What channel model to choose while simulating a system in passband with frequencies over GHz range?

I was wondering as to what kind of channel model do we use to test for BER vs SNR in a passband RF system. The system that I designed is basically a wideband frequency hopped OFDM system where I use Digital up-converter (DUC) to up convert the signal to a frequency in GHz range and use the Digital down converter (DDC) to down convert the signal back to baseband in the receiver side. Now my system works fine. However, when I add an AWGN channel, I get pretty low BER of 10^-5 in passband as compared to around 10^-4 BER in baseband under same AWGN channel of SNR 10 dB. Why is it happening? I thought that there would be more attenuation in passband as attenuation increases with respect to frequency. Or is it that my way of modelling the channel isn't good. Btw, I have tried adding Rayleigh SISO fading channel along with AWGN but in that case I achieve around 0.4 BER which is quite high. I just don't know anything about BER vs SNR plots in passband. All the plots that I found are mostly in baseband.

So my question is

1. How do I model the channel in passband?

2. How does a typical BER vs SNR look like in passband?

Thank you!

1 Answer

However, when I add an AWGN channel, I get pretty low BER of 10^-5 in passband as compared to around 10^-4 BER in baseband under same AWGN channel of SNR 10 dB. Why is it happening?

You've got a bug in either passband channel or baseband channel model, or in how you transform your noise or your signal – end of story!

I thought that there would be more attenuation in passband as attenuation increases with respect to frequency

If your passband model does that, then your baseband model has to do the same – they are mathematically equivalent.

If your passband model is frequency-selective, then your baseband model has to do the same.

• Yeah I didn't model the baseband system similarly so that's why I was getting such weird values. Now it's working fine! May 30 '20 at 9:28