Ah, just set the timezone to UTC-00:05:46
March 7, 2018 2:48 AM   Subscribe

Synchronous electric clocks count the oscillations of the mains current to keep time: After 50 or 60 cycles (depending on where you are in the world), one second has elapsed. This works wonderfully as long as the average mains frequency is constant, or—in practice—adjusted to compensate for errors. Things get a little tricky when 113 GWh of energy somehow go missing and all microwave clocks on an entire continent go slow…

Every day at 8 am, Swissgrid adjusts the daily target frequency of the synchronous grid of Continental Europe to keep the long-term average mains frequency on the European continent at 50 Hz. Since mid-January, the grid hasn’t been able to keep up with these targets. “The deviations originate from the Serbia, Macedonia and Montenegro block, specifically in Kosovo and Serbia, Entso-e spokeswoman Susanne Nies told S&P Global Platts. Kosovo has been using more power than it produces while Serbia, responsible for balancing the Kosovo grid, has failed to do this, Nies said. As a result, Serbia has been free-riding on the system.”

Provided that the political situation between Serbia and Kosovo can be resolved this week and Serbia starts feeding its fair share into the grid again, affected clocks will be made to catch up by speeding up the mains frequency to 50.01 Hz for the next few weeks. For now, they’re running about 6 minutes behind mean solar time.
posted by wachhundfisch (52 comments total) 48 users marked this as a favorite
 
For those wondering why energy going missing causes the power grid frequency to slow down, there's a nice natural language explanation on this Stack Exchange answer, and a fairly decent article on Wikipedia.

But wow, I had no idea that (A) this happened, (B) That the only way to tell that entire countries are free-loading on the European power grid is when the frequency drifts by whole minutes, and (C) that this kind of politically-derived screwup doesn't happen more frequently.

Thanks for posting this, I don't think I would have found out about it anywhere else.
posted by Eleven at 3:07 AM on March 7 [11 favorites]


I'm guessing that frame rates in the PAL and NTSC analogue TV systems were derived from the mains frequencies in Europe and North America.
posted by acb at 3:24 AM on March 7 [6 favorites]


On my tariff, that 113GWh would cost me about £18.7 million.
posted by dowcrag at 3:42 AM on March 7


@acb There is a great YouTube video that explains how exactly the 29.97 NTSC frame rate came to be.
posted by nertzy at 3:45 AM on March 7 [16 favorites]


Such clocks are typically radio-, oven clocks or clocks for programming the heating system.
I wonder if that's why some of the clocks I have of those types are so inaccurate. (I live in 60-cycle land.) Then I wonder why they would build clocks like that. Do quartz regulators really cost that much more?
posted by Kirth Gerson at 4:50 AM on March 7 [1 favorite]


From the last FPP link ...

The decrease in frequency average is affecting also those electric clocks that are steered by the frequency of the power system and not by a quartz crystal: they show currently a delay of close to six minutes.

Microwave ovens would use quartz crystals for their microcontroller-based electronics, including the clock.

The clocks affected by AC mains frequency variation would mostly be motor-based. Some old cheap LED clocks used the AC mains as their time base. Not sure if the new ones still do.
posted by ZenMasterThis at 4:53 AM on March 7 [3 favorites]


From the first link:
If the FERC adopts the NAESB petition, TECs will no longer be utilized in the United States and Canada, and clocks timed by them will likely wander uncontrolled until manually reset; it was noted in a technical paper by employees of the National Institute of Standards and Technology[20] and the U.S. Naval Observatory[21] that, had TECs not been inserted in 2016, there would have been over seven minutes lost by electrically timed clocks over much of the United States and Canada...
posted by Kirth Gerson at 4:59 AM on March 7


Such clocks are typically radio-, oven clocks or clocks for programming the heating system.
I wonder if that's why some of the clocks I have of those types are so inaccurate. (I live in 60-cycle land.) Then I wonder why they would build clocks like that. Do quartz regulators really cost that much more?


This is a real TIL situation (that these types of clocks don't necessarily have quartz regulators), and so now I know why when I had a cheap stove sitting next to a cheap microwave in a cheap apartment and the two gizmos both had digital clocks on them, said clocks were always maddeningly different a) from each other b) from any time standard.

But to give a rhetorical question to your rhetorical question - if the engineering/design budget for a $50 microwave is so small that another dollar or two matters, why have a clock in it at all? Why does everything with a display have to tell time? It's kind of maddening when the time change comes.
posted by randomkeystrike at 5:11 AM on March 7 [11 favorites]


Digital clocks that use the mains frequency as a time base are more common than you'd think - the assumption being that the mains frequency is meant to have less long term drift (in theory, zero) than a quartz crystal.

There is a great YouTube video that explains how exactly the 29.97 NTSC frame rate came to be.

Great video, but he glosses over one key part - my understanding is that mains frequency was used originally because the electronics weren't good enough to filter out the mains hum from the picture, which produced a visible band on the screen. If the scanning frequency and the mains rate were different than the band would jump around the screen with each scan - the obvious fix is to make them the same, so the band at least stays still. I'd guess by the time colour came around this was no longer a concern.
posted by grahamparks at 5:12 AM on March 7 [4 favorites]


Brexit's silver lining..
posted by joeyh at 5:20 AM on March 7


> if the engineering/design budget for a $50 microwave is so small that another dollar or two matters, why have a clock in it at all?

It needs a clock to measure the cooking time. Giving you the time of day is a nearly-free value add.

Microwaves used to be built with mechanical timers. These works fine (as long as the spring holds out) but imposes a hard limit on maximum cook times and makes it harder to offer features like adjustable heating levels because that is provided by switching the emitter on and off on a timed schedule — which, again, needs a clock.

In any event, an electronic clock plus display is probably cheaper to make than an egg timer these days.
posted by ardgedee at 5:21 AM on March 7 [5 favorites]


One reason to use a synchronous motor instead of a crystal is that historically, barring shenanigans like the one here, it's way more accurate - someone upthread has posted a bit about the proposed relaxation of the standard but North American producers were basically required to deliver 60x60x24x365 sine wave periods per year, and if the frequency diverged at some point they had to make it up later (that's why the main article here mentions raising the frequency to 50.001Hz for a bit, if they can get everyone to start behaving again).

But yes, guaranteed a crystal and microcontroller is way cheaper than a synchronous motor these days. I realized the other day when I waved my hand uselessly under a soap dispenser that I've come to expect that "sensor + microcontroller + actuator" is at least as likely to be encountered in the wild as "lever."
posted by range at 5:35 AM on March 7 [4 favorites]


This came up [warning, badly translated Dutch ahead] in the news here this week. We have a monthly test of the air-raid sirens that's the first Monday at noon, except ... it didn't happen this time. Turns out all the clocks being wrong broke the scheduled test.

Fortunately we still have NLAlert, which is that special system that terrifies an entire office at once when all our phones go off at once and only half the people can read the Dutch "this is just a test" message and the other half walked away from their desks without their phones.
posted by sldownard at 5:53 AM on March 7 [4 favorites]


Brexit's silver lining.

Joking aside, we've been here before (and nobody was told then either) ... Seconds out in the miners strike
posted by StephenB at 6:06 AM on March 7


Microwaves used to be built with mechanical timers.

Oh, to twist those satisfyingly heavy dials again!
posted by rokusan at 6:07 AM on March 7 [3 favorites]


Does this result in electric customers getting charged (very very slightly) more for the same power? I think modern sampling electric meters will understand the frequency drop and not charge the customer for cycles that don't happen, but I think old mechanical meters might be vulnerable. Of course their overall accuracy is far below the frequency difference, so it's possible no one really cares.
posted by miyabo at 6:18 AM on March 7


Hmmm, the other day I noticed that my oven clock (located in Germany) was about 5 minutes out of sync with my mobile phone. I wonder if this explains why.
posted by chillmost at 6:47 AM on March 7 [1 favorite]


I blame bitcoin
posted by NoMich at 6:52 AM on March 7 [1 favorite]


I blame bitcoin

A hypothesis that fits that facts very well. Cryptomining is my guess, too.
posted by tclark at 6:55 AM on March 7 [1 favorite]


The connected world is far more fragile than most people would like to believe.
posted by tommasz at 6:55 AM on March 7 [1 favorite]


A good mains frequency clock can be more reliably accurate than cheap quartz-timed clocks, which can drift. I have a 100-year old electric clock that uses mains frequency for timing. It has this neat indicator that turns red if power is ever cut to it. I regularly check it to the second against the National Research Council Time Signal on the CBC. As long as the indicator is not red, it is always spot on.
posted by fimbulvetr at 7:08 AM on March 7 [3 favorites]


I've been reading the parallel discussion over at HN and someone posted a cool link to a realtime map of grid synchronization from UTennessee and Oak Ridge National Lab.
posted by JoeZydeco at 7:08 AM on March 7 [8 favorites]


I've been reading the parallel discussion over at HN and someone posted a cool link to a realtime map of grid synchronization from UTennessee and Oak Ridge National Lab.

That’s pretty neat! Looks like the west coast is running slow this morning.
posted by TedW at 7:14 AM on March 7


I can also add some perspective as an appliance control designer here.

A micro + quartz crystal is indeed cheap. But, it also needs DC current to run. When you're flattening an AC wave into DC, you're watching the wave cross from positive to negative and back.

A zero-cross detection circuit is also necessary when you are switching AC loads. You want to minimize stress on your heaters and relays by switching them on and off at the zero-point, not when the hot leg is putting out maximum current.

And if you're watching ZCs, you might as well count them. Now you have a clock.
posted by JoeZydeco at 7:17 AM on March 7 [16 favorites]


And here is more info about that frequency map project. Scroll down to see if you live in an area where they want someone to host a frequency deviation recorder! I would but the nearest location listed is about 70 miles away.
posted by TedW at 7:18 AM on March 7 [1 favorite]


Yeah, I get that the appliance needs a time-keeping circuit to operate cooking timers, and that electronic is better than mechanical. Having a time-of-day display that is always 3 minutes off is still a bug, not a feature.
posted by randomkeystrike at 7:24 AM on March 7 [1 favorite]


Does this result in electric customers getting charged (very very slightly) more for the same power?

Sync issues cause additional thermal losses due to increased current in distribution systems. It's likely to cause infrastructure to fail sooner, which might be an interesting form of cross-border warfare.

Some sampling meters have proven to be quite inaccurate, as they make assumptions about the power waveform.
posted by scruss at 7:34 AM on March 7



One reason to use a synchronous motor instead of a crystal is that historically, barring shenanigans like the one here, it's way more accurate - someone upthread has posted a bit about the proposed relaxation of the standard but North American producers were basically required to deliver 60x60x24x365 sine wave periods per year, and if the frequency diverged at some point they had to make it up later (that's why the main article here mentions raising the frequency to 50.001Hz for a bit, if they can get everyone to start behaving again).


Yup.

Crystals are cheap, but they are not uniform in their frequency output, and their frequency drifts over time and is a function of the ambient temperature. Even if you're using one, it's convenient to use the grid frequency as a source for synchronizing with the rest of the world.

And would you rather have your whole country drift by 6 minutes, so everyone knows that's what's going on, or would you prefer your clock drift its own way, completely independent of everyone else's?

Note that the integration of renewables into the grid is giving grid operator a big incentive to relax grid synchronization standards.
posted by ocschwar at 7:37 AM on March 7


Did I miss the link in the OP that gives background into what the (no pun intended) power-sharing situation is in Serbia & Kosovo with respect to the electric grid? I know a tiny bit of the fraught history between Kosovo and Serbia, but not enough to know this kind of detail.
posted by Fraxas at 7:45 AM on March 7


Power grid synchronisation seems to be largely above politics, or at least the logistics of changing boundaries are too difficult to bother with. The Baltic states are still synchronised to the Russian grid rather than the Central European one (though the then Warsaw Pact countries switched to the European one sometime after the fall of the Berlin Wall), Ireland is one grid, and if the map in the Wikipedia page is accurate, Israel is synchronised with its Arab neighbours.
posted by acb at 7:50 AM on March 7


Electricity seems to transcend all our differences.

220, of course. 120 is for barbarians.
posted by rokusan at 8:27 AM on March 7 [3 favorites]


Even if you're using one, it's convenient to use the grid frequency as a source for synchronizing with the rest of the world.

In effect, as well as supplying energy, electricity grids have functioned as a non-Internet-dependent NTP since before quartz crystal timing was even a thing.

It kind of happened by accident, too. An alternator is pretty much the same machine as a synchronous motor, so when you connect an alternator to the grid, it will want to sync itself to the frequency of the mains without you having to do anything terribly clever to force it to. Then you start pushing extra mechanical power into its drive shaft, and it will shove that power back into the grid as nicely synchronized AC.

Today's power station alternators are huge bloody things and would blast themselves to bits if not carefully synchronized before being connected, but the point is that once they are synced and connected there are inherent electromechanical effects that keep them that way. Any grid-connected alternator that starts to slow down will automatically start operating as a motor and draw just enough power from the grid to keep itself spinning in sync with the fleet. When the fleet as a whole sees more load than it can supply, the whole lot of them slow down together.
posted by flabdablet at 8:36 AM on March 7 [7 favorites]


Finally! An explanation for why every clock in my apartment (clock, microwave, alarm clock/radio) gradually slows down and gets out of synch. At first I would change the time every couple months manually, lately I am refusing out of spite - As of now they're 11 minutes slow and so far I'm still holding strong against good sense.
posted by matcha action at 9:34 AM on March 7


Note that the integration of renewables into the grid is giving grid operator a big incentive to relax grid synchronization standards.

My experience was that putting renewable energy facilities far away from traditional generation sources uncovered existing grid sync issues that the operators weren't aware of. It was usually cheaper to quietly put in VAr compensation than complaining that the grid was out of spec. If you're gonna be using someone else's transmission facilities for the next 30+ years, best not to antagonize them from the get-go.

Smaller inverters do copy the existing grid waveform for their output power (I suppose they could be seen as a very powerful FM synthesizer), and have been implicated in propagating unpleasant transients. Most of the projects I worked on had best-in-class power quality control, and some even had the ability to regulate local power quality at the system operator's control.
posted by scruss at 9:40 AM on March 7


acb: "Power grid synchronisation seems to be largely above politics, or at least the logistics of changing boundaries are too difficult to bother with."

Though the exceptions to this pattern are significant and interesting: it looks like Quebec and Texas maintain their own independent power grids (although connected to the main North American grids).
posted by crazy with stars at 9:51 AM on March 7


I wonder if that's why some of the clocks I have of those types are so inaccurate. (I live in 60-cycle land.) Then I wonder why they would build clocks like that. Do quartz regulators really cost that much more?

These synchronous motor clocks aren't inaccurate. They are more accurate over long periods than crystal clocks. If you had plugged one of those synchronous motor kitchen wall clocks into the U.S. grid over 100 years ago when they first implemented the time error correction system, it would still be accurate today to within 10 seconds, as long as it wasn't disconnected from the grid.

Most crystal clocks are nowhere near that accurate. They typically lose or gain 10 seconds a month, 120 seconds a year, 1200 seconds in a decade. The synchronous motor clock will never be off by more than 10 seconds.

Quartz crystal clocks are cheap and accurate over short time periods but you have to reset them often. The crystal clocks in your PC and iPads are reset every day via internet. The crystal clocks in your cell phones are periodically reset by the cellular network and GPS. The internet (and GPS, cellular) are replacing the power grid as the time base that has worked so well for a century.

But crystal clocks as stand alone devices are not so great for long time periods. Which is why you have to reset your oven, microwave, DVD player, etc or they get off by minutes. Since most people are resetting twice a year for daylight savings time, they aren't aware that their crystal clocks aren't really that accurate.
posted by JackFlash at 9:53 AM on March 7 [4 favorites]


Finally! An explanation for why every clock in my apartment (clock, microwave, alarm clock/radio) gradually slows down and gets out of synch.

No, if you are in North America, it isn't the grid that's causing the problem. The North American grid is still as accurate as ever. It's because all those devices you mentioned have crystal clocks in them which are notoriously inaccurate over months. You have to manually resync them to the correct time. Most people do this twice a year for daylight savings time.
posted by JackFlash at 9:58 AM on March 7


No, if you are in North America, it isn't the grid that's causing the problem. The North American grid is still as accurate as ever. It's because all those devices you mentioned have crystal clocks in them which are notoriously inaccurate over months. You have to manually resync them to the correct time. Most people do this twice a year for daylight savings time.

I've lived in a variety of apartments as the person who would reset the clocks and I've never had them slow down this much (1 or more minutes per month). I will have to take your word for it that this is not the reason, but there has to be some reason why it's only in the apartment I'm currently in that they get so far off.
posted by matcha action at 10:05 AM on March 7


I'm in the process of building this transistor clock kit (no integrated circuits, no quartz crystal) which uses the wall line for timing. It will be interesting to see if it will hold as accurate time as my ancient mechanical electric clock.
posted by fimbulvetr at 10:11 AM on March 7 [2 favorites]


there has to be some reason why it's only in the apartment I'm currently in that they get so far off.

Could be temperature variations between the different places you have lived. Crystal oscillators are subject to variations in temperature.

None of those devices you mentioned use the grid for a time base. You set the crystal clock and they run on their own until you reset them.
posted by JackFlash at 10:14 AM on March 7


Quartz crystal clocks are cheap and accurate over short time periods but you have to reset them often.

Are these crystal clocks fundamentally different from the quartz-regulated wristwatches I own? I have a half-dozen of them, none very high-end, and when I check them against my phone, they are spot-on.
posted by Kirth Gerson at 11:30 AM on March 7


The frequency of a quartz oscillator varies by temperature. The one in your watch is calibrated to be the most precise when it's strapped to your wrist and soaking in your body heat.

You turn out to be a very important part of the watch's accuracy!
posted by JoeZydeco at 11:56 AM on March 7 [3 favorites]


fimbulvetr: I have a 100-year old electric clock that uses mains frequency for timing. It has this neat indicator that turns red if power is ever cut to it.

You can't just say something that intriguing without providing pics!
posted by xedrik at 11:58 AM on March 7 [2 favorites]


Metafilter: you can't just say something that intriguing without providing pics.
posted by randomkeystrike at 12:19 PM on March 7


Oh, there is nothing exotic about it. It is just an old Telechron like this. Guess I should have said 90-year-old, rather than 100. Telechron clocks are really nifty, easy to find, and extremely reliable.
posted by fimbulvetr at 12:49 PM on March 7 [1 favorite]


I know that Serbia-Kosovo relations are pretty strained at the minute, after the assassination of Oliver Ivanović, so presumably that would be something that's kind of holding back a resolution between Serbia and Kosovo, and with Russia having influence in Serbia and also control over European natural gas supplies, I guess everyone's treading lightly on this.

But any implication that it was hard to track down where the excess load was coming from seems pretty hard for me to believe. Surely that's the kind of thing that's being continually analysed at a grid level: not just the current flowing through interconnects (e.g. 6.07 GW from Germany to France and 2.00GW from France to Britain at time of posting), but also local lags and leads on the frequency.

So I'm assuming that publicly shaming Serbia and Kosovo like this is the last resort, and all kinds of other diplomatic daggers have been drawn beforehand to no avail. I guess also that wherever the excess load is going (stolen by bitcoiners?) is a slightly murky part of the market, rather than an ordinary wholesale electricity consumer. Maybe someone's built an additional transformer onto a 350kV distribution station and they're using that for, err, purposes. It seems extremely odd.
posted by ambrosen at 5:26 PM on March 7


Funny enough, telco timing sources have also become less accurate in recent years. I'm fairly certain the frequency regulation standards in most of the US's grid regions were relaxed somewhat sometime in the last decade or so. Back when they were required to ensure the correct number of cycles had occurred every single day and make any necessary adjustment in fairly short order (less than 24 hours, IIRC), but now they can drift farther and take longer to correct. Same for the voltage standard, though that has always been much less strict than the frequency standard, since there are good reasons why it isn't feasible to ensure all users have precisely the same voltage at all times. I believe it used to be around +/- 5ish volts and now it's +/- 10ish.

(That's separate from the currently active proposal to basically eliminate the use of the electric grid as a timing source)
posted by wierdo at 5:32 PM on March 7


The one in your watch is calibrated to be the most precise when it's strapped to your wrist and soaking in your body heat.

Maybe so, but most of the time. I'm not wearing them. Sometimes I wear the same one every day for weeks. Then I don't wear it for a while, and wear other ones. The watches don't seem to care how often or how long they're on my wrist; they just keep time. So, again -- are their movements different from the clocks being disparaged above?
posted by Kirth Gerson at 5:57 PM on March 7


As for any electronic component, crystal oscillators are available in a range of tolerances, both for frequency accuracy and temperature sensitivity. A wristwatch crystal could easily be over ten times as accurate as one in a typical microwave oven.

Also, even when not on your wrist, a wristwatch will typically be stored indoors i.e. in an environment maintained within the fairly narrow range of temperatures comfortable for human habitation. Crystals in appliances, by way of contrast, will often be subject to local heating from within the appliance itself; an appliance consuming multiple hundreds of watts is going to make much more local heat than a wristwatch that draws microwatts from its battery.
posted by flabdablet at 7:11 PM on March 7


I guess pretty much all future grid interconnects will be HVDC, so synchronous grids aren't going to grow in size any more. Still, one hopes that there are a lot more grid interconnects going to happen in the future. There's no reason for Iceland not to connect to North America and Europe, and, in the long run, for global shipping of electricity to be as cheap as containers have made global shipping of goods.

This is my techno utopia, and obviously I'm assuming that the most profitable energy exporters will be running renewables. There's no reason why this shouldn't happen.
posted by ambrosen at 6:05 AM on March 8


for global shipping of electricity to be as cheap as containers have made global shipping of goods

Though long-distance transmission of electricity inevitably incurs loss, and while it will be possible to send electricity through HVDC interconnects to wherever supply and demand dictate, it will always be more efficient to use electricity produced locally. And the rise of renewables, and grids going from a few large centralised generators serving lots of consumers to large numbers of small solar/wind generators mixed in with consumers, looks like reducing the average distance electricity has to travel. Presumably the interconnects will be used to smooth out peaks (much as battery farms are, only working in space rather than time).
posted by acb at 9:19 AM on March 12 [2 favorites]


Much like container ships vs. supermarket lorries, the long-distance bulk transportation of electricity is surprisingly cheap. The converters for the 1100kV UHVDC lines in China are (sorry, I lost my links for this, but it was from Siemens) 99% efficient, and a 500km 800kV line loses 3% due to resistance and corona losses.

The EU's certainly got targets around this, including the following
The countries surrounding the Northern Seas are currently insufficiently interconnected to make optimal use of existing and foreseen onshore and offshore generation capacity. The Northern seas offer a unique opportunity to supply a substantial amount of low carbon, indigenous energy, produced close to some of the most energy intensive regions of Europe. Its power generation potential may amount to 4-12% of the EU's electricity consumption by 2030.
, and a target (already met by most) for all EU electricity markets to be able to trade ± 10% of their generation capacity. The claimed benefit is a €12-40 billion per year reduction in electricity costs by 2030 through driving production costs down to that of the lowest cost providers.
posted by ambrosen at 5:20 PM on March 17 [1 favorite]


« Older Two hunters bonding   |   Melodifestivalen, and the road to Lisbon Newer »


This thread has been archived and is closed to new comments