1

PROBABILITY THEORY & STOCHASTIC PROCESSES

II B Tech (ECE) II Semester 2023-24
Compiled by Dr K Prabhakara Rao
Professor, Department of ECE
UNIT-V: Noise Sources: Classification of noise

Contents:
Resistive/Thermal Noise Source, Arbitrary Noise Sources, Effective Noise Temperature,
Noise equivalent bandwidth, Average Noise Figures, Average Noise Figure of cascaded
networks, Narrow Band noise, Quadrature representation of narrow band noise & its
properties. Entropy, Information rate.
Noise
Any unwanted form of energy that tends to interfere with transmission and reception of
signals and causes loss of information is called “Noise”

Noise

External Internal

Man-made Natural Thermal Shot

Extra-terrestrial Atmospheric

Solar Cosmic
 2

Thermal/Resistive noise
The noise produced due to the random movement of the electrons in the conducting material
is called Thermal or resistive noise.
• The motion of each free electron inside the conducting medium is due to the
temperature.
• When temperature is 0𝑜 𝐾 there is no random motion of free electron, hence noise is
zero.
• When temperature increases, random motion of free electron also increases.
• Amplitude of the Thermal noise mainly depends on the resistance. So, it is also called
resistive noise also.

Thermal noise is proportional to the temperature in degree Kelvin and the bandwidth.
𝑃𝑛 ∝ 𝑇𝐵
𝑃𝑛 = 𝐾𝑇𝐵 Watts
𝐾 = Boltzmann’s constant= 1.38𝑋10−23 𝐽/𝑂 𝐾
𝑇= Temperature in degree Kelvin
𝐵= Bandwidth in 𝐻𝑧

Equivalent noise temperature

Equivalent noise temperature is a measure of the noise power generated by a device, such
as an amplifier or a receiver, expressed in terms of the temperature of a hypothetical resistor
that would generate the same amount of thermal noise power. This concept is particularly
important in the design and analysis of receivers and other components in communication systems
where noise performance is critical.

It is often used in the context of individual components such as low-noise amplifiers (LNAs),
mixers, or other electronic components. It focuses on the noise contribution of a single
component by equating it to the noise generated by a resistor at a specific temperature.

If a component has a noise figure F, the noise equivalent temperature can be calculated using
the following formula:

𝑇𝑛 = 𝑇𝑜 (𝐹 − 1)

𝑇𝑜 is the standard reference temperature, typically taken as 300 K.

 3

Effective Noise Temperature

The effective noise temperature is a measure of the noise power added by a system or
component to the signal it processes. It represents the temperature at which an ideal
resistor would produce the same amount of noise power as the actual noise generated by
the system or component. Used for entire systems, including antennas and receivers. It
focuses on the total noise performance of a system, including all noise sources within the
system.
The effective noise temperature is given by
𝑃𝑛
𝑇𝑛 = 𝐾𝐵 Kelvin

𝑇𝑒 = 𝑇𝑜 (𝐹 − 1)
Noise equivalent bandwidth
𝑁𝑜
When a source of White noise of zero mean and power spectral density is connected to an
2
arbitrary LPF of transfer function 𝐻(𝜔), then the average output noise power
𝑁𝑜 ∞
𝑃𝑛 = ∫ |𝐻(𝜔)|2 𝑑𝜔
2 −∞
∞
𝑃𝑛 = 𝑁𝑜 ∫ |𝐻(𝜔)|2 𝑑𝜔
0

𝑁𝑜 − Average noise power per unit bandwidth measured at the input of the receiver.
𝑁𝑜
Next consider the same source of White noise of zero mean and power spectral density is
2
connected to the input of an ideal LPF of zero frequency response 𝐻(0) and bandwidth
𝐵 then the average output noise power
𝑃𝑛 = 𝑁𝑜 𝐵𝐻 2 (0)
Equating the two average output noise powers, the noise equivalent bandwidth can be
defined as,
∞
𝑁𝑜 𝐵𝐻 2 (0) = 𝑁𝑜 ∫ |𝐻(𝜔)|2 𝑑𝜔
0
∞
∫𝟎 |𝑯(𝜔)|𝟐 𝒅𝜔
𝑩𝑵 =
𝑯𝟐 (𝟎)

The Noise Equivalent Bandwidth (B) is defined as the bandwidth of an ideal rectangular filter that
passes the same amount of noise power as the actual filter. In other words, it is the bandwidth of a
hypothetical filter with a flat (constant) frequency response that would result in the same total
noise power as the actual filter with its specific frequency response.
 4

Noise Figure:

The noise figure (NF) is a measure of degradation of the signal-to-noise ratio (SNR) as a signal
pass through a system or a device, such as an amplifier. It quantifies the noise performance of an
electronic device and is expressed as the ratio of the input SNR to the output SNR.

𝑃𝑠𝑖 𝑃𝑠𝑜
Network
𝑆𝑛𝑖 (𝜔) 𝑆𝑛𝑜 (𝜔)
𝑯(𝝎)
𝑃𝑛𝑖 𝑃𝑛𝑜

The output noise power spectral density 𝑆𝑛𝑜 (𝜔) = |𝐻(𝜔)|2 𝑆𝑛𝑖 (𝜔)

𝑃𝑠𝑖 𝑃𝑠𝑜
Two port network

𝑮𝒂 (𝜔)
𝑃𝑛𝑖 𝑃𝑛𝑜

𝑮𝒂 (𝜔) − Available power gain

Mathematically, the noise figure is defined as:

Input Signal to Noise Ratio

𝑭=
Output Signal to Noise Ratio
𝑺𝑵𝑹𝒊𝒏
𝑭=
𝑺𝑵𝑹𝒐𝒖𝒕

𝑆𝑛𝑜 (𝜔) −The total noise power spectral density at the output of the system
𝑆 ′ 𝑛𝑜 (𝜔) − The noise power spectral density at the output due to input noise
𝑆 ′′ 𝑛𝑜 (𝜔) − The noise power spectral density at the output due to the noise generated
internally by the system.

If 𝑭 > 1 the system is said to be noisy.

If the noise figure is given in decibels, it can be converted to either way using:
𝐹𝑑𝐵 = 10 log10 (𝐹𝑙𝑖𝑛𝑒𝑎𝑟 )
 5

Noise Figure of cascaded networks:

Consider two amplifiers that are cascaded.

𝑮𝟏 : Power gain of amplifier1
𝑮𝟐 : Power gain of amplifier2
𝑻𝒆𝟏 : Equivalent input noise temperature of amplifier1
𝑻𝒆𝟐 : Equivalent input noise temperature of amplifier2
𝑭𝟏 : Noise figure of amplifier1
𝑭𝟐 : Noise figure of amplifier2

Noise power at the input of the amplifier1 (𝑃𝑛𝑖1 )= 𝑃𝑛𝑖 + 𝑃𝑛𝑎1

𝑃𝑛𝑖1 = 𝑘𝑇𝐵 + (𝐹1 − 1)𝑘𝑇𝐵

𝑃𝑛𝑖1 = 𝐹1 𝑘𝑇𝐵
Noise power at the output of the amplifier1 (𝑃𝑛𝑜1 )= 𝐹1 𝑘𝑇𝐵𝑮𝟏
Noise power at the input of the amplifier2 (𝑃𝑛𝑖2 )= 𝑃𝑛𝑜1 + 𝑃𝑛𝑎2

𝑃𝑛𝑖2 = 𝑃𝑛𝑜1 + 𝑃𝑛𝑎2

𝑃𝑛𝑖2 = 𝐹1 𝑘𝑇𝐵𝑮𝟏 + (𝐹2 − 1)𝑘𝑇𝐵

Noise power at the output of the amplifier2 (𝑃𝑛𝑜2 )= [𝐹1 𝑘𝑇𝐵𝑮𝟏 + (𝐹2 − 1)𝑘𝑇𝐵)]𝑮𝟐
𝑃𝑛𝑜2 = [𝐹1 𝑘𝑇𝐵𝑮𝟏 + (𝐹2 − 1)𝑘𝑇𝐵]𝑮𝟐
𝑃𝑛𝑜2 = 𝐹1 𝑘𝑇𝐵𝑮𝟏 𝑮𝟐 + 𝐹2 𝑘𝑇𝐵𝑮𝟐 − 𝑘𝑇𝐵𝑮𝟐
𝑃𝑛𝑜2 is the total noise power at the output 𝑃𝑛𝑜
We know that the overall noise figure of the cascaded circuit
1 𝑃𝑛𝑜
𝑭=
𝐺 𝑃𝑛𝑖
Where G is the overall power gain of the cascaded circuit 𝑮 = 𝑮𝟏 𝑮𝟐
 6

1 𝐹1 𝑘𝑇𝐵𝑮𝟏 𝑮𝟐 + 𝐹2 𝑘𝑇𝐵𝑮𝟐 − 𝑘𝑇𝐵𝑮𝟐

𝑭=
𝑮𝟏 𝑮𝟐 𝑘𝑇𝐵
𝐹1 𝑘𝑇𝐵𝑮𝟏 𝑮𝟐 𝐹2 𝑘𝑇𝐵𝑮𝟐 − 𝑘𝑇𝐵𝑮𝟐
𝑭= +
𝑘𝑇𝐵𝑮𝟏 𝑮𝟐 𝑘𝑇𝐵𝑮𝟏 𝑮𝟐
𝐹1 𝑘𝑇𝐵𝑮𝟏 𝑮𝟐 𝑘𝑇𝐵𝑮𝟐 (𝐹2 − 1)
𝑭= +
𝑘𝑇𝐵𝑮𝟏 𝑮𝟐 𝑘𝑇𝐵𝑮𝟏 𝑮𝟐
𝐹2 − 1
𝑭 = 𝐹1 +
𝑮𝟏
𝐹2 −1 𝐹3 −1
Similarly for the 3 amplifiers cascaded 𝑭 = 𝐹1 + +
𝑮𝟏 𝑮𝟏 𝑮𝟐

The overall noise figure of the N amplifiers cascaded,

𝐹2 − 1 𝐹3 − 1 𝐹𝑁 − 1
𝑭 = 𝐹1 + + + ⋯+
𝐺1 𝐺1 𝐺2 𝐺1 𝐺2 … 𝐺𝑁−1
It shows that the contribution to overall noise figure is mainly at the first stage.
The equivalent noise temperature of the N amplifiers cascaded
𝑇𝑒2 𝑇𝑒3 𝑇𝑒 𝑁
𝑇𝑒 = 𝑇𝑒1 + + + ⋯+
𝐺1 𝐺1 𝐺2 𝐺1 𝐺2 … 𝐺𝑁−1
Narrow Band noise:

If a White noise is passed through a band pass filter whose bandwidth is 𝑊𝑁 ≪ 𝜔𝑂 the noise
process appearing at the output is called Narrow band noise. The spectral components of the
narrowband noise are concentrated about some mid band frequency ±𝜔𝑂 .

It occupies a limited portion of the frequency spectrum, typically centered around a specific
frequency. It is characterized by its restricted bandwidth, which makes it different from
broadband noise that spans a wide range of frequencies.

Mathematically, a narrow band noise n(t) can be expressed as a modulated carrier:

𝑛(𝑡) = 𝐴𝑐𝑜𝑠(cos 𝜔𝑂 (𝑡) + ∅(𝑡))

Quadrature representation of narrow band noise:

The quadrature representation of narrow band noise involves expressing the noise in terms
of its in-phase (I) and quadrature (Q) components. This is achieved through a process known
 7

as quadrature demodulation. Mathematically, a narrow band noise 𝑛(𝑡) can be expressed

as, 𝑛(𝑡) = 𝑛𝐼 (𝑡) cos 𝜔𝑂 (𝑡) − 𝑛𝑄 (𝑡) sin 𝜔𝑂 (𝑡)

Here, 𝑛(𝑡) and 𝑛𝑄 (𝑡) are the in-phase and quadrature components, respectively. These
components are derived from the original noise signal by mixing it with a cosine and sine
wave of the center frequency ω0. These components are then low-pass filtered to remove
high-frequency terms.
Properties of narrow band noise:
1. The in-phase component 𝑛(𝑡)and quadrature component 𝑛𝑄 (𝑡) of a narrow band
noise 𝑛(𝑡) have zero mean, since 𝑛(𝑡) has zero mean.
2. If a narrow band noise 𝑛(𝑡) is Gaussian, then its in-phase component 𝑛(𝑡)and
quadrature component 𝑛𝑄 (𝑡) are jointly Gaussian.
3. The in-phase component 𝑛(𝑡)and quadrature component 𝑛𝑄 (𝑡) of a Gaussian narrow
band noise 𝑛(𝑡) with zero mean are statistically independent.
4. If a narrow band noise 𝑛(𝑡) is wide sense stationary, then its in-phase component
𝑛(𝑡)and quadrature component 𝑛𝑄 (𝑡) are jointly wide sense stationary.
5. Both the in-phase component 𝑛(𝑡)and quadrature component 𝑛𝑄 (𝑡) of a narrow band
noise 𝑛(𝑡) have the same power spectral density 𝑆𝑁 (𝜔), which is related to the power
spectral density of the narrow band noise 𝑛(𝑡) as follows,
6. Both the in-phase component 𝑛(𝑡)and quadrature component 𝑛𝑄 (𝑡) of a narrow band
noise 𝑛(𝑡) have the same variance.
7. The cross spectral densities of the quadrature components of the narrow band noise
𝑛(𝑡) are purely imaginary.
8. The in-phase component 𝑛(𝑡)and quadrature component 𝑛𝑄 (𝑡) of a narrow band
noise 𝑛(𝑡)are orthogonal, meaning their cross-correlation is zero for a wide-sense
stationary (WSS) process.
Entropy:
Entropy (in bits) tells us the average amount of information (in bits) that must be delivered
in order to resolve the uncertainty about the outcome of a trial. This is a lower bound on the
number of binary digits that must, on the average, be used to encode our messages.
The entropy is a measure of the uncertainty or randomness associated with a source of
information. It quantifies the average amount of information produced by a stochastic source
of data. Higher entropy indicates more uncertainty or more information content.
Mathematically, for a discrete random variable 𝑋 with possible outcomes {𝑥1 , 𝑥2 ,…, 𝑥𝑛 }and
corresponding probabilities {𝑝1, 𝑝2 ,…, 𝑝𝑛 } the entropy 𝐻(𝑋) is defined as:
𝑁

𝐻(𝑋) = − ∑ 𝑝(𝑥𝑖 ) 𝑙𝑜𝑔2 𝑝(𝑥𝑖 )

𝑖=1
 8

where:
• 𝑝(𝑥𝑖 ) is the probability of the outcome 𝑥𝑖
• 𝑙𝑜𝑔2 denotes the logarithm to the base 2, reflecting the measurement of entropy in
bits.
The unit of measurement (when the log is base-2) is the bit (binary information unit --- not
the same as binary digit).
Information rate:
If the source of the messages generates messages at the rate r per second, then the
information rate is defined to be 𝑅 = 𝑟𝐻 average number of bits of information per second.
𝐻 is the entropy of the source, representing the average amount of information per symbol.
𝑟 is rate at which the number of messages generated per second.

Solved Problems:
1. A receiver has a noise power bandwidth of 15KHz. A resistor which matches with the
receiver input impedance is connected across the antenna terminals. What is the
noise power contributed by this resistor in the receiver bandwidth at temperature
270C?
Solution:
B = 15 KHz,
T= 270+30=300K
Noise power 𝑃𝑛 = 𝑘𝑇𝐵 = 1.38𝑋 10−23 x 300 x 15x 103
𝑃𝑛 = 6021 x 10−20
𝑃𝑛 = 6.21 x 10−17 W.

2. A receiver connected to an antenna whose resistance is 50 ohms has an equivalent

noise resistance of 30 ohm. Calculate the receiver’s noise figure in dB and its
equivalent noise temperature.
Solution:
𝑅𝑎 + 𝑅𝑒 50 + 30
𝐹= = = 1.6
𝑅𝑎 50
Noise Figure (𝑑𝐵)= 10 log10 (𝐹) = 10 log10 (1.6) = 2.041 dB
Equivalent noise temperature (𝑇𝑛 ) = 𝑇𝑜 (𝐹 − 1)
Equivalent noise temperature (𝑇𝑛 ) = 300(1.6 − 1) = 180𝐾

References:
[1] Probability Theory and Stochastic Processes, Y Mallikarjun Reddy, University Press,
[2] Probability Random Variables and Random Signal Principles, Peyton Z Peebles, JR,
Tata Mc Graw Hill,4e