High packet error and reception problems with RFX 2400


#1

Hi fellows,

We just purchased USRP from Ettus equipped with daughter board RFX 2400.
We
are trying to test these boards and have following problems:

  1. The benchmark_tx.py and benchmark_rx.py from the GNU radio tar ball
    3.0.3gave close to 100% packet error. During the experiment, we used
    the maximum
    gain for receiver antenna and maximum amplitude for transmitter. We are
    transmitting wirelessly using different settings of center frequency
    from
    2400 to 2475MHz and bit rate of 50k to 500kbps. We are using vertical
    antenna from Ettus for this band. We varied the distance between the tx
    and
    rx from few centimeters to 1m. At 1m, the receiver was unable to
    receive/decode any of the packets at at all. Similarly at low bit rates,
    the
    rx wan unable to receive/decode any of the packets. We appreciate any
    help
    or clue to this problem.

  2. We tried to use usrp_siggen.py and usrp_oscope.py programs to
    transmit/receive a sinusoidal signal. The oscilloscope program did show
    a
    sinusoid wave but the amplitude is weak and noisy. Even the response to
    the
    orientation and distance of the transmitter is random. For some
    orientation
    of the transmitter the received signal gets better for a moment and then
    goes down again.

Have anyone experienced this issue with RFX daughter board and vertical
antenna? Any help will be highly appreciated.

Regards

SFAS


#2

On Apr 3, 2007, at 4:12 PM, Syed Faisal Shah wrote:

50k to 500kbps. We are using vertical antenna from Ettus for this
band. We varied the distance between the tx and rx from few
centimeters to 1m. At 1m, the receiver was unable to receive/decode
any of the packets at at all. Similarly at low bit rates, the rx
wan unable to receive/decode any of the packets. We appreciate any
help or clue to this problem.

Have you tried wiring the radios together with SMA (and 40dB
attenuation)? If nothing else, this should work flawlessly before
you move on to wireless experiments.

What was the signal strength when you had the two boards a few
centimeters from each other? Just curious, What was the maximum gain
you used for the transmitter (i.e. how did you set it?) Another
variable in this problem could be the setting on the Rx PGA. Have
you modified this value? Most people have reported success at half
of maximum.

  1. We tried to use usrp_siggen.py and usrp_oscope.py programs to
    transmit/receive a sinusoidal signal. The oscilloscope program did
    show a sinusoid wave but the amplitude is weak and noisy. Even the
    response to the orientation and distance of the transmitter is
    random. For some orientation of the transmitter the received signal
    gets better for a moment and then goes down again.

What was the gain you used (-g option) for the usrp_siggen.py? Have
you tried using SMA cables?

Have anyone experienced this issue with RFX daughter board and
vertical antenna? Any help will be highly appreciated.

For what its worth, the RFX400 / RFX900 seem to work well with
vertical antenna, and I don’t see why the 2400 wouldn’t either.


#3

Hi David,

No, we haven’t tried connecting the two boards using cable. What do you
mean
by 40db attenuator? Is it something we can set on hardware or from the
software?

We set the signal amplitude to maximum using

./benchmark_tx.py -f 2412M --bitrate 500k -v --tx-amplitude=30000

We didn’t set the gain of the transmitter. How can we set the gain of
the
transmitter?

We tried with different rx gains, e.g.,

./benchmark_rx.py -f 2412M --bitrate 100k -v --rx-gain=75
but does not improve much at highest gain of 90.

We could not find the -g option with usrp_siggen.py. So there must be
some
default settings of tx gain. While looking at usrp_siggen.py, I found
this
line

fg.subdev.set_gain(fg.subdev.gain_range()[1]) # set max Tx gain

and hoping that the tx gain is at maximum. Please let us know how to
adjust
the tx gain in usrp_siggen.py or benchmark_tx.py.

We truely appreciate your help.

Regards

SFAS


#4

Just to clarify for everyone, since I get a LOT of calls about this…

The VERT2400 antenna is BLACK and made out of PLASTIC. If your
“vertical antenna” is copper colored, then you do NOT have a vertical
antenna. What you have is the RIGID COAXIAL CABLE which comes with the
Quad Patch 2400 MHz antenna. This coax needs to be soldered to the
GREEN PCB with 4 squares on it labeled “Highgain Antennas 2.4-2.48
GHz”. See the datasheet for the Quad Patch antennas for more info on
how to do this.

To reiterate –
The VERT2400 is BLACK PLASTIC
The VERT900 is BLACK PLASTIC
The VERT400/1200 is a thin BLACK WIRE ANTENNA

The copper colored rigid line is RIGID COAXIAL CABLE. It needs to be
soldered to the Quad Patch 2400, Yagi 1200, or Yagi 900 GREEN PCB
antenna which came with it.

Thanks,
Matt