[time-nuts] Recommended parameters for Timelab ADEV analysis

Tom Van Baak tvb at LeapSecond.com
Thu Nov 15 03:39:38 UTC 2012


> Hi Edgardo,
>
> however, the 53132A is considerably more sensitive in frequency mode as 
> we have recently discovered.
> I saw an astoundingly low ADEV noise floor of some 6E-13 at t=10 sec as 
> opposed to only 2.5E-11 in TI mode.
> I don't have a 53132A, only a 53131A that I don't use anymore for ADEV, 
> so I can't be more specific.

Adrian,

Part of the reason for your "astoundingly low" results is because the 53132A is rapidly averaging samples internally in order to give a more precise frequency measurement. While this trick is good for determining frequency accuracy it is bad for determining stability. What I mean is that averaging, by definition, partly removes noise (removes "instability") in the samples being collected.

In much of science reducing noise is a good thing. But when making instability measurements of clocks and oscillators the noise is usually what you are trying to measure! So getting rid of it is not always a good thing. This is especially true when the device is removing noise without telling you; it can lead to false conclusions.

As you saw, when you use heavily smoothed frequency data as input to an ADEV calculation you will get unusually low stability numbers. This is not to say the numbers are completely meaningless. But at least it's not correct to compare heavily averaged frequency measurements with actual instantaneous phase measurements.

/tvb





More information about the time-nuts mailing list