I mean you think to yourself "one second of error per million years must be quite enough overkill" and then these beautiful people come to show you wrong. I'm not sure who will ever see the difference but really what a job well done!
I’ve often wondered if a hobby-class atomic clock can be built with “a less accurate gas” that is easy to excite and measure in a feedback loop simply because it’s available in a handy package that lends itself for experimentation without having to mess with melting glass and bottles of pressurised gas. E.g. neon, nitrogen or mercury vapour.
The reason I’m asking is because in RF we often need a stable reference, and these come in a clear $ for phase noise relationship: RC, LC, xtal, TCXO, GPSDO, YIG, Rubidium, …
Price-wise, all atomic clocks come after Rubidium. But would it be possible to build an atomic clock that sits between TCXO and Rb both for price and phase noise, by employing a non-exotic gas in a readily available lamp?
There are different types of atomic clocks, but in most common types, the output comes from a crystal oscillator, or more generally frequency synthesizer, which is then slaved to some spectral feature in the "physics package". That is to say, the phase noise is as good as that of the crystal in the unit, but the longer term frequency stability is much improved by the slaving.
An exception is an active hydrogen maser, which directly outputs the frequency of atomic transition. It has very good phase noise, but is a rare beast, which is only used where it is absolutely necessary.
You can pull the frequency of a crystal resonator circuit by changing the loading capacitance, for example. You may use a varactor or any semiconductor junction. Doing so doesn't really affect phase noise.
I'm probably missing the right terminology. I would have thought the feedback loop creates jitter of its own. Phase locked VCOs are generally noisier than a crystal alone. But maybe I'm overthinking it and the loop bandwidth can be made very narrow with trimmed crystal and long integration times or whatever.
I mean you think to yourself "one second of error per million years must be quite enough overkill" and then these beautiful people come to show you wrong. I'm not sure who will ever see the difference but really what a job well done!
Somewhat off-topic:
I’ve often wondered if a hobby-class atomic clock can be built with “a less accurate gas” that is easy to excite and measure in a feedback loop simply because it’s available in a handy package that lends itself for experimentation without having to mess with melting glass and bottles of pressurised gas. E.g. neon, nitrogen or mercury vapour.
The reason I’m asking is because in RF we often need a stable reference, and these come in a clear $ for phase noise relationship: RC, LC, xtal, TCXO, GPSDO, YIG, Rubidium, …
Price-wise, all atomic clocks come after Rubidium. But would it be possible to build an atomic clock that sits between TCXO and Rb both for price and phase noise, by employing a non-exotic gas in a readily available lamp?
There are different types of atomic clocks, but in most common types, the output comes from a crystal oscillator, or more generally frequency synthesizer, which is then slaved to some spectral feature in the "physics package". That is to say, the phase noise is as good as that of the crystal in the unit, but the longer term frequency stability is much improved by the slaving.
An exception is an active hydrogen maser, which directly outputs the frequency of atomic transition. It has very good phase noise, but is a rare beast, which is only used where it is absolutely necessary.
How do you discipline a crystal without introducing significant phase noise?
You can pull the frequency of a crystal resonator circuit by changing the loading capacitance, for example. You may use a varactor or any semiconductor junction. Doing so doesn't really affect phase noise.
I'm probably missing the right terminology. I would have thought the feedback loop creates jitter of its own. Phase locked VCOs are generally noisier than a crystal alone. But maybe I'm overthinking it and the loop bandwidth can be made very narrow with trimmed crystal and long integration times or whatever.
Time-servers require hash sigs.
My back of a envelope maths suggests it's accurate to about 1 second every 300 billion years.
43.5 ms over the age of the universe (13.8 billion years).
Article in Nature https://www.nature.com/articles/s41586-026-11072-8
Lutetium-176, element 71, better than 1 part in 10^18.
As long as we don’t go building a glass clock…
... And still come home late for supper