Thank you everyone for your comments and suggestions.

I think my best course of action at the moment is to use a QRP2000 as it stands 
and see if it is sufficiently stable for my purposes.

Regarding a GPS-disciplined oscillator, I think probably the best approach will 
be to construct one at (say) 10 MHz. This can then be used for several 
purposes, including calibrating my frequency counters.

John F5VLF

On 19 Mar 2012, at 19:02CET, John Rabson wrote:

> LF,
> 
> I have been wondering about putting the device in a temperature-controlled 
> enclosure and wonder if there's anything to be gained.  What is the 
> frequency-determining element? If it is a quartz crystal or similar, what is 
> the temperature at which the frequency dependence is a minimum?
> 
> Does this temperature fall within the permitted range of operation for this 
> device?
> 
> Alternatively, could one realistically impose GPS discipline on a unit using 
> this device (such as a QRP2000)?
> 
> John F5VLF
>