Hi! I have been running a bunch of oscillator plots using a DMTD unit built for the NBS in the early 80s that I picked up last year and repaired. I am using an offset oscillator with a 1Hz offset (an FTS1200) for the L.O. and select the 1Hz BW on the unit. I have fed both inputs from the following oscillators thru a power splitter to evaluate the baseline. The Allen Deviation at 1 second for three different oscillators driving both inputs at the same levels and having the delay set to the same small value are. Delay 100us Osc 1 - 1.05E-13th Osc 2 - 9.50E-14th Osc 3 - 2.00E-13th At larger delay values the Allen Deviation at 1 second for Osc 3 rises much faster than the other two. Delay 100ms Osc 1 - 3.10E-13th Osc 2 - 3.50E-13th Osc 3 - 3.00E-12th Can I deduce any useful information from these numbers? Obvoisly Osc 3 is not as good as the other two at the 1 second tau! Plots run against other oscillators also show that Osc 3 causes a rise in the value at 1 second. Thanks, Corby Dawson ____________________________________________________________ Click now for low quotes on great group health insurance plans! http://thirdpartyoffers.juno.com/TGL2141/fc/BLSrjpTQb2SOvEObOFBY7GX9x4Nzi53TOHS8B236icJn2YVVYm4qAqfZrSQ/

Corby Dawson wrote: > Hi! > > I have been running a bunch of oscillator plots using a DMTD unit built > for the NBS in the early 80s that I picked up last year and repaired. > > I am using an offset oscillator with a 1Hz offset (an FTS1200) for the > L.O. and select the 1Hz BW on the unit. > > I have fed both inputs from the following oscillators thru a power > splitter to evaluate the baseline. > > The Allen Deviation at 1 second for three different oscillators driving > both inputs at the same levels and having the delay set to the same small > value are. > Delay 100us > Osc 1 - 1.05E-13th > Osc 2 - 9.50E-14th > Osc 3 - 2.00E-13th > > At larger delay values the Allen Deviation at 1 second for Osc 3 rises > much faster than the other two. > > Delay 100ms > Osc 1 - 3.10E-13th > Osc 2 - 3.50E-13th > Osc 3 - 3.00E-12th > > Can I deduce any useful information from these numbers? > > Obvoisly Osc 3 is not as good as the other two at the 1 second tau! > > Plots run against other oscillators also show that Osc 3 causes a rise in > the value at 1 second. > > Thanks, > > Corby Dawson > ____________________________________________________________ > Click now for low quotes on great group health insurance plans! > http://thirdpartyoffers.juno.com/TGL2141/fc/BLSrjpTQb2SOvEObOFBY7GX9x4Nzi53TOHS8B236icJn2YVVYm4qAqfZrSQ/ > > _______________________________________________ > 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. > > Corby The noise contribution of the offset oscillator is incompletely canceled when the delay between the zero crossings for both channels isn't zero. The effect increases with the delay between zero crossings as your tests demonstrate. You can deduce little from the test apart from the fact that Osc 3 has worse performance than that of the other 2 oscillators. Bruce

Hi Corby, > Corby > > The noise contribution of the offset oscillator is incompletely canceled > when the delay between the zero crossings for both channels isn't zero. > The effect increases with the delay between zero crossings as your tests > demonstrate. > You can deduce little from the test apart from the fact that Osc 3 has > worse performance than that of the other 2 oscillators. I agree with Bruce, the delay removes the cross-correlation gain in the suppression of the transfer oscillator. Compare the oscillators pair-wise instead. Cheers, Magnus

Magnus Danielson wrote: > Hi Corby, > >> Corby >> >> The noise contribution of the offset oscillator is incompletely canceled >> when the delay between the zero crossings for both channels isn't zero. >> The effect increases with the delay between zero crossings as your tests >> demonstrate. >> You can deduce little from the test apart from the fact that Osc 3 has >> worse performance than that of the other 2 oscillators. > > I agree with Bruce, the delay removes the cross-correlation gain in > the suppression of the transfer oscillator. Compare the oscillators > pair-wise instead. > > Cheers, > Magnus > > _______________________________________________ > 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. > Another question that arises is the stability/jitter of the 100ms delay itself. How was this 100ms delay achieved? Bruce

Bruce Griffiths skrev: > Magnus Danielson wrote: >> Hi Corby, >> >>> Corby >>> >>> The noise contribution of the offset oscillator is incompletely canceled >>> when the delay between the zero crossings for both channels isn't zero. >>> The effect increases with the delay between zero crossings as your tests >>> demonstrate. >>> You can deduce little from the test apart from the fact that Osc 3 has >>> worse performance than that of the other 2 oscillators. >> I agree with Bruce, the delay removes the cross-correlation gain in >> the suppression of the transfer oscillator. Compare the oscillators >> pair-wise instead. >> >> Cheers, >> Magnus >> >> _______________________________________________ >> 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. >> > Another question that arises is the stability/jitter of the 100ms delay > itself. > How was this 100ms delay achieved? Indeed. The NBS DMTD system I know of doesn't have this feature. Cheers, Magnus

Magnus Danielson wrote: > Bruce Griffiths skrev: >> Magnus Danielson wrote: >>> Hi Corby, >>> >>>> Corby >>>> >>>> The noise contribution of the offset oscillator is incompletely >>>> canceled >>>> when the delay between the zero crossings for both channels isn't >>>> zero. >>>> The effect increases with the delay between zero crossings as your >>>> tests >>>> demonstrate. >>>> You can deduce little from the test apart from the fact that Osc 3 has >>>> worse performance than that of the other 2 oscillators. >>> I agree with Bruce, the delay removes the cross-correlation gain in >>> the suppression of the transfer oscillator. Compare the oscillators >>> pair-wise instead. >>> >>> Cheers, >>> Magnus >>> >>> _______________________________________________ >>> 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. >>> >> Another question that arises is the stability/jitter of the 100ms delay >> itself. >> How was this 100ms delay achieved? > > Indeed. The NBS DMTD system I know of doesn't have this feature. > > Cheers, > Magnus > > _______________________________________________ > 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. > Hej Magnus This could be done by changing the phase difference between the 2 inputs to the 2 channels by 10% of the mixer input signal period or by about 20ns at 5MHz. Even so questioning the stability of this delay is reasonable considering that an instability of 0.3ps or so would be equivalent to the observed noise at 100ms delay between the zerocrossings of the 2 mixer outputs. A temperature fluctuation of 1C would produce a 2ps change in phase shift if ordinary coax were used for the delay. Bruce