[time-nuts] Z3801 lockin behaviour

[Magnus Danielson]

From: "gbus" <gbus at adelphia.net>
Subject: Re: [time-nuts] Z3801 lockin behaviour
Date: Mon, 24 Jul 2006 11:56:40 -0600
Message-ID: <006901c6af4a$84609280$6401a8c0 at adelphia.net>

Greg,

> Measuring a counter's own timebase will show the counter's noise and offset
> *relative* to its timebase. Don't you need to know this limitation *before*
> making assumptions about the accuracy of your counter as locked to your
> house standard?

This is infact not as a good method as you would think it is. You would only
sample the counters performance for that (or those, integer number) position
in time which you get from the output delay, cable delay, input delay from the
timebase which you measure. You can end up in either a very favourble delay,
where the measurement would give you a too optimistic value, or you can end up
in a very non-favourble delay in which case it would give you a too pestimistic
value. What you do want is a source which is slighly offset from the timebase,
either by being asynchrounous or by being synchronously offset by a small
amount. Only under that condition you may start evaluating it. Also, in order
to make any sense, it needs to be in a similar amplitude and frequency as you
intend to measure, such that you have a similar slope-dependend jitter added.

I've made comparable measurements, and it gives me an indication that my
measurement floor is below the wobbles that I am seeing now.

Cheers,
Magnus