As we all know imp resp is the key for channel simulation. Unfortunately, I've seen many simulators that produce less than sufficiently good imp resp. After some study now I can generate up to spec impulse response w/o causality issue, representing accurate channel characteristics, etc. However, still I'm facing one problem: Why the impulse() function from Matlat produces a peak value as high as 10^10?
Right now my workaround is to reduce the ampl down to the correct value based on s21. But would like to see more exquisite way of doing this...