SAMPLING

GROUP MEMBERS 2
 3

INTRODUCTION
• Sampling is the process of measuring the aptitude of a continuous time signal at
discrete instant converting the continuous signal to discrete signals. It converts
analog signals to the discrete values
• In embedded systems, sampling is crucial for converting continuous real-world
signals into digital data that can be processed by microcontrollers or processors.
• Embedded systems often interface with analog sensors that measure physical
quantities such as temperature, pressure, or sound.
• Application: Sampling is essential because it prevents any information loss
during the transmission loss. It also increases the accuracy of the system.
Sampling is used in various processes, such as PAM, PCM, and TDM.
 Sampling /Nyquist Theorem

• Statement: A continuous time signal can be represented

in its samples and can be recovered back when sampling
frequency fs is greater than or equal to the twice the
highest frequency component of message signal. i.e.
Fs ≥ 2Fm

Sampling rate
Sampling rate is defined as the number of samples taken per second from a continuous
signal for a finite set of values. Also defined as a sampling frequency, which is the
reciprocal of the sampling time.

Fs = 1/Ts

• Fs is the sampling frequency

• Ts is the sampling time
It helps in the successful recovery of the digital signal at the receiving end. Hence, a
fixed parameter was defined for the sampling rate, known as Nyquist rate.
NYQUIST RATE (MINIMUM SAMPLING
RATE)
Suppose H is the highest selected frequency. A bandlimited signal is transmitted at the
frequency components lower than W Hz. Thus, for the replication of the original signal,
the sampling rate should be twice the highest frequency.

Fs = 2W
• Fs is the sampling rate
• W is the highest frequency
The sampling at the Nyquist rate does not introduce any distortion.

NYQUIST INTERVAL
Nyquist interval is the reciprocal of the Nyquist rate.

TS = 1/2W

• TS is the Nyquist Interval

• W is the highest frequency
SAMPLING METHODS 6

1. Ideal sampling
Ideal sampling is also known as instantaneous sampling or impulse sampling.
The sampling process multiplies the input signal and the carrier signal, which
is present in the form of train of pulses. It cannot easily be implemented
The working principle that multiples the input signal and the sampling signal
is known as “multiplication principle”
2. Natural sampling 7
Natural Sampling is considered an
efficient multiplexing and practical
method in Pulse Amplitude Modulation,
where the pulse has a finite width. Here,
the analogue signal is multiplied by the
uniformly spaced rectangular pulses.
Fig 2 shows the waveforms of the
message signal, sampling signal, and the
sampled signal

3. Flat top sampling

The design and reconstruction of flat-top
sampling is easy than the natural sampling
process. The pulses in the flat-top sampling
method are in the flat shape at the top and are
held at a constant height by using a circuit. It
means that the samples are flat and have
constant amplitude.
4. Uniform Sampling
8
In uniform sampling, the input signal is sampled at a
constant rate, typically determined by a system clock. This
regular interval ensures that the digital representation of the
analog signal is consistent. Uniform sampling is commonly
used in digital control systems, signal processing, and
communication systems.
The figure shows the time interval T is called the
sampling period, whereas fs=1/T is referred to as the
sampling frequency, which thus denotes the number of
sampling instances per second.

5. Non-Uniform Sampling
Some applications require non-uniform
sampling, where the intervals between
samples vary. This can be useful in systems
where events occur sporadically, such as in
event-driven or asynchronous data acquisition
systems. Non-uniform sampling can optimize
resource usage, especially when system
power or memory is limited.
6. Oversampling 9
Oversampling is a technique where the signal is sampled at a rate much
higher than the Nyquist rate (twice the maximum frequency of the signal).
Oversampling can help improve the signal-to-noise ratio (SNR) and reduce
quantization noise.
The figure above shows a single tone of random noise in the frequency range DC
to Fs/2 for both the frequency domain and oversampling cases
 SAMPLING AND REAL-TIME
CONSTRAINTS 10

Sampling has to be done at precise intervals to meet real-time processing constraints. Failing to meet
these constraints can lead to inaccuracies or system malfunctions.
1. Interrupt-Driven Sampling
Embedded systems often use hardware interrupts to ensure that samples are taken at the right
moment, without missing important data. The ADC can be triggered by a timer interrupt, ensuring
the sampling process is synchronized with the system’s timing requirements.
2. DMA-Based Sampling
In systems with high-speed ADCs, Direct Memory Access (DMA) can be used to transfer sampled
data from the ADC directly to memory without the intervention of the processor. This reduces the
processing overhead and ensures that high-frequency signals are captured without loss.
FACTORS AFFECTING SAMPLING ACCURACY IN 11
EMBEDDED SYSTEMS
1. Quantization Error
The precision of an embedded system’s ADC is limited by its bit resolution. For instance, an 8-bit ADC
can only represent 256 discrete levels, while a 12-bit ADC can represent 4096 levels. The difference
between the actual analog signal and its digital representation is called quantization error. Increasing
the resolution of the ADC can help reduce quantization error, but this may also increase the system's
cost and complexity.
2. Aliasing
Aliasing occurs when a signal is sampled at a rate lower than its Nyquist rate, causing high-frequency
components of the signal to be indistinguishable from lower-frequency components. This can distort
the digital representation of the signal. Anti-aliasing filters are typically used before sampling to
remove high-frequency components above the Nyquist frequency, ensuring accurate signal
reconstruction.
 12
EMBEDDED SYSTEM APPLICATIONS OF SAMPLING
1. Sensor Data Acquisition
Many embedded systems rely on sensors to monitor the physical environment. The data from these
sensors is often analog in nature (e.g., temperature, humidity, or pressure) and must be sampled and
digitized for further processing. Sampling plays a key role in ensuring accurate data acquisition, which
is critical for systems that perform real-time monitoring or control.
2. Digital Signal Processing (DSP)
Sampling is the first step in digital signal processing applications. Once an analog signal is sampled
and digitized, it can be processed using DSP techniques, such as filtering, compression, or modulation.
These processes are widely used in communication systems, audio processing, and radar systems.
 13
POWER AND PERFORMANCE TRADE-OFFS
In embedded systems, there’s always a trade-off between power consumption and performance. Sampling rate, ADC
resolution, and signal processing capabilities all affect the system’s power usage.
• Low-Power Sampling
For systems that need to operate for long periods on limited battery power, low-power sampling techniques are
crucial. This can include strategies like lowering the sampling rate when the system is idle, using low-power ADCs,
or employing event-based sampling to reduce unnecessary data acquisition.
• High-Performance Sampling
Some embedded systems require high-performance sampling, especially in applications like real-time video
processing or high-frequency signal acquisition. These systems need powerful processors, high-speed ADCs, and
efficient data transfer mechanisms (e.g., DMA) to handle the large volumes of data generated during high-speed
sampling.

MERITS OF SAMPLING
 High accuracy
 Easy to implement
 Less time consuming
 Low signal loss
 High scope
 Low cost
THANK
YOU