[time-nuts] New pics of RFTG-m-Rb, and some comparison details

David I. Emery die at dieconsulting.com
Sun Dec 31 16:15:17 EST 2006


On Sun, Dec 31, 2006 at 09:22:50AM -0600, Jason Rabel wrote:
> Pretty sure that the RB is disciplined from the XO through the interface
> cable. I posted the pinout a couple times before so people can make their
> own.

	While I may have too much shoe leather in mouth to think
clearly, it seems that the following reasoning certainly leads to a
disciplined Rb...

	Presumably the design is intended to supply timing and frequency
reference to a cell site just as reliably as possible in the event of
various kinds of failures.

	Clearly the most common failure is loss of GPS (birds,
lightning, moisture in the antenna assembly, damage to or water in
cables, rodents, local interference, problems with the GPS constellation
etc.)

	And clearly even quite a good OCXO isn't going to meet the
holdover spec for more than a very few hours (1 microsecond drift in 1
day is 1.157 parts in 10^11).   If you can't fix it after that you're
screwed and the site goes down.

	So the logical thing to do is use a rubidium reference which can
very easily have less than one part in 10^11 drift per day and typically
much better than that.   This gives much more time to repair, and if the
Rb is doing well that may extend into several days or even a week or
more.  Enough time to get to the site and fix the antenna, cables,
etc...

	But obviously the only way the Rb is useful is if it is
disciplined somehow so it starts out dead on.   Relying on the factory C
field setting to be within parts in 10^11 years later is ridiculous
considering variations in local magnetic field, aging etc... it is
probably not possible to do much better than parts in 10^9 in a factory
shipped without field adjustment situation in fact.

	And it is quite clear that the dual module design is
deliberately engineered for a reason - each module can independently
produce the 1 PPS and 15 MHz needed and they are independent boxes that
can be hot swapped without taking down the cell site.   So if the GPS
(XO) box dies the site can and will continue with 1 PPS from the Rb and
Lits 15 MHz while a new XO box is hot swapped in (and presumably the XO
fails more often due to GPS receiver failures from lightning hits).

	Further, if the more complex and less reliable Rb dies a new
(RB) module can be swapped in without killing the site either.   All
that is lost is holdover performance.

	And at least possibly the phase noise of the OCXO beats the LPRO
so optimum site performance is with the OCXO rather than a Rb reference.
Thus the normal mode may be to use the 15 MHz from the OCXO with the 15
MHz from the Rb as backup.

	The big question is why disciplining the OCXO requires a 16 bit
A/D and disciplining the Rb does not...  of course disciplining the Rb
quite possibly only works with both modules present - it is not
completely clear that the stand alone m-RB module will lock to a random
1 PPS since there is no obvious need for this in the overall system
architecture.   It is interesting that 10 MHz from the Rb is an input to
the XO - whether this is to supply redundancy in the 15 MHz/1 PPS or
just to allow the XO to work from another master source of 10 MHz isn't
clear (or perhaps the 10 MHz input to the XO from the Rb is required to
get the Rb to discipline).

-- 
   Dave Emery N1PRE,  die at dieconsulting.com  DIE Consulting, Weston, Mass 02493
"An empty zombie mind with a forlorn barely readable weatherbeaten
'For Rent' sign still vainly flapping outside on the weed encrusted pole - in 
celebration of what could have been, but wasn't and is not to be now either."




More information about the time-nuts mailing list