One conclusion Peter 'PLX reached on our original 73kHz SlowCW (the
proper term we used then, none of this new fangled Q code stuff) was
that although the 73kHz path was very bursty with non Gaussian
interfernec ewhen looked at in audible bandwidths, by the time it was
averaged out over teh 40 second dot period we used, its properties
were indistinguishable from pure AWGN.

I can't recally if 73kHz was worse or better than 137kHz in the nature
of interference, but I think you can safely assume that in bandwidths
of much less than [30seconds]^-1    the background noise is
near-enough pure random Gaussian.  So normal error curves apply.

3s dots may be susceptible though.

Andy
www.g4jnt.com


> > S/N is an interesting point, I have no doubt people like Andy may have more
> knowledge, but LF noise is not like other radio noise it is very "spikey"
> and "bursty" (I believe the last term is actually an approved definition
> :-)) A lot of tests assume it is random Gaussian and I dont really believe
> that, though I have no data to dispute it.
>
>