HELP! Is it a defective USRP1?

Hi All:

I am experiencing something strange with USRP1 and already tried a lot
of
approaches to diagnose where the problem is, but no luck so far.

We have two USRP1, each of which has a RXF 2400 daughter board and an
antenna connected to the Tx/Rx port. When we run benchmark
(benchmark_tx.py
and benchmark_rx.py), the packet loss/corrupted rate is very random
(from 0%
to 99%).

We’ve tried several approaches to diagnose where the problem, including
(1) connect the Tx/Rx ports of two USRP by a SMA connector
(2) try both sides of the mother board (A and B) ,
(3) use different central frequency from 2.4G to 2.48G,
(4) adjust the carrier threshold (20 to 50), tx amplitude (0 to 1) and
rx
gain (0 to 90),
(5) place the radios in different locations,
(6) use different versions of gnuradio (stable release version 3.3.0 as
well
as the latest repository code)
(7) place a fft sink at the rx to observe the spectrum, I can see the
signal
spectrum showing up when tx is transmitting something.

Unfortunately, nothing really helps.

Is it possible that it is a defective USRP?

I really appreciate if anyone of you could guide me with this problem
because I have a deadline to make in around 2.5 weeks.

Thanks very much!

Rachel

Try a higher rate (spectrally wider symbols). Maybe its an issue of
frequency offset. -Josh

Thanks Josh! Problem solved after I adjust the bit rate.

One more question, what is the maximum bit rate in USRP1? Also, what is
the
default sampling rate in benchmark_tx.py and benchmark_rx.py?

Based on Nyquist theorem, the sampling rate should be at least twice of
the
baseband bandwidth, is the bit rate equal to the bandwidth of received
signal? or there is something in the middle that I am missing?

Thanks,
Rachel

On 11/09/2010 05:34 PM, Rachel Li wrote:

We’ve tried several approaches to diagnose where the problem, including

My suggestion is that you use a more basic diagnostic. Do both USRPs
have roughly the same
sensitivity and output power?

If you transmit on one, and receive on the other with a FFT, with an
attenuator between them,
can the Rx side “see” the other side, use increasing levels of
attenuation until the Rx can’t
see the transmitted signal anymore.


Principal Investigator
Shirleys Bay Radio Astronomy Consortium