From: jbloom@arrl.org (Jon Bloom)
Newsgroups: rec.radio.amateur.misc
Subject: Re: DSP and The Future
Date: 24 Jan 93 12:38:58 EST
Organization: American Radio Relay League
 
In rec.radio.amateur.misc, regnad@hal.gnu.ai.mit.edu (Paul Prescott) writes:
>
>Ah ha!  So other people *have* had the same idea I got about 7 or 8 years
>ago.  :)  That is, apply DSP to the IF of a standard receiver to witness
>the greatest advance in performance since the vacuum diode replaced the
>coherer.  :)  The best possible scenario I can think of would be a 455 KHz
>add on DSP box.  This would make any receiver with a decent front end
>and a 455 KHz IF nothing short of an absolute marvel.  This is one solid
>state device I would not mind being able to hang on my vacuum tube radios.
 
There are a couple of complicating factors here.  One, as you say, is
that adding DSP to the IF may not change the basic strong-signal
performance of the radio.  But more than just "a decent front end" is
involved.  The strong-signal performance is determined by the analog
circuitry in the front end, filters and IF stages.  So adding the DSP
to an existing radio only makes sense if that radio has good *overall*
RF performance to begin with.
 
Consider the following basic receiver design:
 
___
\|/
 |      _______      ________      ________      _________
 |     |       |    |        |    |        |    |         |
 ------| MIXER |----| FILTER |----| IF AMP |----| DSP SYS |--- OUT
       |_______|    |________|    |________|    |_________|
           |
           |
          LO
 
There are two basic dynamic-range consideratins here: 1) response to
strong signals that are rejected by the roofing filter; 2) response
to signals that are *not* rejected by the filter.  For example, if
the filter is 3 kHz wide, a signal separated from the receive
frequency by 10 kHz is well rejected by the filter.  Any spurious-
response problems (nonlinearities) that cause this signal to block
or generate problematic IMD products will have to occur in the front end
or the filter itself; the IF amp and detector aren't involved.  But
if the signal is only 2 kHz from the receive frequency, the IF amp
and the DSP subsystem can respond unfavorably to the strong signal.
Thus, the ability of the DSP system to narrow the *system* bandwidth
down to, say 100 Hz, doesn't alter the fact that stong signals that
appear in the IF can cause problems.
 
In an analog receiver, the use of AGC mitigates this problem.  That is,
the strength of the signal appearing in the IF is measured and used to
reduce the gain of the front end (and the IF amps as well, often) so
that the strength of the signal applied to the IF stages is controlled.
This works because the entire signal that comes through the filter is
used to derive the AGC control signal.(*)  If you were to derive the
AGC after passing the signal through a narrow filter in the DSP
subsystem (or in an analog subsystem, for that matter), it wouldn't
control those signals that appear outside the narrow filter but inside
the roofing filter.  Thus a strong signal of that sort wouldn't cause
the AGC to reduce the front-end gain, and the IF stages might overload.
 
On the other hand, if you allowed the AGC to respond to all of the
signals in the IF, but allowed the DSP to filter out a 500-Hz wide
segment of the IF, strong signals outside that segment would "pump"
the AGC, causing the desired signal to fluctuate in amplitude.  Of
course, the DSP can compenstate for this, but only up to a point.
 
All of this is not to say that there aren't some real benefits to
having the DSP at IF instead of AF.  One such benefit is that it
becomes much easier to demodulate many different types of signals--
"it's only software."  But present DSP technology is not a panacea,
and sticking a DSP subsystem on the IF of a Knight Star Roamer(**) is
not going to make it into an FT-1000.
 
(*) There are, of course, more complex receiver designs that use
multiple IFs and multiple filter bandwidths.  That only complicates
the analysis; it doesn't change the fundamental point.
 
(**) The other OFs already know this, but for you kids a Star Roamer
was a cheap kit shortwave receiver.  It recived signals, and that's
about the best one could say for it.
-------
Jon Bloom, KE3Z              |  jbloom@arrl.org
American Radio Relay League  |     Justice is being allowed to do whatever
225 Main St.                 |     I like.  Injustice is whatever prevents
Newington, CT 06111          |     my doing so. -- Samuel Johnson
 
