International atomic time

from Wikipedia, the free encyclopedia

The international atomic time ( TAI for French Temps Atomique International ) is a time scale that is determined by averaging and scaling the readings of currently (January 2020) more than 600 atomic clocks in around 80 institutes. The BIPM publishes the result monthly in Circular T. The BIPM estimates that the mean TAI second in 2019 deviated from the SI second by 0.24 · 10 −15 . The TAI is the basis for the Coordinated Universal Time UTC, which since 1972 differs from TAI by an integer number of seconds. Currently (as of January 2020) UTC = TAI - 37 s applies.

Since the TAI is only available retrospectively, the UTC (k) time signal transmitted in real time from an individual laboratory or the time transmitted by the GPS / GLONASS satellites must first be used for an accurate time measurement. Only after the publication of the new Circular T can this reading be converted into UTC and TAI using the tables contained there with the differences UTC − UTC ( k ) or UTC − UTC (GPS / GLONASS). For UTC (PTB) the corrections have always been smaller than 10 ns since 2010.

definition

The definition of International Atomic Time was last changed in 2018 by the 26th General Conference on Weights and Measures . Resolution 2 ( On the definition of time scales ) stipulated:

  1. TAI is based on the best realizations of the SI second, the operating times of the contributing clocks on the potential of the earth's gravity W 0  = 62,636,856.0 m² s² converted /;
  2. the zero point of the TAI is chosen so that on January 1, 1977, 0 o'clock TAI in the center of the earth, the relationship TT  - TAI = 32.184 s applies;
  3. the TAI is calculated by the BIPM.

Explanation

It is not clear from the outset that a group of stationary clocks on the rotating earth can be synchronized with one another, since they are in the earth's gravitational potential and move against each other due to the earth's rotation , which means that they are subject to time dilation . It turns out, however, that stationary clocks that are located on the same equipotential surface of the earth's gravitational field run at the same speed; see the section proper time on the rotating earth in the article dynamic time . The rate rates of watches on different equipotential surfaces can be converted into one another. The potential W 0 used in the definition is the best value available in 1998 for the geopotential of the geoid ; it is also used by the IAU to define terrestrial time and by the IERS . More recent measurements, however, show a value that is about 2.6 m² / s² smaller, so the “TAI equipotential area” is about 26 cm below the geoid. Before 2018, the TAI definition referred to the geoid. The change has the advantage that the definition becomes independent of changes in the geoid potential, for example due to a rise in sea level .

The determination of the TAI zero point by the difference of 32.184 s to terrestrial time on January 1, 1977, 0 o'clock TAI was chosen because it is also in the TT definition of the IAU. The IAU, in turn, referred to this date because from 1977 the “steering” of the TAI was introduced, whereby the TAI second coincided better with the SI second; the difference of 32.184 s guarantees that Terrestrial Time emerges seamlessly from the previously used Ephemeris Time .

Relationship between TAI and UTC

The TAI is an atomic time, that is, it is based on an atomic time standard , namely the SI second . In everyday life, time scales that are as synchronous as possible with mean solar time are more of interest, with the coordinated universal time UTC, which in 1961 replaced UT2 , which was previously used as universal time , is of greatest importance. As the length of day increases due to the slowing of the earth's rotation , UTC falls short of TAI. By the end of 1971, this was largely offset by a slowdown in UTC: the UTC second lasted longer than the SI second. Since 1972, UTC has also been based on the SI second, and the difference is made up by up to two leap seconds per year. On January 1, 1958, 0 a.m. UT2 and TAI exactly matched: TAI = UT2 ≈ UT1 , today the UTC is 37 s behind the TAI.

Formation of TAI

In addition to cesium clocks, which play a special role because the definition of the second is based on a transition in the cesium -133 atom, atomic clocks that use transitions in other atoms also contribute to TAI . Be used hydrogen (see hydrogen maser ) and rubidium -87 (see Rubidiumuhr ) whose transition frequencies as with cesium in the field of microwaves is, and for some years ytterbium -171 and other isotopes with transitions in the optical spectral region. Optical atomic clocks have the advantage over conventional atomic clocks that they are much more accurate, even if only for a short time.

Each laboratory k forms the best possible approximation of UTC ( k ) to the as yet unknown UTC with its atomic clocks . At the PTB ( Physikalisch-Technische Bundesanstalt ) it has been implemented by an active hydrogen maser since 2010, the frequency of which is compared and adjusted almost daily with a cesium fountain clock . In Germany, UTC (PTB) is used to derive the legal time by adding one or two hours .

The time scales UTC ( k ) are used for the constant indirect comparison of all clocks that contribute to the TAI. Within a laboratory, the clocks are compared with the respective UTC ( k ). The differences between the various UTC ( k ) are determined using GPS or TWSTFT . PTB plays a central role in this, because since 2018 all UTC ( k ) have been compared directly with UTC (PTB). The clock comparison using TWSTFT usually provides more accurate results than using GPS. This is due to the fact that the calibration uncertainty u Cal , which depends essentially on the signal propagation time within the devices used and which largely determines the stability of the TAI or UTC, is usually several nanoseconds in GPS measurements and is therefore significantly greater than a typical value of 1, 5 ns for TWSTFT measurements. All comparison results are sent to the Bureau International des Poids et Mesures (BIPM) in Sèvres near Paris and published. The differences between the readings and any two clocks i and j can be calculated from the data .

In order to calculate the TAI, a time scale called "EAL" ( French Echelle Atomique Libre ) is first created in an iterative process , which then results in the TAI by scaling. The EAL is in the form of a weighted mean ,

.

Here again is the reading of clock i at time t . It is modified by a term that is quadratic in time , which is derived from the rate of the clock in the previous evaluation interval and with which a linear frequency drift of the clock with respect to EAL can be compensated. is the weight with which the watch i is included in the mean value ( ). In the slowly changing deviation of this mean value from the reading of a clock j ,

,

the known reading differences are included .

They are determined iteratively in such a way that watches with a predictable rate are given a higher weight than watches that deviate significantly from the forecast. For this purpose, a provisional EAL for the current month M is first formed with the weights of the previous month serving as starting values . With it, the amount of the difference between the expected and actual average frequency of the clock i is calculated for each clock . Together with the corresponding amounts , ... of the previous months is it a measure of the inaccuracy of the clock,

,

won. is an average value of the square frequency deviations of the past months, with recent months contributing more than the longer past. can have values ​​between 5 and 12; Clocks for which only data from less than 5 months are available are observed, but are no longer taken into account for the EAL calculation. With the imperfections of all clocks, the weights are improved

calculated, deviating from this relationship in two cases: (a) The weight of particularly accurate clocks is limited to four times the average weight of all clocks; (b) grossly wrong clocks ( receive the weight 0. The new weights are used to calculate an improved EAL. This improvement loop (improved , improved EAL) is run through four times; the last result is the EAL for the current month.

The duration of the EAL second can differ from the SI second. Therefore, the clock rate of the EAL is compared with that of a few (in the months October to December 2019 there were 7, 5 and 8 clocks.) Particularly accurate clocks, the so-called PFS ( English Primary Frequency Standards ) and SFS ( English Secondary Frequency Standards ) , compared. The comparison is made more difficult by the fact that these clocks usually do not run continuously and that the relationship between the time displayed by the comparison clock (another atomic clock of the laboratory with which the PFS / SFS clock is compared) and the EAL due to the clock comparison using GPS and TWSTFT is known only with limited accuracy. The correction of the EAL through a frequency adjustment (“steering” the TAI) is therefore based on comparative data from a longer period and is announced in advance for the current month. In December 2019 the relationship f (TAI) = f (EAL) −6.493 · 10 −13 was valid . The result is the TAI.

The TAI and the UTC, which differs from it only by an integer number of seconds, is published by the BIPM in the monthly Circular T. It shows the differences between UTC − UTC ( k ) and the real time UTC ( k ) of the individual laboratories as well as the deviations of the GPS and GLONASS times from UTC. (These are approximations of the UTC ( USNO ) or UTC (SU) time scales .) There are also estimates of the accuracy of the TAI and a table with the links used for the time comparison.

See also

literature

  • G. Panfilo and F. Arias: The Coordinated Universal Time (UTC) . In: Metrologia . tape 56 , 2019, 042001, doi : 10.1088 / 1681-7575 / ab1e68 (English). Detailed description of the algorithms

Web links

  • Andreas Bauch: How time is made. In: Einstein Online. Max Planck Institute for Gravitational Physics, 2006 . ;Overview article (not always up to date)
  • FTP server of the BIPM Time Department. BIPM;raw data and results of the UTC / TAI calculation: time comparison measurements, clock weights, circular T, annual reports, ...
  • BIPM Time Department Data Base. BIPM;participating laboratories, clock types, interactive plots of the differences UTC − UTC (k) and UTC − GPS / GLONASS time, ...

Individual evidence

  1. Atomic clocks participating in TAI statistics. In: BIPM Time Department Data Base. BIPM, accessed January 29, 2020 .
  2. List of participants to UTC. In: BIPM Time Department Data Base. BIPM, accessed January 29, 2020 .
  3. a b BIPM (Ed.): BIPM Annual Report on Time Activities 2018 . ISBN 978-92-822-2271-3 , pp. 9–30 (English, full text [PDF]).
  4. a b Circular T. BIPM, accessed on January 29, 2020 .
  5. The value results from the 12 monthly mean values ​​for 2019, which can be found in the relevant Circular T's (Section 3 “Duration of the TAI scale interval d”); see Circular T 373-384 .
  6. TAI − UTC (Jan. 1, 1972 - Dec. 28, 2020). IERS , accessed January 29, 2020 .
  7. a b Andreas Bauch, Stefan Weyers, Ekkehard Peik: How does an atomic clock tick? - Realization of the second from 1955 to today . In: PTB-Mitteilungen . tape 126 , 2016, p. 17–34 ( full text [PDF; accessed on February 19, 2020] text page = pdf page + 16).
  8. a b Resolutions adopted at the 26th CGPM. BIPM, November 2018, accessed on January 31, 2020 (English, French, Resolution 2: On the definition of time scales ).
  9. Michael Soffel: Astronomical-geodetic reference systems. (PDF) 2016, pp. 38–41 , accessed on January 31, 2020 .
  10. a b A conventional value for the geoid reference potential W 0 . (PDF) In: Unified Analysis Workshop 2017. German Geodetic Research Institute , pp. 5–7 , accessed on January 31, 2020 (English).
  11. ^ A b Resolution A4: Recommendations from the Working Group on Reference Systems. (PDF) In: XXIs General Assembly, Buenos Aires, 1991. IAU, pp. 12-22 , accessed on January 31, 2020 (English, French, TT definition in Recommendation IV (p. 16)).
  12. ^ B. Guinot: History of the Bureau International de l'Heure . In: ASP Conference Series . tape 208 , 2000, pp. 181 , bibcode : 2000ASPC..208..175G (English).
  13. Atomic Clocks and PSFS terminology for BIPM clock codes. In: BIPM Time Department Data Base. BIPM, accessed February 19, 2020 .
  14. Andreas Bauch et al .: Generation of UTC (PTB) as a fountain-clock based time scale . In: Metrologia . tape 49 , 2012, p. 180-188 , doi : 10.1088 / 0026-1394 / 49/3/180 .
  15. UTC (PTB) or MEZ / CESZ are distributed by the time signal transmitter DCF77 , several NTP servers and the uhr.ptb.de page that can be accessed in any web browser .
  16. GPS time comparisons. PTB, accessed on February 19, 2020 .
  17. Two-way time and frequency comparisons (TWSTFT). PTB, accessed on February 19, 2020 .
  18. a b c d e G. Panfilo and F. Arias: The Coordinated Universal Time (UTC) . In: Metrologia . tape 56 , 2019, 042001, doi : 10.1088 / 1681-7575 / ab1e68 (English).
  19. Circular T 384 [December 2019]. BIPM, January 9, 2020, accessed on February 22, 2020 (Section 5 “Time links used for the computation of TAI, calibrations information and corresponding uncertainties”; Explanation in explanatory supplement ).
  20. Clocks and time transfer files. BIPM, accessed on February 19, 2020 (English, explanation in readme.pdf ).
  21. ^ Reports of evaluation of Primary and Secondary Frequency Standards. BIPM, accessed February 19, 2020 .
  22. Circular T 384 [December 2019]. BIPM, January 9, 2020, accessed February 19, 2020 (Section 2 “Difference between the normalized frequencies of EAL and TAI”).
  23. The time scales TAI and EAL. PTB, accessed on February 19, 2020 .
  24. "SU" is the laboratory abbreviation of the Russian metrology institute VNIIFTRI .