Prev: FFT for ADC output
Next: forward error correction capabilities?
From: alberto.fuggetta on 15 Jun 2010 16:27
Hi,

I'm simulating a DFE equalizer in Matlab for a MSK modulated signal.
The channel I'm testing is Ricean with max doppler shift of 50 Hz.
I noticed that the performance of the DFE deeply improves if I apply the
derotation before sending the samples to the equalizer.
Has anybody ever observed the same behavior?
Does any theoretical explanation for this phenomenon exist?
Thanks

Alberto
From: Vladimir Vassilevsky on 15 Jun 2010 16:30


alberto.fuggetta wrote:

> Hi,
>
> I'm simulating a DFE equalizer in Matlab for a MSK modulated signal.
> The channel I'm testing is Ricean with max doppler shift of 50 Hz.
> I noticed that the performance of the DFE deeply improves if I apply the
> derotation before sending the samples to the equalizer.
> Has anybody ever observed the same behavior?
> Does any theoretical explanation for this phenomenon exist?

Your model of the channel is wildly inadequate.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

From: Clay on 16 Jun 2010 15:04
On Jun 15, 4:27 pm, "alberto.fuggetta"
<alberto.fuggetta(a)n_o_s_p_a_m.gmail.com> wrote:
> Hi,
>
> I'm simulating a DFE equalizer in Matlab for a MSK modulated signal.
> The channel I'm testing is Ricean with max doppler shift of 50 Hz.
> I noticed that the performance of the DFE deeply improves if I apply the
> derotation before sending the samples to the equalizer.
> Has anybody ever observed the same behavior?
> Does any theoretical explanation for this phenomenon exist?
> Thanks
>
> Alberto

Rotation implies a frequency offset

Freq offset implies the phase offset is always changing

Your DFE is chasing a changing phase offset.

So no suprise when you remove the rotation, your DFE works better.

Clay
 | 
Pages: 1
Prev: FFT for ADC output
Next: forward error correction capabilities?