[time-nuts] Low-Cost Rubidium Performance

[Mike S]

On 2/9/2012 11:34 PM, ws at Yahoo wrote:

> You can of course use the ADEV math function for anything, but MY
> definition of "true and accurate ADEV numbers" must include a set up
> that gives repeatable results.
> If you test is done over different or unknown temperature changes that
> have major effects, then I'm calling that poor ADEV data (taking).

Then you had better control acceleration (gravity/shock/vibration), 
voltages, humidity, pressure, electromagnetic fields, et al, if you want 
validity. Again, what makes temperature a unique case?