How stable is the mains voltage?

[Faringdon]

Hi,
Supposing you have 50Hz mains. Suppose you know when every other zero cross occurs...eg, every 20ms.....then can you definetely say that the inbetween" zero crossings are exactly half way between the ones occurring at T=20ms?

[ataradov]

This question is answered by a simple google search "frequency stability of power grid".

[WattsThat]

Your title is wrong. The correct question is how stable is mains frequency.

Define that and you have your answer.

[coppice]

Hi,
Supposing you have 50Hz mains. Suppose you know when every other zero cross occurs...eg, every 20ms.....then can you definetely say that the inbetween" zero crossings are exactly half way between the ones occurring at T=20ms?

You may have more than 2 zero crossings per cycle. If there is serious high harmonic distortion you can get fast ripples on the sine wave big enough to cause multiple crossings. It takes pretty bad distortion to make that happen, as the main sine wave is changing at its fastest near the crossings, but it can happen.

[nightfire]

For reference, and maybe it is hopeful as a reference to others in the future that found this thread by usage of the search function:

The quality of the mains grid never is intended to be a precision instrument. For practical purposes the carrier(s) define in their standards the allowable deviation from some middle value.

In germany, we follow some european standards, sopme quick search (at this time, I am a bit lazy and don't look for maybe newer publications...) unearthed some paper from 2013 from german energy providers:

In the EN 50160, there are the following corridors for service quality:


  Merkmal                                                     Wertebereich
Netzfrequenz                                               49,5 Hz bis 50,5 Hz
Langsame Spannungsänderung                     230 V ± 10 %  (=Slow changing voltage)
Schnelle Spannungsänderung                        230 V ± 5 %    (rapidly changing voltage)
(selten bis 10 %)


For germany, here: https://www.netzfrequenzmessung.de/  is a live display of the current net frequency.

[ArdWar]

It varies a lot between grids. Check your national standards and power company policy.

Most will only mandate long term stability if at all, but rarely guarantee short term accuracy as frequency can be integral to the grid balancing scheme. Eg they might only guarantee that there will be exactly 4320000 cycle in a day but the frequency itself can drift from 49 to 51 Hz at any given moment.

[WimWalther]

It depends on your stability criteria - short-term, long-term, etc.

In most any first-world nation with a national power grid, the long-term stability is exceptionally good. This means, for example, if you have a clock that uses a synchronous motor (like the classic GE Telechron movement) or an IC time base synced to the line (digital clock / radio), you can expect it to hold within a second or two over the course of years.

Short-term, over several hours, you might see a couple seconds error - but these errors are nulled out on a regular basis by occasionally varying the frequency for exactly this reason.

As for cycle-to-cycle, I'm unsure, but I'd imagine it's good enough for GPS to be accurate within a couple of meters.. so plenty good enough for almost any use you're likely to need it for.

[radiolistener]

can you definetely say that the inbetween" zero crossings are exactly half way between the ones occurring at T=20ms?

No, you can't say this. it may vary depends on connected load and mains frequency stability. ±0.02 Hz frequency variations happens very often, sometimes it can be more, so you can estimate that 10 us variations is pretty common for mains even with no effect of high power reactive loads.

[Synthtech]

Domestic solar feed-in systems might perhaps put noise and phase issues into local mains power supplies?

[DavidAlfa]

The correct question is how stable is mains frequency.

I missed checking the author. Then everything made sense!

*sigh*
I'm assuming you're switching some devices, but you should specify the application of this.
One thing is to drive a motor dimmer, a very different one is injecting 100KW into the mains from a huge battery bank, where a small drift might case hundreds of amps flowing in the wrong direction.
Zero cross will be pretty accurate between few cycles,  the mains voltage will be very low within 300us off zero cross (30V @ 230VAC), so switching the device slightly earlier/later won't cause a huge current spike as when hard-switching at main's Vmax.
As long as you sync the zero cross every few cycles, it'll be fine. 10us variation will cause a negligible change in voltage.
Main's won't suddenly drift by 10mS! Might be 50.2Hz-49.8Hz along the day (I don't remember the limits, but recall them being pretty tight), and the power station will compensate any small drift so at the end of the day the average frequency was 50.00Hz.

[Terry Bites]

www.gridwatch.templar.co.uk
But being a bloody bot you should have already found this?

[Siwastaja]

This question is answered by a simple google search "frequency stability of power grid".

Yes, and another relevant term to look for: jitter.

[tggzzz]

Hi,
Supposing you have 50Hz mains. Suppose you know when every other zero cross occurs...eg, every 20ms.....then can you definetely say that the inbetween" zero crossings are exactly half way between the ones occurring at T=20ms?

Firstly, don't confuse voltage (As per thread title "How stable is the mains voltage?" ) with frequency and time (As per the question body). That's a surprising mistake for someone that designs SMPSs do a living.

Secondly, learn to use a search engine such as google.

Finally, the answer will probably depend on whatever other "nearby" loads exist, especially if they are inductive or switched.

Hence any design that relies on a specific answer to your question will probably be fragile and prone to "entertaining" behaviour in an installed system. Your alter-ego, treez, has a lot of issues with his designs, didn't he?

[Jwillis]

http://mainsfrequency.uk/fm-home

"The GB mains frequency is nominally 50Hz. National Grid is obliged by its licence commitments to control the frequency within ±1% of 50Hz so it can fluctuate between 49.5Hz to 50.5Hz. However......"

[SeanB]

Well, here is the current snapshot of the last 24 hours by me. The 2 hour dropout at midnight was Eskom, welcoming SA being World Champions, and back to rotational load shedding again.They ran out of diesel for the full time running peaking plants, and pumped storage is dry.

[alligatorblues]

When commercial electricity users turn huge loads on or off, or anything that uses power factor correction, and a slew of other loads on utility power, are responsible for 90% of the harmonic disruption in the power grid. Near where I grew up there was a building at which the major transmission lines went inside, and came out the other side.

I found out years later from my uncle, who was a lone lineman in northern Michigan for 30 years, and then became a foreman for a team building transmission lines, that that building is a switch to reverse the flow of power between 2 large power plants.

Running that switch on 96,000 V power lines has got to have a noticeable effect on the geometry of the power. But, America developed AC  power transmission, and that was long ago. So, engineers have learned a few things to balance everything out. The frequency is the easiest to keep stable, because it's directly tied to the rotational speed of the generators. 3,600 rpm for fossil fuel plants, and 1,800 for nuclear plants.

It is not difficult to control that precisely. Generators will lead in stabilizing frequency according to how much power they put on the grid.

[Andy Chee]

This guy analysed the stability of his local 120V 60Hz mains:

http://www.leapsecond.com/pages/mains/

The same principle can easily be applied to analyse your local grid.

[ebastler]

Hi,
Supposing you have 50Hz mains. Suppose you know when every other zero cross occurs...eg, every 20ms.....then can you definetely say that the inbetween" zero crossings are exactly half way between the ones occurring at T=20ms?

It seems to me that every reply so far has still missed the point of the question. Which is not so much about frequency stability (be it short or long-term), but about symmetry of the AC waveform. Even for a perfectly stable 50 Hz signal, the alternate zero crossings do not have to be 10 ms apart if the waveform is asymmetrical.

I have no idea whether anything is specified regarding symmetry in the mains network. I think it is not a good question to ask anyway, since the timing of detected positive vs. negative zero crossings will depend on implementation details of your detector.

[Siwastaja]

It seems to me that every reply so far has still missed the point of the question. Which is not so much about frequency stability (be it short or long-term), but about symmetry of the AC waveform. Even for a perfectly stable 50 Hz signal, the alternate zero crossings do not have to be 10 ms apart if the waveform is asymmetrical.

True. I was closest by mentioning jitter but fair enough, symmetry is even more specific, it can be thought as a special case of jitter.

And while anyone can easily google about the frequency stability of grid, that is something measured in time scale of minutes (if you are lucky, maybe seconds), that was not the question.

This treez guy is an interesting psychological experiment; the questions are actually pretty interesting and difficult, yet people jump, blinded by rage, into belittling the question without actually even understanding what was asked.

[radiolistener]

It seems to me that every reply so far has still missed the point of the question. Which is not so much about frequency stability (be it short or long-term), but about symmetry of the AC waveform.

No, I was talking about both - duty cycle and period. Duty cycle is affected by reactive load. And frequency stability also affects stability of duty cycle.

Reactive load distorts sine wave and affects its symmetry. When frequency is changed it also affects duty cycle at least for one sine cycle. All this leads to unstable duty cycle and as result leads to asymmetric sine.

Also when someone plugs some powerful load to the mains it leads to sparks and pulses in the mains. These pulses also affects stability of zero crossing period...

[antenna]

I saw a forensics video once where mains frequency tracking is used to identify the time and date a particular video is created.  Using the AC hum that can sometimes be detected in the audio track, they can tell exactly when the video was created.  The frequency is always varying and that forms a nice fingerprint-like timestamp that is totally unique.

[TimFox]

Six years ago, I posted this about a surprise I encountered in a design:
https://www.eevblog.com/forum/projects/ac-line-fluctuations/msg1265692/#msg1265692
Short explanation:  I naïvely thought that, for a vacuum-tube audio project, I only needed to ensure a low value of 60 and 120 Hz ripple on the DC plate voltage, relying on filter capacitance but no regulation.
For the actual audio signals, that was probably right, but when doing careful measurements I found them disrupted by lower frequency (below, say, 10 Hz) fluctuations on the DC level.
I found that my AC outlet voltage was varying by more than I expected over that short a time interval.
See older post for details:  I had to add a reasonable voltage regulator circuit to get useful results.

[langwadt]

I saw a forensics video once where mains frequency tracking is used to identify the time and date a particular video is created.  Using the AC hum that can sometimes be detected in the audio track, they can tell exactly when the video was created.  The frequency is always varying and that forms a nice fingerprint-like timestamp that is totally unique.

also to see if the recording was modified, like sudden jumps in phase where pieces was cut

[pdenisowski]

Well, from a test and measurement point of view, the AC mains voltage provided by a "standard" wall outlet is often (at least in my experience) not stable enough for some types of testing.

For example, if you want to run standby power tests as per EIC 62301, the voltage and frequency supplying the device under test must be within 1% of the nominal values and there is also a limit on the amount of harmonic distortion -- I think THD must be < 2% for the first 13 harmonics. 

It's not uncommon that running these kinds of measurements in the United States requires the use a "lab grade" AC power source in order to meet these requirements: you can't just plug the power analyzer into the wall.  The usual culprit is ... surprise! ... THD, probably as a result of all the non-linear, non-PFC'ed devices being plugged into the grid these days.

(Coincidentally, I just finished a video on this very topic -- standby power measurements -- which should go live on YouTube in the next couple of weeks)

[tridac]

Here in the UK, the frequency does vary a little, but has long term correction. Restored a couple of the old 1930's (Temco) synchronous motor clocks, one running in the lab for months, that is still within a few seconds of the radio clock next to it.  ~Ñinety years between them. Short term variations, but corrected in the long term