Path delay influence in BER performance using Rayleigh channel

Hi,

I am simulating a DSSS modulation in a Rayleigh channel to get the BER performance. For these purpose I am using the comm.RayleighChannel object.

My problem comes when I change the Paths delays of the PDP. For example:

SamplingRate=50e-9;
fs=1/SamplingRate;
maxDopplerShift=30;
AvGain=[-45 0 -1 -1.9 -45];

In this case the BER seems to be reasonable:

PathDelay1 = [0 50 100 150 200]*1e-9;

rayChan = comm.RayleighChannel( ...
    'SampleRate',          fs, ...
    'PathDelays',          PathDelay1, ...
    'AveragePathGains',    AvGain, ...
    'MaximumDopplerShift', maxDopplerShift, ... %DIFERENT FRECUNCY RESPONSE IF ITS ZERO
    'RandomStream',        'mt19937ar with seed', ...
    'Seed',                73, ...
    'PathGainsOutputPort', true, ...
    'NormalizePathGains', true);

In this case the BER have no sense because it goes to a 0.5:

PathDelay2 = [0 49 100 150 200]*1e-9;

rayChan = comm.RayleighChannel( ...
    'SampleRate',          fs, ...
    'PathDelays',          PathDelay2, ...
    'AveragePathGains',    AvGain, ...
    'MaximumDopplerShift', maxDopplerShift, ... %DIFERENT FRECUNCY RESPONSE IF ITS ZERO
    'RandomStream',        'mt19937ar with seed', ...
    'Seed',                73, ...
    'PathGainsOutputPort', true, ...
    'NormalizePathGains', true);

Why is this happening? I can imagine that the problem comes due to inequality between the channel coefficients and the taps gains, but I don`t understand the reason and I am not sure.

Thanks for read my noob question.

[Mohammad Kadir]

Can you send the BER code? Otherwise it is not possible to comment.