The minimum number of frames, to be sure that your BER results are reliable, is related to the BER at the corresponding SNR value. In fact, a minimum of 300 errors are needed for good BER values, therefore if the BER is 10^-3, you need 300 000 bits to have 300 errors and thus 293 frames are needed.
One thing more if BER is 10^-7, then how will the figures turn out to be in number of frames?
For a BER of 10^-7, you need to simulate 3 billion bits (2929687.5 frames). If you find that your simulations take very long time you can simulate 1 billion bits.
Thank you sir again for apt reply. Can you provide some mathematics here for 2929687.5?
number_of_transmitted_bits=3e9*1024;
ber=1e-8
numnber_of_bits_in_error=ber*number_of_transmitted_bits
number_of_frames_in_error=numnber_of_bits_in_error/1024
In fact how we can determine that for particular BER how many number of transmissions are reliable?
Should this criteria be same for Rayleigh fading channels?
Because in that channel errors are frequent. For example, at BER 10^-2, i need 300/10^-2=30000 bits. if frame length is 1024 then only 29.2 frames. Is this justifiable number of frames?
- Badges
- Science topic