0

So I know that the baseband bandwidth of BPSK needs to be over 1/2 the baud rate, but I can't find any sort of graph that shows the increase of BER as the bandwidth falls below that, such as when the signal is filtered by the frequency response of the channel i.e. skin effect for instance.

The pulses start to merge together, and the nulls of a pulse shaping filter become unaligned, but roughly what bandwidth does a BPSK become totally unreliable? is it not 1/4 baud rate or something.

Lewis Kelsey
  • 259
  • 2
  • 12
  • 1
    can you define "totally unreliable"? And how do you model your channel? Because, clearly, physically bandwidth-limiting channels don't come as ideal brickwall filters, and thus, the ISI takes the time-domain shape as defined by the filter. – Marcus Müller Sep 05 '21 at 14:51
  • This is answered in "Digital Transmission Engineering" by John B. Anderson, IEEE Press. It's the only textbook I know that discusses this question. – MBaz Sep 05 '21 at 15:53
  • @MBaz uff, in essence this might quickly drift into the domain of faster-than-Nyquist signalling (and though I don't have a great textbook for that, I'm sure there's a metric hellton of material on that out there) – hence my insistence that we first define "unreliable"; we really need to scope this. – Marcus Müller Sep 05 '21 at 16:49
  • That being said: I think we can point Lewis in a helpful direction that rectifies his misunderstanding. – Marcus Müller Sep 05 '21 at 16:53
  • 1
    @MarcusMüller I agree with you. I wanted to point Lewis towards that book because it will help him understand what happens when bandwidth is reduced, and also provide him with the tools to analyze this scenario quantitatively. – MBaz Sep 05 '21 at 17:04
  • 1
    And, on the matter of FTN signaling, Anderson has another book, "Bandwidth Efficient Coding". It's a pretty difficult read, but is a good summary of the field. – MBaz Sep 05 '21 at 17:05
  • 1
    @MBaz thanks for the book hint! – Marcus Müller Sep 05 '21 at 17:18

0 Answers0