[time-nuts] Checking accuracy of Rubidium standards

[Bill Hawkins]

You didn't say anything about the precision you need or the
time you were willing to wait to get it. Nor did you mention
the equipment you have available, so I'll assume no fancy
counters or deviation plotters.

Get two clocks, and do what you have to do to run them on
10 MHz. You could even derive 60 Hz from your sources and
run a pair of synchronous motor clocks. Ah, you'll need to
run everything from uninterruptable power. After a year,
you'll be able to see 3 parts in 10E-9 if you can read one
second of difference. It would take 100 years to get down
to one part in 10E-11, where the people on this list start
getting interested.

So you really need to get something that you can read with
100 nanosecond precision, computer connections for two of
them, and a program that will process and record the data. 

Ask yourself why you want to do this. Is it a 'horsepower'
race to get the best number? If you want an accuracy of one
cycle in 10 MHz, you're already there, if the units are
within specs. A simple Lissajous display will show you one
part in 10E-8 in ten seconds, as has been mentioned.

Bill Hawkins


-----Original Message-----
From: Randy
Sent: Saturday, November 08, 2008 12:17 PM

I was wondering if it is worthwhile or even feasible to compare an LPRO
Rubidium standard against a Z3801.  Since their frequencies are probably
going to be extremely close anyway it would seem some special
method/equipment would be required for high precision.  Suggestions?

Randy, W7HR
Port Orchard, WA