Subject: Re: How does a mixer work?
X-Newsreader: Microsoft Outlook Express 6.00.2800.1106
Date: Wed, 23 Oct 2002 12:06:52 -0500
NNTP-Posting-Date: Wed, 23 Oct 2002 11:43:26 EST
"John Woodgate" wrote in message
> I read in sci.electronics.design that John S. Dyson
> wrote (in ) about 'How does a mixer
> work?', on Tue, 22 Oct 2002:
> >Firstly, the real world signal (as we normally perceive it) is in the time
> >Any normal means of spectrum analysis, or fourier transform tends to make
> >the assumption that the function is non-varying in character.
> No, you are using the wrong term and that is very confusing. The
> function is not assumed to be non-varying, but IS assumed to be
> repetitive over time - the pattern of variation repeats.
I was trying to be very clear ( and was accurate.) Note that the character that
we are discussing is the frequency domain. The time domain entity is indeed
varying in the frequency domain. Nonlinear processes make the situation
> > Additionally,
> >normal linear transforms (where spectrum analysis is allied to the fourier
> >kind of transform) don't generally work well WRT nonlinear operations (or
> >with signals that have time varying characteristics.)
> Again, this is confused and misleading. You can certainly, for example,
> do Fourier analysis on a half-wave or full-wave rectified sine wave,
> which is the result of a non-linear process.
That is one of the 'special cases' that I was speaking of. You can make
some technologists views of certain nonlinear operations, but it isn't
really good in an engineering (or scientific) sense, where the situations
> > Spectrum analysis or
> >fourier transforms can still be useful in varying situations, but the
> >needs to be interpreted slightly differently, and with some caveats.
> You need to define 'varying situations'. Your discussion of passing a
> Morse signal through a very narrow filter, so narrow that the output at
> carrier frequency is almost constant shows that the 'interpretation' may
> need to be dramatically rather than slightly different.
If you use the super sharp filter, you have chopped off the sidebands,
and no information is carried.
> Some of the statements in your long post are correct and some are
> misleading at best.
Perhaps you choose to be misled?
Which microwave device is linear? I mean, you can look at spectra
to find 'problems', but a full diagnosis (in general) requires a time
domain view. As a technician, instead of engineer or scientist, you
can do a 'pattern match' from previous work in the frequency domain,
but the real action happens in the time domain (in the normal
The view of 'waves' instead of f(t), can start being more useful as
frequencies increase, but thinking in terms of 'waves' at any low
VHF/UHF/SHF type frequencies will screw you quickly unless you
are doing first-order design. Nonlinearity and 'waves' tend not
to match well in innovation, even though as a 'pattern matching'
technician, they can be useful. Engineers have both a 'pattern
matching' side to their work, but the real understanding needs
to be done in the time domain.
Until there is a full and convienient fourier domain set of nonlinear
operations, where there is indeed a convienient full dual between
frequency and time domain (in a nonlinear sense), the time domain
is the REAL way to solve the problems.
Not ALL diffeqs can be modeled by fourier analysis (esp nonlinear