"Atomic Clocks"

I have an old Motorola counter with an LED display. lots of TTL in it and a crystal oven. I can put a short wire in its 1 Mhz output and beat it against a 10Mhz or 20MHz WWV signal in a communucations receiver, and adjust it to with maybe 1/10th Hz or so, and it stays there for weeks at a time or for about as long as I keep it powered up. That's about 1 part in 10^8. This seems like a bit of overkill for this counter - it will presumably go to 50 MHz, but it appears its major use was to measure tones for audio communications such as beepers, as it has a function to measure up to about 5kHz to a resolution of 0.001 Hz. About the main application I can think of for this much accuracy is to run the output through a bunch of counters to make a clock that's accurate to three seconds per year.

-----

What frequency and what is your definition of "reasonable standard"?

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

with a

10 dB

in 1

automatically

at all

zero-crossings

When I worked at the CBS affiliate in Madison WI, WWVB WAS used for the frequency measurements. The transmitter 1 MHz reference was divided down to 10 KHz. The WWVB carrier was also divided down to 10 KHz. The 2

10K signals drove a paper plotter displaying the phase difference, usually for 2 hours. The reference provided the timing gate for the frequency counter. In the 8 years I was there, the carrier was never more than 0.1 Hz off for a carrier od 61.250 MHz (channel 3). Precision frequency control reduces interference between stations on the same channel. IIRC the offset between stations was 10010 Hz. WISC-TV was zero offset. GG

setting.

be

Traceable"

reasonable

Most of the "Atomic Clocks" use a tiny crystal-filtered TRF receiver with a ferrite loopstick antenna tuned to 60 KHz. The signal it receives is 10 dB amplitude modulated with a 1 bit/sec code sequence that is completed in 1 minute, and works into the clock firmware to set the clock automatically after a few minutes of solid copy. I don't think the 60 KHz signal is at all useful for frequency meter calibration - - too odd, and the zero-crossings are too slow.

Chuck W6PKP

How about using the TV line frequency, I don't know about the states but over here it is actually derived from a tracable standard.

BBC R4 on 198Khz (in the UK) is also derived from a decent clock source and is pretty good.

In some countries the 'phone dialtone is derived from a rubidium source, might be worth looking into? There might be something equally usable in your area?

Just some thoughts....

Regards, Dan.

On Tue, 28 Dec 2004 21:21:38 GMT, "Chuck Olson" wroth:

The 60 KHz may be odd and the zero crossings may be slow, but averaged over time, the 60 KHz is exquisitely accurate. A 15,360 MHz oscillator can be easily divided down to 60 KHz with just flip-flops. Or a 10.000 Mhz oscillator cam be divided by 500 and multiplied by 3 to get 60 KHz. Then you can compare your local oscillator to the 60 KHz NIST signal to get feedback to correct the high frequency oscillator. Phase lock the local oscillator to NIST and you'll get a very useful frequency standard.

Jim