[Date Prev][Date Next][Thread Prev][Thread Next] - [Date Index][Thread Index][Author Index]

Re: Geosynchronous Satellites

>but it could possibly count as a constant envelope form of modulation), I
>think BPSK can do be done in AM, PSK can (as we see on HF).  I'll defer to
>Phil for more definitive answers.  I think there are multiple versions of

Suppressed-carrier BPSK can be seen either as phase shift keying, with
signal points on the unit circle at 0 and 180 degrees, or as
suppressed-carrier double sideband AM. Both are equivalent.

The design of a classic digital modulation scheme involves laying out
a "constellation" of signal points on a two-dimensional plot that
represents the instantaneous state of the RF carrier.  The plot axes
are generally labeled I (in phase) and Q (quadrature), because the
usual way to generate RF from the baseband I & Q data is to run each
channel through a balanced modulator and combine the results. The
in-phase carrier is fed to the I-channel modulator while the carrier
fed to the Q-channel modulator is shifted exactly 90 degrees

The distance from the origin to the signal point is the instantaneous
carrier amplitude for that point, and the angle at the origin between
the +I axis and the signal point is the instantaneous carrier phase.

It is not necessary for the each signal element to be a single point
on this graph. E.g., in FSK the signal vector is in constant motion,
tracing out either a clockwise or counterclockwise circle around the
origin at a speed that depending on the current frequency "shift". In
MSK (minimum shift keying) the vector always rotates exactly 90
degrees (clockwise or counterclockwise) during each bit interval.

If the signal constellation points (or trajectories, as in MSK/FSK)
all fall on the unit circle, then the signal envelope is constant and
it can be amplified by a nonlinear amplifier without distortion. This
is the case for unfiltered BPSK and QPSK as well as MSK and FSK.

It's possible to lay out a constellation where the carrier amplitude
varies as well as the carrier phase, i.e., the points are not all on
the unit circle. QAM is the generic term here.  Generally the points
in QAM are laid out evenly in a square grid. The main reason to do
this is to pack more data bits into each signal point, thus improving
bandwidth efficiency. (Bandwidth depends on how fast you change the
signal point, not on how many bits are represented by each signal
point). But this comes at a cost. Because the points are packed more
closely, noise is more likely to move one point onto another, causing
an error. And any amplitude nonlinearities in the channel will
compress the outer points together, or even squash them all onto the
unit circle (if the amplifier is totally nonlinear). This obviously
makes it not work very well.

QAM is great when SNR is high, the channel is reasonably linear and
bandwidth is limited. But it's a poor choice on a satellite channel
when the SNR is low, the channel is quite possibly nonlinear, and
bandwidth is more plentiful. That's why you see QAM used on telephone
lines and in cable modems, but not on satellites, where BPSK and QPSK
are usually the way to go.


Via the amsat-bb mailing list at AMSAT.ORG courtesy of AMSAT-NA.
To unsubscribe, send "unsubscribe amsat-bb" to Majordomo@amsat.org