[time-nuts] GPSDO using 100Hz

[Bruce Griffiths]

Peter Vince wrote:
>>> B) The variable pulse width of the 100Hz (and 10kHz) outputs do no favours
>>>       
>> to an XOR phase detector, its better to use the leading edges of these signals.
>> We agree again, That is why they MUST be divided by two first, using the
>>     
> correct edge.
>
> I'm not convinced that would cure the problem.  Yes, it would correct for pulses
> that weren't a perfect 1:1 mark-space ratio, but assuming the shape is constant,
> surely that isn't a problem?  However, when a +/- 50ns step was inserted, surely
> the effect of that would at best be halved by dividing the 100Hz by two?
>
> Peter Vince
>
>
>   
Not so as only the divide by 2 transition that occurs on the second will
have its delay with respect to the previous transition adjusted by 50ns.
The subsequent 99 transitions (or 9999 transitions with 10kHz output)
will occur at intervals of 10ms (or 100uS with 10kHz output).

The pulse duty cycle isn't necessarily constant (datasheet specs for the
pulse width tend to be somewhat ambiguous) and usually the pulse width
of the pulse preceding the second differs.
The spec for the M12+T is:
The 100Hz pulses are high for a period of 2-3ms in duration except for
the pulse that immediately preceding the second which is 6-7 ms in duration.
A 1millisec tolerance in pulse width is comparatively large.

A pulse width variation of 1ms or so should be relatively easy to measure.

Bruce