4 minute read time.
In an historic moment on 16th November 2018; the General Conference on Weights and Measures (CGPM) Agreed proposals to redefine the SI System of units including redefinition of four of the base units. These redefinitions come into force today. Here we look at what this might mean for users of measurement beyond the Laboratories of the National Standards Laboratories. 
93737f811d4b078ac9b1989080bfa602-huge-redef.jpg


Today (20th May) sees the launch of a new set of definitions of SI units. These are changes that have been rigorously researched and argued over for many years and are a remarkable achievement, amongst other things, finally laying to rest the last unit based on a physical artefact (the kilogram).   There will be lots of publicity today about what this all means for the Scientific community [For inspiration try https://www.linkedin.com/feed/update/urn:li:activity:6534861707213320192  ;for the detail try  https://www.nist.gov/si-redefinition/turning-point-humanity-redefining-worlds-measurement-system or https://www.bipm.org/utils/common/pdf/SI-statement.pdf ].  However, I wanted to take the opportunity to start a discussion about what it means to the users of measurement and particularly the billions of people who use measurement to make critical decisions of one form or another.

We are entitled to ask if and how these developments help to provide justifiable confidence in those measurements. 

I do not claim to have all of the answers, and hope that this blog sparks a debate to draw out answers from a diverse range of disciplines but here’s a shot at it:

 

Do any of the units change and does it matter?  For almost all users of the system, the answer is a simple no.  The references on your calibration certificates may change but the values, in general, will not. [The exceptions are electrical units where changes up to circa 1 part in 10^7 are involved].  This is because great care has been taken to ensure that the new realisations will  produce nominal values identical to the old.  It will take some time for the use of the new realisations to "wash through" the system of metrological traceability but when they do, there will be no immediately discernible difference within the uncertainty bands of the calibration other than for the electrical units.  However, since the long established values of the units are being "fossilised" in this process, anyone working with large volumes of measured data over multiple years and trending them to detect changes within the uncertainty bands of the primary standards, could pick up a shift in the trend. 

As part of the redefinition, the following physical "constants" have been fixed by definition and will no longer change:
  • The Planck constant h is exactly 6.62607015×10−34 joule-second (J⋅s).




However, prior to the redefinition, the constants were periodically re-evaluated as experimental data improved so people working with these constants over many years to the highest levels of precision (<<1 part in 10^6) might be well advised to check that historical data is not impacted.



Does the change to the realisation of the units improve confidence?  Initially, the uncertainties we will observe will be similar and the shear number of experiments and cross check underpinning the system should result in preserved integrity of metrological confirmation  i.e. no greater risk of faults and mistakes in the calibration system.  In the future, the standards laboratories, as they look to minimise calibration uncertainties over a wide range of values, will be able to take advantage of the new methods of realisation to improve uncertainty at the extremes of scale.  For example,  the previous realisation of the kilogram resulted in the minimum uncertainty at the value of 1kg. At a millionth of 1kg, the uncertainty is heavily compounded. It should be possible in time to develop Kibble balance approaches to realise 1millionth of a kilogram directly with much reduced uncertainty. 


Are there other spin-offs?  Two benefits that come to mind in the longer term:

1. The potential to realise the units locally without reference to National Laboratories means that  local systems could be envisaged that deliver metrological confirmation without external calibration.  This could drive cost savings from equipment down-time and avoided effort correcting non-conformance identified during calibrations.  It could also allow for traceability in inaccessible systems eg long duration space flight.

2. The greater reliance on quantum standards may result in higher integrity systems - i.e. metrological confirmation immune to drift and either correct or in a failed state.


I hope the above has proved thought provoking. There are many, better qualified than I who will be able to improve on these thoughts and I hope that they will accept this in the intended spirit of beginning a dialogue about the expended impacts of the amazing developments in metrology coming into effect today.


If this has been of interest, The International Measurement Community being championed through the IET and NCSL International should be of interest too: https://communities.theiet.org/communities/home/301
https://twitter.com/IntlMeasurement


Pete Loftus
  • The other fun thing is the permeability of free space is now a measured value not 4.pi.10^(-7) H/m, see https://en.wikipedia.org/wiki/Vacuum_permeability
  • Now that Kg are defined via the measurement of Amps, can we make Kg/Amp a dimensionless value please?


    Alternatively can we please properly extend the dimension system to allow the use of Angle and other dimensions (such as Money?) that will help prevent silly errors of computation that could be easily caught by dimensional analysis.


    As someone who worked in Electro optics where radiance/irradiance/etc had lots of confounding angle factors it was a perennial problem that the BIPM (SI) units (and dimensions) system failed to allow such additions which meant software suppliers were loathe to provide such error reducing capabilities.


    Angle is simple arc-Length / radius-Length. Meanwhile Length is not a dimension, rather it is the measurement norm of the 3d spatial space.


    Philip


    PS the SI "angle of time" is 1metre / (c.1second) = 3.3* 10^(-9) radians = 3.3 nano radians! no wonder no one see quantum effects at SI practical scales.