Would this characteristic be similar across all 10811s or would there be as much unit to unit variation as there is for aging and Allan Deviation? Ed On 9/30/2012 11:03 AM, Richard (Rick) Karlquist wrote: > I recently modified an old 10811 to bring out the crystal > leads on miniature coax (instead of having them connect > to the oscillator circuit). This allowed me to measure > the crystal's inherent flicker noise of frequency. > The measurements indicate that the 10811 phase noise out > to at least 100 Hz is entirely due to the crystal. > An interesting aspect of flicker noise of frequency > is that Allan deviation is independent of tau. Thus, > just one number describes the crystal noise. > > Rick > > On 9/30/2012 4:44 AM, Bob Camp wrote: >> Hi >> >> Close in, it looks like it's pretty much the crystal and the loading >> of the bridge oscillator. >> >> Bob >> >> On Sep 29, 2012, at 11:46 PM, Richard (Rick) Karlquist >> <richard@karlquist.com> wrote: >> >>> The E1938A uses a crystal that is basically the same as >>> the 10811 crystal except that it is in a reduced height >>> package. However the phase noise is not as good as a >>> 10811 due to broadband noise in the automatic frequency >>> control circuit. By the time I discovered this, it >>> was too late to try to fix it. >>> >>> Rick Karlquist N6RK >>> E1938A designer

I only measured one 10811. However, John Vig's tutorial (available at IEEE UFFC) rather categorically states that piezoelectric resonators have flicker noise of frequency. What I measured was most closely related to phase noise, as opposed to Allan Deviation. Phase noise of 10811's is more consistent unit to unit than ADEV and certain more consistent than aging. BTW, ADEV at HP was measured against a special 10811 that was 500 Hz off frequency. I was never able to find out how they arrived at this "golden" unit. But it seems clear that it could not have been the best ever unit for ADEV, thus the real "golden" units that came down the pike were simply rated as ADEV too good to measure. I tried to get a project started where we would use a frequency synthesizer to do the offset. Then we could take the best units and compared them against each other. Then, as well accumulated test data, the cream would gradually rise to the top and we would have some true golden units. The problem was that there ADEV at those levels wasn't a "money spec". Rick On 9/30/2012 12:16 PM, Ed Palmer wrote: > Would this characteristic be similar across all 10811s or would there be > as much unit to unit variation as there is for aging and Allan Deviation? > > Ed > > On 9/30/2012 11:03 AM, Richard (Rick) Karlquist wrote: >> I recently modified an old 10811 to bring out the crystal >> leads on miniature coax (instead of having them connect >> to the oscillator circuit). This allowed me to measure >> the crystal's inherent flicker noise of frequency. >> The measurements indicate that the 10811 phase noise out >> to at least 100 Hz is entirely due to the crystal. >> An interesting aspect of flicker noise of frequency >> is that Allan deviation is independent of tau. Thus, >> just one number describes the crystal noise. >> >> Rick >> >> On 9/30/2012 4:44 AM, Bob Camp wrote: >>> Hi >>> >>> Close in, it looks like it's pretty much the crystal and the loading >>> of the bridge oscillator. >>> >>> Bob >>> >>> On Sep 29, 2012, at 11:46 PM, Richard (Rick) Karlquist >>> <richard@karlquist.com> wrote: >>> >>>> The E1938A uses a crystal that is basically the same as >>>> the 10811 crystal except that it is in a reduced height >>>> package. However the phase noise is not as good as a >>>> 10811 due to broadband noise in the automatic frequency >>>> control circuit. By the time I discovered this, it >>>> was too late to try to fix it. >>>> >>>> Rick Karlquist N6RK >>>> E1938A designer > > > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > >

On 09/30/2012 09:49 PM, Richard (Rick) Karlquist wrote: > I only measured one 10811. However, John Vig's tutorial > (available at IEEE UFFC) rather categorically states that > piezoelectric resonators have flicker noise of frequency. > What I measured was most closely related to phase noise, > as opposed to Allan Deviation. Phase noise of 10811's is > more consistent unit to unit than ADEV and certain more consistent > than aging. BTW, ADEV at HP was measured against a > special 10811 that was 500 Hz off frequency. I was never > able to find out how they arrived at this "golden" unit. > But it seems clear that it could not have been the > best ever unit for ADEV, thus the real "golden" units > that came down the pike were simply rated as ADEV too good > to measure. I tried to get a project started where we > would use a frequency synthesizer to do the offset. > Then we could take the best units and compared them against > each other. Then, as well accumulated test data, the > cream would gradually rise to the top and we would have > some true golden units. The problem was that there ADEV > at those levels wasn't a "money spec". With two such offset 10811s you could use DMTD methods or for that matter cross-correlation phase noise measures to more directly measure the units. That way you would have avoided the "golden unit" issue, since the phase noise of those would average out if treated well. Didn't you consider steps like that? That's what I do in my lab these days. Cheers, Magnus

On 9/30/2012 2:20 PM, Magnus Danielson wrote: > > With two such offset 10811s you could use DMTD methods or for that > matter cross-correlation phase noise measures to more directly measure > the units. That way you would have avoided the "golden unit" issue, > since the phase noise of those would average out if treated well. > > Didn't you consider steps like that? I didn't "own" this system. Production engineering was very conservative and didn't want to change anything. Rick

On 10/01/2012 02:05 AM, Richard (Rick) Karlquist wrote: > > > On 9/30/2012 2:20 PM, Magnus Danielson wrote: > >> >> With two such offset 10811s you could use DMTD methods or for that >> matter cross-correlation phase noise measures to more directly measure >> the units. That way you would have avoided the "golden unit" issue, >> since the phase noise of those would average out if treated well. >> >> Didn't you consider steps like that? > > I didn't "own" this system. Production engineering was very > conservative and didn't want to change anything. As they tend to be. But back at the engineering side at least? Anyway, setting that up today isn't rocket science, and Bert and John among others have been able to build such systems. Cheers, Magnus