[time-nuts] Generating a solid PPS from 10Mhz source

[Bruce Griffiths]

On Friday, January 15, 2016 12:13:47 PM Attila Kinali wrote:
> On Thu, 14 Jan 2016 09:50:15 -0500
> 
> Vlad <time at patoka.org> wrote:
> > I was thinking to make a frequency divider by using FPGA. Here is my
> > attempt to implement it using VHDL.
> > This is frequency divder plus D flip-flop which I was planed to use as
> > source of 60Hz for my Telechron clock.
> > However I never implement it in HW. Instead I was using STM32F4 with its
> > timers.
> > The purpose was to divide 9.8304 Mhz OCXO output by 81920 to get 60Hz
> > and use the D flip-flop to keep output in sync.
> > Some day I'll return to this with my soldering iron in hands. ;-)
> 
> Nice!
> 
> As side note: when using an FPGA anyways, it might be good to use
> something like a lambda divider [1,2].
> 
> I'm not so sure whether their explanation why this improves the noise
> floor is the right one, but it definitly helps and is quite easy to
> implement.
> 
> 			Attila Kinali
> 
> 
> 
> [1] " The sampling theorem in Pi and Lambda dividers", by Calosso, Rubiola,
> 2013
> http://rubiola.org/pdf-articles/conference/2013-ifcs-Frequency-dividers.pdf
> 
> [2] Slides to the above
> http://rubiola.org/pdf-slides/2013C-IFCS--Dividers.pdf

A lambda divider is only useful when one wants a sinewave or a triangular wave 
output. In the case of a CMOS divider It reduces output noise by averaging the 
upconverted supply noise produced by sampling the supply noise at different 
phases of the output signal period. When upconverted power supply noise 
dominates merely paralleling outputs that sample the power supply noise at the 
same time will have little discernable effect on the output noise.

Bruce