Obviously they would chose a variable frequency output if a fixed output was not available in the frequency they needed.
And most equipment which can operate off of an external reference phase locks a low phase noise internal reference to the external reference anyway. All of my timer/counters do this.
=====snipped=====
As does my own much modified "Toy" FY6600-60M AWG with its CQE 10MHz OCXO (plus 3N502 clock multiplier) inspired by Arthur Dent's own OCXO upgrade postings into the FY6600 topic thread.
Arthur had chosen a simple mechanical switching of external/internal references but concerns (possibly unfounded) that the inevitable glitches this would introduce onto the FPGA 's 50MHz clock input might cause lock ups, led me to use the external 10MHz reference to injection lock the internal 10MHz OCXO to avoid the possibility of such lock ups (glitchless, smooth transitions to and from the external reference).
I'd totally missed the even better PLL solution in my researching frequency injection locking solutions until
after I'd successfully implemented my own frequency injecting module
The only reason I've not re-implemented the external reference input module as a PLL controller is simply because the injection circuit works so well and ICBA to rip it out and start over
Both methods have the distinction of automating the change-over and eliminate an unnecessary mechanical change-over switch. When the existing internal reference is already a high quality low phase noise oscillator (in this case, the original cheap 'n' nasty SMD 50MHz XO most certainly wasn't!), injection or phase locking it to an external low phase noise atomic standard (GPSDO or RFS) is the obvious "no-brainer" solution - surely all such test gear with an external reference socket implements a PLL or injection locking methodology?
The FE5680A based diy project videos that pollute YouTube have just the same long term frequency stability as the Efratom LPRO 101 units but their choice of DDS to generate the required locked to the XO reference hyperfine transition frequency of the rubidium vapour cell does introduce unwanted spurs and jitter on the XO output making it unsuited for use directly as a reference for use by GHz rated transverters unlike that of the LPRO101's 10MHz output. However, the use of a "Clean up" VC(OC)XO phase locked to the FE5680A's 10MHz output neatly solves that issue.
The complicated, DDS free, method used by the Efratom design to generate the hyperfine rubidium transition frequency from a 20MHz VTXO doesn't allow an exact frequency match, requiring a tiny 'bending' of the rubidium vapour cell's frequency output via the C field control which is a necessary 'evil' to fine tune all of these secondary atomic standards regardless of their design.
Even though in principle the FE5680A could eliminate such fine tuning, courtesy of the DDS technology, they still require the use of the C field for fine tuning the rubidium hyperfine transition frequency - rubidium vapour lamps and cells are at the mercy of the buffer gasses used and their behaviour in regard of temperature and pressure, making them 'secondary standards' in more ways than just the obvious one of not being caesium upon which the SI second unit of time is based upon.
As for generating frequencies other than the 10MHz output from a RFS, the obvious solution is to use a good quality DDS based AWG or RF generator equipped with an external 10MHz reference input socket to phase lock their on-board low phase noise VC(OC)XO to, which permits even the direct use of an FE5680A, spurs and all.
I'm currently in the middle of my own Efratom LPRO101 based RFS project (namely housing it in an instrument case sized enclosure to allow me to tightly control its baseplate temperature (and, ultimately add barometric pressure compensation) using a PWM controlled air re-circulation fan (unventilated enclosure design). I've been on yet another quest for best practical methods on PWM fan speed control which has proved, as with so many previous internet research efforts, to be more like (to mix metaphors) a hunt to pull hens' teeth from out of the jaws of Unicorns. As for the question of RFS baseplate temperature control, forget it! There's absolutely fuck all to be found on
that subject!
Anyway, discovering this topic thread was simply the side effect of my internet search for ideas and enlightenment on how to achieve the best possible stability from my RFS project. You have the staggering lack of interest on the part of DIYers building their own rubidium lab frequency references from used surplus RFS kit to the best possible standard over and above simply acquiring one to plonk into a case and sit on a handy shelf in their workshop/lab to thank (or curse) for my presence here.
John
Home
Help
Search
Login
Register
