So basically, you set up your power distribution hardware at the generation source to vary the frequency of the electricity in the grid by some arbitrary but small range so that any mains powered consumer device can determine the time without maintaining an internal clock.
I don't know the sensetivity, but for arguments sake let's say that from 00:00 to 23:59, the frequency varies at a linear progression from 49hz-51hz. A simple and non-variable comparator circuit could then deduce the time at any point connected with the power grid to within a second or two.
Could be used to remove unreliable PV circuits in public lighting circuits. Could be used so all your clocks stay on time. Could be used by home computers, electric hot water heaters, any number of devices to reliably determine time in order to trigger or control some other process.
Inspired by [kansan101]'s smart bulbs idea.-- Custardguts, Jan 17 2013 Most computers and mobile phones check off an internet- based clock system already, no?-- UnaBubba, Jan 17 2013 Yep, you're right. But many other appliances and systems aren't.-- Custardguts, Jan 17 2013 Could probably do this with a secondary signal on top of the main 50Hz instead of frigging around with the primary frequency - however I figure that distribution hardware and interference, etc would make that more difficult or harder to maintain accuracy.-- Custardguts, Jan 17 2013 Hmm, I thought it was held very stready due to frequency sensetive devices and the voltage-frequency relationship of the alternators, not for any other purpose. I used to work in a regional power station (off main grid) and we had much management equipment in place to maintain stable frequency - but there was still an error range.
So are you saying consumer goods do a sum of cycles to determine the time after a given reset event, or that they use the mains frequency as a heartbeat for their internal clocks?
My idea is intended that a device could tell the time of day *anytime* and *instantly* by comparing mains frequency to a given internal standard frequency - no clock mechanism needed. Similar purpose, but not the same as you describe unless I've very much misinterpreted what you said.-- Custardguts, Jan 17 2013 Isn't there an effort to smooth power output so as not to damage devices?-- UnaBubba, Jan 17 2013 I like it. +-- Kansan101, Jan 17 2013 Go to 60 Hz. You're clearly getting more cycles for your money, which every good marketing agent knows is a better value. Imagine the time savings as your clocks run faster...-- RayfordSteele, Jan 17 2013 [bigs] I assume (hope!) it has not been relaxed - there must be lots of those mains powered wall clocks which use the AC to run a motor, and which need a long term average of exactly 50hz to keep good time? I read that they deliberately sped up or slowed down the frequency overnight to maintain the average.
Just today I was calibrating an old record player with a stripey disc and an incandescent mains-powered lightbulb.-- pocmloc, Jan 17 2013 Data over power is fairly WKTE, but it generally uses very high frequency small amplitude variation stacked on top of the main (50 or 60 hz) frequency. Most equipment filters this out by it's very nature.
Varying the mains frequency does tend to mess up equipment, it can increase the burn-out rate of motors or cause equipment (some of which still uses clipped wall frequency to provide a timing pulse) to run off-speed.-- MechE, Jan 17 2013 I don't think you'd need to vary the frequency by even 1Hz. Cheap timing circuits should be able to measure frequency to within a tiny fraction of a Hertz.
However, the main problem is that mains frequency varies quite widely, depending on load and on which generators are on- and off-line. So, a clear signal would have to vary the frequency by more than the greatest amount of 'random' variation.
Of course, you could use frequency variation to encode things more subtly. For example, five cycles advanced by 100µs and then five cycles retarded by 100µs; this would be detectable and probably discriminable from random fluctuations. But it would be very difficult to inject that kind of short-term variation into the supply, since mains frequency generally depends on the rotational speed of big heavy things.
Probably easier to superimpose a high-frequency, very-low-voltage signal on top of the mains.-- MaxwellBuchanan, Jan 17 2013 Brilliant, would work well for streetlights or even warning sirens etc... [+]-- Brian the Painter, Jan 17 2013 Tweaking the entire power grid to get a weak time reference.... why not ?-- piluso, Jan 18 2013 random, halfbakery