[time-nuts] DOCXO vs. Rubidium medium-term stability

jimlux jimlux at earthlink.net
Fri Nov 19 21:30:37 UTC 2010


WB6BNQ wrote:
> Dave,
> 
> Something is not making sense to me here.  As GPS is generally available around
> the globe and obviously to your reference stations; how is it that the mobile
> will be able to find an area where the GPS is not available ?

If GPS is jammed, you're in a high multipath area, or other reasons... 
Maybe the reference stations are above water, but your mobile unit is 
underwater during most of the data collection.  Or if you're doing 
underground surveying.. not necesarily well logging, but say you're 
doing Electromagnetic surveys,

I can think of lots of scenarios needing this..


> 
> As to the mobile, if it is not going to utilize the GPS for a reference, you then
> need to determine the worst case error you can have over the course of time that
> the mobile is away from its GPS capability.  That factor will dictate the kind of
> on-board reference you will need.  It could be that a very good quality crystal
> oscillator will suffice.

Indeed.. if you're looking at times <24hrs, a good OCXO would probably 
do it.


> 
>>
>> I'm working on an data acquisition application for my company that will
>> require a very stable oscillator. Without going into too many specifics,
>> there will be some reference stations spaced 100's of kilometers apart from
>> each other and 1 mobile station that will be operated in an area where GPS
>> is not available. I need to be able to collect data at all the stations and
>> have the time synchronization be extremely good between the stations,
>> including the one without GPS.
>>
>> For the reference stations it will be sufficient 

Sufficient meaning you need tens of nanoseconds sort of precision/accuracy?


>> The mobile station can be synchronized to GPS initially to synchronize its
>> clock as described above but will then have to rely on a free-running
>> oscillator. The stability of this oscillator will dictate how much drift the
>> mobile station's clock will experience relative to the reference stations.
>> Keeping this drift as low as possible is my goal.

How good does it have to be, and how long will you be "out of touch"...

"as low as possible" is pretty darn low in this crowd.. Do you need 1 
part in 1E13 over a week?

Or 1 ppb over a day? (about 100 microseconds/day)

Are you recording RF/Acoustic signals and need to be able to form 
coherent sums?


>>
>> I had assumed that a Rubidium oscillator would give me the best stability
>> over the course of 8-12 hours. Obviously a Cesium would be better but those
>> are impractical due to cost and power constraints. I've begun evaluation of
>> a Rb oscillator but now I'm being told by some people that a good DOCXO is
>> likely to give me similar medium term stability (with obviously better
>> short-term stability).

Almost certainly... the advantage of an Rb is that you can turn it off, 
then turn it back on days later, and in a relatively short time, have 
decent absolute accuracy.
>>



More information about the time-nuts mailing list