Originally posted by grey580
reply to post by Hellhound604
Incorrect on the battery part. I've had this conversation with a Swiss watch factory owner on batteries while working as a watchmaker for a watch
Watch batteries and watches are engineered to keep running normally as long as there is a certain voltage/amperage in the battery. once that point is
reached the watch stops. watches consume so little power anyways that's why the battery can last for a year or more on a wristwatch.
Temperature plays a huge factor in watch time deviation. Most watch movements do not have a thermocompensator built in. Usually you only find that in
higher end wristwatches.
Krieger Watch had a Marine Chronometer model that was probably on of the most accurate wristwatches in the world giving a plus or minus 8 seconds per
New one for me that the clocks in europe used the frequency of the power to regulate themselves. I would think that wouldn't be very reliable.
In a properly designed watch I would agree with you, but one never knows with very cheap watches, once the battery voltage goes below a specific
threshold, the crystal isn't excited enough, so instead of +- 5ppm it could easily go to +-100ppm. (for critical applications you would use 1.5ppm,
and if more than that you need to go to TCXO's (temperature compensated crystal oscillators) which can be up to .5ppm, or higher you can go to OCXO
(Oven Controlled Crystal oscillators) or higher still GPS-based timers, which give you up to a 340nS resolution).
Most wrist watches uses a diode for regulation of the crystal oscillator, which is calibrated to typically 37C (body temperature), that is why, if you
remove your highly accurate wrist watch, you will see it being less accurate if it is not on your arm. So, your body acts as a nicely calibrated
constant temperature source, effectively turning the cheap crystal into a nice TCXO. That also explained why, in the days before internet, your cheap
wristwatch was more accurate than the clock in your very expensive PC.
Mains frequency, whether 50 or 60Hz, should never deviate more than +-0.2%, so a properly filtered mains time-base is actually more accurate than a
uncompensated low-cost crystal, that is about 0.05ppm/C.
I have noticed 2 distinct differences in this thread. Some posters said that the clocks on their ovens, etc, which most probably uses the mains
frequency as a time base drifted, and that is easily explained by the mains frequency not correct anymore.
Some other posters have stated, that whilst their oven clocks remained on time, it was their wristwatches, and other digital clocks (which may/may not
use crystals), drifted. If more than one of those clocks started to run fast/slow by the same amount, then there is something else going on.
BTW. some more complicated clock systems ( I am now thinking some advanced cars, etc), might use a programmable oscillator. The 32kHz signal is still
used, but instead of driving a simple oscillator, it is a timing input for a microprocessor, or a dedicated RTC device, so if the software has some
bugs, that reprograms the divider or the RTC-registers of the processor or the RTC, then of course, the time will be wrong.
edit on 9/6/2011 by Hellhound604 because: (no reason given)