[time-nuts] Improving the stability of crystal oscillators

[Bruce Griffiths]

Brooke Clarke wrote:
> Hi:
>
> Something I've been thinking about is attaching a large thermal mass of 
> aluminum to the crystal and it's oscillator then surrounding that with 
> insulation.  Then surround that with a thin copper layer.  Temperature sensors 
> on the copper and aluminum.  If the thermal time constant could be made very 
> close to 24 hours, or integer multiples of that, it might be possible to 
> predict the inner temperature allowing compensation.
>
> Has this already been done?
>
>   
Brooke

Why stop there?
Try burying the oscillator under 6ft or so of soil.
The daily temperature fluctuations at this depth are very small.

Bruce