--- Tom Rauch <w8ji@contesting.com> wrote:
> > These kinds of shortcomings are not specific to
> DSP
> > designs. Any RX that has a ton of mixers will
> share
> > these issues. This is precisely why I'm partial to
> a
>
> The point is DSP filters are so far back in the
> system, they might
> as well be at the headphone jack. While I can find
> many radios that
> have narrow selectivity after just second mixer,
> virtually all DSP
> filters are after three or more mixers and a half
> dozen other stages.
>
> > direct conversions scheme along the lines of the
> > Collins 95S-1A. As anyone will tell you a DC RX is
> one
> > of the cleanest sounding designs there is, just
> don't
> > expect any selectivity. But that is where the
> > wonderful new world of ADC and DSP come in.
>
> What is the IM and blocking dynamic range of this
> "wonderful new
> world"?
>
As long as the A/D conversion stage is not being
overloaded, has good S/N, and sufficient bits it
should be quite good. But I'll acknowledge that is
where the challenge lies. Perhaps that challenge alone
is why we haven't seen more attempts at coupling
ADC/DSP to a DC RX design.
> > <snip>
> > > The last thing I want is a receiver that moves
> > > selectivity even further
> > > back in the system running at frequencies my dog
> can
> > > hear so
> > > marketing departments can say it is "true IF
> DSP".
> >
> > I'm still puzzled why low frequency DSP IF designs
> are
> > so frequently derided. With a direct
> conversion/DSP
> > scheme the selectivity is moved to being just
> after
> > the very first (and only distortion producing)
> mixer.
>
> I admit any audio filter will be a nice addition to
> a direct conversion
> receiver, but my concern is working weak signals
> near very strong
> signals. My point is the typical DSP-based radio
> runs the DSP
> system at a few dozen kilohertz well to the tail end
> of a complex
> system.
>
> Moving something from the audio line to an
> additional special last
> IF near a few dozen kilohertz is not actually much
> different that just
> sticking the DSP on the audio line.
>
> This all goes back to my earlier point marketing
> departments are
> selling people on systems that are another step down
> in
> performance, and making a backwards step in
> performance a
> "desirable feature".
Point taken, I'm frequently puzzled why there is a
tendency for radio designers to try and use DSP
techniques to achieve a design goal instead of good RF
practice. I'll be the first to agree that DSP doesn't
replace good RF design techniques, it is nothing more
than a tool to achieve new capabilities in addition to
what can already be done with traditional filtering
techniques. The place where DSP can earn its mettle is
processing signals other than just CW and SSB. There
is no good reason to not handle modes like SSTV,
PSK31, WSJT, PUA43 packet etc. natively in a DSP stage
of the radio, so far no one has done this in a
meaningful in the amateur radio market. Taking the
logic that having many stages between the antenna and
your intended output is a bad thing; it makes no sense
to take those kinds of signals all the way to audio to
only have to re-digitize them in the computer sound
card in order to process them. This kind of native
digital mode processing is one of the main reasons I
repeatedly get onto my soapbox for tighter computer
integration with the radio.
> While an additional DSP filter can be useful in some
> cases,
> depending on the over-hyped currently available DSP
> systems for
> selectivity simply results in a system that does not
> handle nearby
> strong signals nearly as well as conventional
> filtered radios.
>
> Bottom line is I don't want or need 2000 filters,
> especially if the
> price is decreased close-spaced (none of those silly
> useless
> measurements outside the passband of the roofing
> filters please)
> performance. Five selectivity settings that can
> actually handle
> strong closeby signals are enough, thank you very
> much!
One area that DSP filters seem to excel from what I
can tell is that they seem to introduce less phase
distortions across the filter passband, a parameter
that QST and others don't measure. So it is hard to
find meaningful data to either support or dispute that
impression.
Duane
N9DG
__________________________________________________
Do You Yahoo!?
Yahoo! Sports - live college hoops coverage
http://sports.yahoo.com/
|