Hi Fumihiko Sato,
Based on your query, it seems that the bit error is fixed at 0.5. Even if the other quantities are changed, still the bit error is constantly converging to 0.5.
As per my observation, I found that the error rate is converging to the probability of one/zeros that you input. Ideally it should not. If you change the probability of zero to a different value and then run the code, the Bit error will try to converge to that point. So the delay should be properly taken care since the block error rate calculation is carried out with random 1s and 0s as the convergence makes it clear.
with regards,
Gokul Nath S J
