[time-nuts] Sub 20 ps time detection

[Bob Camp]

Hi

One of the nasty issues doing something like this is the voltage tolerances. If you 
are after 0.02 ns on a 2 ns edge, that is at the 1% level. If you want a 10X margin
to take the voltages out of the picture, you are now at the 0.1% level. 

If you have 3.3V logic 0.1% is 3.3 mv. That is the sum of the voltage issues with the supply
on the detector side, the reference to your comparator, and the offset of the comparator. If 
you are in a dynamic environment, things like temperature stability may have to be 
factored into each of those budgets. You may also have ground offset issues to deal with. 

How much this matters is very much a “that depends” sort of thing. Not knowing much
about the system involved does make this a bit of a guessing game. Sorry if the
guess is not applicable to what you are doing. 

Bob


> On Mar 25, 2017, at 12:32 PM, Mohammad-Hadi Sohrabi <mhsohrabi at ucdavis.edu> wrote:
> 
> Hello,
> I have been following the posts here for a while and learned alot. I
> thought I could get some help as well.
> I need to detect less than 20 ps between two incoming pulses from a
> detector. pulses are 5ns wide and the rise time and fall time is around 2ns
> each.
> I am planning to use a constant fraction discriminator, simply a fast
> comparator with less than 500 MHz bandwidth for zero crossing detection and
> providing a very sharp edge. Then this can be fed into a TDC such as GPX2
> or anything of this sort. I think these can be fairly easily implemented on
> FR4.
> Recently I stumbled upon LTC6957-4 logic converter from another discussion.
> Can anyone guide me if this can be used for producing a sharp edge from a
> pulse that I mentioned? Datasheet suggests 0.5 ns rise time is possible but
> I think this is for single tone sine wave and not a short pulse.
> 
> Bests,
> Mohammad
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.