On Sat, 2012-01-07 at 19:13 -0500, Gary Schafer wrote:
> I forgot to put the link in.
> http://www.k6mhe.com/n7ws/Speech_1.PDF
>
Interesting.
I have just been doing some modelling, and it transpires that the peaks
in the RF envelope are (pretty much as expected) strictly limited to
|rf| < sqrt(I^2 + Q^2) where I and Q are the baseband components in a
phasing exciter.
The stack up is due to the 90 degree phase shift in the Q channel.
http://www.exponent.myzen.co.uk/Baseband_I_Q_Mag.png
This shows the band limited square wave (in Red), the resulting
quadrature component (in green) and the magnitude of the resulting
vector (in blue).
This if for a 300Hz fundamental plus the harmonics out to tenth, which
is about all that will fit in a 2.7Khz SSB channel.
As the channel bandwidth (and thus number of harmonics that will fit)
increases the magnitude of the pile up becomes greater.
http://www.exponent.myzen.co.uk/SSB_RF_Baseband_envelope.png
This shows the resulting RF and the baseband magnitude vector magnitude
(in red this time), notice that the RF envelope is always smaller then
the baseband vector magnitude.
http://www.exponent.myzen.co.uk/absolute_RF_Baseband_envelope.png
This shows the same thing, but with the absolute value of the RF used
instead. Note that my RF is much closer to the modulation then would be
the case in practise, so the phase shifts are more obvious.
Further, the stack up at the leading and trailing edges increase with
the width of the post clipper filter, and in a communications bandwidth
are worst case limited to ~6db (You can only fit at most fundamental
plus 4 harmonics in a 2.7K bandwidth starting at 300Hz).
Thus a peak limiter working in analytic space using the absolute
magnitude of the IQ pair can successfully limit the peak RF level
required without needing to be applied at RF.
Sure this is potentially a bit tricky in an analogue radio, but most of
them these days have DSP first IF and have no excuse.
Regards, Dan.
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|
