[time-nuts] Testing frequency using NTP

Bruce Griffiths bruce.griffiths at xtra.co.nz
Thu Oct 2 22:37:52 EDT 2008

Bruce Griffiths wrote:
>> Tom
>> The major source of interchannel phase shift instability will be the 
>> phase shift instabilities of the sound card input RC coupling network 
>> particularly if electrolytic capacitors are used and the input 
>> frequencies are too close to the high pass RC filter 3dB frequency. 
>> Amplifier phase shift instabilities are much smaller and the 
>> interchannel group delay mismatch will be relatively small as the group 
>> delay for each channel is determined by the digital filters used by the ADC.
> For temperatures around 20C a wet electrolytic cap has a tempco of aroun 
> +500ppm/C or so.
> With a 0.5dBinput frequency of 20Hz a typical sound card has a high pass 
> filter input time constant of about 33ms.
> At 1kHz the corresponding phase shift is about 800ns with a tempco of 
> about 0.4ns/C.
> Bruce
The time constant should have been 22ms and the phase shift 1.16us with 
a tempco of 580ps/C.
Since both channels have the same nominal low frequency cutoff the 
differential phase shift tempco will be somewhat smaller.


More information about the time-nuts mailing list