Talk:Atomic clock/Archive 1

Chip scale atomic clock accuracy

While the accuracy of the chip scale atomic clock is mentioned here, it is not mentioned precisely. The referenced commercially available chip scale atomic clock is off by a second every 634.19583967529 years .

This differs substantially from the FOCS 1 clock of Switzerland, which is out by 1 second every 30 millions years .

I suggest that the accuracy of the chip scale atomic clock be listed in either Allan deviation or the time taken for it to be out by a second.

Whole Oats 7:16 am, 1 November 2019 (UTC)

Discussion

"The most accurate time scales are moderated by precise astronomical measurements and the insertion or removal of leap seconds. " seems to be wrong since TAI is just as accurate as anything else, in fact, you could argue that it's more "accurate" than UTC since you can use a precise TAI time in the future and know how long from now that is. With UTC you can't.

---

How did they set the first atomic clock? Did they use astronomical measurements or some other source?

--haavis 12:15, 5 Nov 2004 (UTC)

A: Using astronomical measurements. Does the last paragraph in caesium standard explain it well enough?

I've added some info and references about this. Terry0051 (talk) 01:10, 25 February 2009 (UTC)

---

--tdunc 8:29,10 May 2007

"When the electrons are attracted back closer by the opposite charge of the nucleus, the electrons wiggle before they settle down in their new location. This moving charge causes the light, which is a wave of alternating electricity and magnetism."

This explaination is unacceptable. The author either needs to elaborate or change it all together, why the electron jumps to a lower level. I have reason to believe it is not because of attraction due to charge (which would be constant), rather the electron emits a photon thus lowering its energy level, which could by coincidence mimick charge attraction but I dont feel its the same concept. If so it needs to work both ways, we cant for instance say a repeling charge is responsible for an electron going into a higher energy level. What is with the word "wiggle" anyway?

I have fixed the article to avoid this issue. The explanation is simply quantum mechanics. Chapter 9 of Griffith's undergraduate QM textbook has a cogent derivation. Amcfreely 07:01, 16 May 2006 (UTC)

---

There is also a transit delay of approximately 1 ms for every 300 kilometers (186 mi) the receiver is from the transmitter. When operating properly and when correctly synchronized, better brands of radio clocks are normally accurate to the second.

Don't you mean millisecond instead of second ? 1 ms for every 300 kilometers means you must be 300,000 (300 x 1000) kilometers from the source to cause a 1 second delay, since the circumference of the earth is ~40,000 km, this does not make much sense.

Radioactivity?

How radioactive are small atomic clocks? Is it safe to make handheld devices with them?

A small atomic clock is about as radioactive as any other similarly-sized piece of electronics: that is, slightly less radioactive than the average brick. --Carnildo 09:33, 10 May 2006 (UTC)

atomic clocks are not based on radioactivity or decay but on electromagnetic wave frequencies, and the materials used (such as cesium 133) are stable, otherwise the clock wouldn't work well.81.206.145.191 21:43, 13 May 2007 (UTC)

Right. Atomic clocks are based on the same effect that MRI medical imaging uses. The atomic nucleus in a magnetic field, can transition between spin states (called hyperfine transitions), and gives off microwave radiation. It's got nothing to do with nuclear radation from radiactivity. DonPMitchell (talk) 02:38, 11 August 2009 (UTC)

Cesium-133 is a stable isotope and is not dangerous in a radioactive sense. Rubidium-85 is also a stable isotope, where as Rubidium-87 is vaguely radioactive, decaying with beta-radiation (electron) with half-life of 4.88E10 years, which isn't directly lively. For all practical uses it is safe. Being alkali metals, they would be seriously bad if in contact with skin or water for that matter. Maybe a small text should state what needs to be said: "atomic clocks does not use radioactivity as a mechanism for measurement of time". It is a common misconception and clearly pointing this fact out should be helpful. Cfmd (talk) 00:43, 24 January 2011 (UTC)

Synchronization

How are Atomic clocks synchronized across the world, or are they not? Do they have to be in the same place to be synchronized off each other sort of thing? Even then how do they set them with such prescison accuracy? —The preceding unsigned comment was added by 90.192.138.193 (talk) 21:02, 22 December 2006 (UTC).

they can be synched to each other 1: over a symmetric latency connection (such as radio waves), or 2: by using GPS as a reference. a location can tell how it's time compares to GPS, or a single GPS satellite, say 50 nanoseconds ahead. then another place can use that information.81.206.145.191 21:43, 13 May 2007 (UTC)

Or you can pop a portable atomic clock with a battery backup on a plane. Jim77742 01:25, 20 June 2007 (UTC)

Using airplanes would be unpractical, because Einstein learned us that clocks run slower the faster their speed and the higher their altitude (or the lesser the gravitational field). See [1] and [2] for explanation and experiment with cesium clocks at moderate altitude. Jaho 00:16, 5 July 2007 (UTC)

An answer to the synchronization issue can be found on http://tycho.usno.navy.mil/twstt.html. Jaho 00:29, 5 July 2007 (UTC)

While the clocks may not need to be synchronized, they are often coordinated to roughly have the same time. Use of GPS, GLONASS and TWSTT for such synchronisation is being used. Caesium beam clocks is then usually allowed to count freely and unsteered and then GPS, GLONASS and TWSTT is used for comparison between clocks not being local to each other. Such comparison forms the base of the EAL, TAI and UTC time as being maintained by BIPM. For clocks with not requirement to be part of the BIPM network, they are usually synchronised once and for all using a GPS receiver and then set alone for itself. Many telecom applications does not require the network clock to even have Time of Day but only to give a very stable 2,048 MHz clock to the network. Cfmd (talk) 00:50, 24 January 2011 (UTC)

Mercury Atomic Clock Keeps Time with Record Accuracy

www.nist.gov/public_affairs/releases/mercury_atomic_clock.htm —The preceding unsigned comment was added by 218.102.23.117 (talk) 14:30, 8 January 2007 (UTC).

Radio clocks

"Atomic clocks" for less than $50: This marketing term "atomic clock" refers to a radio controlled device receiving a time signal from one or more transceivers connected to real atomic clocks. Examples at: koolatrononline(dot)stores(dot)yahoo(dot)net/hummer-multi-bank-automatic-clock(dot)html www(dot)ehow(dot)com/how_2099664_buy-atomic-clock(dot)html —Preceding unsigned comment added by 131.130.249.17 (talk) 09:07, 16 February 2010 (UTC)

The section in the article, "Radio clocks", covers these devices. Jc3s5h (talk) 18:00, 16 February 2010 (UTC)

Ytterbium atomic clocks

Someone should add something to this article about the new ytterbium atomic clocks under development. I don't know a lot about them, and, if no one else does it, may end up doing it myself after I read into it more, but if not, at least a mention should be added. —Preceding unsigned comment added by 99.249.74.84 (talk) 21:27, 15 December 2010 (UTC)

Maser

The following passage from this article:

“The first atomic clock was an ammonia maser device built in 1949 at the U.S. National Bureau of Standards (NBS, now NIST).”

seems to conflict with the passage from the “maser” article:

“Theoretically, the principle of the maser was described by Nikolay Basov and Alexander Prokhorov from Lebedev Institute of Physics at an All-Union Conference on Radio-Spectroscopy held by USSR Academy of Sciences in May 1952. They subsequently published their results in October 1954. Independently, Charles H. Townes, J. P. Gordon, and H. J. Zeiger built the first maser at Columbia University in 1953.”

Some explanation to harmonize the dates in these two passages would be helpful. Psalm 119:105 (talk) 09:58, 17 January 2011 (UTC)

Applications

Atomic clocks can also be used as sensors for gravitational and magnetic fields. See: http://www.wired.com/science/discoveries/news/2007/12/time_nist?currentPage=2 72.221.84.128 (talk) 14:56, 11 January 2012 (UTC)

Lord Kelvin reference

The statement that Kelvin proposed a clock based on atomic transitions seems a bit misleading. He proposed using the known distinct spectral frequencies of atoms, but wasn't he about 30 years too early to know about atomic transitions? — Preceding unsigned comment added by 173.206.4.23 (talk) 03:41, 1 November 2012 (UTC)

I noticed that as well. I thought the sentence went beyond misleading to just wrong. --Davefoc (talk) 08:30, 29 April 2015 (UTC)

In my opinion, this statement about Lord Kelvin looks like a hoax. The reference is completely off topic, the given page of Treatise on Natural Philosophy is about the mechanical oscillation of cords. --Pasum (talk) 16:06, 19 December 2017 (UTC)

Assessment

Found the following undated assessment in comment on this page. --Kvng (talk) 19:11, 6 September 2012 (UTC)

Rationale for class: close to B, but it's a bit overly technical in places, the prose could use a little work, and something needs to be done about the power consumption section.

Accuracy vs precision

A number of places in the article include references such as "accurate to one second". These should actually read "precise to one second". Accuracy refers to the central tendency of a measurement, whereas precision refers to the possible range of deviation. — Preceding unsigned comment added by 108.68.83.101 (talk) 03:00, 22 October 2013 (UTC)

Please cite a source for your statements about the meaning of accuracy vs. precision. In common use among engineers, accuracy is often thought of as the difference between a measurement and the true value (if the true value can be ascertained) while precision is the repeatability of a measurement with a given measuring instrument. Jc3s5h (talk) 14:29, 22 October 2013 (UTC)

NIST's mercury ion clock

So, if that's the standard by which new clock technology is measured, why do we hear no more of it? Jim.henderson (talk) 21:08, 2 January 2015 (UTC)

Diagrams

The section Physics package realisations needs several diagrams added to improve clarity. RJFJR (talk) 13:59, 22 October 2015 (UTC)

Nuclear clock

Hi, nuclear clock redirects here while there is not mention of it within this article. Could someone says few words about it within this article or create another article for this kind of clock? Pamputt (talk) 16:56, 19 August 2016 (UTC)

I replaced the redirect with a stub article. — Edgar.bonet (talk) 11:16, 22 August 2016 (UTC)

"According to sources"?

@Francis Flinch: Thanks for checking my edits, when you removed the uncertainties from the frequency table in this edit with the edit comment "according to sources", I don't quite understand.

The uncertainties were all taken from the cited sources. For example, the rubidium uncertainty was taken from the cited source, converted by a simple WP:CALCulation from a relative uncertainty of ±3×10⁻¹⁵ to an absolute uncertainty of 6834682610.904324(20).

Just re-checking in case I made a mistake: ⁸⁷Sr has a frequency of 429228004229873.4 Hz with a relative uncertainty of ±1×10⁻¹⁵, which is 0.43 Hz, so 429228004229873.4(4) is correct. (Unless you'd prefer 429228004229873.40(43), but I think that's excessive precision.)

And the hydrogen maser citation (available at http://cyber.sci-hub.ac/MTAuMTA4OC8wMDI2LTEzOTQvOS8zLzAwNA==/essen1973.pdf if you'd like to check) already says 1420405751.7662±0.003 Hz.

I thought the uncertainties were an informative contrast to the exact-by-definition caesium value, but I can see how someone could disagree. You might find them unhelpful or distracting for various reasons, but "according to sources" confuses me. Could you clarify? 71.41.210.146 (talk) 17:02, 5 September 2016 (UTC)

I think it is a good idea to simply state the recommended value of the frequency in the sources without complicating things with uncertainties. Maybe the involved uncertainties can be mentioned with some extra explanation for the casual reader.--Francis Flinch (talk) 18:13, 5 September 2016 (UTC)

Attosecond as unit of precision

I think the unit "attosecond" is dubious in this passage:

In the future this might lead to redefine the caesium microwave based SI second and other new dissemination techniques to transfer clock signals will be required that can be used in both shorter-range and longer-range (frequency) comparisons between attosecond (sub-1 × 10⁻¹⁷ s) accurate clocks without significantly compromising their performance.

The passage seems to call for a unit of relative precision (10⁻¹⁷ to 1), not a unit of time. Can User:Francis Flinch supply a quote from one of the cited sources to justify this unit? Jc3s5h (talk) 18:16, 14 November 2016 (UTC)

This is obviously wrong. Using units of time for measuring clock accuracy can only make sense in that part of the sigma-tau plot where the slope is −1, i.e. where the accuracy is limited by jitter/phase noise. The quoted fractional accuracies refer to the bottom of that plot, where the slope is zero.

The fractional accuracies can be converted to time accuracies by multiplying by the averaging time (σ_t ≈ σ_yτ, to within some multiplicative factor). If the averaging time was just one second, then a fractional accuracy of 10⁻¹⁷ could be more or less equivalent to 10 attoseconds. But it seems to me that the accuracies achieved are always for averaging times far longer than one second. For example, Fig. 1 of the first ref (PRL 97, 020801), shows a slope −1/2 from 10¹ to 10⁵ s, with σ_yτ no less than 10⁻¹¹ s. The third reference (Nature Physics) says “The record for stability is now held by an optical lattice clock based on ytterbium atoms, which recently demonstrated fractional frequency instability of 1.6 parts in 10^18 for an averaging time of 7 hours”. That puts σ_yτ ≈ 4×10⁻¹⁴ s. There is also this article in arXiv demonstrating “accuracy (6x10^{-18}) better than a single ion-based clock, with vastly reduced averaging times (3000 s).” which implies σ_yτ ≈ 1.8×10⁻¹⁴ s. While all those numbers are really impressive, we are in the tens of femtoseconds range, nowhere near the attosecond.

— Edgar.bonet (talk) 21:30, 14 November 2016 (UTC)

Good point, changed text accordingly.--Francis Flinch (talk) 12:26, 15 November 2016 (UTC)

But then, what does “comparisons between sub microwave accurate clocks” mean? Is “sub microwave” a synonym of “optical”? If so, why not use “optical” instead? — Edgar.bonet (talk) 15:39, 15 November 2016 (UTC)

I hope "better" is a good enough generalization. Tough research groups work on improving optical clocks, we actually can just speculate on what technology future better clocks will be based. A practical improvement of TAI will require a network of many of those better clocks.--Francis Flinch (talk) 16:17, 15 November 2016 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified 5 external links on Atomic clock. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Corrected formatting/usage for http://www.symmetricom.com/media/files/downloads/product-datasheets/DS_SA%2045s_CSAC.pdf
• Corrected formatting/usage for http://www.franklinclock.com/faq.htm
• Corrected formatting/usage for http://www.allanstime.com/Publications/DWA/Science_Timekeeping/TheScienceOfTimekeeping.pdf
• Added archive https://web.archive.org/web/20130408111458/https://gps.afspc.af.mil/gps/archive/2012/nanus/2012034.nnu to https://gps.afspc.af.mil/gps/archive/2012/nanus/2012034.nnu
• Added archive https://web.archive.org/web/20130627155632/http://archive.nrc-cnrc.gc.ca/eng/projects/inms/optical-comb.html to http://archive.nrc-cnrc.gc.ca/eng/projects/inms/optical-comb.html

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 18 January 2022).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 12:55, 11 July 2017 (UTC)

Unsourced

The following huge chunk of text is almost entirely unsourced. moved here per WP:PRESERVE. Per WP:BURDEN, do not restore without finding reliable sources, checking the content against them, and citing them

Physics package realisations

A number of methods exist for using hyperfine atomic transitions. These methods, with their respective benefits and drawbacks, have influenced the development of commercial devices and laboratory standards. By tradition, the hardware that is used to probe the atoms is called the physics package.

One of the Hewlett Packard HP5061A caesium beam atomic clock units used in the Hafele–Keating "flying clock" experiment in 1971 to test relativity theory.

Atomic beam standard

The atomic beam standard is a direct extension of the Stern-Gerlach atomic splitting experiment. The atoms of choice are heated in an oven to create gas, which is collimated into a beam. This beam, consisting of a mixture of atoms in two states, passes through a state-selector magnet A, where atoms of the wrong state are separated out from the beam. The remaining beam is exposed to an RF field at or near the transition frequency. The beam then passes through a space containing a static homogeneous magnetic field before it is again exposed to the RF field. The RF field and the C-field coil will flip the state of the atoms, with a probability depending on how close the microwave frequency is to the atomic transition frequency. After the second RF field exposure the atomic beam passes through a second state selector magnet B, where the atoms that did not change state (are still in the state selected by magnet A) are discarded. This way, the number of atoms which survive magnet B is related to the microwaves' ability to match the atomic transition frequency. After the second state selector, a mass-spectrometer using an ionizer detects the rate of atoms being received.

Modern variants of this beam mechanism use optical pumping to transition all atoms to the same state rather than dumping half the atoms. Optical detection using scintillation can also be used.

The most common isotope for beam devices is caesium (¹³³Cs), but rubidium (⁸⁷Rb) and thallium (²⁰⁵Tl) are examples of others used in early research.

The frequency errors can be made very small for a beam device, or predicted (such as the magnetic field pull of the C-coil) in such a way that a high degree of repeatability and stability can be achieved. This is why an atomic beam can be used as a primary standard.

Atomic gas cell standard

The atomic gas cell standard builds on a confined reference isotope (often an alkali metal such as rubidium (⁸⁷Rb)) inside an RF cavity. The atoms are excited to a common state using optical pumping; when the applied RF field is swept over the hyperfine spectrum, the gas will absorb the pumping light, and a photodetector provides the response. The absorption peak steers the fly-wheel oscillator.

A typical rubidium gas-cell uses a rubidium (⁸⁷Rb) lamp heated to 108-110 °C, and an RF field to excite it to produce light, where the D1 and D2 lines are the significant wavelengths. An ⁸⁵Rb cell filters out the a component of the D1 and D2 line so that only the b component pumps the ⁸⁷Rb gas cell in the RF cavity.

Among the significant frequency pulling mechanisms inherent to the gas cell are wall-shift, buffer-gas shift, cavity-shift and light-shift. The wall-shift occurs as the gas bumps into the wall of the glass container. Wall-shift can be reduced by wall coating and compensation by buffer gas. The buffer gas shift comes from the reference atoms which bounce into buffer gas atoms such as neon and argon; these shifts can be both positive and negative. The cavity shift comes from the RF cavity, which can deform the resonance amplitude response; this depends upon cavity centre frequency and resonator Q-value. Light-shift is an effect where frequency is pulled differently depending on the light intensity, which often is modulated by the temperature shift of the rubidium lamp and filter cell.

There are thus many factors in which temperature and ageing can shift frequency over time, and this is why a gas cell standard is unfit for a primary standard, but can become a very inexpensive, low-power and small-size solution for a secondary standard or where better stability compared to crystal oscillators is needed, but not the full performance of a caesium beam standard. The rubidium gas standards have seen use in telecommunications systems and portable instruments.

Active maser standard

The active maser standard is a development from the atomic beam standard in which the observation time was increased by using a bounce-box. By controlling the beam intensity, spontaneous emission provides sufficient energy to provide a continuous oscillation, which is tapped and used as a reference for a flywheel oscillator.

The active maser is sensitive to wall-shift and cavity pulling. The wall-shift is mitigated by using PTFE coating (or other suitable coating) to reduce the effect. The cavity pulling effect can be reduced by automatic cavity tuning. In addition the magnetic field pulls the frequency.

Although the long-term stability of the active maser is not as good as that of a caesium beam, it remains one of the most stable sources available. The inherent pulling effects make repeatability troublesome and prohibit its use as a primary standard, but it makes an excellent secondary standard. It is used as a low-noise flywheel standard for caesium beam standards.

Fountain standard

Main article: atomic fountain

The fountain standard is a development from the beam standard where the beam is folded back to itself by the Earth's gravity, such that the first and second RF fields are applied during the atoms' upward and downward trips through the same RF cavity, essentially removing phase errors between the two cavities. The slow speed of the atoms also reduces black body temperature shifts. The length of the beam has the same practical limits on vacuum chamber size as a beam clock, but the laser-cooled atoms travel so much slower that the observation time increases about 100-fold (from roughly 10 ms to 1 s) and hence a much higher Q value is achieved in the Ramsey fringes. (The line width is reduced from about 50 Hz to about 1 Hz.)

Caesium fountains have been implemented in many laboratories, but rubidium has even greater ability to provide stability in the fountain configuration.

Ion trap standard=

The ion trap standard is a set of different approaches, but their common property is that a cooled ion is confined in an electrostatic trap. The hyperfine region of the available electron is then being tracked similar to that of a gas cell standard.

Ion traps have been used for numerous ions. ¹⁹⁹Hg⁺ was an early candidate. Quantum logic spectroscopy of a single Al ion became the most precise in 2008. In 2010 an improved setup using a Mg+ logic ion instead of Be was demonstrated

References

-- Jytdog (talk) 15:40, 28 October 2017 (UTC)

1 ns ≠ 30 cm

In this edit Francis Flinch asserts "A timing error of 1 nanosecond or one light-nanosecond results in about a 30 cm (11.8 in) positional error." A source source is provided.

The source says in relevant part "one nanosecond equals to about 0.3 m in distance". However, the 1 ns error is about equal to an error of 0.3 m in the distance between the GPS receiver antenna and the GPS satellite. But GPS receivers do not depend on a single satellite; the solution involves at least 4 satellites, and often more. So the actual positional error will depend on a complex combination of the delays in each satellite and receiver.

In addition, the source explains a procedure for correcting for the clock error by observing the satellite with monitoring station(s). Since these corrections are possible, the error will be eliminated or substantially reduced in the final position calculation. Jc3s5h (talk) 09:30, 15 March 2019 (UTC)

This sentence says: If you have a clock accuracy of +/- 1 ns seconds you have a position accuracy of 30 cm. This is true. If you have a cable between antenna and a receiver, you can correct this antenna delay, if you know the cable length. Nether the less, still you have an inaccuracy of your Position of 30 cm, if you have a time inaccuracy of 1 ns. Also a satelite has only a certain accuracy of the time. It does not have the absolute accuracy. It is as exact as possible. So at the end the Position accuracy depends on all inaccuracy in the chain (Accuracy of satelite, noise at transmission, receiver accuracy). At the end the GPS receiver is very accurate, but not absolute. So this sentence is true. --GodeNehler (talk) 12:31, 15 March 2019 (UTC)

The position error is the error in the position after all computations have been done and all corrections have been applied. If there were an error in the satellite clock of 1 ns, but this error had been detected by a monitoring station and published so that GPS receivers could receive the correction and apply it, that error would not contribute to the position error. The source provided appears to be describing intermediate stages in the calculating the final position, and tells us nothing about the error in the final position. Jc3s5h (talk) 13:06, 15 March 2019 (UTC)

Quesiton

How does an atomic clock work?

WikiProject Time assessment rating comment

Needs more cites to be a B.

Want to help write or improve articles about Time? Join WikiProject Time or visit the Time Portal for a list of articles that need improving. -- Yamara 15:54, 17 January 2008 (UTC)

Hyperfine Transition Frequency

I don't understand what a hyperfine transition is but I think the article should summarize what a hyperfine transition is because the basis of the second involves the concept.

I also think there should maybe be an explanation of the hyperfine transition in the lead, or leave this technical detail for later in the article. ScientistBuilder (talk) 22:16, 12 February 2022 (UTC) The page on the hyperfine structure has a bunch of math and is hard to understand.ScientistBuilder (talk) 22:17, 12 February 2022 (UTC) In trying to understand how atomic clocks work, I went to the page on hyperfine structure, but I don't understand what the fine structure of atom is and what a degenerate state is so I think it would be helpful if someone who has an understanding of this subject helps explain it to ordinary people.ScientistBuilder (talk) 22:19, 12 February 2022 (UTC) The link to the article on Atomic electron transition does not contain a lot of helpful information to help people understand how atomic clocks work when they click on the link in the lead to understand what is going in the belly of an atomic clock. ScientistBuilder (talk) 21:20, 13 February 2022 (UTC)

I agree. The article could also have a more detailed explanation of exactly how an industry standard caesium clock (not the fountain) works. Another point that is not really made clear is that for a primary standard like the fountain there is no more accurate clock to compare it to, so the article should explain how the accuracy estimate is made. I'd also like to see an explanation of why caesium was chosen as the standard. --Chetvorno^TALK 22:21, 13 February 2022 (UTC)

I agree, except, develop it within the article and then summarize it in the lede. I am certainly not an expert, so this is my own, hand-waving, sort of explanation. I offer it only as an intermediate point in the discussion. I am using textbook voice and not encyclopedic voice.

• When you observe the absorption spectrum of a low density, cool gas, you see absorption lines. The energy of photons associated with the wavelength of these lines are the energies needed to promote electrons from one orbital to another orbital. The wavelengths are optical. When we talk about the state of an atom, we mean which orbitals are occupied. In casual speak, we often use state and orbital interchangeably. However, if you look at the spectrum with better resolution, you see that the absorption lines are actually two or more closely spaced lines. This means that what we previously considered to be an electron orbital with a well-defined energy, is instead an orbital with more than one configuration with slightly different but discrete energy levels. The energy difference between the closely spaced lines is the same as microwave photons. Thus, the gas can interact with microwaves.
• Even more handwaving. When we compute orbitals of an atom, we compute each orbital as though it were independent of which other orbitals were occupied. Expect for the Pauli exclusion principal, we assume that the electrons do not interact. That gives us a single well-defined energy for each orbital. However, if we do a more complete analysis that allows for the electrons to interact, we find that some orbitals may have more than one configuration. This allows for the state of an atom to have a finer structure than just which orbitals are occupied. And it accounts for interaction with lower energy, microwave photons.

There you have it. That is my unreferenced, unexpert, unauthoritative synthesis of what I could find recently. It could be all wrong. Constant314 (talk) 04:20, 14 February 2022 (UTC)

Looks good as far as it goes. It should also explain why atomic clocks are primary time standards - why they are much more accurate than other clocks (because the ultimate accuracy of any clock is determined by the Q factor of its resonator, and the atomic transitions of electrons in atoms have extremely high Q) --Chetvorno^TALK 20:28, 15 February 2022 (UTC)

Atomic Clock Definition Caesium-133 Atom at Absolute Zero

The definition of the caesium-133 at absolute zero according to BIPM in 1997 made me wonder how it is possible to measure the second if absolute zero is not attainable? Absolute zero is not possible because of quantum mechanics but can only be approached asymptotically. ScientistBuilder (talk) 16:03, 9 February 2022 (UTC)

This is not really the appropriate forum for this kind of discussion, I guess. The short answer would be that one can measure it within an accuracy bound, with that bound being determined by factors such as uncertainty of the velocity of the free atoms being used. 172.82.47.211 (talk) 14:57, 10 February 2022 (UTC)

one part in 1 × 10⁻¹⁶

this phrasing is probably wrong. we talk about "one part in a million", not "one part in a millionth", and it should either be one part in 1*10¹⁶, or alternatively, lose the "one part in" preamble.

peace קיפודנחש (aka kipod) (talk) 16:36, 25 November 2021 (UTC)

You're right. Fixed. --Roly (talk) 19:01, 25 November 2021 (UTC)

My correction has been reverted. However, I still say that "one part in 1 × 10⁻¹⁶" is wrong. Yes, "1 × 10⁻¹⁶" is a very small number, but "one part in 1 × 10⁻¹⁶" is an extremely large number! Effectively, "one part in" means "one divided by", so "one part in 1 × 10⁻¹⁶" becomes 1 × 10¹⁶; an extremely large error indeed! I spent many years as a metrologist but I will go with the majority if several people disagree with me. --Roly (talk) 10:12, 26 November 2021 (UTC)

Recent additions

Hi, unfortunately I neglected to include a reference. The article discussing construction of "radio" clocks is in a recent EPE magazine and its been covered in "Electronics World" around early 2004. — Preceding unsigned comment added by ‎ 88.81.156.140 (talk • contribs) 07:21, 6 April 2021 (UTC)

Does “a second in XXX million years” really make sense?

In the article, the accuracy of atomic clocks is stated in several places as “expected to neither gain nor lose a second in XXX million/billion years”. I understand that this is only intended to be a more “layman friendly”, or more “intuitive” way to state something like “an accuracy of XXX × 10⁻¹⁶”. However, not only I doubt this way of describing the precision is really more intuitive (since a million years cannot be grasped by the imagination, it's kind of an abstract notion, except for a geologist), I fear it may be plain wrong.

My (limited) understanding is that the Allan deviation of most clocks tends to worsen at very long sample times due to random walk frequency modulation noise and aging effects. Then, it should not be expected for the clocks to achieve their rated accuracy over decades, let alone megayears.

I propose to reword those instances as “expected to neither gain nor lose XXX nanoseconds per month”. I assume a month is, as least as an order of magnitude, a more realistic sample time than XXX megayears. I know the nanosecond is not really an “everyday life” unit. However, in today's highly technological world, it's closer to everyday life than a megayear. At least closer to the life of those reading Wikipedia on an electronic device clocked somewhere in the GHz range.

— Edgar.bonet (talk) 16:19, 25 February 2016 (UTC)

I understand your proposal, but in references the second in many years is quite regularly used to explain the accuracy levels achieved by atomic clocks that actually measure frequency to general readers. By the way national frequency/time standards are not used for decades but are upgraded/replaced when funding and technology allow this. Obsolete frequency/time standards sometimes end up in a museum.--Francis Flinch (talk) 18:55, 25 February 2016 (UTC)

I agree with Francis Flinch above. WP is a general interest encyclopedia, and the "one second in XXX years" format is widely used to express atomic clock accuracy in nontechnical sources and even in science magazines. The fractional accuracy 10^−xx or Allen deviation could be put in parenthesis after it. --Chetvorno^TALK 21:11, 25 February 2016 (UTC)

External links modified

Hello fellow Wikipedians,

I have just added archive links to 3 external links on Atomic clock. Please take a moment to review my edit. If necessary, add {{cbignore}} after the link to keep me from modifying it. Alternatively, you can add {{nobots|deny=InternetArchiveBot}} to keep me off the page altogether. I made the following changes:

• Added archive https://web.archive.org/20150912100324/http://jilawww.colorado.edu/yelabs/sites/default/files/uploads/Ye.Schnatz.Hollberg_JSTQE_2003.pdf to http://jilawww.colorado.edu/yelabs/sites/default/files/uploads/Ye.Schnatz.Hollberg_JSTQE_2003.pdf
• Added archive https://web.archive.org/20080107043308/http://inms-ienm.nrc-cnrc.gc.ca:80/research/optical_frequency_projects_e.html to http://inms-ienm.nrc-cnrc.gc.ca/research/optical_frequency_projects_e.html#optical
• Added archive https://web.archive.org/20090202223847/http://inms-ienm.nrc-cnrc.gc.ca:80/research/optical_frequency_projects_e.html to http://inms-ienm.nrc-cnrc.gc.ca/research/optical_frequency_projects_e.html#femtosecond

When you have finished reviewing my changes, please set the checked parameter below to true to let others know.

Y An editor has reviewed this edit and fixed any errors that were found.

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—^{cyberbot II}_{Talk to my owner:Online} 04:21, 25 February 2016 (UTC)

I changed these to more appropriate links: the same article hosted at tf.nist.gov for the first one, and the pages archived by National Research Council of Canada for the two others. The NRC-archived pages have the images, unlike those at archive.org. — Edgar.bonet (talk) 09:38, 25 February 2016 (UTC)

Early work: discussion of dubious fact

I tagged the reference to a 1949 NBS atomic clock as 'dubious' and 'citation needed': because I can't find it in the histories. Terry0051 (talk) 01:10, 25 February 2009 (UTC)

It's not exactly hard to find; just go to the website of the successor to the NBS, the NIST. The NIST Time and Frequency History mentions the clock, but doesn't say much about it and doesn't mention that in spite of the fact that the ideas originated in the US, Britain is credited with the first real atomic clock. D.B. Sullivan (2001) Time and frequency measurement at NIST: The first 100 years gives the full story. Basically, the 1949 atomic clock was an ammonia maser, but it wasn't even as accurate as crystal oscillators, just a 'proof of concept' device. Then both the US and Britain began working on cesium clocks, but because the NBS split its resources on an alternate approach and its cesium device had problems, Britain got there first with an accurate atomic clock. --Chetvorno^TALK 09:21, 25 February 2009 (UTC)

As the ammonia maser wasn't as accurate as a crystal oscillator, then, given (a) that the function of the cesium or equivalent device is to improve the regulation of the underlying crystal oscillator, and given (b) the underlying definition that "An atomic clock is a type of clock that uses an atomic resonance frequency standard as its timekeeping element", how does the ammonia maser amount to an atomic clock? I would suggest that it is misnamed by calling it so. But I don't want to start into competitive edits. How about this, instead of arguably mischaracterising the ammonia maser, would it not be better to put in an account of the predecessors whose ideas stimulated the building of the cesium standard? Rabi, for example. I'd also suggest that national issues are not relevant to an article about the clock rather than about the personalities. Terry0051 (talk) 02:08, 28 February 2009 (UTC)

I wasn't suggesting putting that bit about national issues into the article, I was just curious like you about why the NIST and other sources don't say much about the 1949 ammonia clock. Sorry, I didn't mean to sound flippant or chauvinistic. Your idea that the ammonia device doesn't fit the definition of an atomic clock is interesting. It seems to me it fits criterion (b); it consisted of a 100 kHz quartz oscillator, with its frequency locked to the frequency of an ammonia maser with a servo system, controlling an electric clock. Apparently the inaccuracy was due to the inherent inaccuracy of the ammonia resonant frequency, caused by doppler broadening 1 2, which is why the ammonia approach was abandoned. Your point about the crystal oscillator is perceptive; ironically the internal quartz oscillator might have been more accurate if it hadn't been locked to the maser, but I feel the device is still clearly an atomic clock since its frequency is controlled by atomic resonance. I enthusiastically agree with including more early contributors, especially Rabi. --Chetvorno^TALK 08:34, 28 February 2009 (UTC)

Thanks for your extra comment (and collaborative spirit!). I guess we're agreed on the underlying technical details; your evaluation of the maser device is interesting (I'll have to think that one through -- not that it should matter much on here what I think, of course). The challenge for us all, I suppose, as usual, is to offer clear, informative, fair description, in not too many words -- not so easy. On this point, there may be something wrong with the opening sentence-definition in the article: "An atomic clock is a type of clock that uses an atomic resonance frequency standard as its timekeeping element." Maybe it would better read: "...uses an atomic frequency standard to regulate its timekeeping element."? (This is not meant to try and artificially exclude the maser, by the way, I think your argument would be unaffected.) The existing words arguably mislead, by suggesting that the atomic standard is of itself a timekeeper i.e. oscillator, where if I'm not mistaken, two things are needed, the oscillator and the regulator. If that's cleared up, then it becomes probably easier to give concise fair information about how the developments progressed. Terry0051 (talk) 13:40, 28 February 2009 (UTC)

It might be more accurate to say the Ammonia maser was a "molecular clock", but that's not really a common term. The cesium clock was proposed by Rabi, the discoverer or nuclear magnetic resonance. Work on a cesium clock began early on in the US and UK, the nuclear frequency was measured by the Americans in 1952, and people at NBS, MIT and NPL (in Britain) began working on a practical clock. NPL was first to get one working as a useable frequency standard in June 1955. The NBS effort was delayed by their being moved from Washington DC to Boulder CO in the mid 1950s and by some funding squables. MIT's project was aimed at the marketplace, and became the Atomichron product, which you could buy about a year after the NPL cesium standard was operational. DonPMitchell (talk) 02:35, 11 August 2009 (UTC)

"Molecular clock" would have been a far better term but unfortunately the term used at the time was "atomic". See this advert: http://www.rockwelder.com/Atom/atomicclock.jpg

George Dishman (talk) 14:15, 12 March 2011 (UTC)

By the way, be sure to look at Louis Essen's own recollections on this history, on the excellent website here: [3]. The section on "Atomic Clock" gives a good history of what was happening in the USA and UK in the 1950s. DonPMitchell (talk) 16:35, 11 August 2009 (UTC)

picture of HP 5071A

i think the article about atomic clocks needs a picture of a HP/Agilent/Symmetricom 5071A, because it is the most common/famous atomic clock (caesium based primary frequency standard) in existence and it has most weight in maintaining UTC - every country's time keeping lab has them.81.206.145.191 21:43, 13 May 2007 (UTC)

Here's a picture of a 5071A on the USNO website; I don't know whether USNO images are public domain. Image:HP_5061A_Cesium_beam_frequency_standard.jpg is a public domain picture of a HP5061A. Is that the commercial version of the 5071? --Chetvorno^TALK 08:49, 25 February 2009 (UTC)

Travel Atomic Clock Sync

I have a very small travel global Atomic Clock and alarm clock. It is Atomic Radio Controlled. It has to be manually set to: US, UK, Japan and EU via a button on the back. It does not change on its own as I fly from time zone to time zone.

If I am out of range of: England, Switzerland, Japan or the US will the clock no longer sync? In the US the clock wants to sync to Mountain Time due to the Atomic Clock is located in Colorado. Colorado is on Mountain Time.

How does a travel Atomic Clock differ from a stationary clock? My stationary clock updates on its own when it unplugged and plugged in again. If the power goes out the clock resets on its own. A travel clock requires a manual reset, selecting the country for the Atomic Clock and sometimes changing Daylight Savings Time.

Why does the page on Atomic Clocks not cover the small travel version as well? Bree25 (talk) 02:47, 30 April 2013 (UTC)

There's some inaccurate terminology here. What you have is not an atomic clock, it is correctly called a "radio clock". It synchonizes by radio with an atomic clock located in a government laboratory. This article is about atomic clocks themselves. Try Talk:Radio clock. --Chetvorno^TALK 03:00, 30 April 2013 (UTC)

Accurate compared to what?

"National standards agencies maintain an accuracy of 10^-9 seconds per day"

So I set up my brand new atomic clock, go away for 3 million years, and come back to find out it is off by one second. Compared to what? Another atomic clock? I don't see what else it could be if, by definition, an atomic clock is the standard. --Chauncey27 19:21, 24 December 2005 (UTC)

Compared to a theoretical ideal. Accuracy of one part in 10^-9 is because the clock can be off by as much as half a cycle per second without being noticed. --Carnildo 08:57, 26 December 2005 (UTC)

Great question, Chauncey27. The accuracy of an atomic clock is calculated indirectly from how wide an interval of time the device will indicate the same time. In the most accurate atomic clocks, the resonance peak is sharper than we can measure, and the accuracy is limited by how accurately the period of the microwave signal from the atoms can be measured by the apparatus. This is limited by how much time the atoms spend in the resonance chamber. The idea of the cesium 'fountain' clock is to get a ball of supercooled cesium atoms to spend as much time as possible (~1 sec) in free fall by tossing them up. The accuracy calculation looks something like this:

accuracy = (period of one oscillation) * (fraction of period electronics can measure) / (time atoms spend in apparatus)

For a cesium atomic fountain this is approximately:

accuracy $=({\frac {1}{9,192,631,770Hz}})({\frac {10^{-5}}{1sec}})\approx 10^{-15}\,$

--Chetvorno^TALK 13:17, 26 May 2008 (UTC)

That's the theoretical accuracy. To check the actual accuracy, you are rightChauncey27, there is nothing to use except another atomic clock. This is one reason why NIST and other national standards organizations operate 'ensembles' of many atomic clocks, to check them against each other. Some of the clocks are kept at widely separate locations, in an effort to make sure that some local error influence doesn't change the frequency of all the clocks together. --Chetvorno^TALK 22:30, 10 March 2009 (UTC)

Correct me if I'm wrong, but that 1 second means the radio clock can be off by as much as 1 second per day, because it uses its own cheap timekeeping device to keep the time, and corrects its time once a day by synchronizing with the signal from the real atomic clock. It would take too much power to be synchronizing all the time. So that 1 second is how far it can get from the atomic clock before it re-synchronizes. Bostoner (talk) 21:54, 24 November 2010 (UTC)

No, the above discussion is about atomic clocks, not radio clocks. Although what you say about radio clocks sounds right to me;) --Chetvorno^TALK 05:07, 25 November 2010 (UTC)

New Section Organization and goal to reach B class status

I am working on improving this article to reach B class status. I am thinking of reorganizing the sections as follows:

1. applications
2. microwave clocks vs optical clocks vs other types of clocks and how each time works
3. scientific applications of atomic clocks such as general relativity and mapping of earth's service.

I don't think Power Consumption should be a section because it is very small and is not a critical but short role in the article. ScientistBuilder (talk) 01:28, 15 February 2022 (UTC)

I agree that power consumption does not need a section. In scientific and technology articles, the first section after the lede is almost always the history section. We should proboaly stick with that tradition. Constant314 (talk) 02:58, 15 February 2022 (UTC)

What section should I merge it with? ScientistBuilder (talk) 00:47, 16 February 2022 (UTC)

I think a history section of the early development of atomic clocks until 1990 should be included.

I do not think a lot of history after that should be put in the history section because atomic clocks become a developed field around 1990 with the launch of NIST-F1.

I then think there should be a section on how atomic clocks measure time.

I also think there should be a section about how scientists are constantly improving atomic clocks to be more accurate and faster at measuring time signals.

I also think there should a section on applications for GPS, general relativity tests and more. ScientistBuilder (talk) 00:50, 16 February 2022 (UTC)

I think the section on how atomic clocks measure time could cover the quantum mechanics and focus less and the construction of atomic clocks.

I think the Mechanism Section is a blend of technical details on the physical basis of clocks right now (up to the sentence: "In practice, the feedback and monitoring mechanism is much more complex.") and the economics of atomic clocks (the rest of the Mechanism section).

I think the section should be split into physics or science of clocks and construction of clocks. ScientistBuilder (talk) 00:53, 16 February 2022 (UTC)

I think the section on Evaluated Accuracy has too many accuracy statistics.

I don't think having lists of how accurate atomic clocks are contributes to understanding of atomic clocks. ScientistBuilder (talk) 00:54, 16 February 2022 (UTC)

I think the clock comparison techniques and evaluated accuracy sections should be merged into a new section. ScientistBuilder (talk) 00:55, 16 February 2022 (UTC)

If we try to respond to that long list of suggestions, the thread will splinter into incomprehensible spaghetti. I suggest that you pick whichever section you think is the most egregious and open a new thread. Be patient, these things take time. Constant314 (talk) 01:43, 16 February 2022 (UTC)

Attosecond, Femtosecond, Zeptosecond, Yoctosecond

I am wondering if it better to use zeptosecond or 10^-21 seconds or if both are acceptable in the article. I think it is helpful to add the power for reference. Which is best? Use International System of Unit prefixed word, power of 10, both at once, or both are acceptable. ScientistBuilder (talk) 01:05, 16 February 2022 (UTC)

My personal preference is to use exponential notation for anything below a femtosecond and perhaps even for femtosecond. Constant314 (talk) 01:39, 16 February 2022 (UTC)

Not all Wired Content is a reliable source

There has been a bit of dubious content added about chip scale atomic clocks that are attributed to an Article in Wired Magazine. If you look carefully, you will see that the content of the article in not from the staff of Wired, but it is "Partner Content" from SYMMETRICOM. In other words, likely marketing hype. The stuff about IEDs, jamming, drones, and undersea exploration needs a reliable source, and I suspect that those applications do not need atomic clock accuracy. Constant314 (talk) 03:20, 15 February 2022 (UTC)

There are a few Wired pieces here, but I can't find any that are clearly sponsored content. Which are you talking about? Headbomb {t · c · p · b} 04:13, 15 February 2022 (UTC)

Ah, nevermind, you already removed it here. Headbomb {t · c · p · b} 04:14, 15 February 2022 (UTC)

I took the liberty of adding 'wired.com/insights' to WP:UPSD so it's easier notice to flag in the future. There are a few instances on Wikipedia as of writing, about 50 or so. Headbomb {t · c · p · b} 04:21, 15 February 2022 (UTC)

Is atomic time required for financial transactions and communications?

Definitely not. I removed this text from the article.

Atomic timekeeping forms the basis of the world's financial system by accurately recording the exact time of financial transactions.

Well, of course, atomic time does underly all time keeping in the modern world. But it required? No. The communications and financial networks are set up to operate plesiosynchronously using quartz time (or worse). What happens when the network operates plesiosynchronously? You get what is called a controlled slip. The world works on 256 bit frames. A controlled slip is the lost or duplication of a frame. In plesiosynchronous operation that occurs about once every six hours. In other words, you get a few bit errors. That is not the only source of bit errors, so the communication and financial networks need some fault tolerance. That can be redundancy to allow the error to be corrected at the receiving end, or it can involve retransmission. But the recovery is built in.

What atomic clocks buy you is that you can spend less man-hours on keeping your clocks synchronized. The old Bell system had one atomic clock and distributed time to the rest of the US and Canada. Each class 1 office synchronized its clock to the master clock. Each class 2 office synced to a class 1 office and so forth down to class 5. As atomic clocks become cheaper, it is economically feasible to push atomic clocks further down into the hierarchy. The whole thing only needed synchronization. The absolute accuracy of the clock was never important. As you push atomic clocks down into the network, it increases resilience. If an office loses it connection to the time source and it has its own already synchronized clock, it can operate plesiosynchronous for a long time without experiencing additional frame slips.

Let's look at what the source says. "Such accurate time is also used to timestamp financial transactions so that we know exactly when trades are happening, which can mean the difference between making a fortune and going broke." That is pretty accurate. Since atomic clock derived time is typically available, the time stamp is typically based on atomic time. Nowhere is it said that atomic time is essential. Nowhere is it implied that the financial system would collapse without atomic time. In fact, I can guarantee that the financial system will keep chugging along on quartz time.

The source also says, "Accurate time is also necessary for synchronizing communications signals so that, for instance, your call isn’t lost as you travel between cellphone towers." Notice that it said, "accurate time" and not "atomic time". Inferring that "accurate time" means "atomic time" is an example of synthesis (WP:SYN) which leads to an incorrect conclusion. That is why we don't like it. That bit about losing your cell phone call is a bit of grand standing. Good old NIST is keeping you from losing phone calls and going broke. Remember the purpose of the blog. In this case it is PR for NIST. Quartz time will be good enough to not lose cell phone calls.

Constant314 (talk) 01:26, 16 February 2022 (UTC)

Are the rest of my edits so far helpful? ScientistBuilder (talk) 01:42, 16 February 2022 (UTC)

Generally, all good faith edits are helpful, even if they are reverted. They help bring attention and dialog to areas of the article that need improvement. You are doing a great job. You are contributing even if every word you write is removed. The article is already better thanks to the attention you are giving it. Constant314 (talk) 01:48, 16 February 2022 (UTC)

I want to change the organization of the article. How should I go about this goal? ScientistBuilder (talk) 02:05, 16 February 2022 (UTC)

The only sure rules are WP:UPFRONT and that the history section is always the first after the lede, although, that might seem to conflict with the first rule. With a consensus, we can change the order as we wish. For example, in the article transformer, the history section is not first. There was a consensus for that. Just go slowly. I lot of people like to think about things before they respond. We encourage bold editing, but also bold reverting. Moving whole sections is not a big deal, but if you make a lot of intertwined edits that are hard to undo individually, someone may revert a whole day's worth of effort in order to undo some edit that you did yesterday.Constant314 (talk) 02:27, 16 February 2022 (UTC)

I am proposing moving applications up because it is less technical than the section on optical clocks. ScientistBuilder (talk) 02:59, 16 February 2022 (UTC)

On the subject of whether atomic time is required for financial transactions: the answer is yes. This is an important new application which should be mentioned, the development in recent years of high frequency trading [4] [5] [6] This means large investment firms using microwave links to exploit millisecond price differences in stocks and bonds between exchanges located in different cities, for arbitrage. To avoid conflicts, exchanges have recently gone to UTC timestamp standards for trades [7]. Due to the availability of GPS and NTP sources this may or may not require actual atomic clocks at trading houses, but definitely requires UTC atomic time standards synchronized to milliseconds in financial firms across the world.--Chetvorno^TALK 02:59, 16 February 2022 (UTC)

Welcome to the conversation. The first two refs look like mostly sales hype. But regardless, none of them come close to saying atomic accuracy is essential. Granted, you must synchronize to UTC and UTC is atomic based, but nowhere does it say atomic accuracy is required. They are looking for discrepancies on the order of a fraction of a second, i.e., 10s of milliseconds. This is trivially within quartz clock capability. Atomic time is used for this purpose, but atomic clocks are used for convenience, not because they are essential. You must sync to UTC. That is a whole lot easier if UTC is steady as a rock. It is also easier if your clock is as steady as a rock. Atomic clocks are so steady, that you just about forget about synchronizing your clocks. A large organization might save the cost of a small maintenance department. Atomic clocks are cool and allow somethings that cannot be done otherwise. Financial trading is not one of those things. Constant314 (talk) 03:52, 16 February 2022 (UTC)

A New York times article supports the need for accuracy at the 100 nanosecond level. Here are two key paragraphs:

Because the orders are placed from locations around the world, they frequently arrive at the exchange’s computers out of sequence. The new system allows each computer to time stamp an order when it takes place.

As a result, the trades can be sorted and executed in correct sequence. In a networked marketplace, this precision is necessary not only to prevent illicit trading on advance information known as “front-running,” but also to ensure the fair placement of orders.

The Journal of Reasearch of the National Institute of Standards and Technology has an article on the topic which mentions a European standard, MiFID II, that requires 0.1 ms synchronization.

I've read other articles that discuss how long a financial location could continue to obey the regulations if their source of syncronization from outside the campus (for example, GPS receivers) was interrupted. My recollection was that for oven-controlled quartz oscillators it was a matter of a day or two, a few weeks for rubidium atomic clocks, and around six months for commercial cesium atomic clocks. Jc3s5h (talk) 05:01, 16 February 2022 (UTC)

The NYT article was written in 2018 about the future undeployed systems (at the time) and should not be taken as a RS for what is necessary now. The NIST article is compelling, and I will take a closer look at it. I agree 100 us/day is getting into the range that is difficult for quartz. I used to run a satellite tracking station circa 1974. My recollection is that our quartz clock was holding 30us/day, but memory is unreliable. Constant314 (talk) 07:07, 16 February 2022 (UTC)

References

In addition to sources above, we have reliable sources saying that some financial institutions like ICE[8][9] are turning to atomic clocks. But you are skeptical and dismisses all these as hype, because extreme accuracy is not needed. I think the resolution is that you may be missing the point of atomic clocks in fintech--it's not that financial folk need nanosecond accuracy (yet!), it is that they need a time service that is so reliable and so well-respected that they trust it and more importantly, just about everyone in the financial industry agrees that they can trust it. Atomic clocks and time services based on them have amazingly good track records at keeping time and are the foundation of that trust. The sources explicitly mention atomic clocks. So I agree with the other editors that that applications of atomic clocks in the financial industry are worth a mention. But we should not outrun the RS and should refrain from editorializing about why they are being used or why they are irrelevant. Talk page assertions about hype or assertions about trust are editor POVs that should not enter the article without RS backing them up. --{{u|Mark viking}} {Talk} 05:07, 16 February 2022 (UTC)

Those sources show that atomic clocks are being used, but not that it can't work without atomic clocks. But, I agree, atomic clocks are the way to go. They are good, reliable and robust. But never confuse what is easy with what must be. Sounds profound, doesn't it. This has sort of been an exercise in critical reading. The sources have been informative, but none has established the essentialness of atomic clocks. But I am feeling the weight of the consensus and now I am standing down. Yes, add the statement to the article. This is the way things should go. When the reliable sources are not explicit, yet a consensus of concerned editors emerges, we get better articles.

The point is that high frequency trading requires more accurate timing of trades, the trend is toward more precise timestamps that must be synchronized worldwide, and UTC time is how this is accomplished. I'm not suggesting adding anything, the existing few sentences are adequate, but the sources support this as a recent application of atomic clocks. --Chetvorno^TALK 19:42, 16 February 2022 (UTC)

Remove Power Consumption Section

How should I remove the section on power consumption? I think it could be merged with a new section on how atomic clocks are built and constructed that makes up the second half of the Mechanism section (see New Section Organization and goal to reach B class status above for explanation). ScientistBuilder (talk) 02:11, 16 February 2022 (UTC)

European Union

Is there an organization similar to NIST or the UK's National Physical Laboratory in the EU? I am looking for other sites that engage in atomic clock research. ScientistBuilder (talk) 02:04, 16 February 2022 (UTC)

Construction Section

I am working on a way to add information about how atomic clocks are built and constructed. Some material I would put in are:

1. chip scale atomic clock material in History
2. reduction in bulkiness in History
3. topics covered by paragraphs starting with "A number of other...", "Often, one...", "The lifetime of ..."
4. Power consumption material
5. masers for satellites and other non lab environments
6. optical clock constructions because optical clocks have their own unique designs that have different hurdles than traditional atomic clocks.

ScientistBuilder (talk) 02:20, 16 February 2022 (UTC)

NASA Drive

I don't think NASA's project is important enough to the history of atomic clocks to be included in the history section. This is specific to NASA and not representative of the field of atomic clocks overall. This is not an active developed program as far as I can tell and more of an narrowly defined subject. I think it should be moved to the Construction section. In place of the NASA project, I propose a paragraph about optical clocks and ytterbium in the 2000s. ScientistBuilder (talk) 13:58, 16 February 2022 (UTC)

Bullet Points in Lead

I think the lead should be organized and rewritten to have no bullet points. Having a list in the lead is not what a good lead has. I propose turning the two bullet points into sentences. ScientistBuilder (talk) 21:09, 16 February 2022 (UTC)

Quantum Mechanics Science Physics Section

I think there should be a separate section explaining how clocks lock on to an atomic frequency transition emission. This would include the following:

1. information on the definition of the second
2. reducing noise and staying as close as possible to 9192631770Hz.
3. an explanation of how optical combs work
4. fine structure
5. hyperfine transitions
6. electron transitions
7. other technical details ScientistBuilder (talk) 02:25, 16 February 2022 (UTC)

Some good info here from Hewlitt-Packard: "How it works", "Construction details", "Flying clocks", "Performance stats", "World wide synchronization using flying clocks" Constant314 (talk) 04:54, 16 February 2022 (UTC)

I like your reorganization plan in general. I assume the above would be in a later advanced section. My concern is that the article have some explanations comprehensible by general readers. There are a lot of complaints that Wikipedia STEM articles are only aimed at technical people (see User:Chetvorno#Wikipedia_is_too_technical). We technically-educated editors like to write about the advanced stuff. That's OK, not everything has to be for all readers. But we should remember that middle-school students, high-school dropouts, ballet majors, and single mothers trying to get through community college come to our article. These may be the majority of readers. They may just want the simplest explanation of how atomic clocks work.

Another point is that while cesium fountains and optical combs are at the cutting edge, the ordinary cesium and rubidium standards are the common atomic clocks. The main deficiency I see in the current article is lack of a detailed explanation of how these work.--Chetvorno^TALK 20:37, 16 February 2022 (UTC)

Just a suggestion: I'd like to see an elementary introductory section on atomic clocks in general, how they work, why atomic clocks are so accurate (high Q), the main types (cesium, rubidium, hydrogen) and why those elements are used, the definition of the second, and mention of the UTC time scale. Then three following sections could describe in detail how cesium, rubidium and hydrogen standards work. --Chetvorno^TALK 20:37, 16 February 2022 (UTC)

==How Atomic Clocks Work==

==Why atomic clocks are so accurate==

==Main Types (Cesium, Rubidium, Hydrogen, Ytterbium)==

==The definition of the second== ScientistBuilder (talk) 21:14, 16 February 2022 (UTC)

==How atomic clocks work and why they are so accurate==

==Main Atomic clock types==

==the definition of time and the second and the UTC scale==

==Detailed Explanations==

==Cesium==

==Rubidium==

==Optical Clocks==

==Optical Combs== ScientistBuilder (talk) 21:16, 16 February 2022 (UTC)

Looks great to me. --Chetvorno^TALK 21:20, 16 February 2022 (UTC)

I'll do a block diagram for a cesium clock, unless someone else has already started on it. Constant314 (talk) 22:48, 16 February 2022 (UTC)

That would be very useful. Another source, if you haven't already looked there, is Wikipedia's image repository: Wikimedia Commons. --Chetvorno^TALK 06:30, 17 February 2022 (UTC)

Fabry-Perot Interferometers

I am proposing to remove the sentence on Fabry Perot interferometers because there are not that many cases of Fabry Perot interferometers in the literature for atomic clocks on NISt and other websites. I think the main improvements were cooling with lasers and optical combs. ScientistBuilder (talk) 03:04, 16 February 2022 (UTC)

I am going to remove the sentence because there is no source for it. ScientistBuilder (talk) 15:07, 19 February 2022 (UTC)

Redundant SI Definitions

There are multiple paragraphs in different sections that cover the definition of the second. I want to remove the material in the Time Measurement section and put any material that is there and not in the How Atomic Clocks Work Section in the How Atomic Clocks Work section. ScientistBuilder (talk) 19:20, 19 February 2022 (UTC)

Microwaves

Atomic clocks with caesium use microwaves to measure time. The atomic clock measures the frequency of the microwaves. The atomic clock sends a microwave to excite the atoms. The microwave is measured and if it not 9192631770Hz, the microwave is corrected. ScientistBuilder (talk) 14:57, 19 February 2022 (UTC)

The more accurate the microwaves are measured, the faster the clock can correct the microwaves to 9192631770Hz. The same principle applies for other frequencies.

I want to add an efn template explaining that atomic clocks measure microwaves. ScientistBuilder (talk) 14:59, 19 February 2022 (UTC)

Sources:

https://www.explainthatstuff.com/howradiocontrolledclockswork.html

https://www.timeanddate.com/time/how-do-atomic-clocks-work.html ScientistBuilder (talk) 15:03, 19 February 2022 (UTC)

https://www.nist.gov/topics/physics/optical-frequency-combs ScientistBuilder (talk) 15:05, 19 February 2022 (UTC)

My objection to sloppy language. Instead of "atomic clocks measure microwaves," try "atomic clocks use the frequency of internally generated microwaves." That is also sloppy, but still an improvement. Atomic clocks measure the absorption of internally generated microwaves by an internal gas chamber. They use the measured absorption to steer the frequency of the internal microwave source. The frequency of the internal source is used to derive a stable standard frequency such as 10 MHz. The standard frequency is used for time keeping and time measurement. Try to use language that is consistent with that narrative. Constant314 (talk) 20:35, 19 February 2022 (UTC)

Thank you for the help ScientistBuilder (talk) 20:58, 19 February 2022 (UTC)

Is the new edit better? ScientistBuilder (talk) 21:03, 19 February 2022 (UTC)

@ScientistBuilder:Your edits on 19 February occupy two pages in the watchlist. Which new wording are you referring to? Jc3s5h (talk) 18:10, 20 February 2022 (UTC)

Remove Accuracy Information

The article has too much information about the accuracy of specific clocks. I propose removing all parts that mention a clock that reached a new level of accuracy unless it achieved a world record such as $10^{-17}$ to $10^{-18}$. The section on accuracy could be simplified. ScientistBuilder (talk) 21:24, 20 February 2022 (UTC)

I agree with that. There is no point in trying to keep the article up to date with a moving target. However, three or four milestones in accuracy might be appropriate. Constant314 (talk) 22:16, 20 February 2022 (UTC)

@Constant314I have updated the Accuracy section to include the advancement of 10^-16. I think the 10^-17 milestone was achieved with an optical clock at JILA> ScientistBuilder (talk) 00:51, 21 February 2022 (UTC)

U_b removal

the accuracy section does not define $u_{b}$. Therefore, I propose removing the use of the term. ScientistBuilder (talk) 21:25, 20 February 2022 (UTC)

Less Discussion about Earth's rotation in article lead

The article on atomic clocks does not need to include a discussion about the differences of UTC and UT1 and Earth's rotation so I am removing UT1. ScientistBuilder (talk) 22:51, 20 February 2022 (UTC)

• Agree. Constant314 (talk) 23:16, 20 February 2022 (UTC)

  With all the edits being made, I can't tell if any change was made based on this thread.

  Page views of UTC are around 4000 per day. Page views of TAI are around 150 per day. Since UTC is evidently much better known than TAI, we should mention UTC in the lead. Jc3s5h (talk) 23:38, 20 February 2022 (UTC)

Simpler First Sentence in Lead

I propose changing the lead's first sentence to "An atomic clock is a clock that measures time by monitoring the oscillations of atoms." The current sentence is pretty technical and I think it would be better to start with a more understandable and intuitive sentence and go into how electromagnetic radiation is used later. ScientistBuilder (talk) 00:54, 21 February 2022 (UTC)

I don't think the article needs to go into a quantum mechanical concept in the first sentence. ScientistBuilder (talk) 00:57, 21 February 2022 (UTC)

Seconds per Day

I propose to simplify the sentence at the end of the lead that states atomic clocks deviate by at most a nanosecond every day (about one part in 10^14). This is because the accuracy of an atomic clock is often evaluated by the deviation from the second but not by the deviation from the second over a day. ScientistBuilder (talk) 21:11, 16 February 2022 (UTC)

I'm not sure if this talk section is about this change, but currently the lead says the NIST-F2 clock has a "precision of 10^-16 seconds", and that ytterbium and strontium clocks "measure time to 10^-18 seconds". I don't understand what this means. Clock precision is normally specified as a unitless ratio, as is done in the Accuracy section ("accuracy of 10^-11", "accuracies of 10^-15 to 10^-18", etc) and elsewhere in the article, or as its reciprocal ("one part in 10¹⁶"). The latter, for example, means that a clock will lose or gain 1 second after running for 10¹⁶ seconds. I don't understand what it means for an accuracy to be specified as a duration of time. These numbers in the lead are not cited, not explained, and are not mentioned anywhere outside the lead. Can anyone explain this? CodeTalker (talk) 20:51, 21 February 2022 (UTC)

@CodeTalker: @ScientistBuilder: Yes, precision, accuracy, relative accuracy, stability, and uncertainty all have specific meanings and are used sloppily in the article and unfortunately by some of the sources. The better sources, that is the official NIST and BIPM are pretty careful, but some of the others are not. After the article settles down, we should have a discussion and settle of some consistent terms. I am slogging my way through BIPM documents now. I think most of the time, the correct term is relative accuracy with many of the sources shortening that to just accuracy. Constant314 (talk) 22:54, 21 February 2022 (UTC)

@CodeTalker: @ScientistBuilder: There is also the issue of whether the sources are talking about time or frequency. Practically measuring time to a precision of 10⁻¹⁸ seems a bit dubious, but measuring frequency is much easier. Constant314 (talk) 00:08, 22 February 2022 (UTC)

BIPM Images of Atomic Clocks

There are some good images of atomic clocks from the BIPM at https://www.bipm.org/en/-/2021-12-21-record-tai. I would like to find a way to put them in the article. I have added a graph of the accuracy of atomic clocks. I am uploading the rest to Wikimedia Commons:

Strontium Optical Lattice Clock at NICT, Japan's timekeeping organization.

I want to add more images of atomic clocks. Adding images of atomic clocks from around the globe contributes to a global perspective and not just NIST.

A ytterbium optical lattice clock at KRISS, the Korean Research Institute of Standards and Science

ScientistBuilder

A ytterbium optical lattice clock at NMIJ, National Metrology Institute of Japan used to measure International Atomic Time (TAI)

(talk) 01:49, 18 February 2022 (UTC)

ScientistBuilder The images on this Talk page, if they come from the web page you linked -- that page has a Copyright statement at the bottom. Are they really licensed by BIPM? I see they are at Commons. Did BIPM really send WP a license release? 73.127.147.187 (talk) 20:41, 20 February 2022 (UTC)

More images sounds like a good idea. --Chetvorno^TALK 23:57, 21 February 2022 (UTC)

73.127.147.187, if you look at the BIPM copyright page the BIPM images are covered by the Creative Commons — Attribution 3.0 IGO (CC BY 3.0 IGO), an open source copyright that allows others to use the images as long as the source is attributed. Looking at the table on the Commons licensing page this is one of the licenses accepted by Commons, so no release is needed. --Chetvorno^TALK 23:55, 21 February 2022 (UTC)

@ScientistBuilder: if the pictures tend to stack up and crowd each other down the R side of the page, you can use the {{breakafterimages}} template at the end of each section to make the correct picture appear opposite its section. --Chetvorno^TALK 00:51, 22 February 2022 (UTC)

The reason I'm editing this article is because I want to get it to good article status and maybe feature article status @Chetvorno ScientistBuilder (talk) 01:07, 22 February 2022 (UTC)

I think that's a great idea. The engineering and science sections of Wikipedia are far behind most of the other sections in number of FAs and GAs. --Chetvorno^TALK 01:42, 22 February 2022 (UTC)

I noticed that some of your edits don't have edit comments (Edit summaries). Even if you leave a note about them on this Talk page, it's a good idea to put some kind of comment about what it was for in the comment box for every edit. The comment stays with the edit acting as documentation. --Chetvorno^TALK 01:54, 22 February 2022 (UTC)

Revising First Subsection

I am proposing the following revision to the first subsection: "James Clerk Maxwell and Lord Kelvin were the first to argue that radiation could be used to keep track of time. The first practical considerations of how an atomic clock might work were made by Isidor Rabi in 1939. He proposed the concept in 1945, which led to a demonstration of a clock based on ammonia in 1949. This led to to the first accurate atomic clock being built at the National Physical Laboratory in the United Kingdom with caesium atoms." ScientistBuilder (talk) 02:31, 21 February 2022 (UTC)

I think it could be mentioned that the 1949 ammonia clock was not even as accurate as existing quartz crystal clocks; it just served as a proof-of-concept. --Chetvorno^TALK 01:36, 22 February 2022 (UTC)

I have found several good sources on redefining the second

I am working on improving the section's narrative about the history of time and the tropical year and now caesium clocks. ScientistBuilder (talk) 02:38, 22 February 2022 (UTC)

I have found several good sources I would like to use to improve the content of the article:

1. https://royalsocietypublishing.org/doi/10.1098/rsta.2011.0237
2. https://www.nist.gov/si-redefinition/second/second-future
3. https://www.nist.gov/news-events/news/2022/02/jila-atomic-clocks-measure-einsteins-general-relativity-millimeter-scale
4. https://www.nature.com/articles/s41586-021-04349-7
5. https://iopscience.iop.org/article/10.1088/1681-7575/ab3a82
6. https://www.researchgate.net/publication/51654829_When_should_we_change_the_definition_of_the_second
7. https://link.springer.com/article/10.1007/s12647-020-00421-1

ScientistBuilder (talk) 02:40, 22 February 2022 (UTC)

The definition of the second is an important topic for atomic clock, but the history of the definition of the second is not so important. Constant314 (talk) 02:42, 22 February 2022 (UTC)

These articles are about the future of the second ScientistBuilder (talk) 02:43, 22 February 2022 (UTC)

@Constant314 The articles mention optical clocks and the Rydberg constant. ScientistBuilder (talk) 02:45, 22 February 2022 (UTC)

Maxwell and Kelvin Reference Removed

For some reason a reference to James Clerk Maxwell and Lord Kelvin was removed. I am wondering why a sentence about the first people to propose the atomic clock is too much to include or why it was removed. ScientistBuilder (talk) 02:45, 22 February 2022 (UTC)

I removed it. Such speculation did not lead to the atomic clock. It is just fill that the article does not need. Constant314 (talk) 02:49, 22 February 2022 (UTC)

Remove Experimental Research from History

I propose the History section should not include optical clock research and other experimental advances since the 1990s like fountains and ion traps. I think its better to put these later in the article. ScientistBuilder (talk) 02:59, 21 February 2022 (UTC)

I appreciate your courteous collaborative approach. But you don't really need to state your intentions first. Most editors follow the bold-revert-discuss process (WP:BRD). Go ahead and edit a section, or several sections, and maybe leave a note here afterward on what you did and why. Then other editors can see it in context. If they object they will speak up, modify or revert it, and then we all can discuss it on this page. Cheers --Chetvorno^TALK 00:21, 22 February 2022 (UTC)

@Chetvorno

I am working on making the article a good place to learn these topics:

1. international system of timekeeping with TAI and UTC
2. GNSS like Galileo
3. caesium clocks and their role as primary standards for the second and how the second has evolved with clocks
4. optical clocks and the quest to redefine the second ScientistBuilder (talk) 13:05, 22 February 2022 (UTC)

Merge or Remove Paragraph in International System of Units Definition to History Section

The paragraph in the International System of Units Definition includes information that is in the History section. I propose moving the information that is not in the History section there and deleting the rest. ScientistBuilder (talk) 15:27, 22 February 2022 (UTC)

@Constant314I propose removing information about other SI units unless they relate to timekeeping measurements. I propose removing information about the interdependence of the base SI units paragraph. ScientistBuilder (talk) 15:29, 22 February 2022 (UTC)

Yes, I think that is a good idea. A GA article should be complete and focused and not become a coatrack. The more focused the article, the less work for the GA reviewer. Constant314 (talk) 18:19, 22 February 2022 (UTC)

Reorganization

I am working on improving the article but there is a lot of redundant information about redefining the second, optical clocks and discoveries. I would like to reduce this redundancy by moving and reorganizing the sections. I propose removing the history section because to understand the history, the mechanism should be explained first. ScientistBuilder (talk) 16:12, 22 February 2022 (UTC)

Major Section:

• Caesium Clocks
• Definition of the Second
• Advances in the Definition of the Second as caesium clocks matured in the 1990s
• How Caesium Clocks are the basis for International Atomic Time and serve as international timekeepers

New Major Section

• What makes a clock accurate
• Stability

transition to optical clock section

Femtosecond Combs Section Laser Cooling Section Applications of More accurate clocks Redefining the second This involves subjects such as fiber optic communication of time, GNSS

Optical clock section ScientistBuilder (talk) 16:14, 22 February 2022 (UTC)

I agree the article needs reorganizing, and agree with most of your prescriptions. Here are some ideas; these are just my opinions:

• I feel the article should have less about the definition of the second. We have the articles Second, Time standard, Time metrology, SI base units, International Atomic Time, Coordinated Universal Time, History of time and History of timekeeping devices for that. The article must have the second's definition in terms of caesium, but the extensive CIPM quotes in the History section are sort of bureaucratic and off-topic.
• I feel the article should focus more on the details of how atomic clocks work. Some questions

What is an atomic transition? Why is it the world's time standard?

Q: All clocks are based on resonance (harmonic oscillators) - resonance width - the accuracy of a timekeeper is proportional to its Q - Q of pendulums and quartz crystals - atoms are (some of) the highest Q resonators we know

Ramsay design and why it increases the accuracy of a caesium clock

Why the caesium fountain? - the idea is to keep the caesium atoms isolated from interactions for the longest possible time

How laser cooling works - Doppler shift - Doppler line broadening

Phase locked loops - dividing down the high atomic frequency to get a clock signal

• There's a big difference, both in how they work and accuracy, between cutting edge standards, research devices like F1, F2 and strontium clocks in metrology labs like NIST and NPL, and commercial caesium, hydrogen and rubidium clocks used in television stations and industry. Maybe it makes sense to describe the details of how they work separately, something like:

• Definition of the second
• How atomic clocks work
• Commercial atomic clocks

• Caesium
• Hydrogen
• Rubidium

• National time standards and research devices

• Caesium fountains
• Optical clocks
• Nuclear clocks

• International atomic time system
• History

Again, these are just some ideas

--Chetvorno^TALK 00:02, 23 February 2022 (UTC)

@ScientistBuilder I agree with Chetvorno, in that you are attempting to make the scope of this article too wide; GPS/GNSS doesn't need more than a mere mention, as one example.

In addition, you keep spreading overlapping proposals over multiple separate talk threads, its becoming impossible to keep track what exactly you are planning or why. Can you please organize your thoughts and try to stick to a few talk threads instead of creating new ones reiterating points you've made previously?

Finally, with all the edits you are making it is critical that you provide edit summaries -- recently you have over 50 edits in a row with no explanation Strangerpete (talk) 12:13, 23 February 2022 (UTC)

Yes I realize edit summaries are important. I had been using the keyboard shortcut Alt+Shift+V to open the Visual Editor, then Alt+Shift+S to save, but then I realized Alt+Shift+V let me save without leaving an edit summary and I liked that better. ScientistBuilder (talk) 13:44, 23 February 2022 (UTC)

When scientists measure the second with atomic clocks, they have to take into account the things in the definition and clarifications of the second, such as temperature, Doppler shifts, magnetism, etc. ScientistBuilder (talk) 13:59, 23 February 2022 (UTC)

@Chetvorno@Strangerpete

The definition of the second is relevant to the subject of atomic clocks. The second has been redefined and clarified as atomic clocks have become more advanced. The second would not have been amended if atomic clocks had not reached levels of accuracy where the current definition was unsatisfactory.

I think the information about the second should be somewhere in the article. ScientistBuilder (talk) 13:58, 23 February 2022 (UTC)

Okay, but atomic clocks alone is a broad topic, with a lot of very technical information to explain, and there are limits to WP:article size. The article's readable prose size is already 51kb, which is approaching splitting size, and there is a lot that needs to be added to it. The definition of the second in terms of caesium should be kept. But the last 4 lengthy quotes in the History section are not needed; they could be summarized by saying the definition applies to caesium atoms free of electric and magnetic fields, at absolute zero, at rest in the lab frame. I think some of the other peripheral sections such as GPS could be edited down too. --Chetvorno^TALK 16:22, 23 February 2022 (UTC)