Ulysse Nardin Marine Diver Chronometer Accuracy

Some of us would never admit of being nit-picky about their mechanical watches' accuracy. Astonishingly, it is observalbe that, the more expensive the watches are, the less important the accurate display of the time of the day seems to be. Tourbillons, repeaters, perpetuals are admired with awe, the question: "How many seconds is it off per day?" is never asked; even if it was, it would be answered shortly: "If I wanted the accurate time, I would have bought a quartz watch for ten bucks!" Just imagine: We celebrate a moonphase display that is accurate to one day per 250,000 years, but have no problem if the mechanical movement it is based upon is off by a mere 29 years in that period! As a principle we can state that the more complications are collected in a watch, the less accurate it is expected to be.


Consequently, it is not surprising that the - relatively few - real fans of accuracy in mechanical watches are concentrated in a class of watches with only few complications; mostly, the display of hours, minutes and seconds is enough, the date already is often considered pure luxury. Within this class, the designation "chronometer" attributes a certain status of nobility; if not for the wearer, then for the watch. Chronometers are watches with movements certified to be accurate within a certain set of tolerances. The inscription "officially certified chronometer" even makes us pay substantially more than for the same watch without chronometer status.


"On my wrist since four months, and only two seconds off!", better even: "Worn since 1991, never seen a service, and still only one second behind per day!" Statements like these are not rare among chronometer owners. and of course met with appropriate admiration by wearers of less accurate watches. Until very recently, I, too, was proud of such an extremely accurate timepiece, my Ulysse Nardin Marine Diver.

Acquired in 2001, frequently abused by wearing it under all conditions, including doing work with various powertools, fallen onto the wooden floor at least three times, never serviced, and it still gained but 1.5 seconds per day after more than seven years! This value was measured against a radio-controlled clock, with the watch worn on the wrist 24/7. Wow, what a great value! I hope that you out there, mere not-highly-accurate-watch-possessing mortals are appropriately impressed!


When I purchased a new certified chronometer about a year ago, I submitted it to a timing test that I normally conduct only rarely. As I had written frequently in my long-time experience test reports, I am satisfied with the one minute per week performance, resulting in a deviation of 8.6 seconds per day. This is not a great value, but one that has proven to be sufficient in everyday practice. To be fully satisfied, though, the watch's performance should be twice as good, so that it needs two weeks to be off by a minute. Any better value is great, but would not stimulate me to celebrate it with a bath in champagne (of myself, not the watch!).


But a new chronometer deserves some more effort, so I submit those watches to a test procedure similar to that executed by the "Contrôle Officiel Suisse des Chronomètres", short: C.O.S.C. Contrary to the popular belief, the accuracy of a mechanical watch movement cannot be measured from its daily deviation alone; especially not when the watch is worn on the wrist, and therefore subjected to countless position changes, sometimes temperature changes, too.


Since Earth's gravitation does affect the interaction of all the wheels and pins, most notably the oscillation of the balance wheel, it is also easily understandable that the system's accuracy varies with its changing positions. Thus, a - really bad - watch movement that gains one minute a day when positioned horizontally, and loses a minute in the vertical position, will appear to be spot on accurate when positioned vertically for twelve hours and horizontally for another twelve; the massive error simply compensates itself. Another watch movement, that gains 4 seconds in the horizontal position and loses 6 seconds vertically, will show a 2 seconds deviation after such a "test period", and will consequently appear to be the inferior movement, in spite of its - objectively - excellent performance.

Therefore, the watch's accuracy on the wrist does not tell the entire story. As a consequence, the C.O.S.C. tests the movements submitted by the manufacturers for two days in every position, plus at three different temperatures. The two days' intervals are necessary to check the movement for substantial fluctuations of the movement within the same position.

So when I get a new watch that is sold with the assumption to offer a good accuracy, as certified chronometers normally do, I submit it to a test procedure similar to that of the C.O.S.C.: The watch is fully wound once a day, and placed face up, face down, crown right and crown left for two days each. Once a day, at about the same time, the values are compared with the radio-controlled clock. Then I place the watch in the refridgerator for three days (the first day goes without checking the performance, to give the watch some time to acclimate to the new condition). Since I do not have an oven that would keep the watch at 38 degrees Celsius for a prolonged time, I omit the high-temperature test.

I spontaneously decided to let my favourite, hyper-accurate Marine Diver join the new chronometer in the test procedure, so the new one could learn from the expereinced one how to keep excellent time for many years.


Unfortunately, the results of the latter caused, to say the least, a certain disillusion: Not only did the deviation between the various positions vary for up to 24 seconds (+ 6 seconds face up, - 18 seconds crown left), but especially when comparing the two 24 hours intervals within the same position, there were differences of up to 13 seconds!

Could it be that seven years without service, abuse with tools, such as power drills and orbital sanders, drops onto the floor, countless hits and bumps left their traces without myself noticing it? What a shame, and I am considering myself knowledgeable about watches????

Off I went to my nearby watchmaker, who placed the poor Marine Diver onto his demagnetizer. A first check there showed the watch to be badly magnetized, and the watchmaker needed six or seven demagnetizing pulses until the instrument showed the field to be neutralized. My satisfaction was of only short duration, though: The following days showed the watch to gain some fifteen seconds per day, on the wrist, and still massive variations when left on the table in changing positions.

With my grand self-delusion shattered to pieces, I had no other choice than sending my Marine Diver for a factory overhaul. After I received it back, I wanted to enjoy the piece on my wrist, and therefore did not spend the eleven days necessary for a full timing test. Only three days were dedicated to a quick test, two days face up, and one day crown right. the decisive value was the variation within one position, and that was down to a second, and the variation between the two positions (normally, the single position's average values are taken there), and this was down to four seconds. On the wrist, the watch now gained two and a half seconds per day, so it was worse than before the service - seemingly.


But, as always in life: What seems to be simple, is not. Accuracy is accuracy is NOT accuracy ...

Marcus