Problem with berfading function: Ideal BER curve given by MATLAB does not match with actual Probability of error expression for Equal Gain Combining diversity

Hi,

I am trying to simulate Equal Gain Combining diversity scheme through a rayleigh fading channel with independent fading paths. I am using QPSK modulation scheme. I used the command:

berrlideal = berfading(EbN0_dB,'psk',Mod_Order,Nr); %Mod_Order = 4, Nr = 2

to get ideal case curve of BER. But this curve does not match with my experimental results. I have also checked this curve with the expression for probability of error given by Zhang in the paper cited. My experimental results coincide with the probability of error curve (as it should, since Probability of error is the limiting case of BER), so I think my experiment is properly designed. But this does not match with the matlab curve.

The only reason I can think of is that the diversity curve given by MATLAB is for some other diversity scheme (possibly Maximal ratio combining?). If this is the case, is there any way to get the ideal curve of Equal gain combining? I need to check this for a diversity order of upto 7, so I do not have probability of error expressions for them.

My program is attached.

Sincerely,

Abhay Mohan

Paper: Probability of error for equal-gain combiners over Rayleigh channels: some closed-form solutions Zhang, Q.T. Communications, IEEE Transactions on Volume 45, Issue 3, Date: Mar 1997, Pages: 270 – 273