Dear Jan-Martin, LF group
At 11:37 24/01/2003 +0100, you wrote:
Suppose I could use an FF, preferably CD4013 for the mixer.
But I can't find any reference to similar constructions, only have a vague
memory that it has been mentioned in TT some time in the last 35 years
A D-type flip-flop does work something like a mixer. If you feed 1MHz into
the clock input, and 1.001MHz square wave into the D input, the output will
approximately be a 1kHz square wave. The interesting thing is that you will
also get the 1kHz square wave if the D input is 0.001, 0.999, 1.999, 2.001,
2.999, 3.001... nMHZ +/-1kHz. The D-type flip-flop effectively "samples"
the logic level at the D input the moment the clock transition occurs - it
works like a sample-and-hold with a square wave input, and no anti-aliasing
filter (this analogue scheme is used in things like sampling scopes,
modulation meters, vector voltmeters, etc., or optically with a
stroboscope). The trouble is that, since it is effectively mixing the
square wave input with a train of impulses at the clock frequency, and low
pass filtering the result, the output spectrum contains many harmonic and
non-harmonic components as well as the desired output. These are easy to
filter out if the difference frequency is much smaller than either input
frequency, but if you try to use the sum frequency output, or the output
frequency is comparable to the input frequency, the output waveform is
roughly a square wave but with substantial pulse-width modulation, which
would be difficult to filter to get a clean VFO signal. So for your
mixer-VFO, a conventional mixer would work better.
Cheers, Jim Moritz
73 de M0BMU
|
