[time-nuts] Time-Nutters-- Adding Rubidium to a Thunderbolt...?

[Bob kb8tq]

Hi

So backing up a bit: 

What are you trying to do? 

GPS signals by their nature are noisy short term. They get better the longer
you stretch out the observation time. If you are running a single band / uncorrected
GPS ( = most GPSDO’s and most GPS modules) you will hit a bunch of issues 
related to day / night sorts of things. 

Rb’s have pretty good stability and it also improves as you stretch out the observation
time. If you have a stable environment, they may keep getting better for quite a ways.
How much better for how long depends very much on the exact unit you have and 
how you treat it. 

If the objective is simply to get the Rb close to frequency and then shut down the
GPS stuff, that is a pretty simple thing to do. 

If you want to run GPS + Rb and not significantly degrade a fairly good Rb, things 
get a bit complicated. 

Bob

> On Mar 2, 2020, at 12:32 PM, Skip Withrow <skip.withrow at gmail.com> wrote:
> 
> Hello Mike,
> Replacing the OCXO in a Thunderbolt (or one of the many telecom
> cousins) is very easy to do.  Just yank out the OCXO, hook the
> Thunderbolt EFC to the Rubidium EFC (the Thunderbolt range is 0-6V),
> and hook the Rb 10MHz where the OCXO 10MHz went.
> 
> That is the easy part.  Understanding how to tune the Thunderbolt, and
> what kind of performance to expect is a little harder.  Rubidiums need
> a much longer time constant and the Thunderbolt only goes to 1000s.
> This can be circumvented by doing some math and 'fudging' the
> Thunderbolt gain and damping numbers.  There was a discussion on how
> to do this in time-nuts many years ago.  Also, a word of warning, some
> of the telecom cousins do strange things when this technique is used
> (and some don't).
> 
> The other thing to consider is the Thunderbolt DAC and Rb resolution.
> A PRS-10 takes the EFC and runs it into an A/D such that the minimum
> step in frequency is about 1x10E-12.  You will never get an ADEV below
> this value.  Many of the other rubidiums (X72, SA22c, FE5680A,
> FE5650A) use DDS in their loops and have similar resolutions.  The
> LPRO is one unit that is around has an analog EFC.
> 
> The Thunderbolt DAC steps in 10uV steps IIRC.  And it steps its DAC
> voltage every second.  With GPS signals jumping around you can still
> (will) end up with poorer short term performance than if the Rb was
> left to its own devices.  Hint, let the Thunderbolt do its thing with
> the Rb, but use Lady Heather and disable disciplining when making
> measurements/comparisons.
> 
> Regards,
> Skip Withrow
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.