- Posts
- 2,321
- Likes
- 6,656
Deafboy
··His Holiness Puer SurdusI would like to share my method for measuring and regulating a recently received Omega Seamaster « Titane » (a.k.a. Polaris).
This much maligned 1980’s Gerald Genta design has a unique gold inlay titanium case. The dial is of a dark matte gray with white minute markers and cabochon lume hour markers. It is pretty much the only date dial design that I find attractive. My watch has the reference (TR 396.1121) which dates from 1988, and came with either the 1441 or 1438 movement. I got lucky because I obtained the 1441, which is a high grade caliber using a thermally compensated quartz crystal. This movement uses a second quartz crystal as temperature sensor and is capable of 2 seconds/year accuracy. It has two switches on the movement's electrical circuit board that permits adjustments in increments of 0.33 seconds/month (i.e. 4 seconds/year).
Being of a scientific nature I wanted to analyze the accuracy of this watch. I don’t have dedicated watch instrumentation but I do have a Hewlett-Packard 5335A timer (well… who doesn’t?). It has two channels so I used one channel for the watch seconds electrical signal from the movement’s circuit (see open watch picture) and compared it to a GPS signal (which provides an accurate 1 pulse per second signal - pps) on the other channel.
The figure below shows three curves of results. The one traced in Blue is the watch’s error compared to the reference 1 pps signal. We see that the movement uses a variant of the Inhibition that @Archer described in a recent post. Every few minutes the watch makes adjustments to the rate based on the temperature and regulation setting. The deviation from perfect time is a loss of 2.5 milliseconds in the hour of measurement, which corresponds to -23.2 seconds/year. After tapping the switches on the movement several times and obtained Red curve, which has a slope of -3.2 seconds/year. One final adjustment I obtained -1.2 s/year over the hour time span (Green). As a sanity check I did a longer test (60 hours) to see if the accuracy would be maintained. The results are shown in last figure, where I obtained a deviation of 6 ms or +1 second/year. During this period the ambient temperature varied from 17 to 29 degrees Celsius.
Not bad for a 30 year old movement! 😎
Michel
This much maligned 1980’s Gerald Genta design has a unique gold inlay titanium case. The dial is of a dark matte gray with white minute markers and cabochon lume hour markers. It is pretty much the only date dial design that I find attractive. My watch has the reference (TR 396.1121) which dates from 1988, and came with either the 1441 or 1438 movement. I got lucky because I obtained the 1441, which is a high grade caliber using a thermally compensated quartz crystal. This movement uses a second quartz crystal as temperature sensor and is capable of 2 seconds/year accuracy. It has two switches on the movement's electrical circuit board that permits adjustments in increments of 0.33 seconds/month (i.e. 4 seconds/year).
Being of a scientific nature I wanted to analyze the accuracy of this watch. I don’t have dedicated watch instrumentation but I do have a Hewlett-Packard 5335A timer (well… who doesn’t?). It has two channels so I used one channel for the watch seconds electrical signal from the movement’s circuit (see open watch picture) and compared it to a GPS signal (which provides an accurate 1 pulse per second signal - pps) on the other channel.
The figure below shows three curves of results. The one traced in Blue is the watch’s error compared to the reference 1 pps signal. We see that the movement uses a variant of the Inhibition that @Archer described in a recent post. Every few minutes the watch makes adjustments to the rate based on the temperature and regulation setting. The deviation from perfect time is a loss of 2.5 milliseconds in the hour of measurement, which corresponds to -23.2 seconds/year. After tapping the switches on the movement several times and obtained Red curve, which has a slope of -3.2 seconds/year. One final adjustment I obtained -1.2 s/year over the hour time span (Green). As a sanity check I did a longer test (60 hours) to see if the accuracy would be maintained. The results are shown in last figure, where I obtained a deviation of 6 ms or +1 second/year. During this period the ambient temperature varied from 17 to 29 degrees Celsius.
Not bad for a 30 year old movement! 😎
Michel
Edited:
