Atomic clocks are so last epoch, it's time someone nailed down the nuclear clock An international team of researchers has, for the first time, coupled an atomic nucleus to an atomic clock to compare differences in their timekeeping frequencies. The breakthrough promises to ease the development of the next generation of ultra-precise timekeepers based on nuclear physics and help study the fundamental …
Comparing clocks is relatively easy, mostly they operate in the MHz or GHz region so you can look at the phase difference and that can be measured with great precision.
For example, in my own work I compared two 10MHz frequency references by mixing them with a 10MHz + 15Hz oscillator (also OCXO based) and then the two 15Hz results I filtered and then used cheap-ish reciprocal counting frequency counters (TTi-AMI TF930) to measure the offset to 6 digits (so 0.000015 Hz or so) and as that is relative to 10,000,000 MHz I could see differences of the order of 1 second in 21,125 years in terms of short-term wander. And that was all for under £1k in hardware!
If you ever have to adjust the reference oscillator in some bit of test equipment then the simple choice is to put that and an off-air GPS disciplined reference in to an oscilloscope in XY mode and watch the Lissajous figure. If it takes 5 seconds to flip that is 10 seconds for one cycle of error, and for the usual 10MHz that is 0.01ppm error, etc.
Now doing the same for state-of-the-art sources is a lot harder, but the same principle applies!
Pretty much what I had expected, thanks - but those methods only give a a relative accuracy between two clock sources.
If you're trying to claim that a clock is accurate (going all the way back to Harrison!) then I suppose the task is to make multiple clocks and see how they drift relative to each other? And a clock is more accurate if it drifts less (or a bunch of them drift less against each other? i.e. how do you know which one is right?
Apropos of nothing in particular: in my broadcasting days forty years ago, we indirectly phase locked local references to a rubidium master clock in London. It was strongly rumoured that it was affected by the phase of the moon... (in those days, everything was based on a 4.43361875MHz signal - the colour burst frequency.)
Strictly speaking the comparison is only about short-term (as in fractions of a second to days, etc) variations, and would not show systematic long-term drift if it were something common to all oscillators. To check for environmental effects (temperature, air pressure, magnetic field, etc) you would need to torment one example oscillator and see what it did relative to its unmolested brothers.
Comparing two does not tell you which is best, they might be equal, or one near-perfect and the other flaky, etc. So you need 3 oscillators to compare in what is known as the "3 corner hat" test http://wriley.com/3-CornHat.htm and then you can work out roughly how they are all ranked.
As for lunar effects I'm not sure, but if you used a LF source for the frequency reference (e.g. the MFS time code on 60kHz or the Radio 4 on 198kHz) in the pre-cheap-GPS days then there is a strong diurnal impact as the ionosphere changed day/night ands so the LF propagation path varies a bit.
> how one compares clocks?
Is it a fashionable, expensive brand? Does some "celebrity" have one? Is it trendy enough? Is the material expensive expensive looking enough? Does some "celebrity" have one? Is it covered with enough diamonds?
Does some "celebrity" have one?
"...Though... how about a description of how one compares clocks?...
A suspicion is that the only thing one need do is wait for the decay of a proton--approximately some 1034 years.
-----------------------------------------
"Every--seemingly--very hard problem has a very easy solution if you just look at it the right way." (paraphrase)--Douglas Adams
10³⁴ years? That's nothing! When I were a lad, we didn't have protons, we literally had to wait forever for something to happen, anything at all, just so we could set our watches, and it never did ... we were constantly late to our appointments!
Doubtful. 99 percent of science, ever since the first caveman chewed a rock and wondered WHY chewing a rock made his teeth hurt, has been to reach the conclusion the guy paying for the research wants. That actual science ever does happen is a happy accident, and is usually not accepted until long after the discoverer has been run out of the halls of learning and censured by his peers.
So, it's not just in my sad corner of the world that a professor "asked" that he is co-author on each published paper? Of course, there's no hard evidence of his instruction to do so. But "surprisingly", he was a co-author on nearly every paper published before this got media attention. And he is far less contributing since it got into the news...
Sure, this is good work and all, but at US R1 universities, doctoral students typically publish original work, present at conferences, etc. In some technical fields, a doctoral dissertation is often basically a collection of a few published papers lead-authored by the candidate.
as a physics fan-boy, about the fine structure constant. It is used to compute the speed of light. And the rate of nuclear decay.
Both of these look to me like fundamental ways to measure time. I guess what I'm thinking is that a change in the fine structure constant might change the time axis. If so, attempts to measure its drift are only going to be meaningful only if there is a way to define time that doesn't end up falling back to this constant.
Maybe I've had too many of these ---------------------------------->
Nice article on the APS site Shedding Light on the Thorium-229 Nuclear Clock Isomer.
The extraordinary balance of the massive strong and electromagnetic forces which uniquely gives Thorium-229 a transition ~8eV reminds me of the Douglas Adams scripted The Pirate Planet Calufrax's Queen Xanxia held in near death stasis.
Quanta Magazine has a lot of rather curious background https://www.quantamagazine.org/the-first-nuclear-clock-will-test-if-fundamental-constants-change-20240904/ including Cold War vats of U-233 containing solutions.
I noticed the ribbons of time image used for this article have IV for 4 rather than the traditional IIII which is preferred I believe for asthetic reasons.
The rules how to write roman numerals, in either subtractive notation ("IV") or additive notation ("VI") have apparently always been a little fuzzy: You can find mentions of "XIIII", but also of "IIX", as far back as the actual Romans, so it's not just an error due to ignorance (IIRC the 4th Colosseum entry shows "IIII" instead of "IV").
In short, anything goes.
> a little fuzzy:
No, clockfaces have a smaller subset of variants, because all the glyphs are in one small circle. An idiosyncrasy that could be overlooked as you walk around a Colosseum gate by gate might stand-out on a clock face. The usual dodge is to use 'IIII", an illegal number, for 'IV', because it sits across from 'VIII", to balance the stroke-density.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
