OT: Radio-controlled clocks

[Hugh Newbury]

Like everyone else, I have some RC clocks and a wrist watch which all keep good time. I have aligned my bedroom clock with the window through which presumably the signal will come. I put my watch about 6 inches from the clock, but I've noticed that the signal it picks up is often about 0.5 sec in front of the clock. Lately I've experimented with different positions for the watch and found that when it is aligned exactly in line with the clock they both have exactly the same times to some minute fraction of a sec.

I don't understand the physics of this. The signal can't be different for different distances from the window, can it? Or is something else going on?

Hugh

[Harry Bloomfield]

Hugh Newbury used his keyboard to write :

It needs to be aligned on the transmitter, the window makes little or no difference to the signal.

There is a minute difference between the signal arriving at one, versus the other if they are different distances from the transmitter, but it really will be minute. The signal travels at the speed of light,so the difference is several orders of magnitude smaller than you could possibly detect.

The difference in time sync will be down to processing delays within the two devices.

They use laser light bounced back off a reflector on the moon, to measure the distance between the earth and the moon, which allows them to measure down to around a meter or so.

[Harry Bloomfield]

Harry Bloomfield submitted this idea :

A good processing strategy, in a decent system will allow for the delays in processing the sync speed.

Your computers clock is synched over the Internet. Part of the standard synching process is to measure the delays to and from the time server and make allowance for the delays over the Internet.

[Martin Brown]

Putting two receivers so close together may result in one blinding the other from seeing the MSF signal correctly. The MSF time signal clocks should be synched locally to within roughly separation/speed of light.

Well it is delayed by about 3ns for every 1m further from the transmitter you place the receiver.

Way back in the 1980's when Duffet-Smith used MSF Rugby to synch remote aerials for low frequency long baseline interferometry they phase locked local rubidium clocks to Rugby at each antenna and discovered that the MSF Rubgy transmissions lagged reality when there was a heavy dew on the ground. This is a later paper but the first free one I found that more or less describes their technique.

[Bob Henson]

The stone wall in which my windows are set does though - hence mine need to be aligned with the transmitter and/via the window. It's worth people bearing that in mind if they have trouble.

[whisky-dave]

I think it's more accurate than that as they have measured that the moon earth distance is inceasing by about 3.8cm a year. So they must be measuring it to within a few millimeters

[Hugh Newbury]

It looks as if the distances/alignments/etc are microscopic, so it is not easy to see why the differences should be so relatively large.

But thanks all for your contributions.

Hugh

[John Williamson]

I think someone mentioned the possibility of interference between the two receivers, which may be deceiving one of them into showing the wrong time. Do you get the same time difference if the receivers occupy the same spot at different times? They should freewheel accurately enough to make the experiment feasible.

[Dave Liquorice]

within

Yeah but is a £6.99 MSF wall clock going to bother. It also takes a whole minute to send all the information. The carrier transitions are accurately timed though.

But that has a two way dialog even if it's only a response to a ping. IIRC NTP also assumes that the total round trip time is divided equaly between send delay and return delay, it might not be. Just because packets take one way to a host it doesn't mean packets from that host take the same route back. Or may there is more taffic in one direction than the other...

[Johny B Good]

Actually, the accuracy is sub centimetre (they're aiming towards mm accuracy with technology improvements) and have determined that the moon is spiralling away from us at a rate of 3.8cm per year.

[dr.s.lartius]

It is possible, but perhaps unlikely, that two signals are available at you r location, perhaps MSF from Anthorn and the German equivalent, and that th e local radiation from one of your clocks interferes with your other clock in a separation-dependent manner that causes it to choose a different stati on. The transmissions leaving the stations should, however, be in exact ti me agreement, and each subject to about 1ns/ft delay for their journey. Bu t transit delays cannot be the explanation, since the radio signals would n eed to travel a difference of 3.5 times round the Earth to give an 0.5 sec discrepancy.

P.S. Duffett-

[Brian Gaff]

Do you know which time transmission they are on, as I believe Germany has one as do the UK. Maybe in certain alignments the internal circuits interfere with each other. Or it could just be the polling for the signal is handled slightly differently. Brian

[polygonum]

I have found my MSF alarm appears slightly inaccurate. And have now found out why.

On its own, it syncs regularly without any problems. But if my iPad is nearby, it fails to sync. Because it actually keeps quite good time, pretty much the only occasions on which this is any sort of an issue are hour-change weekends and new battery times. In between, I too can see a difference of a few seconds between this one and another that syncs all the time if I bother to look.

[Dave Liquorice]

They probaby are but MSF is on 60 kHz and DCF on 77 kHz.

I think most "radio clocks" are built for MSF or DCF not switchable either automatically or (user) manually. One interfering with the other and corupting the data stream is a possibilty.

[Colin Stamp]

Are you sure they're both syncing every time? If one isn't seeing a signal, then it'll fall back to free-running on its internal crystal and it could easily drift by half a second or more in a day.

Cheers,

Colin.

[Harry Bloomfield]

Nail, head!

[djc]

+1

My analogue display radio-controlled clock only syncs every 24 hours at around 01:15.

[Graham.]

Perhaps they see about 170 cm since 1969 and divide by 45

[Hugh Newbury]

Ditto.

Hugh

[Martin Brown]

No it is a lot more accurate than that. The outgoing laser pulse is about 2.5cm long (ie an inch in US "English" units) and they get a precision close to 1mm after only a few minutes operation. See:

The moon is in an elliptical orbit that varies considerably so that the apparent size of a full moon varies by more than 10% in diameter.

It is for this reason that sometimes you get annular solar eclipses.