Receiver Link Budget Analysis

Because processing gain reduces as bit rate increases, receiver sensitivity must be
determined across all possible data rates and for a required Eb/No (briefly, the ratio of
energy per bit to the spectral noise density; we will discuss this further shortly). The
calculation needs to comprehend the performance of the demodulator, which, in turn,
is dependent on the level of modulation used. Other factors determining receiver sensitivity
include the RF front end, mixer, IF stages, analog-to-digital converter, and baseband
process (DSP). (See Figure 3.19.)
Let’s look at a worked example in which we define receiver sensitivity. For example,
let’s determine receiver sensitivity at three data rates: 12.2 kbps, 64 kbps, and 1920
kbps at a BER of 1 in 106.
The noise power is dimensioned by Boltzman’s constant (k = 1.38 × 10-23 J/K) and
standardized to a temperature (T) of 290K (17° C). To make the value applicable to any
calculation, it is normalized at a 1 Hz bandwidth. The value (k × T) is then multiplied
up by the bandwidth (B) used.
The noise power value is then -174 dBm/Hz and is used as the floor reference in sensitivity/
noise calculations. The receiver front end (RF + mixer) bandwidth is 60 MHz,
in order to encompass IMT2000DS license options. The noise bandwidth of the front
end is 10log10(60MHz) = 77.8 dB. The receiver front noise floor reference is therefore
-174 dBm +77.8 dB = -96.2 dBm. In the DSP, the CDMA signal is despread from 3.84
Mcps (occupying a 5 MHz bandwidth), to one of the three test data rates—12.2 kbps,
64 kbps, 1920 kbps—and can be further filtered to a bandwidth of approximately:
 Modulation bandwidth = Data rate × (1+ α)/log2(M) (where α = pulse-shaping
filter roll-off and M = no of symbol states in modulation format)
 For IMT2000, α = 0.22 and M=4 (QPSK)
Thus, reduction in receiver noise due to despreading is as follows:
 = 10log10(IF bandwidth/modulation BW)
 = 10log10(5 MHz/7.5 kHz) = 28.2 dB for 12.2 kbps
 = 10log10(5 MHz/39 kHz) = 21.1 dB for 64 kbps
 = 10log10(5 MHz/1.25 MHz) = 6.0 dB for 1920 Mbps

The effective receiver noise at each data detector due to input thermal noise is thus:
SOURCE RECEIVER NOISE EFFECTIVE RECEIVER
DATA REFERENCE NOISE
12.2 kbps -107 dBm-28.2 dB = -135.2 dBm
64 kbps -107 dBm-21.1 dB = -128.1 dBm
1920 kbps -107 dBm-6.00 dB = -113.0 dBm
The real noise floor for a practical receiver will always be higher because of filter
losses, LNA and mixer noise, synthesizer noise, and so on. In a well-designed receiver,
5 dB might be a reasonable figure. The practical effective noise floor of a receiver would
then be
12.2 kbps -135.2 dBm + 5 dB = -130.2 dBm
64 kbps -128.1 dBm + 5 dB = -123.1 dBm
1920 kbps -113 dBm + 5 dB = -108.0 dBm
Using these figures as a basis, a calculation may be made of the receiver sensitivity.
To determine receiver sensitivity, you must consider the minimum acceptable output
quality from the radio. This minimum acceptable output quality (SINAD in analog systems,
BER in digital systems) will be produced by a particular RF signal input level at
the front end of the receiver. This signal input level defines the sensitivity of the receiver.
To achieve the target output quality (1 × 10-6 in this example), a specified signal (or
carrier) quality is required at the input to the data demodulator. The quality of the
demodulator signal is defined by its Eb/ No value, where Eb is the energy per bit of
information and No is the noise power density (that is, the thermal noise in 1 Hz of
bandwidth). The demodulator output quality is expressed as BER, as shown in Figure
3.20. In the figure, a BER of 1 in 106 requires an Eb/ No of 10.5 dB.
Because receiver sensitivity is usually specified in terms of the input signal power
(in dBm) for a given BER, and since we have determined the equivalent noise power in
the data demodulator bandwidth, we need to express our Eb/No value as an S/N
value. The S/N is obtained by applying both the data rate (R) and modulation bandwidth
(BM) to the signal, as follows:
 S/N = (Eb/No) × (R/BM)
 For QPSK (M=4), BM ~ R/2, thus:
 S/N = (Eb/No) × 2 = 14.5dB for BER = 1in 106
Assuming a coding gain of 8 dB, we can now determine the required signal power
(receive sensitivity) at the receiver to ensure we meet the (14.5-8) dB = 6.5 dB S/N target.
RECEIVER SENSITIVITY
DATA RATE EFFECTIVE NOISE FOR 1 IN 106 BER
12.2 kbps -130.2 dBm -124.7 dBm
64 kbps -123.1 dBm -116.6 dBm
2 Mbps -108.0 dBm -101.5 dBm

There is approximately 22 dB difference in sensitivity between 12.2 kbps speech and
2 Mbps data transfer, which will translate into a range reduction of approximately 50
percent, assuming r4 propagation, and a reduction in coverage area of some 75 percent!
83