[TenTec] Jupiter Audio vs. DSP

Robert & Linda McGraw K4TAX RMcGraw@Blomand.Net
Mon, 30 Dec 2002 09:11:16 -0600


Reading recently about concerns regarding some distortion artifacts in
receive audio and in the suggestion that the attenuator should be used
brought me to disclose a few facts with regard to DSP based receivers.

As with any digital processing system there is a specific dynamic range that
the system will handle.  It's simply a mathematical relationship.  Once the
signal exceeds the dynamic range, unlike analog circuits which have a broad
or round knee into the distortion range, the digital system goes suddenly to
overload mode, thus to a sudden and high level of distortion.  In general,
it sounds "ugly".  In the case of DSP for ham receivers, regardless of IF
DSP or AF DSP the same factor applies.  While we would like to think
otherwise, designers of circuits have yet to produce ACG circuits which will
absolutely prevent a wide range of front-end signals from, at some point,
producing overload of the DSP system.  We didn't realize this with
conventional analog signal chains as the overload or distortion point was
typically a smooth transition from clean, to very slightly distorted, to
moderately distorted and finally to total distortion.  In the case of
today's modern receivers and digital signal chains, the attenuator must be
used to extend the dynamic range of a receiver.

The dynamic range of a digital system is based the number of binary bits in
each word.  Each bit corresponds to a doubling of the resolution, or roughly
6dB of dynamic range capability.  Telephone systems typically use 8 bits
which is approximately 48 dB dynamic range linear, although effectively more
when companded.  Today's CD's and most digital recorders use a 16 bit system
for some 96 dB or dynamic range.

Look at it this way, making the assumption that 1 S unit is ~6dB, then S-9 +
40 dB is 94 dB.  Today's receivers can easily hear a signal of -125 dBv and
are expected to handle a signal to +40 dBv.  This is a dynamic signal range
of  ~160 dB.  Hence the need to apply attenuation ahead of the DSP system to
allow it to operate within its limited dynamic processing range.