|
|
|
|
|
|
|
|
|
|
|
|
|
|
Split band
audio signal processor 2007 for SSB and AM |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
I build my first split
band speech processor already in 1976. Since then I have used it. |
|
|
|
|
|
For some reason that
external processor has given better results than the integral one's in
my |
|
|
|
|
Icom rigs. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The idea in split band
processor concept is to avoid harmonic and intermodulation products, minimize
distortion. |
|
|
Smaller distortion is
achieved by clipping the signal in narrow frequency bands and having
efficient band-pass |
|
|
|
filters after each such
clipper. Threfore filters are needed before and after the clippers. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The challenge in making
a split band processor is in getting the 200mH level coils. Those are not
easily available. |
|
|
Some time ago my
favorite outlet had Vogt 2x39mH common mode filter coils at bargain price, 1
Euro each. |
|
|
|
So I bought 50 of them
and begun to build another split band processor. My idea was to make the
filter bands |
|
|
|
more narrow than I had in
the vintage version, this time exactly one octave each. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The block diagram and
schematics are enclosed. As only one inductance value was available |
|
|
|
|
different impedance
levels had to be used for different frequency bands. On lowest band two
inductances |
|
|
|
were connected in series
to avoid 200 ohm impedance ( which might work as well). So 24 coils, 156mH each, are used |
|
|
in this processor. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Each inductance was
measured. As there was variation +/- 10% in inductance, the capacitors had to
be |
|
|
|
matched individually in
each filter in order to reach the targeted responses. The component values
shown |
|
|
|
in the schematics are for
ideal case. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The same settings as with
the 30 year older version seem to be the best: Clipping amount on the |
|
|
|
|
lowest band can be up to
13 dB, on the mid band a bit more, up to 15 dB and on the highest band
between |
|
|
|
13 and 21 dB, which is the practical maximum. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
In summing up the bands
best results are achieved by attenuating the lowest band 3-4 dB and |
|
|
|
|
the mid band 1-2 dB. The highest band is not attenuated.
The full 10 dB tuning range is not quite needed. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The microphone gain is
set to give 4-5 Vpp during speech peaks in the output of LM386. |
|
|
|
|
|
That is about the point
were overdrive indicator LED starts blinking. With my Heil headset the
setting |
|
|
|
is -6dB from maximum. |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Pekka, OH1TV |
28.12.2007 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Block diagram |
|
|
|
|
|
|
|
|
|
|
|
Clipper unit |
|
|
|
|
|
|
|
|
|
|
|
Microphone amplifier |
|
|
|
|
|
|
|
|
|
|
|
Filter responses |
|
|
|
|
|
|
|
|
|
|
|
Front panel |
|
|
|
|
|
|
|
|
|
|
|
Sound
sample |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Back to MyProjects
page |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|