[time-nuts] OCXO pressure sensitivity before TCXO
jimlux
jimlux at earthlink.net
Mon Aug 5 17:16:19 UTC 2019
On 8/3/19 7:22 AM, ew via time-nuts wrote:
> Are wee missing something. Reading Bill Slade post and old NASA attachments the attached plot to me was an eye opener. Our work has always been focused on Rb pressure compensation. Are we overlooking something. Has any one done some work on this subject
> Thanks Bert Kehren
>
Those papers are reporting on designs from the 80s - things have
improved a bit since then.
The thing that potentially causes problems in all "sealed" oscillators
are mechanical stresses - whether from a hermetic enclosure flexing due
to pressure changes or from thermal expansion.
The latest USOs from APL take pains to minimize all of those - One thing
that is interesting is that the reason for a dewar is partially for
thermal isolation - in space, you're in a very hard vacuum, which is a
pretty decent insulator so the dewar isn't improving things - and also
for mechanical consistency - you can mount everything on a subchassis
that sits inside the dewar and is mechanically and thermally isolated
from the surrounding dewar. It also lets you test in air on Earth.
The vacuum in the dewar is probably worse than the vacuum in space
(depending on where you are "in space"), but nobody has figured out a
good way to reliably "vent" things - and then you still have the "test
on the ground" which is in a not very good vacuum: 1e-5 torr (1e-3 Pa)
is typical for thermal vacuum testing.
One other thing to know about those USOs - they start about 1000 quartz
blanks that are mounted in holders, and wind up with a couple dozen
oscillators in dewars, and then run them all for a long time (many
months) and watch them. 2 or 3 out of a batch will be the ones selected
for use - based on their observed stability -
This is a very, very different manufacturing model from something where
you make 100s a month - and that's why they cost $1M each and take 2-3
years to make. (well, there's a heapin' helpin' of paperwork that goes
with the USO, too, and that isn't inexpensive to produce)
More information about the Time-nuts_lists.febo.com
mailing list