[time-nuts] Question about effect of spurious frequency modulation on Allan Deviation

[Richard (Rick) Karlquist]

I am thinking that I should simply make a very simple notch filter
that suppresses the spurious line by 10 or 20 dB.  I can then do
an A/B comparison of the measured ADEV with and without the filter.
If there is no change, I'm done.  Otherwise, I add filtering in
steps until the ADEV reaches an asymptote.

Rick

On 8/6/2018 6:39 PM, John Miles wrote:
>> Again, this is analogous to the 5071A having 1 MHz
>> spurs of about -90 dBc, yet the ADEV is not much
>> different than an open loop 10811.  How high would
>> the 1 MHz spurs have to be to affect ADEV?
> 
> With ADEV, you have to consider the measurement bandwidth before you can ask
> questions like this, much less answer them. :)  1 MHz spurs are far outside
> the restricted measurement bandwidth that would be desirable when measuring
> a high-performance source like a 5071A.  I suppose there might be some
> influence if you use a DC-to-daylight sampling process such as a counter,
> but the instrument noise floor will probably obscure any interferers in the
> -90 dBc range.
> 
> With a TimePod or similar device, a general rule of thumb is that a spur a
> few Hz away from a 10 MHz carrier at -120 dBc will cause ripple in the 1E-11
> to 1E-12 range.  This coincides with typical levels of coupling between
> nearby RG-58 cables -- see page 38 of
> http://www.miles.io/TimePod_5330A_user_manual.pdf for instance. That
> particular situation wouldn't have shown up on a traditional counter-based
> measurement, but others might.
> 
> For instance, in your 1 MHz example, I doubt you'd see any effect at exactly
> 1 MHz, but if your spurs are at 1.000001 MHz and your counter is capable of
> 1-ns single-shot resolution, I can imagine that there will be some
> corruption.  The exact numbers will depend on so many factors  that it's
> almost easier to set up a test and  observe the effects firsthand.
> 
> -- john, KE5FX
> Miles Design LLC
> 
> 
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
> 
>