Resolution of F vs t measurement

[Excom]

Hi all

I have measured the resonance frequency of a resonator as a function of time by mean of a Laser-Doppler vibrometer and a frequency counter. In this measurement there are some fluctuations in the measured frequency as time goes by.
My question is: How do I determine the resolution of this frequency measurement? Or in another way: How precise can I resolve the frequency? Do I take the standard deviation for a period and then define this as my frequency resolution?

I hope someone can help me.

Best regards
Excom

 

[sophiecentaur]

That would be the RMS error. A pretty universal way of stating errors - I say go for it.

 

[Excom]

Thanks:-)

 

[Excom]

A more correct way of doing it is to calculate the Allan deviation as a function of integration time and then take the minimum value.

 

[f95toli]

Indeed. the standard deviation is not a good measure in these situations. The reason is that it only converges for white noise, for all other noise types the STD will give you the "wrong" result which is why the Allen deviation should always be used for this type of analysis.

Have a look at the NIST handbook of frequency analysis (free download). Also, there are free Matlab routines for calculating the ADEV available on Mathworks file exchange.