PLL Experiments

[Don Tillman]

   Date: Sat, 19 Aug 1995 08:38:06 -0700
   From: chordman at ix.netcom.com (Scott Gravenhorst, Synthaholic)

   I was playing around with a 4046 micropower PLL (connecting it to lock 
   to a VCO output) and found that while it will track a VCO very nicely 
   without a counter in the loop,  when a divide by N counter (I used div 
   by 3) it will lock and frequency multiply only when a rather low freq 
   lowpass filter is used.  It tracks, but with a large amount of over and 
   undershoot.  The settling time is enormous at something around a half 
   second or more.  For my needs, the settling time is too long.  I am 
   hoping to reduce this effect to not more than a short chirp.

This is difficult to diagnose completely since you haven't given us
any circuit details, but basically the problem is that you've changed
some parameters of the loop transfer function and now it's ringing at
a low frequency.  This "rather low freq lowpass filter" sounds
especially suspect.

You say the original circuit tracks fine, right?  Then just insert the
divide-by-three counter into the original circuit *and* use a VCO
capacitor of 1/3 the original value, and you should be jammin'.

   Date: Sat, 19 Aug 1995 11:03:28 -0500
   From: Ric Miller <rmiller at merlin.magic.mb.ca>

   You could use a decade counter, picking off the Divide by 3 from it's 3rd 
   step.  Of course using this method, you'd only have a square wave output, 
   unless you used a comparator to obtain a pulse wave.

Huh?  What decade counter is this that divides by three and provides
square waves?

  -- Don