CCL-CCTF 2021 adjustment of reference frequencies
Information on the initial publication of results - 25 March 2022
Input measurement file
The input measurement file of the adjustment contains 105 measurements, including 33 frequency ratios and 72 absolute frequency measurements.
Each line contains the following information:
- Internal reference number. Note that q51 is not used.
- Identification of the transition (for absolute measurements) or pair of transitions (for ratios).
- Measurement value in Hz for absolute measurements or without dimension for ratios.
- Measurement uncertainty in Hz for absolute measurements or without dimension for ratios.
- Identifier of the reference within brackets with optional additional information (see below).
A few measurements require additional information regarding their use in the adjustment:
- Uncertainties multiplied for q1 (x3), q31 (x1.5), q52 (x6), q74 and q78 (x3), q88 and q105 (x2).
- Uncertainty expanded from 1.5x10-16 to 1.65x10-16 for q73 and from 1.3x10-16 to 1.6x10-16 for q98.
- A few other remarks clarify the use of the indicated reference and do not change the stated uncertainty.
These uncertainty changes are included in the values of the input measurement file and are also reflected in the computation of correlation coefficients.
Input correlation file
The input correlation file of the adjustment contains 483 coefficients.
396 coefficients account for the correlation between absolute measurements due to using the same Primary or secondary frequency standards to access the SI second.
87 coefficients were determined in a series of different analyses.
See more details in this Technical Memo.
Results
The adjustment's results consist in the values of recommended frequencies and a full covariance matrix that will be made accessible with the final publication.
An interim publication is through entries tagged "Update: 2021" on the Recommended values web page.
In these entries, the recommended standard uncertainty is twice the square root of the corresponding diagonal entry in the covariance matrix of the adjustment.
The multiplication by two was recommended by the CCL-CCTF WGFS to allow for as-yet-unidentified sources of uncertainty or correlation,
in keeping with the purpose of the recommended values [Riehle et al. 2018],
and is separate from any coverage factor used to calculate expanded uncertainties
or confidence intervals, see the GUM.
The recommended value is the direct result of the adjustment, rounded as deemed adequate with respect to the recommended uncertainty.