Re: [tz] CCTF survey on Time and Frequency Metrology - Telecom
Dear Colleague, The last time the definition of the SI second was changed was in 1967 when Caesium atom hyperfine frequency transition was chosen as the reference value. Since then first the primary thermal beam caesium clocks and then caesium fountains have been providing the SI second realization values to steer the International Atomic Timescale. The current fountain uncertainties are at the low 1E-16 level. The frequency metrology research in recent years has surpassed the microwave clocks, and optical frequency standards based on trapped ions and atoms now provide frequency measurements with evaluated uncertainties at the level 1E-18. The frequency comparisons have also been advancing both through improved techniques in Global Navigation Satellite Systems signal analysis and the Two Way satellite Time and Frequency transfer as well as with fibre based frequency links and portable optical clocks. Frequency combs used to compare frequencies between clock lasers of different optical clocks and also between optical frequencies and microwave 133Cs frequency of 9.192631770 GHz are now routine laboratory equipment and are commercially readily available. These factors now provide a solid foundation to prepare for the redefinition of SI second based on the optical frequency standard. The exact approach is currently under discussion and it may be selecting one atomic candidate or building an ensemble. The specific requirements and the conditions for the redefinition to take place are also being finalized. The Consultative Committee for Time and Frequency (CCTF) has started preliminary work on this subject, to investigate possible scenarios and their possible impact. It is within this framework that we contact you an expert in the field of Telecommunications, in order to make sure that the needs and wishes of your community are integrated into the general debate that will take place and may, eventually, lead to the redefinition of the second. With the new definition, the frequency measurement accuracy will surpass that of Caesium-based timescale by at least one order of magnitude and will provide an adequate tool for current and future research into fundamental physics as well as technology exploitation. The redefinition is not expected to affect any of the existing commercial frequency standard uncertainties such as those of caesium beam clocks or hydrogen masers, because their uncertainties are generally orders of magnitude higher than the expected accuracy of the new definition. Same is true for caesium fountains. The computation and dissemination of UTC will also not change, but UTC will become more stable and accurate based on the improved optical realizations of the SI second. We would like to ask your expert opinion on whether the redefinition may have any impact be it negative or positive on your field of study and operation. The issues we would like you to consider are not just scientific and technological, but also regulatory and legislative. Your answers will be carefully examined and taken into account by CCTF in their analysis of Time and Frequency metrology current and future needs and SI second and UTC evolution. You can find the survey at https://www.surveymonkey.com/r/CCTF_Survey_2020 and in the attached letter from CCTF president Noël Dimarcq you will find additional information on other activities at CCTF for your reference. Best regards, The CCTF Task Force on the redefinition of the second, subgroup A Co-chairs: Marina Gertsvolf (marina.gertsvolf@nrc-cnrc.gc.ca<mailto:marina.gertsvolf@nrc-cnrc.gc.ca>) Gaetano Mileti (gaetano.mileti@unine.ch<mailto:gaetano.mileti@unine.ch>) Secretary: Frederic Meynadier(frederic.meynadier@bipm.org<mailto:frederic.meynadier@bipm.org>)
I replied to that survey, basically saying that currently our work is based on two timescales (TAI and UTC) and that I hope they don't turn it into four (TAIold, UTCold, TAInew, UTCnew). Or would it be three? That would be even more confusing....
On 2021-01-24 11:53, Paul Eggert wrote:
I replied to that survey, basically saying that currently our work is based on two timescales (TAI and UTC) and that I hope they don't turn it into four (TAIold, UTCold, TAInew, UTCnew).
Or would it be three? That would be even more confusing....
My reply suggested they divorce the time scales to allow business systems to use approximate POSIX like common law legal solar time as they ignore leap seconds one way or another; allow scientific systems to use atomic time and ignore legal solar time; and just ignore the ITU as broadcast radio time signals (DCF/JJY/MSF/WWV/B/H) are unavailable most places, may need equipment no longer commonly available, and too low accuracy for modern needs; other commonly available sources are delayed to uselessness by digital processing. -- Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada This email may be disturbing to some readers as it contains too much technical detail. Reader discretion is advised. [Data in binary units and prefixes, physical quantities in SI.]
On 2021-01-24 18:53, Paul Eggert wrote:
I replied to that survey, basically saying that currently our work is based on two timescales (TAI and UTC) and that I hope they don't turn it into four (TAIold, UTCold, TAInew, UTCnew).
Or would it be three? That would be even more confusing....
No new time scale and no new time scale designations would be created. The change would just be a redefinition of UTC, to be applied beginning with a date in the future (and probably a redefinition or abolition of DUT1 and possibly of dUT1). The last such redefinition of UTC applies since 1972. No new time scale and no new time scale designations were created then, nor would any be created with a future redefinition -- much the same as for the redefinition of the SI units kg, A, K and mol in 2018, where the old names and symbols have been retained. As far as tzdb is concerned, the leap second table would just end for good (perhaps with TAI = UTC + 42 s), and the "right" tzdb timezones would become unnecessary for time stamps from then on. As would the zoo of "leap second smearing" schemes in computer system time demons. Michael Deckers.
On Jan 25, 2021, at 2:45 AM, Michael H Deckers via tz <tz@iana.org> wrote:
As far as tzdb is concerned, the leap second table would just end for good (perhaps with TAI = UTC + 42 s), and the "right" tzdb timezones would become unnecessary for time stamps from then on. As would the zoo of "leap second smearing" schemes in computer system time demons.
In C++ it would mean that std::chrono::utc_clock (the clock that portably includes leap seconds, https://en.cppreference.com/w/cpp/chrono/utc_clock) could now do constexpr computations (compute at compile-time). This would be a good thing, making some things more efficient both in cpu time and code size. Howard
participants (5)
-
Brian Inglis -
Gertsvolf, Marina -
Howard Hinnant -
Michael H Deckers -
Paul Eggert