[time-nuts] Zero-Crossing Detector Design?

Magnus Danielson magnus at rubidium.dyndns.org
Thu Jul 19 22:15:58 UTC 2012


On 07/19/2012 11:53 PM, ehydra wrote:
> On the Bruce page there is a table with increasing stage amplification
> from low-level to the output.
> If this is the optimum for low jitter how does it connect to the
> well-known rf design philosophy to have the highest amplification at the
> first stage, not the last stage, to have maximum S/N ?
>
> Any idea?

You balance noise bandwidth with slew-rate gain. Normally you just look 
at the noise of the amplifiers and comes up with the traditional gain 
formula. Here you only want the first amplifier to have the bandwidth 
that supports the slew-rate it will have, in the same way the next 
amplifier's bandwidth and gain is set to optimum. The goal becomes to 
achieve optimum slew-rate gain with least added noise. The formulas in 
the article is derived for same amplifier noise, where as Bruce 
generalized them for the case where the amplifier noises may be different.

So, different design goals makes for different solutions. Makes sense or 
should I go into more detail?

Cheers,
Magnus




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