root@petro.UUCP (Petty System God) (12/28/90)
There have been several programs posted recently to allow one to have his system call the Naval Observatory clock and have the time set all without human intervention. Having decided that this is a good idea I coerced some code into working only to discover that there appears to be a discrepancy between the time reported by WWV and that reported by the dial up number at the Naval Observatory. The difference is about 12 seconds which I consider significant since both these guys are supposed to be time standards and should do better than my Timex. I noticed, after spending a couple of hours hacking code to call the Naval Observatory and set the clock, that the time reported by the Naval Observatory was not the time I expected. I was expecting something close to the drumming that WWV was producing in the corner but to my surprise the Naval Observatory was reporting time about 12 seconds behind WWV. To confirm the observation I called back the Naval Observatory with cu and let it report while WWV was reporting over the shortwave -- sure enough when WWV hit 124900 the Naval Observatory was reporting about 124848. OK, I give up, who's right ? ---- Gilbert B. Andreen - bruce@petro.UUCP cs.utexas.edu!swrinde!petro!bruce
bob@rel.mi.org (Bob Leffler) (12/29/90)
In article <63354@petro.UUCP> root@petro.UUCP (Petty System God) writes: >the Naval Observatory was reporting time about 12 seconds behind WWV. >To confirm the observation I called back the Naval Observatory with cu and >let it report while WWV was reporting over the shortwave -- sure enough when >WWV hit 124900 the Naval Observatory was reporting about 124848. My first guess for the difference is the delay intransmitting the data via the different transports. i.e. How long does it for the data to transmit through the numberous switches, land lines, satelites, etc to reach your modem? I'm sure that it is different than shortwave. bob -- Bob Leffler, (bob@rel.mi.org), (313) 696-2479 Opinions expressed Electronic Data Systems, GMC Truck SBU may not be those Box 7019, 5555 New King Street, Troy MI. 48007 of my employer.
wht@n4hgf.Mt-Park.GA.US (Warren Tucker) (12/29/90)
In article <63354@petro.UUCP> root@petro.UUCP (Petty System God) writes: > >... the Naval Observatory was reporting time about 12 seconds behind WWV. >OK, I give up, who's right ? Most assuredly it is WWV. Certainly propagation through the phone network is no 12 seconds, so the Naval Observatory time must be just plain wrong. The WWV time over radio is absolutely correct (some will disagree, but really, what's a few femtoseconds among time freaks?). The propagation delay is a factor, but my clock ticks are 60HZ :-). ---------------------------------------------------------------------------- Warren Tucker emory!n4hgf!wht or wht@n4hgf.Mt-Park.GA.US "I was 35 years old before I knew a pie was meant to be eaten." - Moe Howard
barry@tanstaafl.edgar.Mn.Org (barry) (01/01/91)
root@petro.UUCP (Petty System God) writes: > [ confusion as to who is correct deleted... ] > > ...the Naval Observatory was reporting time about 12 seconds behind WWV. > To confirm the observation I called back the Naval Observatory with cu and > let it report while WWV was reporting over the shortwave -- sure enough when > WWV hit 124900 the Naval Observatory was reporting about 124848. > > OK, I give up, who's right ? > > ---- > Gilbert B. Andreen - bruce@petro.UUCP They both are. By Act of Congress the following relationships are true F != 1 / T and T != 1/F. In its great wisdom Congress has defined as part of each organization's mission the following. That the Bureau of Standards be responsible for keeping the frequency standard, and the Naval Observatory is responsible for keeping the correct time. Of course the scientists at both organizations understand the physics and mathematics of the situation. Something that Congress does not (what does reality have to do with Congress anyway, have you ever watched C-SPAN? [0.5 :-) ] ) They achieve their mission the following way. The NBS maintains radio stations WWV WWVH and WWVB. These time broadcasts are synchonized to the Master Atomic Clock in Boulder CO. An Atomic Clock BTW is a container of Cessium Gases which has the proportions to maintain resonance. This is a very accurate frequency source for long intervals, but things like air pressure variations (such as people talking in the room) make minute changes in the cavity dimensions which will affect frequency stability. This gives a very poor short term frequency standard. The output of the Cessium Cavity is phase-locked to a 10 Mhz crystal oscilator which has good short term accuracy, but poor long term accuracy. This 10 Mhz oscilator then drives a rather conventional clock that is broadcast on WWV/WWVH/WWVB. Each of these clocks are synchronized, with propagation delay accounted for for both the voice and data portions. WWVB sends out time in BCD on 60Khz (via power drops of 10dB) WWV and WWVH send out not only the time in voice (Male for WWV Female for WWVH) but also in BCD via a 100Hz sub-carrier. In addition to the time the current UT1 and UT2 in milliseconds are broadcast. Once a year, (I believe in the spring) at the Naval Observatory a special telescope with a pinpoint aperature is uncovered. When a particular star lines up exactly with this telescope a signal is sent to the NBS over a dedicated circuit. This circuit has a known path length, and an exact propigation delay. The signal marks midnight of the siderial year. The propigation delay is added to the signal and the Master Time Clock is sync'ed to X miliseconds after midnight (where X is the propigation delay) The current values of UT1 and UT2 are updated, and WWV is now set by the Naval Observatory. Note that the Master Atomic Clock is not a clock but a very accurate (1 part in 10 ** -12) "crystal" oscilator. The Master Time Clock is a digitally sync'ed tape deck (or so I've been told). The Naval Observatory in Wahington DC has an Atomic Clock sync'ed to the NBS Master Atomic Clock in Boulder CO. So what happend to the 12 Seconds, well propigation delay on the phone line added some to that, then there was the time sharing system or whatever you access at the Naval Observatory, its processing time, and other things. Still 12 seconds seem to be excessive. The correct answer acording to Congress is the Naval Observatory, but the real answer is WWV is right. If you want to load the correct time into your system, you may want to do what I do. Heathkit makes a clock that recieves the WWV signals in BCD and decodes them. The clock has an RS-232 connector, and so I just read it when I want to set the time. I have no other connection with Heath other than a long time user. Note this clock is somewhat expensive for personal use, but it is neat to have around, when the power goes out and comes back, I have the correct time in 4 minutes. Also, the NBS has a bit that they send for Daylight Savings in effect, so that is also covered. The accuracy is +/- 250 miliseconds. You can save a little money by building it yourself. It will be a challenge to a 1st time kit builder, but if you go slow, and don't rush you'll do fine. Hope this helps. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Barry Berg UUCP: barry@tanstaafl.edgar.mn.org -- Too much of the world is run on the theory you don't need road manners if you are a five-ton truck. -- #include <std.disclaimer> -30- {sometimes older ;-) }