[hpsdr] Collecting Gibraltar thoughts

[Christopher T. Day]

Brain dump coming...

 

I want to collect the various ideas for the Gibraltar board so I can get
clearly in my mind what is being proposed. I come up with roughly three
sets of things.

 

1)       Producing a high-quality frequency reference.

2)       Using a high-quality frequency reference.

3)       Producing a high quality absolute time reference.

 

[I'm imagining two possible platforms, one as a single card in an
Atlas-based SDR and the other is the same card in a separate Atlas-based
box as a test instrument. The latter could have additional distribution
cards in it, and maybe a low-cost USB interface card.]

 

Case 3) would be useful for coherent CW and the like where the two
contact ends are not synchronized by the signal but by absolute clocks.
Cf. long baseline interferometry. Presumably, the frequencies need to be
accurately locked as well. I assume that this takes a GPS module with a
1 PPS output and a decoding of the NMEA time message or equivalent for
the clock part. A distribution card like the TADD-3 might also be useful
in an Atlas format. 

 

Case 2) would be using a reference to lock a VCXO with a PLL to make a
LO [don't ya' love these abbreviations?] for a VHF/UHF XCVR. Or to clean
up a DDS for a VFO. Or to measure the frequency of the SDR-1000's
reference oscillator and provide a correction value so the software
could accommodate the drift.

 

Case 1) seems to cover the widest range of possibilities. You might have
an external reference source like a cesium or rhubidium source that you
want to discipline against the GPS signal. This, as I understand it,
gives fantastic long-term stability but not so hot phase noise. Or one
could take an extremely good VXCO, like the Crystek. This gives slightly
worse long-term stability, though a few parts per ten trillion is
nothing to sneeze at, but much better phase performance. Either of these
would be good for microwave and lower frequency locking via 2). Or one
could use a less expensive VCXO or make one's own, discipline it to the
GPS and have something adequate for HF work.

 

What seems to be in common here, and which I presume goes on Gibraltar
is

 

1)       A good GPS module with 1 PPS output, preferably a timing model
that can go into fixed position mode and use any redundant satellite
signals to correct the time, not the position.

2)       A very low frequency PLL, minutes to hours, for disciplining
the reference to the GPSS 1 PPS.

3)       A more regular frequency PLL to use the reference to lock
something else.

 

So, footprints for a couple of GPS modules. A fair sized FPGA/CPLD for
the PLL chains. Some conditioning circuitry. Place for a very good or
pretty good VCXO/OCXO [where does the 15V come from?]. I/O for
reference, PLL control outputs, I/O for controlled oscillator, 1 PPS
I/O. Antenna feed with power to same. Does this seem like a reasonable
core for Gibraltar? Plus. TADD-1 and TADD-3 style fan-out boards.

 

 

            Chris - AE6VK

 

[Now to see what Steve actually has in mind...]

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20060523/531ebd96/attachment-0003.htm>