Digital Signal Processing Reference
In-Depth Information
point on, the remaining part of the baseband receive chain is no different than
this used in a regular narrowband qpsk radio receiver with a digital signal
processor in the back-end.
In essence, the reader should consider the pulse-to-baseband conversion pro-
cess as being the equivalent of the passband-to-baseband conversion in a
heterodyne receiver. In combination with the coarse interference suppres-
sion of the receive window, this type of
bandwidth compression
forms one
of the biggest advantages of the pulse-oriented system over an ofdm-based
approach of low-psd wideband radio. Since power consumption has been in-
troduced early in the design process, the parallel processing topology of pulse-
based radio system is - by design - far more energy-efficient than the ofdm
topology.
Low-pass filtering from an energy point of view
From a mathematical signal processing (dsp) point of view, rejecting
all signal components beyond the symbol sampling frequency is al-
most a criminal act, since a lot of valuable signal power is wasted. An
analog designer sees things somewhat different though. A low-pass
filter is typically implemented as a resistor in series with a capacitor.
In this particular case, the resistor is determined by the real part of
the output impedance of the downconversion mixer (see Figure 6.5).
The low-pass characteristic is obtained by adding a capacitor in paral-
lel with the output node of the mixer circuit. The capacitor is charged
and discharged by the output current of the mixer, but the only thing
that is important for the snr of the signal is the amount of
signal en-
ergy
that gets stored on this capacitor. No matter what the value of this
capacitor may be, half of the signal energy is always lost during the
transfer from the mixer to the capacitor [Tse95]. This implies that the
size of the capacitor - and thus also the location of the pole of the fil-
ter - is of no importance whatsoever for the loss of signal energy. It is
demonstrated that caution is needed in the interpretation of mathemat-
ical results without taking the elementary principles of electronics into
account.
I
charge
DSP low-pass
R
C
frequency
analog low-pass
