Weird (or maybe not) received signal amplitude behaviour

Hi List,

I’m transmitting a pure sinusoid of 50kHz using the tx_waveforms UHD
example and receiving the same using rx_samples_to_file example. I’m
using N210/WBXs for Tx and Rx at 250MHz, at 500 ksps, --ampl=1.0,
buffer size of 200 at Tx, and 0 dB for Rx,Tx gain.

I plot the received file (the I stream, after de-interleaving the
complex shorts) using gnuplot and here’s what I see. The peak to peak
amplitude is 40. The wave’s dancing nicely about 0 and its a
spectacular sine wave at almost precisely 50 kHz. But here’s the
catch: I can see this happening only after about sample # 8000
thereabouts in the graph. Before this, the wave starts out (at sample

10 ) at 1400-1440 peak to peak and gradually makes its way

down(maintaining its sinusoidal shape during the descent) to settle at
-20 to +20 peak to peak. It takes 8000 samples for this to happen, or
8000*2us = 16ms.

Why is this happening?

Arya

P.S. At Tx/Rx gains of 15 dB, same thing happens but the signal starts
at around 1000-2000 p2p and settles at -500 to 500 at around sample
#6000.

On 02/24/2011 10:13 PM, Arya S. wrote:

Hi List,

I’m transmitting a pure sinusoid of 50kHz using the tx_waveforms UHD
example and receiving the same using rx_samples_to_file example. I’m
using N210/WBXs for Tx and Rx at 250MHz, at 500 ksps, --ampl=1.0,
buffer size of 200 at Tx, and 0 dB for Rx,Tx gain.

Can you repeat the results with the default amplitude? An amplitude of
1.0 is maximum, and could lead to clipping in the FPGA DUC.

-josh