>Hi all,
>
>I tried to simulate the bit error rate performance using bpsk with
>convolutional code: R=1/2, K=7, generator polynomials (133,171) in AWGN
>and the performance using bpsk ofdm with the code in AWGN. I got the
same
>ber curve for both cases, which showed some advantage or gain after
using
>the coding
>
>However, when I simulated the bpsk ofdm system using the conv. code in
>flat rayleigh, I didn't see any improvement and the plot was pretty much
>parallel with the ber curve of bpsk in flat rayleigh. 
>
>But when I used the conv.code with bpsk only in flat rayleigh, I saw
some
>improvement (gain). I supposed I should get the same ber performance for
>both cases in flat rayleigh, right?
>
>Anyone has suggestions over this issue? Thanks very much.
>
>Yohan
>
>
>

I would first check if Eb/N0 is calculated correctly in the Rayleigh case.

Hi all,

I tried to simulate the bit error rate performance using bpsk with
convolutional code: R=1/2, K=7, generator polynomials (133,171) in AWGN
and the performance using bpsk ofdm with the code in AWGN. I got the same
ber curve for both cases, which showed some advantage or gain after using
the coding

However, when I simulated the bpsk ofdm system using the conv. code in
flat rayleigh, I didn't see any improvement and the plot was pretty much
parallel with the ber curve of bpsk in flat rayleigh. 

But when I used the conv.code with bpsk only in flat rayleigh, I saw some
improvement (gain). I supposed I should get the same ber performance for
both cases in flat rayleigh, right?

Anyone has suggestions over this issue? Thanks very much.

Yohan