Why your mains powered clock is running 6 minutes slow

[qubit]

It's all to do with the mains frequency being 49.996Hz and not 50Hz. Get your nerd on and check it out.

https://uk.news.yahoo.com/electrica...low-thanks-missing-electricity-141147945.html

EDIT: please stay away from the political stuff. See @Norton's post below.

ANOTHER EDIT: here's a better link to this story: http://www.theregister.co.uk/2018/03/09/europeans_losing_minutes_smm_fingered

 

[CAPSLOCKSTUCK]

6 minutes over what period of time? the article doesnt make it clear.

If its 6 minutes over a week it would be annoying but if its a million years who cares apart from yahoo

 

[basco]

quote\ A ‘dip’ in electrical power in the region has had an impact on the national power grids of 25 countries in the Continental European Power System \ quote

wow that show´s how fragil this system is-awaiting the next big sunbursts

 

[PLAfiller]

Yeah, I heard about this one on the news. The reason is Kosovo/ Serbia. The disbalance spreads from there to the whole network. The problem is political, not so much technical. Very interesting thing though, I didn't know about the electrical networks being connected. The frequency drop by itself is insignificant, but the period the network has stayed this way - that's what is worrying. It's the first time this has ever happened.

@basco what it has to do with sunbursts? It's about Kosovo using more energy than it can produce and Serbia not wanting to "lend" some to balance the area.

 

[basco]

serbia is the only one who can lend some energie to kosovo?

quote from you:
Very interesting thing though, I didn't know about the electrical networks being connected.

so if sunbursts(sorry i meant solar flares) go into one it fries down all of the network?

 

[Norton]

meh...

Not too many clocks are powered by an AC synchronous electric motor these days (chart recorders are though).

Ready to lock this thread up if it devolves into a geopolitical discussion

 

[Ferrum Master]

they all have compensation circuit for that... even the soviet clocks did have that.

EDIT.

It seems I've missed special cut down versions, i've better use a pole and and sun than those without quartz.

 

[basco]

sorry but this seems like precrime from minority report

 

[Aquinus]

Aren't most RTCs driven by crystal oscillators these days? That would be independent of mains frequency. The article sounds more political than scientific.

 

[qubit]

Ready to lock this thread up if it devolves into a geopolitical discussion

Everyone: please keep it to the technical nerdy stuff, peeps. I don't want mod action in this thread.

I've got just such a mains locked alarm clock. Got it in the late 90s with a beautiful big red 2" LED display. You have to set it manually of course, but after that it just doesn't drift. Ever. Like a radio controlled clock. It's because the mains frequency itself is syncronized to the atomic clock. Kinda weird and awesome at the same time.

 

[newtekie1]

Not too many clocks are powered by an AC synchronous electric motor these days (chart recorders are though).

Yeah, pretty much what I thought too. I know around me the only places I see AC driven clocks are government buildings and schools. Pretty much everything residential has gone digital with crystals.

But it could be different in other countries.

 

[FordGT90Concept]

It really doesn't matter though because all internet-connected devices sync to time servers at least once a week. Just have to adjust all of the ones that aren't.

 

[RejZoR]

Mains powered clocks are the ones on microwaves, ovens, stand alone stoves and mains powered alarm clocks with radio (with glowing numbers). All these devices use mains frequency as timing oscilator and not a quartz crystal oscilator.

@FordGT90Concept
Windows is connected and has capability to use internet time, but it's a miserable failure. Just open atomic clock on time.is and be amazed how off your time is... I have to use alternate time service to sync all my computers, otherwise they drift off by up to 6-7 minutes because microsoft's NTP service is garbage and hasn't worked right for literally decades now.

 

[FordGT90Concept]

2 whole seconds. Not worried about it.

 

[newtekie1]

Windows is connected and has capability to use internet time, but it's a miserable failure. Just open atomic clock on time.is and be amazed how off your time is... I have to use alternate time service to sync all my computers, otherwise they drift off by up to 6-7 minutes because microsoft's NTP service is garbage and hasn't worked right for literally decades now.

The ability itself isn't a failure, the problem is the Microsoft ran server that it tries to sync to is constantly overloaded. That's why they built the ability to use the alternate government one right in. Switching to sync to the government time server is one of the first things I do.

 

[RejZoR]

The ability out of the box is non existent. I mean, if you offer the feature, buy some damn servers for it. It has been overloaded for what, 15, 20 years? It's stupid.

 

[dorsetknob]

Windows is connected and has capability to use internet time, but it's a miserable failure. Just open atomic clock on time.is and be amazed how off your time is... I have to use alternate time service to sync all my computers, otherwise they drift off by up to 6-7 minutes because microsoft's NTP service is garbage and hasn't worked right for literally decades now.

One of the "Services " i always disable

 

[FordGT90Concept]

I changed the time server to my literal server and now it's <0.1 second off. Because my server is a domain controller/time server, it naturally syncs to internet time more frequently. On top of that, syncing to it will never fail because it's never that busy.

 

[95Viper]

Latest info:

Deviations affecting frequency in Continental Europe have ceased; ENTSO-E working on step 2

ENTSO-E has now confirmed with the Serbian and Kosovar TSOs, respectively EMS and KOSTT, that the deviations which affected the average frequency in the synchronous area of Continental Europe have ceased.
This is a first step in the resolution of the issue. The second step is now to develop a plan for returning the missing energy to the system and putting the situation back to normal. Register to ENTSO-E news alerts for updates.
ENTSO-E, in close interaction with the European Commission, is trying to identify a sustainable long term solution that will avoid that this happens again.
The situation experienced is unprecedented in the Continental European Power System. The European transmission system operators interact constantly, across the borders and through ENTSO-E, to ensure that security of supply is maintained in one of the world's largest synchronous area.

Some more interesting data:

National grid UK (check out real-time data links on this page)

System Data


Demand: 33735 MW
13:42:00
Frequency: 50.05 Hz
13:42:00
System Transfers

Northern Ireland to Great Britain: 36 MW
Ireland to Great Britain: 0 MW
France to Great Britain: 1495 MW
Netherlands to Great Britain: 1004 MW
10/03/2018 13:42:00

North-South: 838 MW
Scot - Eng: 1746 MW
10/03/2018 13:42:00

Data last updated on : 10/03/2018 13:44:00

 

[qubit]

Yeah, pretty much what I thought too. I know around me the only places I see AC driven clocks are government buildings and schools. Pretty much everything residential has gone digital with crystals.

But it could be different in other countries.

Note that my alarm clock that I was talking about was built in the late 90s. If it was made today, it might use a crystal; I really have no idea.

btw, I remember that back in the day, around 1997 when I bought it, the time did drift on it, so clearly at some point after that, the mains frequency was synced to an atomic clock. I don't know why they run it quite so accurately, either. I wouldn't have thought a few millihertz difference would matter for devices. Perhaps it has something to do with syncing generators when they come online, but still it's not critical.

I changed the time server to my literal server and now it's <0.1 second off. Because my server is a domain controller/time server, it naturally syncs to internet time more frequently. On top of that, syncing to it will never fail because it's never that busy.

I remember when I was teaching myself Active Directory a few years ago, the AD server deliberately didn't sync over the internet to improve security. It had to be enabled in Group Policy somewhere, although I never did figure out where the setting was. I was using Server 2008, ie the server version of Vista. Maybe you're using a later version and the default behaviour has changed?

 

[FordGT90Concept]

Pretty sure enabling the Active Directory role also enables W32Time. I just checked and it was running syncing off of time.windows.com. While I was in there, I changed it to time.nist.gov:
https://support.microsoft.com/en-us...n-authoritative-time-server-in-windows-server

I'll have to check in a few days because NtpServer doesn't broadcast that it updated.

 

[newtekie1]

The ability out of the box is non existent. I mean, if you offer the feature, buy some damn servers for it. It has been overloaded for what, 15, 20 years? It's stupid.

The ability out of the box is built right in. It defaults to using Microsoft's time server(which haven't been overloaded and have worked fine for a good couple years now), or they already include the national time server built in as an option as well as an alternative if you want to switch to that. It isn't like the national time servers aren't overloaded at times too.

 

[FordGT90Concept]

time.nist.gov does a round-robin of requests so they recommend people use that. time.windows.com likely does too for that matter. That said, I tried using time.windows.com before switching to my server and it failed to sync. Windows still gets overloaded frequently, it seems.

The difference between an NtpServer (e.g. Windows Server) and NtpClient (e.g. Windows) is that NtpServer will keep trying until it gets it because it presumably has NtpClients connecting to it for domain time. NtpClient, on failure, often just waits until the next scheduled sync.


When I changed to time.nist.gov earlier today, it was 0.6 seconds off. Now it's 0.3 seconds off. time.windows.com was <0.1 seconds off. Beginning to think time.windows.com is more accurate than time.nist.gov.

 

[RejZoR]

They are using atomic clocks. They are all equally accurate. It's how much your computer goes off in between syncs...

 

[eidairaman1]

I dont care about an additional service running for this. If WUD doesnt throw a fit about the time or date I dont care