[time-nuts] GNSS-disciplining a Ball FRS-C Rubidium Oscillator

[Francis Grosz]

Hi, John,

     Bob's suggestion is exactly on point, but if you have the right toys
you can automate much of the process.  As Bob says, get a GPS/GNSS module
that has a PPS output.  Then feed the 10 MHz from the Rb into one of TvB's
picDIV chips to get a 1 PPS (more or less) by dividing the Rb output by 10
million.  If you have a counter with the Time Interval function (e.g., HP
5334B), feed the two 1 PPS signals into that and it will output the time
between the two pulses every second.  Then, if the counter has a GPIB (IEEE
488) interface, you can feed that into a computer and collect the data into
a file for as long as you want and process it to get the frequency offset
between the Rb and the GPS, and other statistics if you want that.  Doing
it this way makes it easy to collect very long data sets and the resolution
of the counter is more accurate than eyeballing a 'scope.  Anyway, it works
for me.  Have fun!

                    Francis Grosz

-----------------------------------------------------------------------------
Original Message:

Message: 2
Date: Sat, 14 Jan 2023 19:47:14 -0500
From: Bob  Camp <kb8tq at n1k.org>
Subject: [time-nuts] Re: GNSS-disciplining a Ball FRS-C Rubidium
        Oscillator
To: Discussion of precise time and frequency measurement
        <time-nuts at lists.febo.com>
Message-ID: <69E2DAAB-2BF7-4FFC-9EA2-0400384FCCE9 at n1k.org>
Content-Type: text/plain;       charset=utf-8

Hi

Turns out that getting one set on frequency is not as insane as you might
think:

Grab a fairly cheap GPS module that puts out a PPS output.

Feed one channel of a scope with it ( or at least trigger the scope with
it).

Watch the 10 MHz sine wave move relative to the trigger

You can (eventually) adjust the Rb to the point that the 10 MHz does not
move.

Simple math:

The Rb is likely to already be within 1x10^-8 of the correct frequency.

If it’s that far off, it will move 1/10th of a cycle per second ( 10 MHz
period is
100 NS and 10 ns would be 1x10^-8 ….).

Watch it for 10 seconds and you can be reasonably sure it’s within 1x10^-8

Watch for 100 seconds and you are pretty sure of 1x10^-9 ….

Since your GPS will have some jitter, things get exciting if you start to
try to
go below 1/10th of a cycle. This isn’t a process that you can speed up a
lot.

At a bit over a day ( if you keep at it ) you get to 100,000 seconds and a
cycle is
going to be 1x10^-12. For most folks, that is into the “good enough” range.
For
the insane, you can keep at it for days / weeks / months. At the month
level, the
resolution of your adjustment pot will become a pretty big issue….. Yes,
there are
ways to attack that problem … :) :)

Bob

> On Jan 14, 2023, at 4:49 PM, John Miller via time-nuts <
time-nuts at lists.febo.com> wrote:
>
> Hello Everyone!
> A while back, a friend of mine who knows I'm into this hobby found a Ball
FRS-C RuXO at an electronics surplus shop local to him. It works great, and
outputs 10MHz with as many zeros as the best frequency counterI have access
to can measure. I don't really know what to do with it, but I have thought
about building a system to discipline it with a GNSS receiver. Any
recommendations, thoughts, etc.? I suspect this thing may be out of my
scope and I should find a new home for it.
>
> Here are photos of the unit:
https://photos.millerjs.org/?f=Ball+Efratom+FRS-C
>
> Thanks,
> John
> KC1QLN
> _________