# Embedded Computer Systems Elec 296 - MyWeb at WIT by jianglifang

VIEWS: 0 PAGES: 42

• pg 1
```									Elec 471 Embedded Computer Systems
Chapter 7, Data Acquisition Systems

By Prof. Tim Johnson, PE
Wentworth Institute of Technology
Boston, MA
Theory and Design for Mechanical Measurement
by Richard Figliola
Content
• Sampling concepts
–   Sample Rate
–   Alias Frequencies
–   Amplitude Ambiguity
–   Selecting sample rate and data number
•   Analogue Converters: ADC and DAC
•   Data Acquisition System Overview
•   Data Acquisition Component parts
•   Data Communications
–   Single vs. Double Ended
–   Serial: RS232, SPI, I2C
–   Parallel: GPIB,
–   Wireless
Chapter introduction
• Concept: analogue signals are continuous in
amplitude and time while digital signal are
discrete (meaning countable).
• Digital signal are non-continuous in amplitude
and time however individual data signals are valid
representations for amplitude and time.
• Sampling is the process of making a continuous
signal discrete.
• The problems and solutions for digitization are
presented.
Introduction to sampling

• The question that confronts the digital engineer is how
well does the sampling represent the analogue signal?
• The answer depends upon :
– the frequency content of the signal
– The size of the period between sample (ADC frequency)
– The total sample period of the measurement
Sampling Concepts
• A Discrete Fourier Transform (DFT) conveys all
the information needed to reconstruct a
Fourier series of a continuous signal from a
digital signal.
• Fourier analysis will provide some guidelines
for sampling continuous data.
• Let δt be the interval between samples so the
sample rate (sample frequency, fs) is
fs = 1/δt
Nyquist Frequency
• The sampling theorem states that to reconstruct
the frequency content of a signal accurately the
sample rate fs must be more than twice the
highest frequency content of the signal.
• The highest frequency of a signal that can be
sampled at any particular sampling rate is called
the Nyquist frequency .
• Let fm represent this maximum frequency.
• fs > 2 fm substituting for fs we get 1/δt > 2 fm
• Inverting the equation reverses the > sign:
δt < 1/2fm
Effect of sample rate

Lost amplitude content   Lost frequency content
Alias Frequencies
• When the signal is sampled at a sample rate that
is less the 2fm the higher frequency assumes the
identity of a lower frequency.
• An example of this is when the wagon wheel on a
covered wagon in the early days of television
would turn backwards.
• The folding point of the aliasing phenomenon
seen in the next slide is at the Nyquist frequency:
fN = fs/2 = 1/2δt
Frequency aliasing diagram

Frequency values seen will increase until the signal frequency reaches the Nyquist
frequency then they will begin to decrease and can even go back to zero until they
start to increase again.
Amplitude Ambiguity
• As long as the sampling rate is twice the sampled
signal’s highest frequency the frequency content
remains valid for the digital signal but this rule
doesn’t apply to amplitude fidelity.
• Consider N an integer. If for some value of N,
N*δt equals the time period, Ts, of the frequency
of the sampled signal then the DFT (the digitized
signal) will contain the correct values for the
amplitude of the signal.
• This can be seen in time diagrams and frequency
spectrum diagrams of a signal on the next slide.
This phenomenon is
known as signal leakage

The truncated wave creates a
modulated signal thus creating
sidebands that suck the power
from the correct amplitude value
shown at the sampled frequency
Sampling summary
• The number of data points and the sample
rate should be chosen to satisfy the criteria of
fs > 2 fm

δf = 1/Nδt = fm/N

In general, set fs ≥ 5 fm
Digital to Analog Converter
DAC schematic

This is a design you can
build yourself. The
output bits control the
switching of the source
voltage to each resistor.
Analog to Digital Conversion
• An ADC converts a voltage into a binary value by a
process known as quantization.
• The analog side of the ADC is characterized by its full
scale voltage range, VFSR.
• Resolution, Q, is the smallest voltage increment that
will cause the LSB to change. Q = VFSR/2M
• Quantization error is caused by a voltage value falling
between two different digital values. It behaves as
noise imposed on the digital signal.
• The error value can vary between the LSB step value
and ½ LSB. The span of the quantization error is the
same as the resolution.
Quantization, Bias, & Saturation

Making the quantization symmetric about the input shifts the dotted line, the
presumed linear input voltage thus changing the bias constant v in the linear
equation y=mx + b. The bias only needs to be moved a negative ½ LSB voltage.
Conversion errors
• Sources of errors in the output voltage can be
resolved into hysteresis, linearity, sensitivity,
zero and repeatability errors . These were
covered in Chapter 2.
• The extent of the errors depend on the
• Other factors include settling time, signal
noise during sampling, leakage, and
temperature.
Successive Approximation Converters
• In a SAC ADC a comparator compares the input voltage to a
reference voltage derived from a DAC value based upon the binary
evaluation using the MSB first and setting each subsequent bit
higher (1) or lower (0) until all the bits in the ADC register are filled.
• In this scheme the starting reference value is ½ FSR or 0x80 for an
eight bit converter.
• If the input voltage is higher(lower), the MSB of the register is set at
1(0) and then the next bit is evaluated after the register value is
converted into the reference voltage by the DAC converter.
• The time of conversion is Mbits*bit conversion speed. This
conversion can be effect by noise at the instance of conversion
giving weighted value to small noise.
• The converter requires the input voltage be held constant during the
conversion process so a Sample and Hold is required on the input.
A Successive Approximation Converter
Ramp (integrating) Converters
• Low-level voltage measurement use this converter for
its low-noise features.
• The main components are a comparator, a voltage
generator that increases linearly that the name ramp,
and the counter and the M-bit register.
• The basic principle of this converter is that as a
reference voltage is increased on a timed basis, the
comparator flips thus stopping the comparison. The
counter is then used to determine the input voltage
based on time taken by the ramp generator to cross
the input voltage.
Ramp Converter block diagram

Using a capacitor of known fixed value charged by a fixed current source the time to
charge can be calculated. The coulombs charge C, is related to the current flow
(q=It) divided by the voltage (C=q/V). Since the charge is known, the rate of current
flow is know, substituting in It solves for V with the value of the count known (the
count stops when the comparator flips determining the time, t): V=It/C
Parallel Converters
• Parallel are fast as one
comparator is used for
each bit.
• A highly stable reference
voltage divides the voltage
through a precision
resistor network with the
MSB resistor dropping the
voltage by ½ VFSR and so
on down the line by ½ at
each branch.
• Time for conversion is only
the time for one
comparator to switch.
• Very fast but expensive.
Delta Sigma Conversion

• The sigma-delta (Σ-Δ) ADC is the converter of choice
for modern voiceband, audio, and high-resolution
precision industrial measurement applications. The
highly digital architecture is ideally suited for modern
fine-line CMOS processes, thereby allowing easy
addition of digital functionality without significantly
increasing the cost. Because of its widespread use, it is
important to understand the fundamental principles
behind this converter architecture. *

*http://www.analog.com/static/imported-files/tutorials/MT-022.pdf
One-bit comparator output

Source http://www.analog.com/static/imported-files/tutorials/MT-022.pdf

Source http://en.wikipedia.org/wiki/Sigma-delta_converter
Oversampling filter effects

Source http://www.analog.com/static/imported-files/tutorials/MT-022.pdf
Signal to Noise Ratio in ADC
• The SNR is improved by higher resolution as seen
in the table below.
• Digital SNR is defined as the ratio of power in the
signal to the noise power thus the larger the
number the smaller the noise and the clearer the
signal: SNR[dB] = 20 log 2M
Note: If there is an amplifier in
resolution in Table 7.2 modifies
the formulas to 20 log(Gain*2M)
and Q = VFSR/(Gain*2M)
Data Acquisition Systems
• This is a systematic look at the usage for an
ADC converter after the value is sensed.
• A DAS is that portion of a measurement
system that quantifies/store/displays/controls
the value.
Block Diagram of a DAS

DAS Components
• A voltage divider (lower right) is a very handy and
simple two that is used for two different purposes.
• One—converts a current source sensor into a voltage
limiting resistor.
• Two—converts a high voltage into a lower voltage
• A shunt circuit (lower left) also converts a current
source into a voltage source if the current is “low”.
Nulling Circuit
• In order for an ADC to work correctly the signal has to
have extraneous voltages removed.
• The nulling circuit sets a sensor to zero if it’s output
should be zero. Its use is like the adjustment knob on
a bathroom scale when no one is standing on the scale
except the scale mechanism.
This circuit is essential a
Wheatstone bridge circuit
that has a potentiometer
across three terminals. It
is used to balance the
potential between CH+
and CH- (equal to zero).
Single and Double Ended sources
• A single-ended connection has one signal line (+High)
measured relative to ground. Single ended are the
usual connections we use; not obvious when using
local grounds is all the devices share a common
ground.
• A double-ended connection uses two signal lines due
to different ground potential at either end. A signal
from the local grounds (prevent ground loops).
Pull-up/Pull-down Resistor
• In circuits where an input level is being detected,
have a pull-up resistor can save the battery power.
• In the accompanying diagram of a pull-up resistor,
having the switch open allows the logic gate to
sense the input voltage level without having to
supply the voltage. When the switch is closed, the
ground can be seen by the logic gate because of
the voltage drop across the resistor.
• Switching the position of the switch and the
resistor will change what value the logic gate will
sense when the switch is in the open position. In
this case that would be a low so that is called a
pull-down resistor.
• In actual circuits the switch is replaced with a
CMOS device resulting in an open-drain terminal.
• This arrangement is very useful in 12C circuits.
Data Communications
• Once the data has been acquired and digitized it is
moved to either: data memory or to immediate data
display or to data communication output.
• Modern devices require the capability for all three
destinations.
• The ADC value goes onto an internal data bus inside a
device.
• Software controls distribution nowadays rather than
hardware as seen in Figure 7-14 and the following
schematic of a 1990’s style ADC/DAC board from NI.
• Depending on the type of bus communication selected
different controllers are used to host the bus.
Data acquisition plug-in board

Note: not
shown in this
plug-in board
is the UART
for serial I/O
UART
• A Universal Asynchronous Receiver/Transmitter, abbreviated UART (
/ˈjuːɑrt/), is a type of "asynchronous receiver/transmitter", a piece of
computer hardware that translates data between parallel and serial forms.
The universal designation indicates that the data format and transmission
speeds are configurable and that the actual electric signaling levels and
methods (such as differential signaling etc.) typically are handled by a
special driver circuit external to the UART.*

• A UART is usually an individual (or part of an) integrated circuit used for
serial communications over a computer or peripheral device serial port.
UARTs are now commonly included in microcontrollers. A dual UART, or
DUART, combines two UARTs into a single chip. Many modern ICs now
come with a UART that can also communicate synchronously; these
devices are called USARTs (universal synchronous/asynchronous
• A USI is Universal Serial Interface is a stripped-down UART that specializes
in hosting two wire communication protocols: SPI, I2C, UNI/O, and 1-Wire.

*Source: http://en.wikipedia.org/wiki/UART
RS232 protocol
• One of the first major data communication
buses was RS-232C protocol for serial digital
communication over telephone lines.
• There is still a huge number of installed devices
that utilize this serial protocol characterized by a
DB9 connection on the back of units. This is a
slow connection (< 9600 bps).

There is a plethora of
assistance on this connector
and RS232 on the internet.
Universal Serial Interface
• SPI—Synchronous Peripheral Interface handles a three-
wire communication protocol using separate
transmitter, receive, and clock lines to a dedicated
device
• I2C—Inter-Integrated Circuits is a two-wire
communication protocol utilizing a data line and a
clock line. Control signals are passed on the bus when
data is not being transmitted for up to 8 different
devices of 16 different types (7 bit addresses).
• SMBus and PMBus are derivatives of 12C bus.
• No license is required for I2C so it will be around for
quite some time. The dropping resistors on the bus
mean that small microcontrollers operating on
batteries do not have to power the bus.
An I2C data communication sequence
• The Universal Serial Bus can connect to 128 devices at transfer
rates 1.5-12 Mbs on USBv1.0 and up to 450 Mbs on version 2.
Hot swapping (aka, plug and play) is allowed.
• Other high speed protocols include Firewire which can be
• Ethernet is a serial data communication link when it connects
to a LAN for control from a HMI software package such as NI
LabView.
• Serial communication using double-ended connections using
twisted wire for Direct Memory Access using Low Voltage
Differential Signaling (LVDS) is now commonplace. The
devices don’t even wait for the voltage rise to complete, they
just look for transitional voltage swings, low to high…
Parallel communications
• GPIB—General Purpose Interface Bus is a high speed
parallel bus is still used and has a large installed base of
equipment. This bus is the IEEE 488 standard that used
8-bit parallel, a 3-wire handshake bus (Data Valid, Not
Ready for Data, and Data Accepted), and 5 lines for bus
management (Service Request, Interface Clear,
Attention, End Of Transmission, Remote Enable).
• While the bus is more expensive than serial, multiple
instruments share the cost. Most scientific devices rely
on software drivers controlled by software programs.
Modern devices—Texas Instrument