[time-nuts] A first MDEV plot...

Bruce Griffiths bruce.griffiths at xtra.co.nz
Tue Mar 16 22:21:29 UTC 2010


John Miles wrote:
>    
>> That won't work as one of the sources is the PPS output from an Oncore
>> GT which has a PPS error spec<  500ns with SA on.
>> Whilst it may be a little lower without SA, it may be still be somewhat
>> greater than 100ns.
>>
>> Dividing the 10MHz down to 1MHz or perhaps 100kHz would be better.
>> The 5370A timebase noise contribution to the measurement error will
>> still be insignificant.
>>
>> The plot doesnt appear too discordant with the Oncore GT spec.
>>      
> Ah, interesting... I stand corrected, then.  Not at all familiar with that
> particular piece of hardware.  Looks like the jitter is more like 600 ns or
> so (6E-7 at t=1s), so the measurement is in the right ballpark after all.
>
> I hadn't tried TI in a 1-pps versus 1-pps measurement for some time, so I
> ran a measurement overnight.  The graphs were fine but the frequency-count
> chart is buggy (to put it charitably)...
>
> -- john, KE5FX
>
>
>
>    
If the dominant noise source is white phase noise the ADEV(1s) value 
should be divided by SQRT(3) to give the rms noise.

ADEV(1s) = 7E-7
Rms noise ~ 400ns

Bruce





More information about the Time-nuts_lists.febo.com mailing list