Thursday, September 15, 2005

Kew Observatory Trials Versus COSC - And why things will never get better

At Kew Observatory, the accuracy of a watch was measured by observation of the variations of its mean daily rate. For an 'A' certificate the trials lasted for 45 days, and included tests in temperatures varying from 40 to 90 deg F (4.4 to 32.2 deg C), in every position with dial vertical, face up and face down.
  1. The average daily departure from the mean daily rate could not exceed 2 s/d
  2. Maximum absolute deviation from mean rate in any position could not exceed 5 s.
  3. The errors should not increase more than 0.3 s/d per deg F. (0.54s/d per degC)
From the COSC standard ISO 3159 1976: Of the seven COSC test measurements taken, the closest we have to the Kew trial is:
  1. Mean variation in rates ‘V’ is the arithmetic mean of the five absolute values of variations in rates obtained for the five positions of the watch during the first 10 days of the tests. Note: The variation in rate is the difference between two consecutive daily rates in identical environmental conditions.: Max of 2s
  2. Greatest deviation in rates ‘P’ is the absolute value of the greatest of the ten differences between one of the first ten rates and the mean daily test rate.: Max 10s
  3. Thermal variation ‘C’ is obtained by subtracting the rate at 8 deg C from the rate at 38 deg C, the whole being divided by the temperature difference, expressed in degrees Celsius. : error 0.6s per deg C
I would tend to draw the following conclusions about the Observatory Trials:
  1. Deviation in mean rate: Essentially identical to COSC
  2. Greatest devation from mean rate: Harder than COSC by a factor of '2' (which is really quite a lot tougher)
  3. Thermal control: Very similar to COSC
As the Japanese had dominated the two preceding events, in 1972 some Swiss watch manufacturers demanded the end of the observatory competitions. Now, all chronometers are equal on a pass/fail basis. Apparently, this was the one condition demanded by the Swiss watch industry when COSC was founded in 1973. So who performs best under current COSC testing? Well, we'll never know. The COSC does not reveal this information. Were COSC to introduce any sort of ranking by test results, Swiss watchmakers would be competing on the basic timekeeping qualities of their watches which is not necessarily linked to brand hierarchy! So don't expect any change anytime soon.

No comments: