Re:'Temperamental' Big Ben was ringing up to six seconds late
by VLM on Wednesday August 26, 2015 @09:06AM
(#10122593)
> It might require changes to the pendulum
Not necessarily, tweak it to run slightly fast and rely on the natural dampening ability of a magnetic field to generate a slowing force on moving conductors, then apply a (an extremely) modest (AC to demagnetize) and controlled magnetic field to slow it down to dead on accurate (probably as determined by a $10 GPS module, or I guess a 7 or so pound gps module). Also you can sense pendulum position with the usual optical technologies, so no contact with historical relics needed there either.
Given the "fun" of controlling precision currents vs the ease of controlling precision times/durations its probably easier to run a standardized constant current thru the correctional electromagnet for a varying X nanoseconds just before the top of the hour rather than trying for a continuous 8 bit precision field that'll never really settle down.
A better design than running the correction routine hourly would be continuous position/time monitoring and run the correction field every time the phase variance between the clock and a GPS receiver exceeds a tenth of a second absolute difference or whatever, then monitor how often that correction happens (and alert when broken, etc).
The best part of doing it non-contact is other than dust not settling where the apparatus is sitting, hundreds of years from now researcher types will have no evidence the apparatus was there. Which I guess would be good if they're researching 1800s clocks and bad if they're researching 2000s microcontrollers... hmm.
Not necessarily, tweak it to run slightly fast and rely on the natural dampening ability of a magnetic field to generate a slowing force on moving conductors, then apply a (an extremely) modest (AC to demagnetize) and controlled magnetic field to slow it down to dead on accurate (probably as determined by a $10 GPS module, or I guess a 7 or so pound gps module). Also you can sense pendulum position with the usual optical technologies, so no contact with historical relics needed there either.
Given the "fun" of controlling precision currents vs the ease of controlling precision times/durations its probably easier to run a standardized constant current thru the correctional electromagnet for a varying X nanoseconds just before the top of the hour rather than trying for a continuous 8 bit precision field that'll never really settle down.
A better design than running the correction routine hourly would be continuous position/time monitoring and run the correction field every time the phase variance between the clock and a GPS receiver exceeds a tenth of a second absolute difference or whatever, then monitor how often that correction happens (and alert when broken, etc).
The best part of doing it non-contact is other than dust not settling where the apparatus is sitting, hundreds of years from now researcher types will have no evidence the apparatus was there. Which I guess would be good if they're researching 1800s clocks and bad if they're researching 2000s microcontrollers... hmm.