If I make a digital clock which first detects the AC frequency roughly (whether it is 50, 60, 100Hz, etc), then uses it as its clock pulse source, how precise would it be? Would it work world wide precisely?
For instance, in USA, what is the mean and standard deviation of the 60Hz AC frequency?
Grid frequency will change throughout the day due to imbalance between electricity generation and consumption.
Grid operator is obligated to keep the frequency within +/- 0.5Hz of 50Hz (frequency and range depends on the country's grid codes). When you measure time based on the assumption that every period is equal to 20ms you will have a time deviation error that. This error is kept within certain bounds by the network operators. If the time deviation is too big, a frequency compensation is scheduled to compensate it.
You can find public data about current grid frequency online e.g. Swiss and UK grids. The Swiss website also shows the current time deviation.
This post on hackaday links to this interesting article where someone investigates the accuracy of 60Hz mains frequency over time, I hope it might help.
One thing it shows for certain is that the AC line is definitely not a steady frequency, but it varies over time. It also shows that the frequency seems to flutter a lot more abruptly during peak hours of the day. At night, it changes much more slowly.
As a possible explanation, I've heard that electric utility companies know that many customers depend on that 60Hz to keep their clocks going accurately, but it's difficult to keep the 60Hz exact when energy demand is fluctuating rapidly.
As a solution, they will adjust the frequency slightly during off-peak hours to compensate for any mishaps during the day.
The conclusion seems to be that the frequency probably averages out quite accurately, but will fluctuate depending on the time of day, as a result of the load on the network.
There's some interesting links at the bottom of the article to further reading and experiments. I also seem to remember another article posted via hackaday which also looked at the same topic, but I can't seem to find it now.
There was a fascinating news story on this topic recently: http://www.bbc.co.uk/news/science-environment-20629671
In short, the London Metropolitan Police forensic lab has recorded the variations in AC frequency for the last 7 years in the UK National Grid, and use it to validate any claimed audio recordings were recorded contiguously during the stated period.
"While the frequency of the electricity supplied by the national grid is about 50Hz, if you look at it over time, you can see minute fluctuations in the order of a few thousandths of a hertz."
It'll be something like +/-0.5 Hz maximum, and during the day I noticed a clock drift by a minute or two, but the long term average error will be zero; the utilities presumably reference it to time standards.
Observing the instantaneous frequency is one way to compare supply and demand. If the frequency drops, the utilities will bring more power on line to maintain it.
See also this question and the current UK grid frequency
Wikipedia indicates that the US grid operators intervene to bring the count back when it's drifted by 2, 3, or 10 seconds (West, Texas, and East grids) from the 60hz long term target used for time keeping. In 2011 US regulators authorized an experiment to see what would happen if the adjustments stopped being made; however at least through 2021 (see slide 12) they were still being done.
My Romanian (or was it the Polish guy.. Quite some time ago) friend claimed they would reduce the frequency to slow things down and save power. He might have been telling a joke, but it was pretty dead-pan.
