[time-nuts] No State Of The Art Counter

Tijd Dingen tijddingen at yahoo.com
Mon Jan 10 21:42:39 UTC 2011

--- On Sat, 1/8/11, Magnus Danielson <magnus at rubidium.dyndns.org> wrote:


> >> How will the input side work? How will you handle
> input
> >> signals of various kinds? In particular sine of
> various
> >> amplitudes and frequencies. Slew-rate can be a
> limiting
> >> factor as white noise will convert into jitter if
> hitting a
> >> straight comparator. Choice of trigger point can
> be done to
> >> achieve lowest jitter, so just AC-blocking and
> trigger on
> >> the 0V may not be the best solution if shape is
> not well
> >> known. Noise of input stage comes into play.
> >
> >
> > I am glad you asked. :) This is still one big To Be
> Determined at this point.
> >
> > As you say, the choice of trigger point has a large
> impact on jitter. Do you have any suggestions on how to
> achieve this? Either for the situation where the signal
> shape is well known or not so well known.
> >
> > ...
> Well, you do have comparators and programmable trigger
> level. There is 
> no magic to that side. But do care about bandwidth. It
> might be needed 
> to actually not use a comparator up-front but rather let
> there be one or 
> two stages of gain, where the first one DC-shifts the
> signal with the 
> trigger level. This way you gain yourself to a higher
> slew-rate. This is 
> really what Collins-style ZCD do. The trick is to balance
> the gain and 
> noise-bandwidths such that you get optimum slew-rate
> bandwidth for 
> lowest added noise.

Magnus, thanks for your reply. The Collin's style ZCD was a useful hint. Didn't know that one yet. I read up on it on Bruce's site. At least I think it is Bruce's site, what with the "/~bruce/" in the url. ;)

Especially the bit about the Generalised Collins Hard Limiter. Checked the paper, looks useful ... at a later date (it's on the list for rev 3 now). For now I have decided to keep the analog frontend as simple as I can get away with. Which in essence ranges from "nothing at all, straight up connect, better make sure what you connect is LVCMOS25/33 or LVPECL25/33" to "comparator + DAC, and just learn to live with the extra jitter until a later revision".



More information about the time-nuts mailing list