I'm not convinced that my real question relates directly to dropped tick
rates.  I have no disagreement with anything you say, but take a look at
the following:

>macbsd-840av:hotz {101} date
>Wed Mar 25 14:52:18 PST 1998
>macbsd-840av:hotz {102} ntpq
>ntpq> rv
>status=06f4 leap_none, sync_ntp, 15 events, event_peer/strat_chg
>system="NetBSD", leap=00, stratum=3, rootdelay=7.63,
>rootdispersion=961252.94, peer=5636, refid=hotzsun.jpl.nasa.gov,
>reftime=b8c40909.501b1000  Wed, Mar 25 1998 14:51:53.312, poll=6,
 						^^^^^^^^^
>clock=b8c40927.77e5f000  Wed, Mar 25 1998 14:52:23.468,
					      ^^^^^^^^^
>phase=961071.659, freq=512000.00, error=943.71
>ntpq> quit
>macbsd-840av:hotz {103} date
>Wed Mar 25 14:52:26 PST 1998

Why are these two times different?

And where does it say that the PLL has a problem of *any* kind?  Granted
the rootdispersion is huge, but the hex status code is identical to my
correctly-running (I think) Sun, and I thought the sync_ntp flag indicated
that everything was OK.  It looks to me as though everything is working
just fine, but it somehow never gets around to updating the main clock.

Should I be asking these questions on comp.protocols.ntp instead of here?

Signature failed Preliminary Design Review.
Feasibility of a new signature is currently being evaluated.
h.b.hotz@jpl.nasa.gov, or hbhotz@oxy.edu