[time-nuts] Digital PLL ICs, alternatives and digital loop filters

[Magnus Danielson]

From: "Stephan Sandenbergh" <stephan at rrsg.ee.uct.ac.za>
Subject: RE: [time-nuts] Digital PLL ICs, alternatives and digital loop filters
Date: Thu, 28 Dec 2006 23:31:49 +0200
Message-ID: <000a01c72ac7$9c7377e0$a2143e0a at mycomputer>

> Hi Magnus,

Hi Stephan,

> Sorry for my delayed response, but I was a bit under resourced on e-mail
> facilities during the festive season. 

No worries. You answer in the same decade it's good.

> > Go and get Floyd Garners "Phaselock Techniques". The reference and very
> > worth reading. I have other books to recommend too. The popular BEST book
> > isn't as great, but can sometimes be a good reminder too, but Garners book
> > beats it in my opinion.
> 
> I haven't got any of the books you mentioned but I worked through Wolaver's
> "Phase-Locked Loop Circuit Design" some time ago. It is probably a bit out
> dated but it was the thinnest on the shelf :) Nonetheless, I thought it is
> quite good and serve as a good intro to PLLs.

It is not a bad book, but I view it as a complementary book to the others. You
will however find it in my bookshelf both at home and at work.

The Wolaver book is actually very good in describing certains aspects better
than some other books I've seen. So it can be a good book to grasp some of the
aspects, but the Gardner book is _very_ readable and is really strongly
recommended since is also such a deep book in what it covers even if it is only
sligthly thicker than Wolavers book (which uses larger pages).

> I've seen two books by William Egan: "Phase-Lock Basics" and "Frequency
> Synthesis by Phase Lock" on our library's shelves. Are you familiar with
> these? Are they worthwhile?

I haven't looked in them. There are gazillions of books, but if you have
Wolaver, Gardner and Best you should be able to get clues from Wolaver and Best
while Gardner rules.

> > I prefer phase detectors which is more continous such as mixers, XOR or
> > S-R style phase detectors.
> 
> In this case the phase detector is purely digital so I guess the loop filter
> could also be a purely digital IIR filter (or something?) I guess the only
> limitation is the frequency at which the digital system could run.

If your phase detector is a TIC, then IIR filter is certainly the way to go.
Be carefull about the dynamics. The processing you need should usually be very
small so it should not be a hard problem to solve. If you process of each PPS
and only needs a handfull of multiplications and additions you can do almost
arbitrary bitlengths even in a very poor CPU. My Cs beam is controlled by a
6809 CPU, which also takes care of the serial line I/O. Whiee!

Cheers,
Magnus