[time-nuts] Zero-Crossing Detector Design?

Magnus Danielson magnus at rubidium.dyndns.org
Sat Jul 21 02:11:55 UTC 2012


On 07/21/2012 03:30 AM, Hal Murray wrote:
>> albertson.chris at gmail.com said:
>>>> Maybe, but it is absolutely needed if there is any noise on the signal.
> A
>>>> perfect comparator with zero hysteresis would dither on every zero
> crossing.
>
>> On 07/21/2012 01:41 AM, Hal Murray wrote:
>>> Hysterssis will eliminate spikes or double pulses that are caused by noise
>>> on a signal with low rise time as long as the noise isn't too big.
>
> magnus at rubidium.dyndns.org said:
>> ... which is what he called dither.
>
> OK, I need a few hints.
>
> What is dither in this context and what does it have to do with hysteresis?
>
> I think of dither as being related to noise at the sampling level.  In that
> context, hysteresis just shifts the switching point.  It doesn't change
> anything (much) related to noise.  How does that translate into eliminating
> spikes or double pulses?

The jump about around the sampling point which non-schmitt triggered 
comparators shows results in a signal which is _similar_ to a dithered 
signal. It causes multiple spikes. The schmitt trigger will shift it's 
trigger level depending on state to avoid this.

Exactly what to call the forrest of spikes for a non-hysteresis trigger 
is to the best of my knoweldge not really well established. Dither is 
similar enough to "work".

Oh, and the dithering noise does not need to be applied specifically, 
you may use the noise that is there. It works for GPS receivers. It's 
only when you don't have enough noise that you need to apply it, so the 
Wikipedia article is incorrect in that aspect.

But if you object to the term dither to be applied for this jumping 
around, then propose a better term.

Cheers,
Magnus



More information about the time-nuts mailing list