[time-nuts] Digital divider phase noise

John Miles jmiles at pop.net
Wed Feb 7 03:33:23 EST 2007

Nice collection of articles they have there... lots of fun stuff if you
navigate up a directory or two and look around.

I have a synthesizer that uses an LT1016 comparator to square up the 10.7
MHz signal coming out of a post-DDS crystal filter, prior to driving the
reference pin of an Analog Devices PLL chip.  The comparator tends to
improve the loop's inband noise by a couple of dB by feeding the PLL a much
stronger reference signal than it would otherwise get.  After reading that
article, I found myself wondering if there'd be any benefit to running an
identical crystal filter between the comparator output and PLL reference
pins.  It seemed possible that the PLL's R-divider and phase detector might
benefit from a band-limited input, since the comparator is a
relatively-broadband high-gain device.

It turns out not to help in this case (
http://www.thegleam.com/ke5fx/test.gif ).  Seems to be little or no
difference in the loop's inband noise regardless of the presence or absence
of the post-comparator filter.

I did verify that the (obviously mismatched) filter wasn't costing any
signal amplitude.  Because the PLL's input Z was so much higher than the
comparator output Z, the peak-to-peak filter output voltage is actually
slightly higher than its input voltage.

So I guess this effect shows up only when you're dealing with an
already-well-optimized divider chain, rather than a simple x10 R-divider and

-- john, KE5FX

> -----Original Message-----
> From: time-nuts-bounces at febo.com [mailto:time-nuts-bounces at febo.com]On
> Behalf Of Dr Bruce Griffiths
> Sent: Tuesday, February 06, 2007 1:32 PM
> To: Discussion of precise time and frequency measurement
> Subject: [time-nuts] Digital divider phase noise
> For those who would like to improve the phase noise of the output of a
> digital divider or who would like a good reason to using DDS instead see:
> http://tf.nist.gov/timefreq/general/pdf/1380.pdf
> Bruce

More information about the time-nuts mailing list