[time-nuts] questions on uncompensated crystal oscillators

[Magnus Danielson]

From: Hal Murray <hmurray at suespammers.org>
Subject: Re: [time-nuts] questions on uncompensated crystal oscillators
Date: Wed, 05 Jul 2006 13:02:00 -0700
Message-ID: <20060705200201.D7762BDF0 at ip-64-139-1-69.sjc.megapath.net>

> 
> > As I recall there is a hysteresis effect on the frequency vs.
> > temperature curve that needs to be compensated for as well. - Mike 
> 
> Is there really hysteresis, or does it just look that way due to time lag 
> between where the temperature is being measured and that actual temperature 
> at the crystal?
> 
> If there is, what's the mechanism?

For any particular temperature, the crystal has an "optimal" distribution of
mass when driven as an oscillator. When you change the temperature, you begin a
process of mass movement, as you do that it changes frequency. The frequency
drift continues over a long time and aging have been measured as a continous
process over 5 years for a set of oscillators.

For OCXOs you can avoid it at powerup by keeping the oscillator circuit turned
off until the oven reached temperature.

There are naturally other sources of aging.

I'm sure certain members can both correct me and add much more knowledge.

Cheers,
Magnus