How are you going to power the clock?
Is the clock going to be located permanently indoors?
There is so much 60Hz AC signal being radiated that you can pickup the signal without a direct connection.

You can build a solar powered clock. I never completed mine. I had laid out the PCB but never got around to having it fabricated.
It would keep track of solar time and was designed to run for 1000 years without requiring adjustment.

Powering from battery
I have never heard of that solar method, I'm going to read on it
Thanks

 

Powering from battery
I have never heard of that solar method, I'm going to read on it
Thanks

I am not sure how much info you will be able to find because I came up with the idea after having been inspired by this:

http://longnow.org/clock/

 

Precision is normally quoted in terms of percent error or drift.
For "precise" timing applications, a quartz crystal oscillator is commonly used.
Accuracy is of the order of 10ppm for a typical crystal oscillator.
1000ppm = 0.1%
100ppm = 0.01%
10ppm = 0.001%
1ppm = 0.0001%

There are 1,036,800 seconds in 12 days.
Therefore a crystal oscillator with 10ppm drift will drift by about 10 seconds in 12 days.

If you want long term accuracy for a time-of-day clock, use the 50Hz or 60Hz line frequency.
The short term stability of line frequency is about ±0.03%. However, the total number of cycles is corrected daily so that long term error is practically zero.

You will discover that any RC oscillator will result in very unsatisfactory time-keeping.

I recall building a divider (1 Hz output) with two chips ( CD4000 series) and a 2" wire "antenna" to catch the 50 Hz radiated by the mains.
Another option is just disassembling an alarm clock. Most of them use a small circuit which usually gives a 2 Hz (not 1) output pulse.

 

The guy who would come over and tune my father's piano used a set of tuning forks.

Hmmm ... a 32768 Hz watch "crystal" actually is a tuning fork inside, so that could meet the requirement.

https://en.wikipedia.org/wiki/Crystal_oscillator#Operation

Or you could use an actual 440 Hz tuning fork. Way in the back of my brain is a glimmer of a magazine article about this, but I have no idea which mag.

Back to the oscillator. An oscillator circuit needs a reactive element of some kind, and the value of that element is going to change with temperature. A tuneable L-C or R-C circuit that drifts less than +/-1% as air conditioning comes and goes will be difficult to achieve, and that is an error band of +/- 14.4 ***minutes*** over 24 hours. The long-term average error will be less than that, but the RMS error value is almost 1 minute per hour.

"Classical era" oscillators were L-C, both because that's all they had and because those guys knew a lot more about inductors than your average EE today. BUT, they were not 60 Hz or anything close to that low of a fundamental.

OK, not all oscillators. Some were alternators, an AC generator with lotsa poles, driven at very high speed by a motor with a mechanical governor. IIRC, the first AM voice transmitter was at 60 kHz.

https://en.wikipedia.org/wiki/File:Fessenden_alternator_transmitter_(1906).jpg

ak

 

Terrible idea!

Ha ha, they did it for years