[time-nuts] Improving the stability of crystal oscillators

[Bruce Griffiths]

Don Collie wrote:
> Hi Hal,
>     I was thinking of attaching a temparature sensor [AKA Star Treck] to the 
> cold side of a Peltier [what`s the other type? Are they available/better?] 
> pile. and driving the pile from the output of some sort of servo loop to 
> maintain a temparature of ,say , 0 Degree C.
>      If you wanted a double oven, you could heatsink a small oven, 
> containing the crystal, the oscillator, and buffer[s] to this, and use a 
> second servo loop to raise the temparature of this to 25 Degrees C working 
> against the Peltier. In this way, you could maintain the crystal, and 
> circuitry at 25 Deg C., over an ambient temparature of 0 to ,say 70 Deg C.
>     Yes, you can cut a crystal to have an inversion temp at 25Deg C. [well 
> certainly with an AT cut - I`m not sure about the SC cut.]
> Cheers,............................................Don C.
>
> ----- Original Message ----- 
>   
Don

With such a wide swing in the temperature differeence across the Peltier
stack, you will need to adjust the PID loop parameters to maintain
optimum control.
A fixed set of parameters is only useful for a temperature range of up
to 10C. The Peltier stack performance varies significantly with the
temperature difference across the stack.
Consequently you also need to sense the peltier device heatsink temperature.

Bruce