[time-nuts] LN RB

[John Miles]

(Forwarded message from Said below)

-----Original Message-----
From: Said Jackson [mailto:saidjack at jackson-labs.com] 
Sent: Tuesday, February 05, 2019 4:25 PM
To: 'John Miles'; 'Discussion of precise time and frequency measurement'
Subject: RE: [time-nuts] LN RB

Gents,

We have to also put this into perspective. We put a large number of
configuration capabilities into the unit to be able to optimize it just for
this type of thermally challenging application, and those options have not
been explored yet.

In particular the loop time constants can be simply set with the

SERV:MODE <SLOW|MEDium|FAST|AUTO>

commands (user manual section 3.10.2), and by default the setting is SLOW as
we assume stationary laboratory conditions with only typical AC excursions
and want to provide the absolute maximum ADEV and PN performance under
typical laboratory conditions. As Johns testing shows, the unit is extremely
stable even when not shielded from airflow and experiencing windows with -2C
outside air temp being opened in the room etc.

We do not optimize the unit to be thermally shocked from 25C to 65C+ by
default, but alas it is very easy to use the above pre-configured settings,
or to even come up with more aggressive  loop settings to accommodate such
extreme environmental conditions.

In this case, none of these options were explored, and the results should be
predictable based on knowing that the OCXO has a thermal stability of
+/-10ppb from -20C to +70C and the default loop time constants go out to 30+
minutes, and the unit was thermally shocked with much faster time constants.
The math becomes very simple with this information about the default
settings.

We also describe much more aggressive loop settings than the FAST setting in
Table 2.6 in the user manual. These could be explored to guarantee that the
loop time constants remain significantly faster than the expected thermal
gradients, which guarantees that the loop is able to keep the OCXO on-phase
with the Rubidium oscillator during thermal excursions.

We cannot cheat physics here, but with almost 10x difference in the loop
time constant options between the default SLOW and the FAST settings as
described in Table 2.6 it is not surprising that the unit does not perform
well when not properly configured for a particular environment. Also, as
John mentioned - there are potentially other issues in that particular unit,
and it's hard to judge without having access to comprehensive log file
information.

Bye,
Said
----------------------

-- john, KE5FX
Miles Design LLC



> -----Original Message-----
> From: time-nuts [mailto:time-nuts-bounces at lists.febo.com] On Behalf Of
> AC0XU (Jim)
> Sent: Monday, February 04, 2019 7:31 PM
> To: time-nuts at lists.febo.com
> Subject: [time-nuts] LN RB
> 
> John-
> 
> The LNRB had been powered on continously for several weeks prior to my
> tests. Also, the same antenna/LNA is feeding multiple GPSDOs (4 others)
> with Lady Heather connected and running continuously, and none of them
> exhibited any kind of strange behavior at the times of the anomalies in
the
> LN RB timing.
> 
> It certainly appears that Jackson Labs did not design the LN RB to be
stable in
> the presence of external temperature fluctuations. To be useful, this
seems
> to imply that it needs an external environmental temperature control.
> Furthermore, it doesn't seem to be stable near the top end of its
supported
> temperature range even when the external temperature is constant, so the
> external temperature needs to be fairly low (probably under 40C or so).
One
> application I was interested in is a low SWaP timing reference for
smallsats.
> Its sensitivity to temperature seems to make that application unlikely,
> because adding external temperature control would make the electrical
> power prohibitive, even if size and weight can be kept under control. As a
> lab reference, I expect I can implement an external temperature control
and
> it may work well enough. Overall, I am very disappointed with the product.
I
> have written to Jackson Labs about my concerns but I have not heard back
> from them.
> 
> Jim
>