Thoughts on accuracy of mechanical watches, take 1January 31 2003 at 4:14 PM
|Marcus Hanke (no login)|
Response to Frequently Asked Questions - FAQ - and Reference Articles on GO
I posted this 'article' on December 20, 2001:
The adjustment of mechanical watches is a never ending story ... From the beginning you should know that all mechanical watches which are meant to be really accurate under reproducable conditions are clocks which are not "worn" by their owner but are left in one position most of the time (marine chronometers or wall regulators). The problem is that each mechanical movement is influenced by gravitation differently, depending on the movement's position in relationship to the gravitation's direction. Therefore the movement of a wristwatch will produce different accuracy rates depending on its position. The implementation of a tourbillon does not help here since it is constructed to reduce the different gravitational effects on the moving balance wheel only. When the movement itself is moved, the tourbillon's positive effect is negatively affected again.
A well adjusted wristwatch should show a performace which differs as little as possible depending on the movement's position. For its wearer, however, this might not be noticed immediately, since also bad performing watches can be highly accurate when worn on the wrist. If the watch for example loses ten seconds a day in the face up position and gains ten seconds in the face down position, it is considered a bad performing watch. In spite of this it will run perfectly accurate if it is kept twelve hours face up and another twelve hours face down, since the error neutralizes itself.
Even a perfectly adjusted wristwatch, however, will show a slightly different behaviour in different positions and it is up to its wearer to neutralize the difference by changing the position as often as possible. This is where you as the watch's owner are entering the game: If you have a sitting job at a desk or driving around a lot the watch's movements will be entirely different from those if you are a car mechanic, for instance. As a consequence one and the same watch will perform differently on different persons' wrists. This is a part of the fascination with mechanical watches: they are "calibrated" to their wearers.
If you send in a watch for adjustment now means asking the watchmakers there first to check the watch's current performance. Because of what I said before you can imagine that this again is like flying blind through night and fog. All the watchmakers can do is setting the watch on a watch mover and check its performance over at least a week (don't forget that sending the watch in a parcel subjects it to bumps and shocks which can change its accuracy dramatically; the watch needs some time to settle down to its usual behaviour). And then the watchwinder is not you, the movements it makes are perfectly regular without those variations which are inevitable in everyday's life.
Therefore it always is useful if you can set up a very detailed accuracy chart of your watch in which its daily performance is written down for at least two weeks. Additionally you should do your best to describe your everyday business during which you are wearing the watch. All this helps the watchmakers evaluating in how far the watchwinder can replicate the moving the watch has on your wrist.
After the week the watchmaker will be able to see how (in-)accurate the watch is and will open it for its first adjustment. What follows now is a delicate mix of long experience and pure trial-and-error: He tries to estimate which manipulation will be enough to assure the best accuracy. To check that makes another three to seven days on the watchwinder necessary, since each watch is reacting somewhat "offended" to the operation and needs some days until it regains a regular performance. Only in very rare cases the watchmaker will be satisfied with the result of his first adjustment and repeats the procedure several times.
You see that without the time it takes to send the watch to and from the atelier at least three to four weeks are needed for a well made adjustment. And this is calculated only for the case in which the watch does not behave erratically, which would make necessary a complete diagnosis and a long search for the troublespot. Maybe ten weeks turnover time really are a bit long, but take into accout holiday times with lack of good staff and such long times are explainable, yet not nice for you as the customer, of course.
That the watch then is not as accurate as you want it to be after you get it back again has its main reason in that the watchmaker's machinery cannot replicate you as the wearer. The watch might have been perfectly adjusted to the watch mover used in the repair shop, but you as a person behave differently and therefore your watch does so, too. And finally keep in mind that the watch again has been put into a parcel and you cannot know how it was treated on its way back to you.
The only way to get a really perfect adjustment would be to do it yourself at home ... but who of us can do it? the next best alternative would be to have a good watchmaker near your place where you can bring in the watch without having to send it in a parcel, and pick it up again when it is ready. this is what I do. My watchmaker knows me and how I wear the watch, additionally I get the watch without having to pay for adjustment after he adjusted it one or two times. Then I wear it for about two weeks, constantly checking its performance, until these data are used as a base for the final adjustment process. Only when I'm satisfied I pay for it. Of course all this lasts about two months, but it is worth the wait.
Well, I am sorry for my overlength post, but I sometimes come to think that many wearers of mechanical watches still think of them as binary computers which are either true or false. Due to their complexity and their dependence on their wearers one cannot await from them "perfect" performance from the beginning and under all circumstances.