Much of my experience comes from the 20 years of general repair work we used to do, long before I ever thought about manufacturing watches. So while you might know us as watchmakers, we had a lot of fun playing with everyone else’s toys first.
Occasionally we would have a modern, chronometer-certified watch in for overhaul. Often only a few years old, these watches could cause trouble. No matter how carefully we tested them there was a significant difference between what the watch was supposedly capable of and any hope of passing a chronometer test.
One of the reasons for this is that the chronometer test is just that, a test for supreme timekeepers, which are tested in static positions. Each movement is manually wound every day. Only after testing and certification does a movement have its auto-winding fitted. With auto-winding, a watch spends much of the day fully wound and never quite runs down. With manual winding, the whole length of the anticipated 24 hours’ worth of spring is used.
When the spring is wound to its tightest, a normal lever escapement will run fractionally slow. So in an automatic watch, that will happen whenever the auto-winding mechanism has been efficiently used. A watch that has only just passed a chronometer test can run slow when cased up and worn on the wrist. It is an impossible quest to have it test as well as it did when manually wound.
The critical aspect you need to understand is that a chronometer certified watch may not be the most accurate thing on the wrist. Englishman Stephen Forsey said, after Greubel-Forsey won the international chronometry competition of 2011 using an Arnold-type double tourbillon escapement in the GF02S: “You have to study the rules of the contest and learn to adjust the watch accordingly. We were used to adjusting our watches for the wrist.” The GF02S has, naturally, an overcoiled hairspring.
SPRINGING INTO ACTION
Back in the 1770s, irregular transference of power from the mainspring to the escapement was taxing the great minds of the day much as it continues to do still. The English chronometer maker John Arnold was the first person to successfully adapt the two final curves of the hairspring and overcome this error: what we now refer to as the overcoil.
No doubt readers will complain but this is the “Breguet” overcoil. Breguet was merely one of many styles and I am not aware of a manufacturer fitting “Breguet-type” overcoil these days. Even early 20th century attempts at mass-production of overcoils, such as those undertaken by Hamilton in the 1920s had already modified and further improved further on the Arnold/Breguet principles. Modern overcoils are the fascinating results of millions of pounds of investment and many PhD theses. It is a subject that fascinates pure mathematicians because of the sheer complexity of the shapes involved in one deceptively simple-looking spring.
Interestingly, the original Arnold and Son chronometer business was taken over by the firm of Charles Frodsham, which is still in business today; the eponymous Swiss watchmaking firm of Arnold and Son has no connection. John Arnold and Abraham Louis Breguet in fact, saw themselves as fellow seekers, who stood above the petty boundaries of nation. They freely exchanged ideas and sent their own sons on a prolonged exchange visit, each having been apprenticed to the other’s father.
What does the overcoil do? You need to know that it performs two distinct functions. One is to help control the timekeeping of a balance wheel, in spite of the watch running with more or less torque (fully wound or run down). The other is to help normalise the overall mass of the balance-complete (the balance staff, the balance wheel, the roller and the hairspring assembled together) no matter what angle the watch is running at. These are, probably, of primary and secondary importance to the owner.
Primarily, the overcoil helps to equalise the action of the balance wheel so that it keeps the same time (is isochronous) either fully wound (the balance has a large amplitude) or run down (and the balance has a small amplitude). Thus the problem of the errant automatic watch is radically improved.
It is rarely possible to achieve perfect isochronism and most manufacturers set their overcoils so that, if anything, the watch will end up running slightly fast if completely run down. That way a watch which may be still running but not yet completely run down is, if anything, fast rather than slow. Watch escapement designers are very thoughtful people you know, they don’t want you to be late.
Thus, the modern overcoiled Rolex seems to be a perfectly well-designed and implemented overcoil. I have seen it working well in a number of watches. This is one of the best balance completes I know.
There is a secondary function to the overcoil, developed rather later and first perfected in the 1910s by well-known overcoilers – Lossier, Grossman, Phillips and others. This related rather to the constant concentricity of the coils of the hairspring and particularly applies to the wrist-watch. In a standard lever escapement as the spring tightens, the weight of the spring shifts to one side. Not by much, probably around 0.00005 grammes. It is enough however, to affect the performance of the watch.
The demise of skilled manufacturing watchmakers is linked with the improvements available in micro-lasering hairsprings in place and making accurate factory jigs, a lot of the skill has gone out of watchmaking. If however, you are hand-finishing a watch, then you dynamically poise the balance complete.
Poising, once again, is vital to pass that chronometry test. The balance is poised out of the watch to ensure it has no heavy spot one side that might play havoc with timekeeping. All quality watches are poised. Dynamic poising is repeating these tests, with the movement assembled and ticking, to see how the balance performs with the hairspring engaged. In theory, this could be a job of weeks but a skilled springer can dynamically poise a balance in an hour or 10. As I said before, the action of the balance pushes the spring from side to side as it coils and uncoils. The advantage of the overcoil is that it lessens the extent to which that happens and thus lessens wasted side friction on the balance-
A skilled springer can perfect that poise to ensure the balance runs correctly in the four horizontal planes (leaving dial up and dial down adjustment to the angle of bluntness on the balance-staff pivots). It is skilled work and time consuming. Occasionally we see such watches in the workshop for repair and I am sure the owners have no comprehension of how hard the designers worked to improve on a system that, as my remarks on the chronometer test demonstrate, have no official stamp of approval or measurement.
BEYOND OVERCOILED SPRINGS
So in essence yes, an overcoil hairspring is probably always a good thing if fitted to an automatic; it helps to equal out fast or slow running because of erratic winding. And yes again, an overcoiled hairspring is probably a good thing as regards chasing accurate timekeeping in a wristwatch but beware watches of unknown origin or provenance.
Once again, some of the 21st century technology employed in the larger watchmaking concerns has somewhat surpassed the need for an overcoil. Vaucher, a Swiss firm closely linked with watchmakers Parmigiani, has implemented some very simple, if difficult, improvements to the standard lever escapement to produce a significantly more
In university tests in Hong Kong, a watch assembled with great care using a flat hairspring could overcome the poising errors just as well as well as an overcoiled spring did. Great care means just that however, because our Eastern chums are just as transfixed by the lever escapement as we are. Again, we are moving beyond chronometry tests here but moving, up the notch, to chronometry competition watches. You will find in competition watches that some have overcoils and some do not. In theory, the poising advantage of the hairspring matters. In practice it seems more relevant to the independent makers who persist with slow (usually 18,000 beats per hour), heavy, balance assemblies. The faster beating (often 28,000 beats per hour) mass-produced escapement uses the advantage of the high-beat to compensate for the lack of overcoiling.
The Swiss watchmakers Tissot won a chronometry prize a few years ago with a very standard ETA 2824/2 movement watch. There is a tale, I do hope it is true, that they tested a dozen, chosen at random, and picked the best one to send for international competition. Nowadays they – perhaps spurred by that story – prepare their watches with special care for chronometry, using silicon hairsprings. My point is that this was a standard, flat springed balance and it out-performed every overcoil in its class.
This article first appeared in the October issue of WatchPro magazine. To read the full digital version, click here.