Page 73 - Complete Wireless Design

Modulation

72 Chapter Two

As stated above, this peak-to-average power ratio, if high, can cause exces-
sive inefficiency since the amplifier must be backed off by this amount to allow
for the occasional peaks. This ratio, however, will vary with the symbol pat-
terns and clock speeds, as well as the channel filter and bandwidth. General
peak-to-average ratios, nevertheless, are at least 5 dB for QPSK, 8 dB for 64-
QAM and orthogonal frequency division multiplexing (OFDM), and up to 15 dB
for code division multiple access (CDMA). This means that for some QPSK
modulations, the intermittent peaks will rise above the RMS power by 5 dB.
To measure this peak-to-average ratio, first measure the power peaks with
a fast-acting, digital-modulation-capable, peak-reading power meter (such as
the Boonton 4400) over a time of at least 10 seconds. This will give a reason-
ably accurate indication of the signal’s peak power. To measure the average
power, see “Measuring digital signal power” in Sec. 2.6.2, or use a special dig-
ital-modulation-compliant average power meter. Now subtract the average
from the peak, in dBm. This equals the peak-to-average ratio in dB.
To recap this very important concept of digital signal power: Digital signals,
because of their nonrepetitive, random nature, as well as the fact that all of
their power is spread over frequency rather than condensed as it is in analog
signals on all sides of the carrier, have no predictable, recurrent peak power
points to measure. Since it is difficult to measure these peaks in a nonstatis-
tical manner, we are forced to take an average measurement of the digital sig-
nal’s power over its entire bandwidth. But how much bandwidth does the
normal digital signal consume? It is considered to be the 30-dB bandwidth
where most of the power of the digital signal resides, rather than just the 3-
dB bandwidth used in most analog signal measurements.
Any digital baseband signal we work with will also have been filtered, since
any unfiltered (ideal) digital signal would theoretically take up infinite band-
width. But filtering a digital signal will cause the signal to go from a square
wave to a more rounded signal. This allows the signal to be placed into a nar-
rower bandwidth; but will also increase the required power that the power
amplifier (PA) of the digital transmitter must occasionally transmit. In fact,
this filtering is mainly responsible for the peak-to-average problem we just
discussed (along with the signal passing through the origin), so the more fil-
tering of the digital square wave we employ to decrease the bandwidth, the
more we create a higher peak-to-average ratio. Thus, the PA must be backed
off in power to allow these occasional peaks in power (Fig. 2.31) to be sent
through the PA without causing nonlinear (IMD) performance and a widening
of the bandwidth. More detail on filtering of digital signals is in Sec. 2.4.4,
“Digital modulation issues.”

2.4.4 Digital modulation issues
In digital communications, everything is geared to not only supplying reliable
communications at the lowest transmitted power and bandwidth practical, but
also to maximizing the data rate. In fact, bandwidth, power, noise, and infor-
mation capacity are all interrelated by Shannon’s information theorem, which

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)
Copyright © 2004 The McGraw-Hill Companies. All rights reserved.
Any use is subject to the Terms of Use as given at the website.