Return to Index  

Thoughts on accuracy of mechanical watches, take 2

January 31 2003 at 4:17 PM
Marcus Hanke  (no login)

Response to Frequently Asked Questions - FAQ - and Reference Articles on GO

Originally posted on January 29, 2003:

Well, mechanical watch movements are strange things: On the one hand, there is a massive effort in development and production, tolerances become computer-optimized, parts are made to narrowest specifications on CNC machines. And still there is that element of uncertainty, of inaccuracy that dwells in every complex mechanism. There are hundreds of parts that make out a watch movement, and a variation of a thousandth of a second in the movement of one part might sum up to an error of several seconds per day. This variation can be caused by a manifold of factors, which lie far from the production and assembly phases. Therefore, watch enthusiasts tend to think of their watches as organisms, each of them interacting with its wearer.

As John has already mentioned it below, even a mechanical sports watch is still a very delicate thing, compared with a G-shock, which does not feature any moving parts aside its clasp. Each bump, or shock experienced during transportation (and most watches have to travel thousands of miles from the factory to their final destination) can introduce a variation like that I mentioned above. Additionally, like most car enginges need a certain period during which they do not reach their full capability, a watch movement has to run for several weeks, even months, until all the lubricants have dispersed evenly, and the accuracy settles down to a certain value. Moreover, watch movements will eventually react to their surrounding conditions: A watch worn by somebody who frequently is working hard in high temperatures, will show other accuracy values as one worn by somebody who misuses a computer keyboard in an airconditioned office all day, even if both watches had been regulated to the exactly same values at the factory.

This is the reason why every serious dealer is not only selling a mechanical watch, but normally offers to regulate it (mostly for free) after a certain period. When I bought my watches, most dealers did not say simply "Good bye", but they said "We'll meet for regulation in some weeks". Not one of them did guarantee to me that the watches sold would be highly accurate already from the beginning, and I must say that only a few of my watches were accurate enough, regardless of their manufacturers. After about a month, I brought them back to my watchmaker, who regulated them. And this was also a lengthy procedure: Based on the performance values I had written down, he adjusted the watch. Normally, watches react 'insulted' to such a procedure, which means, that during the first two or three days after the adjustement, their accuracy is going mad, until it settles. Only then the watchmaker can evaluate if his adjustement has been enough or not. Often, this has to be repeated several times, resulting in at least a week, before I can pick up my watch again. And even then, the regulation is not absolute, since it is based on the watch's performance on a winding machine. This is very different from my own wrist. In one instance, a friend of mine, who is a very good watchmaker, really try to make the perfect adjustement for me. In three weeks, he regulated a chronometer so that it gained less than three seconds within a complete week. On the winder, of course. As soon as I had it in everyday use on my wrist, it gained two seconds per day, that are fourteen seconds per week. Once again it is proved that a 'laboratory' can never simulate real life.

After some years of being an accuracy fanatic, writing long performance lists of all my watches, under changing conditions (summer, winter, humid climate etc.), I have become a lot more pragmatic now. For me, the limit of my error tolerance is a minute per week. Everytime my watch is wrong by a minute, I reset it. If this happens more than once a week, I have it regulated (or serviced, if necessary). Sometimes, this is necessary following a slight mishap: After a chronometer, which was very accurate until that day, fell on the floor, it lost eight seconds per day. My watchmaker checked it for damage and adjusted it, now it is gaining three seconds per day.

Finally, the wearer of a mechanical watch should always be aware of the fact that accuracy is but a fluent state. Lubricants wear out, shocks, humidity and magnetism can affect the movement. All this can make the accuracy change over time. Sometimes a watch performs differently after it has been lying in the drawer for a longer period. Personally, I found that the famous high-beat El Primero movement by Zenith is such a movement: Similar to a sophisticated sportscar, which has to be moved on a regular base, this watch wants to be worn often. Otherwise it answers with a different accuracy value every time I wear it after a period of unfunctionality.

Please excuse this long referate, all I tried to explain is that mechanical movements are different. In spite of their high prices, a customer normally does not get some kind of "out-of-the-box-performance", as in a DVD player. The watch has to be tweaked and adapted to its owner, and this is the reason why we are so fascinated by them. They are much more personal than most other belongings, because as soon as we strap them to our wrist, we make it a part of our physis, bringing their mechanism into a 'resonance' with our body and our life.

Therefore, a mechanical watch that is not highly accurate out of the box is something normal, regardless of its brand and price. Adapting it to your personality as its wearer is part of the acquiring - and part of the fun using it.

I hope I could express myself in a somewhat intelligible manner ...


 Respond to this message Home Page