Dave Platt wrote:
In article ,
Jim Lux wrote:


I would venture to say that with today's technology, one should be able
to get within 0.5dB of the Shannon limit without too much trouble. So,
since a BPLish signal can be many MHz wide, and the putative interfering
signal is going to be pretty narrow band in comparison, they can code
around it.


For a relatively weak interfering signal, that's probably true.

That's not likely to be the case in a lot of situations, though. A
strong interfering signal - say, a few watts of HF coming from a wire
antenna, a few houses away from a BPL receiver - is likely to be
strong enough to saturate the RF front-end of the BPL receiver. This
will result in a severe "de-sense" problem - it'll wipe out most other
signals within the receiver passband.

Ham (and other narrow-band) HF receivers deal with this problem by
limiting their receiver passband. Single-band filters prior to the
first mixer or first RF amp will keep out interferers outside that one
band, and narrow-bandwidth multi-pole filters after the first mixer
can reduce the impact of interfers that are closer to the desired
signal.

This approach doesn't work with systems which have a "wide-open"
broad-bandwidth front end, such as a typical BPL receiver, as the
receiver's front end *has* to be left open to the entire bandwidth of
the desired incoming signal. Many modern ham HTs have similar
problems... their "DC to daylight" front ends are easily desensed, or
driven into severe intermodulation by nearby VHF transmitters (e.g.
police, fire, paging, and so forth).

I imagine it's possible to reduce the severity of a BPL receiver's
desense problems by using a wide-dynamic-range front end... but these
take more power and aren't as suitable for large-scale chip
integration, and are thus going to be more expensive to build.

You can do pretty well these days.. Consider if you're using a direct
conversion receiver using a mux driven by quadrature clocks into a low
pass filter like the SDR1000.. it has fairly good out of band strong
signal rejection. Granted, in band, the performance is limited by the
following audio stages.

There's also a variety of receiver designs intended for GPS in a strong
signal environment that avoid desensing, while still retaining wideband
receive for the desired signal.

One can also do adaptive analog cancellation.

I would imagine that first (and second) generation clunky BPL receivers
would have all sort of problems, but the wide open receiver with strong
interferer problem is such a prevalent one, they'll have to come up with
a solution that fits in their price bracket.

(or, they won't, because they don't really need BPL to actually work..
They just have to keep the BPL industry alive, and speculative future
performance might be enough for that...)