I've read through some of the replies and didn't see what I thought was a good
answer to "where can I find a good explanation". We've been doing a series of
technical seminars at work, and one of the first ones covered AM and FM
modulation. (FYI...we build equipment that is very good at analyzing spectral
content of signals, so it's an area we care quite a bit about.) We used a
vector diagram that I think is fairly easy to understand. Wish I could draw it
here! I'll try to describe it verbally in a way you could draw it yourself,
and think about it.

For AM: Draw a vector starting at the origin and going one unit right. This
is the carrier, at time=0. It rotates counterclockwise (by convention) at the
carrier frequency. Now consider, say, 50% modulation with some sinewave, maybe
1/1000 the carrier freq. To represent this, draw two more vectors. The way
we've done it is to start them both at the right end of the first (carrier)
vector. Both are 1/4 unit long. To start, at time=0, draw them both further
to the right from the carrier. Since they are both adding to the carrier, the
net at that point in time is 1.5 units long. Now if the carrier didn't move
(zero freq), one of the little vectors would rotate clockwise and one would
rotate counterclockwise, at just the same rates. (Careful here! The one going
clockwise represents your "negative freq" if you will, but there is NO MATH,
just a picture, so don't let your mind lock up on this one!) They'd get to be
both pointing to the left at just the same time, and at that time they'd
subtract from the carrier and leave you with a vector 0.5 units long. But
before you got to that point, you'd have one of them pointing straight up, and
one pointing down, and they'd cancel out, leaving just the carrier. Now just
imagine all that happening as the carrier rotates them around... it's all just
the same but produces the carrier plus the two sidebands. One key thing to get
from this picture is that the two modulation vectors always sum together to a
vector which is parallel to the carrier vector (or else zero length).

For FM: Draw the same picture, but now the modulation vectors both start
pointing up, at 90 degrees to the carrier. As they rotate around, they always
sum to something that is perpendicular to the carrier vector. Hmmmm...but
notice that if they are very short, the net result is practically the same
length as the carrier vector all the time, but if they are a bit longer, you'd
have the carrier amplitude changing. Draw the picture to see that! Let's say
that each of the two are 1/10 as long as the carrier, so that the result is a
right triangle with the carrier 1 unit long and the modulation 1/5 unit long.
So the net in that case would be sqrt(1^2 + 0.2^2) = 1.02. But this is FM, and
the amplitude is not allowed to change. So we have to put in a correction.
One way to do that is to add a couple more vectors which correct this initial
error. If you think it through, you'll see they have to rotate twice as fast
as the initial two modulation vectors. So the initial ones represent the first
sidebands, and the next pair represent the second sidebands...and if you draw
it out right, you'll be able to see how the whole set of sidebands comes about.
So...why is it FM? Because the sidebands rotate the carrier phase. In fact,
that's how you have to draw the set of modulation vectors: to sum up to a
carrier whose phase is modulated (which is the same as FM, of course, for this
single sine freq modulation).

But notice that if the modulation is low enough, practically all the modulation
energy is in those initial two sidebands, represented by the first two vectors.
Now if you transmitted ONLY those two and removed the carrier, and someone on
the other end inserted the carrier at t=0 pointing UP instead of to the right,
why you'd have -- AM! Or at least something very, very close to AM. So, I
think it should be clear from that, that single sideband FM (assuming very low
modulation index) should be practically equivalent to single sideband AM.

By the way, back several years ago there was a lot of interest in finding ways
to make more efficient AM broadcast transmitters. If you use a class C power
amplifier, you can get good RF-generator efficiency, but the modulator running
class AB or B is inefficient. And if you do the modulation at a low level, you
have to run the RF chain AB or B. So one of the ways invented to get AM was to
generate two FM signals, which of course can be amplified by class C power
amps, whose modulation was generated through a pretty special DSP algorithm, so
that when you combined the RF outputs of the two FM transmitters you got,
ta-da, AM! I always thought that was pretty cool, but I don't think it ever
caught on in a big way, because folk have come up with other ways of
efficiently generating AM.

Cheers,
Tom


(Bruce Kizerian) wrote in message
. com...
Can anyone direct me to some good understandable references on single
sideband frequency modulation? I have no real practical reason for
wanting to know about this. It is interesting to me in a "mathetical"
sort of way. Of course, that is dangerous for me because my brain gets
very stubborn when I try to do math. Such ideas as "negative
frequency" kind of send my mental faculties into total shutdown.

But I read schematic very well. It is a visual language I can usually
understand. Seems like years ago there was an article on SSB FM in Ham
Radio. That would probably be a good start. If anyone can send me a
copy of that article I would be much appreciative.

Thanks in advance

Bruce kk7zz

www.elmerdude.com



Cheers,
Tom