This week I was in Rotterdam for a RIPE meeting. On Friday morning I gave a lightning talk called where does my computer get the time from? The RIPE meeting website has a copy of my slides and a video of the talk; this is a blogified version, not an exact transcript.
Where does my computer get the time from?
from NTP - here’s a picture of an NTP packet
and here’s a picture of David Mills who invented NTP
simple question, easy answer, end of talk? No!
let’s peel off some layers…
stratum 3 NTP servers get the time from stratum 2 NTP servers,
stratum 2 NTP servers get the time from stratum 1 NTP servers,
stratum 1 NTP servers get the time from some reference clock
maybe a radio signal such as MSF in Britain or DCF77 in Germany
but in most cases the reference clock is probably a GPS receiver
here’s a GPS timing receiver
and here’s a GPS satellite
where does GPS get the time from?
Schriever Space Force Base in Colorado
they look after a lot of different top secret satellites and other stuff at Schriever, as you can see from all the mission logos
so you can’t get close enough to take a nice photo
Where does Schriever SFB get the time from?
the US Naval Observatory Alternate Master Clock is on site at Schriever in Colorado
the US Naval Observatory Alternate Master Clock gets the time from the US Naval Observatory in Washington DC
there are three answers
the first answer is atomic clocks, lots of atomic clocks
in the background there are dozens of rack mounted caesium beam clocks
in the foreground the black boxes house hydrogen masers
these shiny cylinders are rubidium fountains
the USNO has so many atomic clocks they have entire buildings dedicated to them
When I was preparing this talk I noticed on Apple Maps that there’s a huge building site in the middle of the USNO campus. It turns out they are building a fancy new clock house; the main limit on the accuracy of their clocks is environmental stability: temperature, humidity, etc. so the new building will have serious air handling.
the second answer is that UTC is a horrible compromise between time from atomic clocks and time from earth rotation
so the USNO gets the time from the international earth rotation service, which is based at the Paris Observatory
twice a year the IERS sends out Bulletin C, which says whether or not there will be a leap second in six months time; leap seconds are added (or maybe removed) from UTC to keep it in sync with earth rotation
the IERS is spread across several organizations which contribute to its scientific work
for example, you can subscribe to IERS Bulletin A, which is a weekly notice with precise details of the earth orientation parameters
Bulletin A is sent out by the US Naval observatory
they need to know the exact orientation of the earth under the GPS satellites, so they can provide precise positioning
the third answer is, how does the USNO know its atomic clocks are working well?
that information comes from the international bureau of weights and measures in Paris, who maintain the global standard UTC
how does the BIPM determine what UTC is?
the BIPM collects time measurements from national timing laboratories around the world, and uses those measurements to determine official UTC
periodically they send out Circular T which has information about the discrepencies between official UTC and UTC from the various national time labs
the BIPM is responsible for maintaining the international system of units, which is defined by the general conference on weights and measures
the CGPM is an international treaty organization established by the metre convention of 1875
UTC is an implementation of the SI unit of time, based on quantum measurements of caesium atoms
where did this magic number, about 9.2 GHz, come from?
in 1955, Louis Essen (on the right) and Jack Parry (left) built the first caesium atomic clock
the current definition of the second came from the calibration of this clock
before atomic clocks, the definition of the second was based on astronomy, so Essen and Parry needed help from astronomers to find out how fast their clock ticks according to the existing standard of time
they got help from the astronomers at the US Naval Observatory
the way it worked was William Markowitz measured time by looking at the skies, and Louis Essen measured time by looking at his atomic clock, and to correlate their measurements, they both listened to the WWV radio time signal broadcast by the national bureau of standards in Washington DC
this project took 3 years, 1955 - 1958
Markowitz was measuring the “ephemeris second”
in 1952 the international astronomical union changed the definition of time so that instead of being based on the rotation of the earth about its axis, it was based on the orbit of the earth around the sun
in the 1930s they had discovered that the earth’s rotation is not perfectly even: it slows down and speeds up slightly
clocks were now more precise than the rotation of the earth, so the ephemeris second was a new more precise standard of time
the ephemeris second is based on an astronomical ephemeris, which is a mathematical model of the solar system
the standard ephemeris was produced by Simon Newcomb in the late 1800s
he collected a vast amount of historical astronomical data to create his mathematical model
it remained the standard until the mid 1980s
here’s a picture of Simon Newcomb
he is a fine-looking Victorian gentleman
where did he work?
at the US naval observatory!
(and the US nautical almanac office)
I have now run out of layers: before this point, clocks were set more straightforwardly by watching stars cross the sky
so, to summarise my talk, where does my computer get the time from?
it does not get it from the Royal Greenwich Observatory!