[Discuss-gnuradio] BPSK constellation failing

[Jason Uher]

Hi all,

I am currently trying to get plot the received bpsk samples and look
at the constellations as received by the USRP, but all I get a big
'ole blob of points in the middle.

I think the main issue probably has to do with the interpolation and
decimation rates so I included that code at the bottom, the whole
source is attached along with a symbol plot.  I

In the symbol plot you can clearly see the received bits on the 'unit'
circle, I was wondering why there are so many bits clustered at the
crossing point; I was expecting a big clump above/below 0 (which I
have) and almost no points at 0.

My questions are:
    1.  Is this the right way to find the decimation and interpolation
rates for a given bit rate (I tried to base it off of pick_bitrate.py)
    2.  Using only one USRP, will there be significant (any?)
constellation rotation?
    3. Why are there so many ambiguous bits (maybe same as #1)


data_rate = 250000
bits_per_symbol = 1 # 1:bpsk 2:qpsk 3:8psk, etc

#Calculated generic parameters
bit_rate = 8 * data_rate
symbol_rate = bit_rate / bits_per_symbol
arity = pow(2,bits_per_symbol)

#User defined transmit parameters
tx_samples_per_symbol = 1
tx_converter_rate = 128e6  #Default DA converter rate
tx_interp_rate = int(math.ceil( tx_converter_rate /
tx_samples_per_symbol / symbol_rate ))

rx_samples_per_symbol = 1  #How many rx samples are devoted to one symbol?
rx_converter_rate = 64e6 #Default AD converter rate
rx_decim_rate = int(math.ceil( rx_converter_rate /
rx_samples_per_symbol / symbol_rate ))

capture_bpsk_symbols.py
Description: Binary data

bpsk_constellation.jpg
Description: JPEG image