[time-nuts] Cheap stable frequency sources
sandeenpa at yahoo.com
Thu Sep 23 03:22:12 UTC 2010
Using 4 surplus Lucent rubidiums compared to a Lucnet GPS receiver I find a consistent error of between 110 and 860 pico-seconds using a HP 5370B after running them for about two weeks. By measuring the weekly drift one could calculate the drift. I measure the offset at about 20 milli-Hz.
To phase lock it use the NIST 60KHz by diving it down to 10KHz and using the comparator circuit used by James Miller G3RUH in his Rockwell GPS circuit for 10MHz crystal oscillators.
You’d probably want to inhibit the circuit during the two periods between dawn and dusk to avoid what I believe is called the diurineal (sp) shift. In that small amount of time the Rb oscillator shouldn’t drift out of spec.
There was a circuit published in 73 magazine called “the worlds most accurate clock” and a PC board kit for it is available from FAR circuits. There was also a similar one published in Ham Radio magazine in the 70’s IIRC.
10 foot diameter loop antennas made out 3/4 inch PVC pipe and orientated towards Boulder, CO would give reliable reception and are inexpensive to make.
Total cost is less than $800. The trick is borrowing a TIC for set up.
Hope this might help.
More information about the time-nuts