FPGA Central

I have forget to say that this is a simulation of WCDMA
to see the influence of processing gain with increasing number
of users

----------------------------------------
SINR = P/(N0 + In); % signal to interfernce noise ratio
with P = 1 (in my case) % we normalized the power for all users
N0: noise
In: Interference with => Interference = (number_Users - 1)/PG.
PG: processing gain.

in the simulation i do so:
for PG = 8 and PG = 16;
%------------------------

SINR = 0:2:20; %In dB loop of SINR
for N0 = 10.^(-SINR/10) - (Users - 1)/PG; % (as in the formula)

I generate data for X users, spreading, transmitting

n = (randn(1,N+ j*randn(1,N))*N0 %generate noise
desired_user = n + sum_other_users % add noise and mai to desirde user)

calculating BER, and ploting

But the result are not correct according ti the theory;
can some one tell me what is the problem, and if the generation of
noise is correct.

thanks
maca

[email protected] (maca) wrote in message news:<[email protected] com>...

> ----------------------------------------
> SINR = P/(N0 + In); % signal to interfernce noise ratio
> But the result are not correct according ti the theory;
> can some one tell me what is the problem, and if the generation of
> noise is correct.
>
> thanks
> maca

I don't know what kind of theory you used in that case.
In case of the theory to estimate BER of QPSK,
if your simulated BER is equal to Q[sqrt(2*SINR)],
that BER is definitely right.
Thus, I want to say that the BER from SINR is not regarding to any
interference, if SINR is estimated correctly as in your case.

--
Best regards,
James K. ([email protected])
- Private opinions: These are not the opinions from my affiliation.