[time-nuts] Time Dilation tinkering

[jimlux]

On 3/21/17 7:03 PM, Tom Van Baak wrote:
> Yes, MAC and CSAC do show environmental sensitivity. But that should not be a surprise to anyone that works with precise time & frequency.
>
> The factors include voltage, temperature, temperature gradient, pressure, humidity, acceleration, tilt (orientation), and who knows what else. Maybe even radiation? One looks for the *coefficient* of course, but also linearity, repeatability, hysteresis, and even interaction amongst factors. It's not an easy task to do this, nor can one assume each unit of a given make/model will be the same. I don't have comprehensive data or a nice report.
>
> But attached is one of the first plots where I put a SA.32m in a home-brew vacuum chamber and pulled down to a few inches of Hg for a few hours to simulate the low pressure of a flight up to 50 or 90,000 ft. For a high altitude relativity experiment -- where you'd like your clock to remain stable to parts in e-13 and not accumulate too many stray ns -- it's not a good sign when your clock changes by 2e-11 (that's more than 1 ns per minute) just because of ambient pressure changes.
>


I wonder if the effect is due to vacuum, or something like thermal 
gradients (no convective heat transfer in vacuo)