I want to do a direction finding application in the 462 Mhz spectrum
(FRS,
GMRS), and need a coherent set of receivers to process the signals.
Preferably from 3 or four antennas. The bandwidth is only 200 Khz for
the
whole band.
My questions are: Can I use 4 receivers with a USRP?
What receiver solution can I use that will maintain the phase
relationship
between the antennas? Is anyone working on such a thing. I
unfortunately,
am a software guy who knows little about RF design. I gather the
current
TVRX won’t work because of the individual local oscillators.
I gather the current TVRX won’t work because of the individual local
oscillators.
No such luck.
You are correct, two TVRX’s wouldn’t work because their LO’s are not
coherent. Also, you wouldn’t get 4 receivers (the USRP only takes two
RX boards and two TX boards.)
There is a firmware load for the FPGA that allows you to treat the two
inputs on the basic RX independently, and use two basic RX boards to get
four independent streams. Since these boards don’t do down-conversion,
the sampling clock on the motherboard is coherent across all the
streams.
Alas, the input bandwidth to the ADC sample-and-hold bandwidth is 3db
down at 200 MHz, and of course gets worse from there. Also, there is no
gain on these boards. So theoretically you could build an external
receive chain that filters out the image bands and amplifies the 462 MHz
band you want, then feed four of these into the two basic RXs, and
undersample to get your four coherent streams of data.
Unfortunately, that’s a lot of analog electronics in front of the ADCs
that must be closely matched in amplitude and phase delay in order to
preserve the accuracy of phase relationships between the DF antennas.
I am working on the design of one of these to cover the 2M ham band and
plug into the USRP, as a follow-on to my current pseudo-Doppler
controller that uses an external radio and the GNU Radio code base.
On Sun, Oct 29, 2006 at 05:42:46PM -0800, Johnathan C. wrote:
that must be closely matched in amplitude and phase delay in order to
preserve the accuracy of phase relationships between the DF antennas.
The usual approach in this case is to add “pilot tone” injection,
electrically close to the antenna (or even as part of the antenna
system). The pilot tone can be continuous or switchable. Then you
can calibrate all that analog electronics in-situ.
As long as you have to build custom electronics, you will get much
better
results doing an actual downconversion, than you will relying purely
on undersampling.
- Larry
This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.