[time-nuts] GHz output from fe5680a
magnus at rubidium.dyndns.org
Tue Jan 31 19:16:34 EST 2012
On 31/01/12 07:27, Javier Herrero wrote:
> El 31/01/2012 02:52, Magnus Danielson escribió:
>> On 31/01/12 02:20, Bob Camp wrote:
>>> Simple answer is don't bother.
>>> More complex answer - of course you can. You will get a forest of
>>> spikes (every 5, MHz or so ).
>> Also, since the 5,3 MHz is modulated, 1,4 kHz apart is the two
>> signals, and then a skirt of side-bands as the frequency modulation is
>> relatively quick.
>> It's not clean.
>> BTW. A fun hack would be to hook up the 63,8976 MHz OCXO in
>> replacement of the 60 MHz and then re-adjusting the DDS to 2.35645
>> MHz, as the 107s overtone of the OCXO minus the new DDS frequency hits
>> the Rb resonance.
> And then we will have a not less fun 10.6469MHz and 0.939Hz outputs :)
> However, if the DDS can easily be reprogrammed to large offsets (and the
> output filter is a simple lowpass...), that idea is very useful if you
> need an strange frequency and have an OCXO at 6x that extrange frequency ;)
We seems to get those OCXOs alongside anyway, so I was just toying with
the idea to see what kind of performance that would give us. A quieter
OCXO. For more ordinary frequency output, an additional DDS like the
original 5680 would be nice.
More information about the time-nuts