System block is following

1.bernoulli binary generator - probability of a zero : 0.5
                                        sample time : 1e-7
                                        samples per frame : modulation
order(4)*1024
2.convolutional encoder - trellis structure : poly2trellis(7,[171
133])
3.random interleaver - Number of elements : modulation order(4)*1024*2
                               initial seed : 12345
4.rectangular QAM modulator baseband - M-ary number : 2^4
                                                 - Input type : Bit
                                                 - Constellation
ordering : Gray
                                                 - Normalization
method : Average Power
                                                 - Average power,
referenced to 1 ohm(watts):1
                                                 - Phase offset(rad) :
0
5.multipath rayleigh fading channel - Maximum Doppler shift(Hz) : 50
                                                 - Doppler spectrum
type : Jakes
                                                 - Discrete path delay
vector(s) : 0
                                                 - Average path gain
vector(dB) : 0
                                                 - Complex path gains
port : checked!
6. AWGN channel - Eb/No(dB) : EbNo(0:2:30)
                         - Number of bits per symbol : Modulation
order(4)
                         - Input signal power, referenced to 1 ohm
(watts) : 1
                         - Symbol period (s) : (1e-7)*Modulation
order(4)/2

Signal diveded by complex path gains after awgn channel in order to
compensate fading effect, then go to rectangular QAM Demodulator
Baseband.

7. Rectangular QAM Demodulator Baseband
- Decision type : Hard decision
- The others is same as the Rectangular QAM Modulator base band block
8. Random Deinterleaver : this is same as the Random Interleaver
9. Viterbi Decoder - Trellis structure : poly2trellis(7,[171 133])
                          - Decision type : Hard decision
                          - Traceback depth : 34
                          - Operation mode : Continuous
10. Error Rate Calculation - Receive delay : 34
                                    - Computation delay : 0

After simulation, i got ber curve.
But, there are a distinct difference between my expectation and
simulation curve.
If i use convolutional coding, performance has to become better than
theoretical rayleigh fading curve.
But, simulation curve and theoretical curve are almost same.
i don't know why normal performance doesn't come out.
so i am writing to ask for your help. please help me

i configure EbNo as SNR in awgn block, then SNR is EbNo+10*log10(M*(1/2)). 1/2 is code rate and M is modulation order(here, this is 4).
this setting makes performance become worse. There is about 3dB degradation compared to theoretical curve. Actually,,,convolutional coding makes performance improvement, right????? i wonder where is  wrong.

jamjomjaja <ygdizzy007@hanmail.net> wrote in message <c35e36d7-5108-41f7-bee8-58213f07e8b8@x18g2000pro.googlegroups.com>...
> System block is following
> 
> 1.bernoulli binary generator - probability of a zero : 0.5
>                                         sample time : 1e-7
>                                         samples per frame : modulation
> order(4)*1024
> 2.convolutional encoder - trellis structure : poly2trellis(7,[171
> 133])
> 3.random interleaver - Number of elements : modulation order(4)*1024*2
>                                initial seed : 12345
> 4.rectangular QAM modulator baseband - M-ary number : 2^4
>                                                  - Input type : Bit
>                                                  - Constellation
> ordering : Gray
>                                                  - Normalization
> method : Average Power
>                                                  - Average power,
> referenced to 1 ohm(watts):1
>                                                  - Phase offset(rad) :
> 0
> 5.multipath rayleigh fading channel - Maximum Doppler shift(Hz) : 50
>                                                  - Doppler spectrum
> type : Jakes
>                                                  - Discrete path delay
> vector(s) : 0
>                                                  - Average path gain
> vector(dB) : 0
>                                                  - Complex path gains
> port : checked!
> 6. AWGN channel - Eb/No(dB) : EbNo(0:2:30)
>                          - Number of bits per symbol : Modulation
> order(4)
>                          - Input signal power, referenced to 1 ohm
> (watts) : 1
>                          - Symbol period (s) : (1e-7)*Modulation
> order(4)/2
> 
> Signal diveded by complex path gains after awgn channel in order to
> compensate fading effect, then go to rectangular QAM Demodulator
> Baseband.
> 
> 7. Rectangular QAM Demodulator Baseband
> - Decision type : Hard decision
> - The others is same as the Rectangular QAM Modulator base band block
> 8. Random Deinterleaver : this is same as the Random Interleaver
> 9. Viterbi Decoder - Trellis structure : poly2trellis(7,[171 133])
>                           - Decision type : Hard decision
>                           - Traceback depth : 34
>                           - Operation mode : Continuous
> 10. Error Rate Calculation - Receive delay : 34
>                                     - Computation delay : 0
> 
> After simulation, i got ber curve.
> But, there are a distinct difference between my expectation and
> simulation curve.
> If i use convolutional coding, performance has to become better than
> theoretical rayleigh fading curve.
> But, simulation curve and theoretical curve are almost same.
> i don't know why normal performance doesn't come out.
> so i am writing to ask for your help. please help me