Standard Frequencies

[Nick]

What about the MSF Signal at 60 kHz - Caesium stabilised to 2 parts in 10^12

Also see http://www.npl.co.uk/science-technology/...ime-signal

MSF is better than that. That's just the short term stability. It's ultimately referenced to the national standards at the NPL and hence to the interntational standards.

...which is why I referenced the NPL documents.

Looking at the latest monthly bulletin, MSF was within 6.5uS of UTC (as determined by the NPL) every day (about 1 part in 2x10^15). NPL's UTC timekeeping is accurate to 1 part in 5x10^15, or about 1 second in 158 million years.

Probably good enough for most people.

[ppppenguin]

When you're using a transmitter which sends time signals (as against just a standard frequency) then remember that each kilometre from the TX is about a 3us delay. If you know your location you can in theory allow for that. if you really need that level of absolute timekeeping. Frequency accuracy is of course unaffected.

I am still amazed by PTP (Precision Time Protocol) which allows sub microsecond timekeeping over TCP/IP netwroks whose delay isn't very predictable.
https://en.wikipedia.org/wiki/Precision_Time_Protocol

One application of PTP, not yet widespread, is the synchronisation of TV systems. Unlike PAL or NTSC which need nanosecond accuracy genlock between sources for accurate switching etc, digital TV is happy with microseconds of error since all inputs to studio equipment have at least a 1 TV line buffer.