Re: Combined constellation modulator/Decoder do not seem to work

[George Edwards]

Hi Mike,

Again, thanks for the great help and sorry to bother you again. 

The Polyphase Clock Sync seems to have serious problems. It works well at 2 sps, but does not work for 4 or 8 sps. I replaced it with the recommended Symbol Sync block, which appears to work at the different sps based on constellation plot observations. However, when I tried to test the validity of its output against the input data, I couldn't find any delay to get a match. So I am left wondering if it is working well???

Any suggestions?

Thanks!

George

On Fri, Mar 12, 2021 at 8:23 PM Mike Markowski <mike.ab3ap@gmail.com> wrote:

Hi George,

You used differential encoding but forgot differential decode.  I added
the block in the attached.  (I disabled your file creation and displayed
to gui, easier for me to check.)

Little things are the hardest to spot.  :-)

Mike Markowski

On 3/12/21 4:46 PM, George Edwards wrote:
> Hi Mike,
>
> Thanks for your help!
>
> As I mentioned in my email, my flowgraph consists of some of the blocks
> in the example. I dropped some of the blocks because I am doing concept
> testing to ensure I know how to set parameters for the blocks. In
> terms of delays, that is not a problem. I am dumping the outputs into 2-
> file sinks, so in post processing I can compensate for that. The problem
> is that the tx and rx data are not the same, so I am missing in setting
> parameters.
>
> Thanks again!
> George
>
> On Fri, Mar 12, 2021 at 3:31 PM Mike Markowski <mike.ab3ap@gmail.com
> <mailto:mike.ab3ap@gmail.com>> wrote:
>
>     I can't make out details in your image, George, but attach one I made
>     when working through the tutorial and works just now under gnuradio
>     3.8.1.0.  Like someone else mentioned in another reply, I found the
>     delay to be 58 and set it as the default in the attached flowgraph.
>     Zoom in on the bit stream plot and allow a second or two after changing
>     delay to see the effect.
>
>     Hope this helps,
>     Mike Markowski
>
>     On 3/12/21 10:37 AM, George Edwards wrote:
>      > Hello,
>      >
>      > I cannot get the combined Constellation Modulator/Decoder as
>     shown below
>      > to work. The figure below is an example from the online Guided
>     Tutorial
>      > PSK Demodulation.... I leave out the Equalizer, Costas Loop
>     blocks, the
>      > channel was set to pass the signal unimpaired and the Poly Clock
>     Sync
>      > was set for Output sps=1. The constellation plot of the output of
>     the
>      > Poly Clock Sync is perfect. Now, in the flowgraph below, the signal
>      > entering the Constellation Modulator and that leaving the
>     Constellation
>      > Decoder are routed into Time Sinks for display. I route these
>     signals
>      > into File Sinks so I can print them and compare 1's and 0's. The
>     results
>      > show they are not the same. I passed Constellation Rect.Object
>      > parameters to both the Constellation Modulator and Decoder, which
>     seems
>      > logical to me, but I could be wrong???
>      > Will appreciate any help you can provide because what enters the
>      > modulator should be the same as what is leaving the demodulator.
>      > Regards,
>      > George
>      >
>      > image.png
>