Michael Coslo wrote:
Jim Lux wrote:


This is only true if you're doing a ARQ type protocol. If you're
expecting interference (and BPL certainly is), a Forward Error
Correction type strategy would be a much better bet. For instance, ECC
memory uses a rate 8/11 code (3 syndrome bits for 8 databits) to
correct any single bit error and detect double bit errors. For BPL,
one would probably choose some form of multiple carrier encoding
(since interferers are likely to be narrow band) with some form of
interleaving (since interferers are bursty), and a fairly robust code.



Are you talking about OFDM?


That would be one possible implementation. There are others. For
instance, MultiTone FSK is another. Multitone PSK like Link-11/TADIL-A
(which has been used for HF comms for decades) is another.



There are plenty of examples to the contrary of the digital
signal as robust entity. Experiments have been made in which a 5 watt
signal in a car with a mobile antenna will knock out nearby BPL
signals. 100 watts in a car will do even more damage, and a base
station yet more. Other experiments have shown the so called notches
being abandoned after the system was unable to send good packets.
Kind of like it was getting desperate almost.


But this is merely an example of a specific bad implementation of BPL.
There's no physics reason why one couldn't make it work (whether it's
cost effective is another story).


Sure! They could start running our power into the house through
coax! ;^)

More, I was thinking that the early BPL implementations used fairly
simple receivers and transmitters which aren't particularly interference
immune. Technology, particularly for digital processing, advances quite
quickly.




Seeing as how qrp levels into inefficient antennas can cause
problems, I have to wonder what will happen during the next sunspot max.


The signal level received via ionospheric paths are so low, that the
BPL receiver isn't going to have any trouble regardless of how good
propagation is. OTOH, the interference radiated by BPL will spread
that much wider.


Has asny testing been done on the level of signal needed to start
affecting a BPL type signal? So far, we've gotten the stories like I've
given above, which are practical if anecdotal.

This would be complicated by the fact that there's not just one "BPL
type signal" or, more accurately, one "BPL implementation". One could
design a system using, say, BPSK that is hideously unrobust and another
system, also using BPSK that is very robust.

For instance, BPL doesn't need to have good doppler tolerance, and could
rely on recovering a very stable frequency reference, something that
wouldn't be practical in, say, a cellphone.


But it if it provides access to capital markets for the relatively
small number of people working in the BPL industry, so they can get
paid their salaries and bonuses, then it's a good thing for them.


Sounds like a plausible rationale for the crystal meth industry, too!

Probably more people in that business than in BPL.