[time-nuts] Calibration and temperature

Poul-Henning Kamp phk at phk.freebsd.dk
Fri Nov 28 19:38:39 UTC 2008


In message <6D3AC52128424986A16AB823763F4AEE at cyrus>, "Bill Hawkins" writes:

>Seems to me that all physical quantities have to be calibrated against
>standards, because they're all defined by humans, starting with the
>length of the king's arm.

For several decades now, the metrological community has been on a steady
march away from "artifact based" standards to "physical phenomena"
standards.

For instance the measure of length is no longer a piece of metal in Paris
but a given number of wavelengths of a particular spectral line under
certain conditions.

The biggest missing piece is the kilogram, where the two best contenders
are both too unmanageable still:

One contender is a perfect sphere of silicon-28 is so state of the
art that it takes russia to separate the isotopes, germany to make
a crystal and australia to polish it into sphere.
	http://www.csiro.au/science/ps35k.html
The point is that if the crystal is perfect and the spherical form
is prefect, you can calculate how many atoms there are.

The other contender is the "Watt-balance", where in essense you link
energy spent in an electro magnet to force, and from force to
mass using newtons 2nd law:
	http://en.wikipedia.org/wiki/Watt_balance

So overall, the strategy is to find physical phenomena which are locally
reproducible, exactly like the second is derived from Cs-133 atoms,
instead of comparing to "random artifacts".

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.



More information about the time-nuts mailing list