MIME-Version: 1.0
Date: Thu, 3 Dec 2015 09:25:44 +0000
Message-ID: <CAA8k23Tetsrqo=RJf57bzcLiGY6WmMQar5kKA3EODBzAf+U6Vw@mail.gmail.com>
From: Andy Talbot <andy.g4jnt@gmail.com>
To: rsgb_lf_group@blacksheep.org
Subject: LF: Frequency References GPSDOs etc. (was EbNaut etc...)
Content-Type: text/plain; charset=UTF-8
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Status: O

Rubidium sources are not absolute references, they always have a
twiddle / calibrate control on them.  Typical adjustment range is
hardly ever more than 1 part per billion, and modern Rb sources will
easily stay at a couple of parts in 10^10 from cold turn on.   But
I'll leave you to work out the implications of even 1 part in 10^10
for several hours at 137kHz

GPS on the other hand makes use of Caesium (or at least properly and
regularly calibrated very stable Rb) on board the satellites and in
the LONG TERM is error free.    How long, 'Long Term' is, is debatable
but certainly over several hours you should expect no drift wrt. real
time / UTC.  However, the vagaries of propagation and the movement of
the satellite constellation and your GPS receiver introduce random
short term errors.

I once made a GPSDO based on a GPS module that delivered 10kHz (many
designs exist for these) and kept the time constant really short - a
few tens of seconds - so I could see what the GPS system itself was
capable of in real time.
See http://www.g4jnt.com/10MHz_Reference_Source_Stability.pdf   for
this and a comparison of several other 10MHz references
Also http://www.g4jnt.com/freqlock.htm for more on frequency locking

So what GPSDOs are not good for, unless they are really high spec, is
short term drift.  Low cost GPSDOs use a fast PLL for rapid lock up
and will be seen drifting 1 or 2, perhaps even a bit more, PPB over
minutes or tens of minutes.

The drift on GPSDOs is noticeable on 10GHz as a slowly wavering audio
tone, and on 47GHz it is quite objectionable I find.  Hence up there,
Rb is preferable as the possibly 47Hz absolute error is unimportant,
but an audio wobble is.

Somewhat ironically those low cost wobbling GPSDOs may even be the
best for LF long duration signalling.  Several PPB over the space of a
few tens of seconds or minutes it not terrible in terms of phase shift
at 137kHz, and we know that averaged out over longer periods it is
'exact'

Of course, if you were to GPS discipline a Rubidium source....
But don't forget the time constant to make that worth the effort.  A
poor GPSDO will ruin Rb's brilliant short term capability.

Andy  G4JNT

On 3 December 2015 at 09:01, edgar <edgarjtwining@virginbroadband.com.au> wrote:
> Hi Paul,
>
> GPSDO for the local oscillator?
>
> And why is a GPSDO better than using just a Rubidium Oscillator for EbNaut?
>
> Regards, Edgar
> Moonah, Tasmania.