[time-nuts] Recording mains frequency

[Bill Hawkins]

Just a thought about interpreting data:

The mains frequency changes as the loads on the system change.
Dispatchers bring more generators on line or drop them.
Some of this is predictable, depending on business hours, weekends,
holidays and the seasons.

Just as one can take the crystal aging out of the frequency vs. time
display, it should be possible to subtract an average of the variations
caused by predictable load changes. This leaves a display that makes
unusual deviations much more visible.

The line frequency is not like a disciplined crystal oscillator. There
is no phase locked loop adjusting the line frequency, just dispatchers
trying to produce the same number of cycles per day. The length of a day
is precisely known from national time standards.

Short term plots (e.g. daily and weekly) of line frequency should be
referenced to time of day, not elapsed time.

It would be interesting to see how the daily load averages vary from
region to region.

Bill Hawkins