Bpsk Simulation

GNU Radio
1

Hi,

I’m new to gnu radio. To get familiar with this project, I’m trying to
simulate a dbpsk link.

My flow graph look like this :
stream of bits => Modulation => frequency up translation => Noisy
channel
=> frequency down translation => Demodulation

But, there is something going wrong!! If the input stream looks like
this
: 0 0 1 1 0 0 1 1 0 0 1 1… the output will be ( after Costa’s loop
and
MM are stabilized ) : 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1…
I looked for an undesired decimation, but didn’t find any. Do you Have
any idea of what is going wrong?

Axel

My code :
class my_top_block(gr.top_block):
def init(self):
gr.top_block.init(self)
fg = gr.flow_graph()

sps = 8
symbol_rate = 1625000.0 / 6.0
sample_rate = sps * symbol_rate
p_size = 1024
lo_freq = 7.5e5
lp_cutoff = 2.5e5
lp_tw = 5e5
noise_mag = 0.0
#Source
       self.src_data = (big stream of : 0 0 1 1 0 0 1 1 0 0 1 1 0 0

1 1)
self.src0 = gr.vector_source_b (self.src_data)
######

###Modulation####
       self.objMod = dbpsk.dbpsk_mod(sps,0.35,False,False,False)
#################

### Canal########
# Mix to IF
self.lo = gr.sig_source_c(sample_rate, gr.GR_SIN_WAVE, lo_freq,1.0,

  1. self.mixer = gr.multiply_cc()
    self.connect(self.lo, (self.mixer, 1))

    Simulate noise in the channel

    self.noise = gr.noise_source_c(gr.GR_GAUSSIAN, noise_mag)
    self.air_noise = gr.add_cc()
    self.connect(self.noise, (self.air_noise, 1))

    Mix to baseband

    self.chan_taps = gr.firdes.low_pass(1.0, sample_rate, lp_cutoff,
    lp_tw,
    gr.firdes.WIN_HAMMING)
    self.chan_filter = gr.freq_xlating_fir_filter_ccc(1,
    self.chan_taps,-lo_freq, sample_rate)
    ###################

    Demodulation######

    self.objDeMod=dbpsk.dbpsk_demod(sps,0.35)
    ######################

    ######Connection des blocs#####
    self.connect(self.src0,self.objMod, self.mixer, self.air_noise,
    self.chan_filter, self.objDeMod)
    #####################

################# Observation entree/sortie########
self.sink1 = gr.vector_sink_b()
self.connect(self.objDeMod,self.sink1)
self.sink2 = gr.vector_sink_b()
self.connect(self.src0,self.sink2)

def print_data(self):
print "demod: ",self.sink1.data()
print len(self.sink1.data())
print "check: ",self.sink2.data()
print len(self.sink2.data())
################################################
if name == ‘main’:
try:
tb=my_top_block()
tb.run()
tb.print_data()
except KeyboardInterrupt:
pass

2

On Thu, Mar 11, 2010 at 4:09 AM, Axel B. [email protected]
wrote:

: 0 0 1 1 0 0 1 1 0 0 1 1… the output will be ( after Costa’s loop and
MM are stabilized ) : 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1…
I looked for an undesired decimation, but didn’t find any. Do you Have
any idea of what is going wrong?

Axel

This is a bit confusing, but it makes sense in the way we meant for
the digital modulators to work. The problem is that you are trying to
put in “bits” into the modulator, which actually takes a “packet”
stream of bits. In other words, when you think you are putting in the
bits “0 0 1 1,” you are actually telling it that you are putting in
the bytes “0 0 1 1” that gets unpacked into the bit stream “0 0 0 0 0
0 0 0, 0 0 0 0 0 0 0 0, 0 0 0 0 0 0 0 1, 0 0 0 0 0 0 0 1” or “0x00
0x00 0x01 0x01”.

What you want to put in as the source is a vector of “0x33” instead.
You can also use the gr.unpacked_to_packed_bb on the demod side to
repack the bits. I’ve edited your code below to show how to do this.
Now, when you do this, you’re still going to see some random stuff
going on in the output stream for the first few symbols. These garbage
bits are due to the filters’ group delays.

One final thought. You’re doing a frequency shift operation on the
complex baseband signal and then almost immediately downconverting it.
This doesn’t really do anything of any consequence to your simulation
except require you to use a larger sampling rate than really required.
If you want to test frequency offset errors, you can use the features
in the channel model that we have (blks2.channel_model), which takes
in a noise voltage, frequency offset, timing offset, and a vector of
taps to simulate multipath.

           lo_freq = 7.5e5
           lp_cutoff = 2.5e5
           lp_tw = 5e5
           noise_mag = 0.0
           #Source

              N = <something>
              self.src_data = N*[0x33, ] # make an N-long vector

with 0x33 in every element

           self.mixer = gr.multiply_cc()
           ###################

           ### Demodulation######
           self.objDeMod=dbpsk.dbpsk_demod(sps,0.35)

             self.pack = gr.unpacked_to_packed_bb(1,

gr.GR_MSB_FIRST)

           ######################

           ######Connection des blocs#####

            self.connect(self.src0,self.objMod, self.mixer,

self.air_noise,
self.chan_filter, self.objDeMod, self.pack)

           #####################

   ################# Observation entree/sortie########
           self.sink1 = gr.vector_sink_b()

              self.connect(self.pack, ,self.sink1)
3

Hi,
I’m also having issues with the use of this block if I try the same
concept. I’m not concerned with the noise generation and just want to
modulate and then demodulate data. I am using the vector creation and
packing as stated above. The results I’m getting are deterministic,
just incorrect. Here is what i’m using:

099 self.src_data = 3 * [0x40,0x41,0x42,0x43,]
100 self.src0 = gr.vector_source_b(self.src_data)
101 self.mod = blks2.dbpsk_mod(8,0.35,False,True,False)
102 self.demod =
blks2.dbpsk_demod(8,0.35,0.1,None,0.5,0.005,False,True,False)
103 self.pack = gr.unpacked_to_packed_bb(1,gr.GR_MSB_FIRST)
104 self.connect(self.src0,self.mod,self.demod,self.pack)
105
106 self.sink1 = gr.vector_sink_b()
107 self.connect(self.pack,self.sink1)
108
109 self.sink2 = gr.vector_sink_b()
110 self.connect = (self.src0,self.sink2)
111
112 def print_data(self):
113 print "data: ",self.src_data
114 print len(self.src_data)
115 print "demod: ",self.sink1.data()
116 print len(self.sink1.data())
117 print "check: ",self.sink2.data()
118 print len(self.sink2.data())

Here is the output:

gr_fir_ccf: using SSE

Modulator:
bits per symbol: 1
Gray code: False
RRC roll-off factor: 0.35

Demodulator:
bits per symbol: 1
Gray code: False
RRC roll-off factor: 0.35
Costas Loop alpha: 1.00e-01
Costas Loop beta: 2.50e-03
M&M mu: 0.50
M&M mu gain: 1.00e-01
M&M omega: 8.00
M&M omega gain: 2.50e-03
M&M omega limit: 0.01

data: [64, 65, 66, 67, 64, 65, 66, 67, 64, 65, 66, 67]
12
demod: (1, 239, 1, 5, 9, 13, 1, 5, 9, 13, 1)
11
check: ()
0

Thanks in advance for any help