It's hard to generalize about all digital communication. I think
BPL is some kind of phase modulated OFDM as Frank says, so in that
case you could use essentially rectangular pulses (in practice there
is probably some roll-off and guard time to boot). Each individual
tone would actually occupy a bandwidth much greater than its keying
rate, but since each tone's keying rate is so low compared to the
total bandwidth, the net effect is minor, again exactly as Frank
says.
For single carrier high date rate systems however, the last thing you
want to use is rectangular pulses. The spectrum won't have discrete
harmonics but it will look like (sin(x)/x)^2 in frequency with
significant energy beyond the Nyquist frequency. In those applications a
waveform that falls off in time as t^2 is generally used, though there
are other options, like minimum-shift keying, which can be looked at
either as continuous phase FSK or QPSK using smooth shaped pulses.
Continuous phase modulation has some complications though.
Oz
Brenda Ann wrote:
"Frank Dresser" wrote in message
...
A square wave, itself, won't convey much information. It needs to be
modulated, and the modulation would have to effect the symmetry and
result in both odd and even harmonics.
I don't know what sort of modulation BPL is using. I can imagine
hundreds of low amplitude sine wave carriers from 2 to 60 Mhz, all of
them phase modulated. In that case, I don't think there would be much
harmonic output.
Digital comms are purely square waves. The modulation is FSK or similar
(generally)... in other words, the on-state is one frequency, the off state
is another. This creates a chain of square waves which themselves are not
modulated. The bandwidth, in this case 75 MHz, is how many on/off states
there are in one second. This is also concurrent with bitrate. Compression
schemes can raise the apparent bitrate, however the actual bitrate is the
same as the frequency used. I'm not sure how they do the band notching that
Japan tried before they tossed out the idea completely.
|