[time-nuts] 60hz disciplined watch follow up
bruce.griffiths at xtra.co.nz
Fri Apr 22 01:28:38 UTC 2011
tom jones wrote:
> My orignal post that 60hz light flicker received at citizen ecodrive solarcells is used to
> discipline the watchs rate, needs a little correction to my post as brought to my attention by the replies.
> Yes it is actually 120hz flicker 60 positive cycle flickers and 60 negative cycle flickers per
> second thus 120 flickers per second.
> But I will suggest the idea that positive cycle flicker and negative cycle flicker probably presents different voltage levels for positive and negative cycles at the solarcell? Not to mention probable phase changes in the light itself caused by the positive and negative cycles?
Not likely unless in the absence of substantial third (and other odd)
harmonic distortion the light output from an incandescent bulb or a
flourescent source will not differ between positive and negative mains
> Anywhy let me continue to reference 60hz disciplineing rather 120hz disciplineing.
There will be little or no 60Hz signal component other than that due to
magnetic or capacitive coupling.
> Now imagine an algorithm for disciplineing the watchs reference (32767 xtal), which simply looks for stable 60hz.
> Lets say for example 120 stable consecutive cycles of 60hz (2 seconds) occur consecutively. The watch could easily detect this stable instance and deposit its sample
> in a high quality storage bin.
Idle conjecture without reasonable estimates of the of the likely
accuracy isnt particulalry useful.
> Conversly the watch with its inexpensive reference oscillator could also detect unstable 60hz samples at its solarcell and deposit those samples in a trash bin for disposal.
Method of measurement used??
> My estimate of the citizen ecodrives automatic rate adjustment occures at 10 day intervals.
> How many consecutive cycles of precise 60hz ac occur in 10 day of sampleing....no one knows.
Statement is meaningless without associated error bands etc.
> maybe it happens thousands of times(in 10 days) but each time only for 1 second duration or 60 consecutive cyles... who knows. One would assume this happens sometime.
> Maybe in 10 days of 60hz sampling from the watches solarcell there are instances where 600 consecutive precise cycles occur ?
> And what if the watch could store those longer consecutive cycles in a higher quality storage bin (for latter averageing).
> Eventually the watch says I've observed many instances of (short duration) precise rate flickers. But none at exactly 60.0000? hz.
> Then the watch says its my reference that is off frequency not the precision instances of 60hz that I've observed.
Without a detailed measurement scheme such speculation isnt particularly
> So the watch corrects its reference, empties its storage bins and starts a new 10 day cycle.
> No one could be more amazed than myself to discover that my citizen stainless skyhawk holding time within 13 milliseconds
> day by day for weeks (measurements taken approximately at 11am +- 2hours).
> I noted handling this watch with my hot hands caused a 1 milisecond per minute (estimated) additional error as the watch temperature increased.
> Thus during all subsequent measurements I handled the watch by its band as briefly as possible to eliminate temperature errors.
> The only thing citizen watch company could do to make me more happy is a 60hz frequency alarm feature to alert me of disturbances on the power grid !
More information about the time-nuts