Can anybody make it?
After many hours in the laboratory, the concept of the required mechanical oscillator is clear. We need to improve the 'Q' factor and yet preferably have a system capable of being tolerant to wrist movement. The answer lies in the form of running a low inertia balance wheel with an ultra high speed escapement. Hitherto impossible, the solution arises from the adoption of a predominantly co-axial escapement combined with silicon/ruby parts giving both stability, low friction and particularly low inertia of the pallet fork. I then decided on the superclock philosophy of a free balance only concerned with keeping time rather than releasing the escapement.
With spring barrels similar to those now appearing in most long duration watches, three days running can be achieved with a balance wheel from a typical modern 8 day movement (around 4 mg cm2) oscillating at 43200 vph.
The master balance sits next to the slave. The master balance is entirely free and has no function to perform such as counting or releasing an escapement. It oscillates in solitary spendor preferabl in a vacuum. It would eventually come to rest of course as there is no perpetual motion, and therefore it must receive an impulse to keep it in motion. This is delivered every thirty seconds and just in time by a light arm. This itself is re-set by a heavier arm. The triggered reset of the heavy arm releases the impulse which is fed back to the slave balance.
The slave balance commands the watch timing, entirely self contained and is of the form of any high standard timekeeper that takes one’s fancy, but it must be a highly accurate one and needs to be so as it feeds an impulse to unlock the master light arm and just in time. The phase of the master and slave balances are directly compared, and if the slave is late, it is given an extra force to push it back in phase with the Master. If it is early, no action occurs, but the slave is deliberately set to run a few seconds slow a day, so that hits and misses occur regularly, thus ensuring that the phase of the slave is locked to the master. This is an incredibly sensitive action, with a timing interval of only some mili-seconds. The slave is kept in synchronisation with the master in a phase locked loop. The only engineers I can imagine that could pull this off in a wristwatch would be JLC. I await it.