[time-nuts] syncronized clocks

Chris Caudle chris at chriscaudle.org
Fri Sep 18 14:06:43 UTC 2015


On Thu, September 17, 2015 8:26 pm, Can Altineller wrote:
> b. If we put two rubidium sources, and have them warm up and stabilize,
> and then hook up a 2ch scope and observe the pattern, will there be a
> phase difference between the 10mhz signals? (without the scope doing any
> syncing)

Any two sources which do not have a way to force synchronization to some
external reference at startup will always have a phase difference.
Did you really mean a phase difference (i.e. the oscillators did not start
oscillating at exactly the same time) or a change in the phase difference
over time (i.e. the oscillators are not at exactly the same frequency)?

The answer in both cases is yes, but the reasons are different.  You can
eliminate the starting phase difference by having a gate circuit in the
oscillators to allow holding the output off until a synchronizing signal
is received.  That will get them started together, but if the
synchronizing signal is a one time event then the two will drift apart
over time, either because of differences in the environment due to
influence of temperature, magnetic field, barometric pressure, etc. on the
oscillating mechanisms, or just imperfections in the mechanisms themselves
that prevent operating at exactly the theoretical absorption frequency of
rubidium.

To eliminate the drift in phase difference you have to discipline the
oscillator frequency with an external reference of some kind.  Of course
that is talking in absolutes, and there is probably a point where you no
longer care because the differences are very small.

The typical language used to describe the difference of one frequency
referenced to another is parts per million, meaning 1 Hz frequency error
of offset per MHz output, so 1Hz off for a 1MHz oscillator, or 10 Hz off
for a 10MHz oscillator, etc., sometimes written in scientific notation as
1 x 10^-6.
Similarly for parts per billion, 1 x 10^-9, and parts per trillion as
well, 1x10^-12, but it seems like the scientific notation is much more
common than the English language description when you get to those levels
(possibly due to differences in American and historical British definition
of billion and trillion, see Wikipedia references to long scale and short
scale).

-- 
Chris Caudle







More information about the Time-nuts_lists.febo.com mailing list