[time-nuts] Ultra Stable Rb

[Skip Withrow]

Hello Time-Nuts,
This is a subject that I have been interested in for quite some time.

If you do some searching on Rb oscillator aging, there is a paper from
FEI that showed GPS units aging in vacuum  (and space) differently (as
in opposite sign) than at sea level.  My thought was that there should
be a pressure where the aging should go to zero.

I have been told by several people (that should know) that this has
been studied in the industry and has never come to anything.

But I'm a sucker for punishment.  In April of 2018 I acquired a very
nice vacuum chamber surplus from the aerospace industry.  I
contemplated which Rb oscillator to use and finally decided on the
LPRO-101 since it had no DDS (and thus no discrete steps) and a wide
supply voltage range (and I had several of them).  The unit was
mounted to the heavy aluminum lid of the chamber (which had five
61-pin electrical feedthru connectors, so no problem there).  Kapton
heaters were applied to the outside of the chamber and connected to a
temperature controller, and a low noise power supply that could be
varied from 18-32 volts was used to power the LPRO.  A modified NTBW50
is used to monitor the output of the LPRO.  A UPS and line conditioner
were also added to the system.

I did not want to  drive the EFC (to remove a many variables as
possible).  The C-field was set to get the unit about on frequency at
around 20Torr, then the supply voltage was tweaked to put it exactly
on frequency.

I have found that, indeed, the aging direction can be changed with
pressure.  And there is a pressure that you can get the drift to zero.
However, another fly in the ointment is that changing the supply
voltage to put the unit on frequency also changes the aging.

At this point (Jan 2019) I connected the NTBW EFC drive to the LPRO.
Now the LPRO could be disciplined, but as we all know GPS degrades the
short-term performance.  So, I run the unit with discipline disabled
and just manually change the DAC voltage to keep the LPRO on
frequency.  The supply voltage, chamber pressure, and chamber
temperature have not been touched since that point.  By knowing the
EFC gain I can calculate the aging.

For the 238 day period ending 3Nov19 the aging was -3.76x10E-14 /day.
Not bad as far as Rb goes, but I can certainly do better.

The next step I would like to take is to move the C-field adjustment
outside the chamber (and increase its resolution) so that I can put
the oscillator on frequency without any changes in supply voltage.
And again disconnect the EFC (since there is a temperature dependence
on the DAC value).  Then I should be able to get back to finding the
exact pressure the chamber should be set at.

The whole system is contained in a very short rack with the chamber on
top of it and an insulation shield over it.  I call it RUFUS (RUbidium
Frequency Ultra Stable).  It lives underneath the stairs going to the
basement.

I have also considered just building a box to drive the EFC and
increment the voltage at the proper rate for whatever the drift might
be for the temperature, pressure, and supply voltage that the
oscillator might be at.  Too much fun!

Sorry for the long post.  I'm hoping to have a detailed paper with all
the details at some point, but getting all the data of course goes
very slow.

Regards,
Skip Withrow