Just 1 count: either you see -1, 0 or 0, +1. From -1 to +1 there are 2 counts and, unless the clock is affected by a really bad jitter, it is hard to see more than 1 count of wander in a setup like you did. Anyway, you need a trusted reference, feeding the counter with a 10MHz and counting the same 10MHz divided by 2, doesn't tell you anything.
On Mon, Jul 15, 2013 at 5:09 PM, Bob Stewart <[email protected]> wrote: > In this case, I was using the 10MHz output from the GPSDO as the reference > and measuring the 5MHz output from the GPSDO. So, even if the GPSDO is off > in frequency, the relationship between the 10MHz reference and the 5MHz > signal are mathematically fixed. It just seemed strange to me that under > such a circumstance the final count would wander only in the minus direction. > I expected either to see a solid 5.0000000, or to see it wandering between > 4.9999999 and 5.0000001. > > Bob > > > > > >>________________________________ >> From: Azelio Boriani <[email protected]> >>To: Discussion of precise time and frequency measurement <[email protected]> >>Sent: Monday, July 15, 2013 4:47 AM >>Subject: Re: [time-nuts] GPSDO - Does it work? >> >> >>It is normal for a counter to have a +/- 1 count of wander. Then you >>have to consider the reference for that counter. When you connect a >>GPSDO on the input of a counter with its internal reference used as a >>timing source, you are measuring that internal reference accuracy and >>stability, not the GPSDO's one. To verify a new GPSDO it is necessary >>to have another (know good) GPSDO or a Cs reference. Missing that, you >>can use the PPS output of the GPS receiver but then an average process >>is needed: say, a digital 'scope with the display persistence function >>and... wait a lot of time (or use a high resolution TI counter and the >>TimeLab software). >> >> >> > _______________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
