On second thought, having slept on the idea, I mean about using a coil
to convert something such as 10 MHz to baseband.  I do know of a
possible way.   I surely would simplify allot of things also, no chips
or transistors nor would we need a LO IN section either, at least not
in a hybrid SDR receiver scenario, in a Kenwood or Yaesu, etc.

The experiment to try then, to see if we can go straight from RF to
baseband without LO IN mixing is to use a simple diode detector on
both the I and Q outputs of a phase splitter coil.  And thus see if
the 90 degree phase angle on the 10 MHz split signal is reproduced in
the baseband?

"....Hence it should be a matter of the phase shift only and not the
method of demodulation...."

First Thing Though:

I mean if it is a possible to go from a phase splitter coil in the
10.7 MHz range straight down to audio without a RF mixer involved.  We
would not have to worry about the coil's resonance shifting around
with thermal variations.  

The phase splitter coil, does not effect the I and Q much if the
resonance is varied within a region.  You can test that in LTspice and
see that the coil is very wide banded.  Thermal variations will not
mean much at all in a very wide banded 10 MHz coil such as this kind
of coil.  The higher the phase splitter coil's center frequency, the
wider the bandwidth is, in MHz terms.  Hence the phase splitting coil
is ideal in this regards, for the possibility I am looking at.

The Coil:

The +/- 1 dB variation of the I and Q over the bandwidth from left of
the spectrum to the right can be 1 or 1.5 MHz of the HF band with
these kinds of coils (transformers).  And the +/- 3 dB region is
exceedingly wide.  Perhaps as much as 5 MHz at the 10.7 MHz coil
resonance frequency.  From what I recall off the top of my head from
looking at the coil earlier in analysis plots.   Hence we do not have
to worry about a perfectly exact center of resonance being maintained
at all times.

The Ways:

There are two ways we use in radio to take the baseband audio off of a
signal carrier.

1.  The beat frequency mixer method such as is used with Sideband and
in the Quadrature demodulator ideas.  

2.  The audio is demodulated off of the RF signal carrier from the
i.f. by means of a diode.  Where the diode is not being used as a
mixer but rectifier.  Keep this one in mind now.....

As I was looking at the way RF signals mix in software simulations
with phase shifted carriers and audio.  I noticed that if you split
and phase shift a RF signal and demodulate the audio off, that the RF
phase angle gets converted down into the audio modulation when taken
off the signal carrier.  

You see, we have perhaps all commented here allot about the mixing of
a phase shifter pair of RF signals with a LO IN, or of using a phase
shifted LO IN pair with the RF input to demodulate off the audio into
a phase shifted pair.  Hence the 90 degree offset between the two RF
I/Q signals gets transfered down to the audio waveform phase angles in
the baseband region after demodulation.

"...The experiment to try then, to see if we can go straight from RF
to baseband without mixing is to use a simple diode detector on both
the I and Q output of a phase splitter coil.  And thus see if the
phase angle on the 10 MHz split signal is reproduced in the baseband?..."

A 10.7 MHz coil then followed at its pair of outputs by two
demodulation diodes might work ...and thus be very cost effective.

All the right and left channel information should still be present
after diode detection of the I and Q channels.  And the SDR software
should see no difference of any kind with such a diode demodulated I-Q
pair of signals.  And the didoes should allow for some wide band width
in the baseband region, and thus cover all of the desired spectrum we
need for the sound card input.

* ...It should not be a matter of the method of demodulation that
makes things work but merely the phase shifted pair of signals that
are the end result of the demodulation.  Hence it should be a matter
of the phase shift only and not the method of demodulation.

Hence simple AM diode demodulation might be all we need,...

If this method can work, then we do not have to have anything in our
hybrid SDR receiver set ups but a resonant phase shifting coil and two
diodes.  And that would be inexpensive and simple.  And so, no LO IN
input for mixing would be required.

We could then after this idea, if it works, use the Gilbert Cell to
amplify the signal pair on up for output into the sound card.  However
in a receiver we might already have enough RF signal level and hence
demodulation level to dismiss the need for a audio amplifier section
between the demodulated diode outputs and the sound card.   Just a
little simple capacitor filter is perhaps is all that follows.

If this idea can work, it sure would be nice wouldn't it?

Why you could make a purely crystal set version of a SDR receiver with
this idea.

Dan

Reply via email to