Hi to all,
I am a beginner in using GNU radio(GRC). I have made a DQPSK transmitter
using the blocks of GNU radio (attached Figure QPSK_1). when I connected
the constellation plot with it, It gave me the a very scattered
Constellation Plot Instead of four Clean Constellation Dots (attached
Figure QPSK_2). I have checked with all forms of signals but the
constellation remains distorted.(i.e. vector source, signal source etc.)
Is it normal plot or am I missing some thing? Kindly help.
Regards
Salman D.
On 01/14/2015 04:56 AM, Salman D. wrote:
Hi to all,
I am a beginner in using GNU radio(GRC). I have made a DQPSK transmitter
using the blocks of GNU radio (attached Figure QPSK_1). when I connected
the constellation plot with it, It gave me the a very scattered
Constellation Plot Instead of four Clean Constellation Dots (attached
Figure QPSK_2). I have checked with all forms of signals but the
constellation remains distorted.(i.e. vector source, signal source etc.)
Is it normal plot or am I missing some thing? Kindly help.
My guess is yes, it’s a normal plot, and you have no time sync in there.
M
It’s a constellation plot for sure - says it right on the block in the
PNG, but it’s all internally generated anyway. There should be 1
perfect phase, then 3 out of phase points. No noise - no
phase/frequency offset. Timing shouldn’t be an issue here even if it
were a normal plot.
I think the issue is no matched filter on the other side. RRC
filtering won’t look very appealing to the eye, but RC filtering
should match what you expect.
Try adding a matched RRC filter before you do the constellation. You
should see the 4 samples/symbol and their trajectories as they
converge onto the 4 nice points.
Brian
On Wed, Jan 14, 2015 at 8:50 AM, Brian P. [email protected]
wrote:
Try adding a matched RRC filter before you do the constellation. You
should see the 4 samples/symbol and their trajectories as they
converge onto the 4 nice points.
Brian
Yes, indeed, Brian is correct. The modulator block uses an RRC
pulse-shaping filter that also interpolate to some number of
samples/symbol. So what you are seeing is due to the ISI (inter-symbol
interference) as a result of the RRC filter. Applying a matched RRC
filter
will remove (most of) the ISI.
Tom