The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
From: Chuck Simmons
Organization: You jest.
X-Mailer: Mozilla 4.61 [en] (X11; U; Linux 2.0.33 i586)
Subject: Re: Phase locked loops and confusion.
Date: Sat, 23 Nov 2002 05:49:35 GMT
NNTP-Posting-Date: Fri, 22 Nov 2002 21:49:35 PST
John Larkin wrote:
> On Sat, 23 Nov 2002 01:06:45 GMT, Chuck Simmons
> >PLLs are fairly common and seem fairly simple but I found over the past
> >week that confusion abounds and since I have seen this confusion before,
> >I wonder if it is common.
> >Our digital team (yes, 2 engineers) has a need for a PLL but for various
> >reasons, an all digital loop could not be considered and, therefore, I
> >was dragged in as resident analog stuff guru. The y showed me the loop
> >they were proposing which consisted of digital equivalents (pretty exact
> >really) of a phase detector driving up and down charge pumps dumping
> >into a capacitor followed by a filter followed by the VCO (which is
> >analog). I was asked what the filter needed to be and I mumbled that
> >they had two integrators so it should probably be a classic lead/lag and
> >proceeded to draw response curves on the whiteboard indicating what was
> >The next day, I was told that all of the papers they could find on the
> >web used a low pass filter. My comment was an unenthusiastic "Oh." I was
> >correctly interpreted and a printed article soon appeared on my desk.
> >Suddenly I saw the confusion. The proposed loop had a phase detector,
> >charge pumps and capacitor which is Kp/s. The article used a phase
> >detector that was Kp. So the article loop has Kp*Kf/s while the proposed
> >loop has Kp*Kf/s^2. A low pass is perfectly fine for the article loop
> >but not workable for the proposed loop.
> >I saw this confusion some years ago as well because the phase detector,
> >charge pumps and capacitor is used a lot in practice but many articles
> >don't mention this variant. This leads to confusion about what sort of
> >filter is actually needed. The charge pumps and capacitor type has the
> >advantage that it will drive the steady state phase error to zero even
> >when the desired frequency is greatly different from the VCO center
> Yeah, the charge pump adds another integrator to the loop. I don't
> know why the charge pump thing is so popular; most of them have nasty
> deadbands. A charge pump pd doesn't inherently drive the loop in the
> right direction when the freq error is large; it takes a bit of
> additional logic (like in the 4046) to do that.
We have an ironclad tested strategy for setting the frequency. There is
a pattern we can look for that can tell us if the frequency is off by
huge amounts well beyond the operating range. We get the frequency to
about 0.1% before we attempt phase lock (overkill).
> My favorite phase detector is a d-type flipflop operating in bang-bang
> mode: input signal on clock, reference osc on D. This is really
> interesting to analyze and to design a filter for. Gain is infinite,
> sort of.
We looked at an equivalent because the digital team considered that
simple. I pointed out that it would limit cycle with larger amplitude
than the deadband from somewhat more linear methods. The simulations of
the bang bang were unimpressive given the fact that we are locked to
edges coming at most at f/6 and on the average of f/12. We are
essentially in harmonic lock but the harmonic varies from 6 to a maximum
of 28 from edge to edge. Harmonic locking with the additional
complication of a loop divider of around 200 makes it really
interesting. The bang bang seemed to have trouble with that so we
decided to change our approach.
Aside from implementation problems, the fundamental confusion of the
Kp/s type of phase detector and the Kp phase detector. Is the real
problem. It had us speaking different languages for a couple of days.
... The times have been,
That, when the brains were out,
the man would die. ... Macbeth
Chuck Simmons firstname.lastname@example.org
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup