I will briefly explain why some receivers are double and
even triple conversion and why that has nothing to do with
selectivity.

	The superheterodyne design mixes a local oscillator with
the desired frequency to make the first IF or conversion. You
can either feed the local oscillator in to the mixer below the
desired frequency which is called low-side injection or you can
feed it in above the desired frequency for high-side injection.
If, for example, you have a receiver tuned to 28 MHZ with a 455
KHZ first IF, the local oscillator can either be 28.455 MHZ or 27.545
MHZ and you still hear 28 MHZ.

	Now for the bad news. Let's say your 28 MHZ receiver
uses low-side injection. The mixer is built to allow 28 MHZ in
from the antenna by means of a tuned tank circuit set to 28.0
MHZ exactly. There is, however, a nasty fly in the ointment.
Remember studying about the Q factor in tuned circuits? That is
the way the quality of a tuned circuit is measured and the
technique is to sweep a signal through the pass band of the
tuned circuit and determine how far off frequency you have to be
for the circuit to be down 3 DB from what it is in the center.

	One of the laws of physics is that in a tuned circuit,
the Q factor is a percentage of the frequency so a LC tuned
circuit at 1.8 MHZ will be rather sharp compared to the pass
band of a voice signal and it will not pass much more than one
AM signal which is not too bad.

	By the time you get to 28 MHZ, however, that tuned
circuit may pass frequencies between 26 and 29 MHZ fairly well.

	So, what's wrong with that? Everything! You have a 455
KHZ first IF. You are listening to 28.0 MHZ by injecting a
27.545 MHZ signal in to the mixer so here's what else happens.

	If you didn't have that tuned circuit in the mixer to
select 28.0 MHZ, you would still hear 28.0 MHZ and you could
also hear 27.545 - 455 KHZ or 27.09 MHZ which is right in the
Citizens' Band. In other words, that is the potential with every
superheterodyne receiver. The first mixer converts the local
oscillator + the desired frequency to the IF and it can also
convert the local oscillator minus some other frequency to that
same IF. That other frequency is the image.

	okay, let's put back that tuned circuit in the mixer. If
we tuned the receiver down to the AM broadcast band, it's not
too bad. We hear pretty much only what we wanted to hear. Now,
let's go back to ten meters and 28.0 MHZ. If you are listening
in AM mode, you should maybe hear a few CW signals as popping
sounds but there's something else. There is a strong image of
whatever CB channel is at 27.085. That darn tuned circuit in the
mixer is worthless by the time you get to ten meters or VHF.

	Okay. That's a single-conversion receiver. Let's make it
double-conversion by putting a second oscillator and mixer at
the 455 KHZ IF. The second conversion oscillator is at 355 KHZ so
our second IF is at 100 KHZ. This would work just fine except
that we still have that crappy front end in which the images
from the CB channels are all over ten meters. We will get a
little more selectivity because we combine the selectivity of
the 455 KHZ if and the 100 KHZ second IF but the receiver is
still a piece of junk.

	Now, let's make a decent receiver.

	The first mixer mixes a local oscillator that is maybe
70 MHZ higher than the desired frequency with that frequency to
produce a first IF of 70 MHZ. The difference between the image
frequency and the desired frequency is now 140 MHZ which is
much, much easier to build a tuned circuit for.

	If you put a crystal filter at the 70 MHZ first IF that
has a fabulous Q factor, then your second conversion mixer
doesn't have to try to filter everything as much. You end up
with good selectivity and virtually no images over the whole
tuning range.

	There are a lot more problems in building a good
receiver and I don't even begin to know them all, but I hope
this helps in the discussion as to whether single conversion or
double conversion receivers are best. There are some good ones
of both kinds but generally a double-conversion receiver with a
high first IF beats the pants off of a single-conversion
receiver with a fairly low-frequency first IF.

	Everything I have said here is generalities and some of
the older gear used lower first IF's with higher-Q multiple
tuned circuits in the first mixer and achieved beautiful
results.

Martin McCormick WB5AGZ  Stillwater, OK 
Systems Engineer
OSU Information Technology Department Telecommunications Services Group