[time-nuts] Allan Variance

Bruce Griffiths bruce.griffiths at xtra.co.nz
Tue Mar 29 18:21:30 UTC 2011


Martyn Smith wrote:
> Hello,
>
> The answer to my original question is that Allan Variance is not to be 
> used
> but rather Allan deviation.
>
> However, David Allan does not mention Allan Deviation on his website, 
> only
> Allan Variance.
>
The distinction is obvious to anyone with sufficient exposure to 
statistics theory.
> Also I spoke to the designer of the PRS10 and he definitely said his 
> specs are
> Allan Variance.
>
If that were literally correct the PRS10 would be too noisy for use as a 
local clock in VLBI.
Whereas in fact as long as the other antena(e) use hydrogen maser(s), a 
PRS10 suffices at least for short (~15minute) integrations for signals 
in the waterhole region.
> Also I looked at Wenzel and they only quote Allan Variance.
>
> Also NIST quote Allan variance and the list goes on.
>
> So I'm still confused, as it seems is everyone else.
>
HP/Agilent correctly label the frequency stability measures quoted in 
datasheets originating in the 1970's.
> Regards
>
> Steve
>
Bruce




More information about the time-nuts mailing list