Richard Harrison wrote:
FM usually occupies about twice the bandwidth of AM.

Thanks for adding "usually" to your original statement.
I was just pointing out that if the FM peak deviation
is equal to the maximum modulation frequency, then the
FM signal occupies the same bandwidth as AM. The S/N
ratio advantage usually enjoyed by FM over AM occurs
when the FM peak deviation is *greater than* the maximum
modulation frequency. FM seems to have been the original
"spread spectrum" mode. :-)
--
73, Cecil http://www.w5dxp.com

In article ,
Cecil Moore wrote:

FM usually occupies about twice the bandwidth of AM.

Thanks for adding "usually" to your original statement.
I was just pointing out that if the FM peak deviation
is equal to the maximum modulation frequency, then the
FM signal occupies the same bandwidth as AM.

I believe that you're mistaken on this point, Cecil.

As I understand it, the spectrum of a frequency-modulated carrier
will have energy at the carrier frequency (except at certain very
specific modulation indices), and energy at offsets from the carrier
which are equal to the modulating frequency and all of its harmonics.

It's entirely possible (and in fact common) for an FM signal to have
energy at frequencies which are further away from the nominal carrier
frequency, than the maximum instantaneous deviation of the carrier
would suggest. For example, a carrier which is modulated with a 1000
Hz tone, but only to the level of having a maximum peak deviation of
500 Hz from nominal, will still have sideband energy out 1000 Hz and
2000 Hz away from the nominal carrier frequency.

This is *really* counter-intuitive, and I haven't been able to fully
wrap my brain around the question of just how the math works... but
the math says that it's true, and my own spectrum measurements show
that it's true.

An FCC-approved NBFM phone signal, which has a modulation index of no
more than 1.0 at the highest modulating frequency, *will* have energy
out further than the peak deviation would suggest. There will be some
amount of sideband energy located out at twice the highest modulating
frequency, and a bit at three times.

The levels of these further-away-from-nominal-carrier sidebands will
be relatively low - they don't start to become appreciable until you
get to a higher modulation index.

An AM signal being modulated by the same intelligence signal would not
have any energy out at the multiples, *if* it was generated and
transmitted in an entirely-distortion-free manner.

My guess is that in practice, NBMF (per FCC regs), and ham-grade AM,
probably have very similar bandwidths. All it would take for the AM
signal to be spread out as far as the FM signal, would be a bit of
nonlinear distortion in the audio chain, mixer, or amplifier chain...
this will create a second-harmonic sideband.

I did an introductory presentation to a local hamclub last year (with
some math, admittedly simplified and approximated and covered up with
moderate amounts of hand-waving) which explains some of these
concepts. I included some plots of actual measured spectra, generated
by an HP signal generator and captured using an old Systron Donner
spectrum analyzer.

http://www.radagast.org/~dplatt/hamr...modulation.pdf

--
Dave Platt AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!

Richard Harrison wrote:
. . .
Terman says about twice the bandwidth is required for FM as compared
with AM on pages 589 and page 590 of his 1955 opus.
. . .

You really should try to understand the context of the various
quotations from Terman.

In his _Radio Engineering_, Third Edition (1947), he points out that
"When the modulation index is less than 0.5, i.e., when the frequency
deviation is less than half the modulating frequency, the second and
higher order side-band components are relatively small, and the
frequency band required to accommodate the essential part part is the
same as in amplitude modulation." This is, of course, what is considered
to be narrow band FM.

Unlike an AM signal with its one pair of sidebands containing replicas
of the modulating signal, any FM signal contains an infinite number of
pairs of sidebands. However, as Terman and any other communications text
points out, the relative strengths of some of those sidebands can be
made to be very small by the choice of modulation index. In the case of
NBFM, all but the first pair are small. That first pair are spaced the
same distance from the carrier as AM sidebands, so the bandwidth is
essentially the same as for AM.

You can, of course, increase the modulation index which increases the
bandwidth by increasing the amplitudes of higher order sideband pairs,
making wideband FM. The advantage of doing this is that you can improve
the signal/noise ratio of the received signal as a trade for the
increased bandwidth.

So FM can be as narrow in bandwidth as AM, or any greater bandwidth, all
depending on the modulation index. Saying that "twice the bandwidth is
required for FM as compared with AM" is simply not a correct statement
and, if said by Terman, was taken out of context which surely qualified it.

Roy Lewallen, W7EL