[time-nuts] Another "atomic" clock question
Tom Van Baak
tvb at LeapSecond.com
Sat Mar 1 21:31:13 EST 2014
Everything about time & frequency is simply a matter of degree, of decimals points. If all you require is 1 second accuracy then any old WWVB RC clock will work.
If you want 0.1 second, or 10 ms, or 1 ms accuracy a PC running NTP should work.
As you push closer to the microsecond level you need a correspondingly better internal stable frequency source (e.g., rubidium) or external accurate time source (e.g., GPS). Most of us use GPS one way or another, achieving 100 ns accuracy with no effort and 10 ns with extreme effort.
Listening to WWV makes a nice example. Where I am near Seattle, say 1000 miles from NIST, the radio wave delay from Ft Collins (due to speed of light, 1 ns/foot, or 5 us/mile) is about 5 ms. The delay from the WWV radio speaker to my ear (due to the speed of sound, 1 ms/foot, or 5 s/mile) is about 5 ms.
----- Original Message -----
From: "Bob Albert" <bob91343 at yahoo.com>
To: "Discussion of precise time and frequency measurement" <time-nuts at febo.com>
Sent: Saturday, March 01, 2014 6:04 PM
Subject: Re: [time-nuts] Another "atomic" clock question
All this is very interesting. However, my interest is frequency. In other words, I want to know that my standard oscillators are as close to desired frequency as possible, and how close that turns out to be.
Yes, the Internet gives me time of day as close as I care to know. I have an 'atomic' clock from LaCrosse that resets itself nightly, although it's fussy about where in the house I put it. If I put it where I'd like, it won't receive WWVB, so I put it across the room. I called the company inquiring about augmenting the internal antenna but they were of no help.
While watching the clock and listening to WWV, it seems the clock is a fraction of a second behind. Even that doesn't matter, but calibrating the counter time base is another kind of thing.
I am trying to understand how this is done. Should I ever get a rubidium standard, I'd want to check its calibration, and that's not a trivial exercise.
On Saturday, March 1, 2014 4:56 PM, Paul Alfille <paul.alfille at gmail.com> wrote:
There are WWVB clocks with serial output. Arcron made one that I added
linux ntp support for some years back.
As I recall, it was under $100, quite nicely styled, and is sitting here on
my desk. (Reception on the East Coast can be spotty, so I've switched to
standard internet net time source).
On Mon, Feb 24, 2014 at 7:44 AM, Bob Camp <lists at rtty.us> wrote:
> Ok, so 0.1 second at the sync point is indeed a reasonable estimate. If
> that's all you need to deal with (you correct out the crystal offset one
> way or the other) then:
> At 1 day you have 11.5 ppm accuracy. Roughly a 100 Hz beat note with WWV
> at 10 MHz.
> At 10 days you have 1.15 ppm. Roughly a 1 Hz beat note at 10 MHz.
> At 100 days you have 0.115 ppm. That would be about a 10 second period
> beat note.
> None of that is to say that a beat note is all there is to getting
> accuracy off of WWV or that the two approaches deliver the same net
> accuracy. Yes I've done the 10 second beat thing, it can be done with care
> and a good stable WWV signal.
> On Feb 23, 2014, at 5:21 PM, Tom Van Baak <tvb at LeapSecond.com> wrote:
> >> Now that you have brought up this subject, do you know of any way to
> use these LaCrosse clocks to calibrate frequency standards?
> > I suggest using a direct electric (1.5 VDC high-Z) or indirect magnetic
> (high gain) pickup on the coil to get the +/- pulse per second. Compare
> this time with your local frequency standard and over several days you
> should get accuracy better than 10 ms per day (1e-7). Here's an example of
> a raw phase plot:
> > http://leapsecond.com/pages/Junghans/
> > /tvb
More information about the time-nuts