The last NPL MSF Bulletin showing their errors was over 4 years ago: http://resource.npl.co.uk/time/bulletins/bulletin_archive.html http://www.npl.co.uk/science-technology/time-frequency/products-and-services/time/msf-radio-time-signal
The error at a receiver over an interval of a couple of days is nearly completely due to propagation effects. The variations in propagation delay far exceed the error in transmitter frequency/phase control unless there are large changes in the local weather at the transmitter site. The phase of LF signals at 60 kHz is affected by the effective height of the ionosphere, which forms a 60 kHz waveguide with the Earth's surface. Unless you are very close to the transmitter, the amplitude and phase of the received signal change significantly through each 24 hour period. WWVB clocks should always measure the signal during the dark path (when the propagation path between Ft Collins CO and your location is fully dark), since that is when the signal is strongest and tends to have a more stable phase. WWVB clocks can get UTC with an uncertainty of about 100 microseconds, according to NIST: http://www.nist.gov/pml/div688/grp40/wwvb.cfm -- Bill Byrom N5BB On Sun, Nov 22, 2015, at 04:16 PM, Alan Melia wrote: > Hi Bob I have just realised that MSF may work diffently?? The Anthorn > signal is monitored by NPL at Teddington, West of London and frequency > off-sets twice a day are published in parts in 10^12 on their > web-site.....involving lot of averaging I think. They do not recommend > using the signal after dark. > > You certainly could predict roughly the the skywave phase change during > the > day and the variation as the sun angle changes on the path over the year. > It > gets more difficult if there are flares or geomagnetic storms. If two > independent stations at slightly different distances collect information > it > could be corrected even more accurately. > > The same ionospheric problem occurs of course with Loran, and the now > closed > Decca system.. > > Alan > G3NYK > > > > ----- Original Message ----- > From: "Bob Camp" <[email protected]> > To: "Nick Sayer" <[email protected]>; "Discussion of precise time and > frequency > measurement" <[email protected]> > Sent: Sunday, November 22, 2015 7:34 PM > Subject: Re: [time-nuts] KD2BD WWVB receiver/decoder in QEX > > >> Hi >> >> The most basic gotcha with WWVB is that propagation can (and does) shift >> the >> carrier more than a full cycle over the course of a day. That’s at 60 KHz, >> so one >> cycle is a lot (as in 16.666 ppm). Even at one second with a not so great >> receiver >> and a poor antenna, GPS should give you ~0.01 ppm. Right up front, you >> have a bit of a problem. >> (Yes I’m mixing measurements in that comparison, but the point is still >> valid). >> >> What I keep wondering is - There is no big mystery about the WWVB >> transmitter's >> location. You likely know your own location as well. Part of demodulating >> the data >> gives you day of the year. From that you can figure out some of the basics >> of the >> propagation effects (sunrise is at X:XX sunset is at … etc). You also >> could grab stuff >> like weather data fairly easily (no idea if that actually helps). If you >> fit out the basic >> propagation impacts, WWVB could get a lot better. At the very least, you >> would know >> when to ignore the signal. >> >> So yes, you could do better today than they could back in the good old >> days in terms of >> the propagation coarse effects. >> >> Unfortunately, there also is data on 24 hour comparisons of WWVB carrier >> (same time of >> day, one day apart). If you pick your time right (noon or midnight), the >> variable propagation >> can be reduced quite a bit. Based on that data, you are doing well at 100 >> ppt over >> 24 hours. Might the new modulation help that by 10X? ..maybe. GPS over a >> 24 hour >> period should be giving you something in the 0.1 to 0.01 ppt range (same >> sort of pick a likely >> stable ionosphere time slot and compare). >> >> Does that make a WWVB device un-interesting? Not by any means. If you >> stretch out the >> time, both systems get down into the “I have nothing else that good” >> range. Checking one >> against the other is indeed an interesting thing to do. You just need a >> **lot** of time to do it. >> >> Bob >> >> >>> On Nov 22, 2015, at 12:41 PM, Nick Sayer via time-nuts >>> <[email protected]> wrote: >>> >>> >>>> On Nov 22, 2015, at 7:47 AM, paul swed <[email protected]> wrote: >>>> >>>> As mentioned a nice answer to the wwvb modulation change. >>>> I looked up the parts and it seems that they have gone into the NOS >>>> state. >>>> Though you can get some from digikey and such especially in the SOIC >>>> package. Also the VCO isn't available. >>>> It appears that the Chinese sight has the lmc6484 and LM387n at >>>> reasonable >>>> prices for small quantities. Most likely will order from there. >>>> Have not checked out the PIC chip yet. >>>> The 74HCXX are common and reasonable. >>> >>> That’s kind of a shame. I’m sure a redesign with modern SMD parts could >>> be accomplished. >>> >>> The big question is how the stability of a wwvb disciplined oscillator >>> would compare to a GPS disciplined one (all other things being equal). >>> Well, it’s a big question for me, since I have no idea, but I imagine >>> simply asking here will give an immediate answer. :) >>> >>> I’d have to guess that the PLL would behave better given a 60 kHz >>> reference rather than a 1 Hz one. But how stable is that 60 kHz reference >>> after going through, what, a thousand miles of ionosphere or so? >>> _________________________________________________ >>> time-nuts mailing list -- [email protected] >>> To unsubscribe, go to >>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >>> and follow the instructions there. >> >> _________________________________________________ >> time-nuts mailing list -- [email protected] >> To unsubscribe, go to >> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts >> and follow the instructions there. > > _________________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
