Problem with QAM data spreading due to LPF

[silver_sparrow]

QAM Spreading due to LPF

Hi

I am modulating OFDM symbols,
sending it out to filters,
performing low pass filtering with proper BW specs. ,
then sampling and demodulating the signal by taking the FFT
the plotting the QAM symbols on the constellation.
--no channels are involved.

If you would please have a look at the constellation plot attached....why
does the data spread more for the QAM data points further away
from the center due to low pass filtering? can this be mathematically
explained?

I think this stuff is fundametal knowledge, but I am having troubles finding
the answer on the internet.

thnx!
sparrow

 

[zorro]

QAM Spreading due to LPF

Hi sparrow,

It doesn’t seem to be intersymbol interference.
Are the data really random?

Z

 

[flatulent]

Re: QAM Spreading due to LPF

One thing to notice is that the spreading is proportional to I and Q signal amplitude and not angle or radius from the center.

Also the spreading is in all directions equally for all points. That is, it is totally circular and not elliptical shaped.

Is this a simulation or measurement of real hardware?

 

[dora]

Re: QAM Spreading due to LPF

Hi silver_sparrow,

As the FFT is perfectly reversible the only reason which I see for the constelation spreding is
the low pass filter you have used.

Can you describe your model parameters more in details.
If your model runs on matlab and if you can share it please uploade it.

Best Regards
dora