Problem in executing benchmark_tx.py

GNU Radio
1

Hi,

I am trying to run benchmark_tx.py with USRP1 from one PC, and receiving
data with another USRP1 from another PC, with this some .dat files are
getting generated with some junk data.
My approach:
At TX PC: ./benchmark_tx.py --modulation=gmsk --size=1500
–megabytes=1.0 --from-file=payload.dat --freq=423e6 --bitrate=500000
–tx-subdev-spec=A --tx-amplitude=12000

At RX PC: ./benchmark_rx.py --modulation=gmsk --freq=423e6
–bitrate=500000 --rx-subdev-spec=A

Can anyone tell me what is wrong in my approach, (why junk data is
getting written) and how to resolve this issue.

Thanks,
Ashwini

2

On Mon, Feb 1, 2010 at 5:22 AM, ashwini [email protected] wrote:

Can anyone tell me what is wrong in my approach, (why junk data is getting
written) and how to resolve this issue.

We need more information on your set up to get an idea of what might
be going wrong
(http://gnuradio.org/redmine/wiki/gnuradio/ReportingErrors).
For example, what do you mean by ‘junk data being written’? The
benchmark files don’t really write any data, they simply check if the
frames were received correctly and erase them. Even if you use the
–log option I think you will only get the demodulated bits back in a
binary fashion, not a direct copy of ‘payload.dat’, but you aren’t
even doing that according to your commands.

I would spend some time looking through the appropriate files to
understand what the receiver is doing so you can find out where to
appropriately save your file to disk:
http://gnuradio.org/redmine/repositories/changes/gnuradio/gnuradio-examples/python/digital/benchmark_rx.py
http://gnuradio.org/redmine/repositories/changes/gnuradio/gnuradio-examples/python/digital/usrp_receive_path.py
http://gnuradio.org/redmine/repositories/entry/gnuradio/gnuradio-core/src/python/gnuradio/blks2impl/gmsk.py

The logging is done in the modulation specific files (gmsk.py), so you
don’t really have anything like framing or symbol->bit conversion done
yet, you will probably have to add a file sink later in the chain to
get the final data out (maybe there is an existing option for this?).

Jason