I’m trying to “carve off” about 250Khz of bandwidth around 38.5MHz,
(actually any frequency
between about 25MHz and 40MHz) using a Basic_RX and a USRP2.
In GRC, I created a simple flow-graph with a USRP2 source, and an FFT
sink.
I set the tuning parameter of the source to 38.5MHz, and what I get on
the FFT is a flat-line at
-410dB! Now, granted, I don’t have the inputs connected to anything
yet, but I would expect there
to be some noise, and -410dB seems, well, like bad physics 
Cribbing from hfx2.py in gnuradio/examples/python/apps it seems that
this should “just work”.
Am I missing something? I thought that the FPGA would “do the right
thing” in the case of the Basic_RX,
and use the DDC to pull the “target” frequency down around DC, etc,
etc.
–
Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
On Mon, May 31, 2010 at 15:03, Marcus D. Leech [email protected]
wrote:
I set the tuning parameter of the source to 38.5MHz, and what I get on
the FFT is a flat-line at -410dB! Now, granted, I don’t have the inputs connected to anything
yet, but I would expect there to be some noise, and -410dB seems, well, like bad physics
The Basic RX has no gain. It’s very likely your noise floor is either
below the LSB of the ADC, or the DDC is filtering the noise floor to
below the LSB of the filtered sample stream. Add a file sink and look
at the actual sample values.
Or, try connecting something to the input
Johnathan
You are seeing the lower bound clipping in the nlog10 block. Try looking
at it with the scope sink, you should see some bits twiddle in the wind.
-Josh
On 06/01/2010 11:38 AM, Josh B. wrote:
You are seeing the lower bound clipping in the nlog10 block. Try
looking at it with the scope sink, you should see some bits twiddle in
the wind. -Josh
II puut 40dB of 50MHz-low-pass-filtered gain in front of the Basic_RX,
and now it’s “rational”.
But I do have another question. Is it possible to treat the two
inputs to the Basic_RX as
two largely-independant “real-sampled” streams? If I wanted to use
one of the inputs as
a real-sampled signal of the “outside world”, and the other input as a
real-sampled reference
stream, is that possible? The two streams would be at the same center
frequency. Will
the DDC and half-band filtering interfere with the “independence” of
the two streams?
–
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org
On 06/01/2010 02:47 PM, Josh B. wrote:
go with the dual usrp source in grc, setup the antennas and sides
correctly and it should work
-josh
Hmmmm, OK. But this is a single BASIC_RX card. Is that still going
to work?
–
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
go with the dual usrp source in grc, setup the antennas and sides
correctly and it should work 
-josh
On Tue, Jun 1, 2010 at 11:47, Josh B. [email protected] wrote:
go with the dual usrp source in grc, setup the antennas and sides correctly
and it should work
I think he’s using a USRP2. Isn’t the dual source USRP1 only?
What he’s asking for is a dual, coherent DDC for the I and Q inputs of
a BasicRX in a USRP2. This is a common enough request that I think it
would be useful to put into the stock FPGA functionality of UHD, using
VRT to separate the samples into two separate streams.
Johnathan
What he’s asking for is a dual, coherent DDC for the I
Discuss-gnuradio Info Page
If there is going to be a discussion on adding more functionality to the
USRP2, then I would like to endorse the development of another
independent DDC in the FPGA of the USRP2. I currently use 2 LFRX
daughter cards in the USRP1 to obtain 4 independent real channels; I
would love to be able to do the same with 2 USRP2’s connected via a MIMO
cable.
Thanks,
eric
On 06/01/2010 03:15 PM, Johnathan C. wrote:
Yes, that’s precisely what I want. I want to be able to treat the two
inputs of a Basic_RX as
real-sampled data that are largely independant, but they’re both tuned
to the same frequency.
This is for precision radiometry, with a “outside world” channel and a
“reference” channel.
There are a couple of approaches–one using a reference channel, and the
other switching the
entire gain chain between “outside world” and “reference”. I want to
explore both options.
The switching option would use a single input, with the input
switched.
–
Principal Investigator
Shirleys Bay Radio Astronomy Consortium