[Elecraft] Re: SSB - Fooie!

[Ferguson, Kevin]

The bandwidth considerations are correct, but there
is another, easy to understand, contributing factor:

The PEP to average power ratio for SSB is somewhere
around 10 dB. Yes, compression can lower that number,
by increasing the average value. Including more 
bandwidth (thinking 80 M hifi types
here) can dramatically raise the ratio by increasing
the peaks. (multiple non-harmonicly related frequencys 
will periodically add constructivly, producing higher 
peaks think two-tone test with 10 tones instead!)

While keyed the peak to average ratio for
CW is close to 0 dB. (Exactly zero only if you permit nasty 
clicks)

If only the peaks are loud enough to distinguish from the noise,
you can't understand voice, but you can copy CW.  To understand
voice, the average needs to be a bit above the noise so that
most of the softer voice sounds also come through. (the receiving brain
can fill in some gaps if they are small enough)  

Enter narrow bandwidth and compression, the SSB DXer's friends.

Anyway, between the bandwith considerations and the peak-average
differences, I think a 20 dB advantage for CW using the _same_ PA IS
a reasonable estimate. Compression might shave 3-4 dB off that number.

The only way I could get down to 9 dB were if I were only considering the
S-meter reading at the recieving end...and then I would say, that yes
you probably do need to see the S meter peaking 1-1/2 S units or so
above background for solid SSB copy, while it needn't move to copy CW. 
But keep in mind that this background
noise is already around 10 dB higher due to the increased reciever
bandwidth.