Rician bit error rate problem

I am trying to simulate simple BPSK system using simulink with a ricain fading channel. When I run my model everything seems ok accept that I am not getting the bit error rate that I was expecting. My modeling very simple it consists of a BPSK modulator from the communication toolbox that is feed with a binary input source. Then the output of the BPSK modulator is feed into a Rician fading channel and the into an AWGN channel then into a BPSK demodulator. The output of the BPSK demodulator is then compared to input binary source using the error rate calculator block. I am using the MATLAB command berfading(3,’psk’,2,1,2) to determine the theoretical value of the BER when k = 2,EbNo=3,M=2. The value the above command says I should receive is about 0. 0677, but I am receiving something around 0.1. I know I am calculating the error rate correctly because when I remove the rician channel everything works according to theory. So I suspect that I am not setting up the rician channel correctly. Here are the parameters I used to set up the rician channels

K-factor (scalar or vector) = 2

Doppler shift(s) of line-of-sight components(s) (Hz) = 0

Initial phase(s) of line-of-sight component(s) (rad) = 0

Maximum diffuse Doppler shift (Hz) = 1

Doppler spectrum type: Jakes

Discrete path gain vector(s) = 0

Average path gain vector (dB) = 0

Normalize average path gain vector to 0 dB overall gain

Now the reason I have used these parameters is that I they seem very basic and the berfading command I used to calculate the theoretical rate did not have any input to specify things such as the discrete path gainvector or dopplershift of line of sight component so I left them blank. If there is anyone out there that has some experience with this type of simulation I would really appreciate any input you can give me.

Thanks in advance

Matt