with hot brass stamping for maximum finish, however this would have required expensive machining and the internal detailing would have been difficult to achieve but not impossible. TIMEX Dundee and Feltham were world leaders in this field and it would have been interesting to use this older technology.
CNC machining of copper electrodes for cavity die sinking and spark erosion (EDM) had to be carried out in 3 portions due to limited machine memory. This was my suggestion after the toolmakers at Cleveleys had given up. I visited them and said how about machining the electrodes in 3 portions, one starting where the previous one stopped, using the stop as the definitive process register. After some humming and hahing they said it has not been done before but yes ok lets have a go. This stretching of technology and innovation in process development was very typical of the FM Watch development. I certainly developed a reputation as a maverick with numerous toolmakers throughout the world. Usually they came round to seeing things my way after a while, once they realised I respected their skills and craft tremendously and only wanted the world to move forwards.
The Watch Circuit
Electronically this was a standard watch circuit which I sourced from a company in Hong Kong. It had a special graphics for the display and physically it had to fit the watch case and provide switching for the programming and setting of the watch and alarms as well as switching the watch on and off. The timing switching was done by three finger pads and a silicone keypad with silver loaded pills and tactile feedback. Actuation was achieved by three diecast buttons through apertures in the watch case.
The Switch Contact
The on off switch and alarm1/alarm2 functions were actuated by cams on the watch switch shaft. The switch itself was a compact double spring contact manufactured by etching beryllium copper sheet and folding the component in a press. It was then copper, nickel and selectively gold plated in the contact and solder areas. The spring elements describe S curves in the available volume in order to provide sufficient length of spring for the required deflection and spring rate.
This design was very novel and proved extremely robust in service. Operating the watch switch knob produces a very satisfactory feeling of quality with indents and tactile feedback for each of the three stops. The watch switch shaft has seven separate functions which is quite a tall order for such a small component. The flat version of the etched component looks great, like a mysterious space alien.
Standard 60 degree thread profile created too much torque as the thread was biting into the very hard surface skin of the die casting. This caused some screw heads to shear off on assembly. A new thread profile was designed with shallower thread angle which enabled the thread to penetrate surface without over-torqueing which completely solved the problem. My apologies to Mr. Joseph Whitworth on that one!
Stainless Steel Casebacks
Plastic tools which I machined in acrylic were used to prototype press formed brass components. These were then complete by hand folding and filing and assembled with miniature brass piano hinges using a reflow soldering process in a jig to ensure accurate alignment during the application of heat, solder flow and cooling down, quite a tricky process to control and it had to be right first time! This one off prototype allowed us to test the watch mechanically and electrically. The tests proved that the design and the assembly worked fine and gave us 100% confidence before spending £ 38,000 on production tooling.
The multiple progression tooling was running at 600 hits per minute and it was a great feeling to stand next to the press and seeing my components dropping into a box at this speed! I remember thinking that the design and dimensions had better be right! I need not have worried everything was fine, no adjustments were necessary.
Watch Straps and attachment pins
Moulded from Quinn a polyurethane material and heat welded to seal in aerial and flexicircuit conductors. New design for attachment pins allow the conductors to remain in the neutral plane to avoid work hardening and fracture. The pin was patented and much cheaper to manufacture than standard watch strap pins.
Harboro Rubber Company. Compression moulded from neoprene rubber, 0.3mm thick. 96 cavities per tool. The bellows provided protection for the flexi circuit and made the transition between cases visually continuous and attractive.
This project was a golden opportunity for me and I am exceedingly grateful to have been offered this. It was without doubt a formative period in my career and very rich in learning and experience. As a process it was also fairly complete in that it started with an idea/creativity and concluded with a product in production, being shipped abroad to the USA. Along the way it touched on virtually everything that goes into a complex design process.
The product is still there today and has appreciated significantly in value, the ultimate measure of lasting quality. The Sinclair FM Radio Watch is highly regarded by many which is appreciated. The patents are just about exhausted after their 20 years in force.
We stretched materials, technology and manufacturing to the limits - and sometimes beyond - in our pursuit of the product and despite the high ambition and problems encountered we succeeded.
This was not a trivial exercise, I worked closely with the suppliers to achieve this. These collaborations represent some of the most interesting technology developments which I have been involved in. I would particularly mention the technologies developed with 3M, TIMEX and Dynacast. At Dynacast the process development was aimed at improving the cast finish and involved development of injection gating design which proved satisfactory and which Dynacast was able to transfer to other manufacturing projects.
The collaborations, partnerships and camaraderie was phenomenal and a great testimony to teamwork and co operation which has been a very positive experience and informed much of my subsequent work.
Packaging electronic circuits in a high density wearable consumer electronics product which has a high intensity human interface was a tremendous challenge. The watch design made technological and human factors progress which fed directly into the development of the mobile telephone technology. My last project at SRL was to develop visual concepts for the mobile phone and assembly concepts which built directly on experience from the watch.
This process has been the prototype for much of what I have done and taught. I hope many other young designers get the opportunity and resources to do similar things.
Above all it was great fun and I am indebted to all those I worked with and their individual efforts and specialist knowledge and skills which made the Sinclair FM Radio Watch a reality.
Dagfinn Aksnes - Product Design Manager SRL
Gary Shaw - Draughtsman SRL
Peter Nutting - Freelance prototyping engineer
John Logan - General Manager, Watch manufacturing TIMEX Dundee
Nn - Timex Feltham
Clive Sinclair - Uncle and funder SRL
Mike Pye, MD - Sinclair Telecomms (Ex MD, Sinclair Radionics) SRL
Graham Beasley - Electronics Engineer SRL
Phil Holliday Electronics and Sofware Engineer SRL (Later Holliday Electronis in New Zealand)
Roger Hance - Dynacast
Steve Bullen - Dynacast
and many others!