WHITE PAPER

The Complete Guide to

Building a Measurement
System
WHITE PAPER

It’s time to get the actionable insight you need from your data. Building a measurement
system that will take you into the future involves much more than choosing a DAQ
device, you have to consider each part of your system. That includes physical components
such as sensors, DAQ devices, and computers, along with software and software
components including drivers and reporting tools.
This guide walks walk you through the top seven things you should consider to
make the most of your DAQ investment, ensuring that your system is ready for the
measurements of today and tomorrow:

03 How to Choose the Right Sensors

12 How to Choose the Right DAQ Device: Specifications

16 How to Choose the Right DAQ Device: Bus and Form Factor

22 How to Choose the Right Computer

25 How to Choose the Right Driver Software

28 How to Choose the Right DAQ Software

33 How to Choose the Right Data Management Software

The Complete Guide to Building a Measurement System

How to Choose
the Right Sensors
Overview
Before you design a measurement system, you need the right sensors (or transducers).
Today’s market offers a multitude of sensors that measure different phenomena—
you’ll even find multiple sensors that measure the same phenomena, including
thermocouples and resistance temperature detectors (RTDs) for temperature
measurements, for example.
This chapter categorizes and compares the most common sensor types for measuring
seven of these phenomena to help you choose the best options for your application:

04 Temperature

04 Strain

06 Sound

07 Vibration

08 Position and Displacement

09 Pressure

10 Force

03
The Complete Guide to Building a Measurement System

Temperature
The most common temperature-measurement sensors are thermocouples, thermistors, and RTDs.

Temperature
Required Signal Conditioning Accuracy Sensitivity Comparison
Sensor

Self-powered
Amplification
Inexpensive
Thermocouple Filtering Good Good
Rugged
Cold-junction compensation
Offers a large temperature range
Amplification
Very accurate
RTD Filtering Best Better
Very stable
Voltage excitation
Amplification
High resistance
Thermistor Filtering Better Best
Low thermal mass
Voltage excitation

TABLE 1
Common Temperature Sensors

Thermocouples
Strain
Thermocouples, the most popular temperature sensors,
are effective in applications that require a large Strain typically is measured by a resistive strain gage.
temperature range. They are inexpensive ($1 to $50 These flat resistors usually are attached to a surface
USD) and have a response time of fractions of a second. that is expected to flex or bend—for example, an
Due to material properties, however, it’s hard for them airplane wing. Strain gages can measure very small
to achieve a temperature accuracy of less than 1 °C. surface twists, bends, and pulls. Wiring more than one
resistive strain gage together creates a bridge.
RTDs
Using more strain gages can help you achieve a more
RTDs are nearly as popular as thermocouples and sensitive measurement—for example, you can use up
can maintain a stable temperature reading for years. to four active strain gages to build a Wheatstone bridge
In contrast to thermocouples, RTDs have a smaller circuit in a full-bridge configuration. You also can choose
temperature range (-200 to 500 °C), require current to use half-bridge (two active strain gages) and quarter-
excitation, and have a slower response time (2.5 to bridge (one active strain gage) configurations. The
10 s). RTDs primarily are used for accurate temperature more active strain gages you use, the more accurate
measurements (±1.9 percent) in applications that your readings are. Table 2 highlights the different
are not time critical. RTDs can cost between $25 and bridge benefits and drawbacks.
$1,000 USD.
Strain gages require current or voltage excitation and
are susceptible to temperature drift, bending strain,
Thermistors
and axial strain, which can give false readings if you
Thermistors have a smaller temperature range (-90 don’t use additional resistive strain gages. See Table 2
to 130 °C) than the previous sensors. They have the for common strain gage mounting options.
best accuracy (±.05 °C), but are more fragile than
thermocouples or RTDs. Thermistors involve excitation Axial bridges measure material stretching or
like the RTD; however, the thermistor requires voltage pulling apart.
rather than current excitation. Typically, a thermistor Bending bridges measure a stretch on one side of
costs between $2 and $10 USD. a material and the contraction on its opposing side.
Torsional and shear bridges measure material twist.

04
The Complete Guide to Building a Measurement System

Strain is measured with a dimensionless unit (e or ε), which is equivalent to a small change in length divided by the full
length of the object under measure.

Sensitivity MV/V
Strain Gage Setup Bridge Type Details
@100 Ue

Good: Simplest to implement, but must use a

¼ 0.5 dummy gage if compensating for temperature.
Responds equally to axial strain.

Better: Temperature-compensated, but

½ 0.65
sensitive to bending strain.

Axial
Better: Rejects bending strain, but not
½ 1.0 temperature. Must use dummy gages
if compensating for temperature.

Best: More sensitive and compensates for

Full 1.3
both temperature and bending strain.

Good: Simplest to implement, but must use a

¼ 0.5 dummy gage if compensating for temperature.
Responds equally to axial strain.

Better: Rejects axial strain and is

Bending ½ 1.0
temperature-compensated.

Best: Rejects axial strain and is temperature-

Full 2.0
compensated. Most sensitive to bending strain.

Good: Gages must be mounted 45 degrees

½ 1.0
from centerline.
Torsional
and Shear
Best: Most sensitive full-bridge version
Full 2.0 of previous setup. Rejects both axial and
bending strains.

TABLE 2
Common Strain Gage

05
The Complete Guide to Building a Measurement System

Sound
Microphones measure sound. There are several types of microphones to consider when you choose a sensor.

Impedance
Microphones Price Environment Sensitivity Comparison
Level

Most used
Prepolarized Condenser Medium Tough Medium Best condenser designs
Best in humid environments
Most used
condenser designs
Externally Polarized Condenser High Tough Better Good
Best in high-temperature
environments
Low quality
Carbon Microphone Low Average High Good Used in early basic
telephone handset design
Better with
Electret Low Average Low Better
high frequencies
Suitable for shock
Piezoelectric Medium Tough High Good and blast pressure
measurement applications
Resistant to moisture
Dynamic/Magnetic High Tough Medium Better Not good in highly
magnetic environment

TABLE 3
Common Sound Sensors

Condenser Microphones Dynamic/Magnetic Microphones

Condenser microphones, the most common In addition to the piezoelectric microphone, dynamic or
microphone, are either prepolarized (meaning that magnetic microphones function in tough environments.
a power source is included within the microphone) They rely on movement to magnetically induce an electric
or externally polarized. Externally polarized condenser charge in a way that makes them resistant to water—but
microphones require an additional power source, obviously, these microphones are not very useful in highly
adding cost. Prepolarized microphones are preferred magnetic environments.
in humid environments, where a power supply’s
components could be damaged; externally polarized Electret Microphones
condenser microphones are preferred in high-
Electret microphones are small and effective at
temperature environments.
detecting high-frequency sound. They are used in
millions of computers and electronic devices around
Piezoelectric Microphones the globe. While relatively cheap, their only drawback
Robust piezoelectric microphones are used in shock is the lack of bass they provide. In addition, you can
and blast pressure measurement applications. These use carbon microphones (which are less common)
durable microphones can measure high-amplitude in applications in which sound quality is not an issue.
(decibel) pressure ranges. Their disadvantage is that
they pick up high noise levels.

06
The Complete Guide to Building a Measurement System

Vibration
Ceramic Piezoelectric Sensor or Accelerometer
Vibration or acceleration is most commonly measured using a ceramic piezoelectric sensor or accelerometer.

Natural Damping
Vibration Sensors Number of Scale Factor Comparison
Frequency Coefficient

Ceramic Piezoelectric Requires Used in vibration and

>5 kHz Up to 3 Small
(accelerometer) high output shock measurements
Limited to steady-state
Linear Variable Differential acceleration or low-
<80 Hz Up to 3 Medium Varies
Transformer (LVDT) frequency vibration
measurement
Limited to steady-state
acceleration or low-
frequency vibration
Proximity Probe <30 Hz Up to 3 Medium Varies measurement
Spring mass attached to
the potentiometer wiper
Output exists only when
mass is in motion
Variable Reluctance <100 Hz Up to 3 Medium Varies
Used in shock studies
and oil exploration

TABLE 4
Common Vibration Sensors

Three major factors differentiate vibration sensors: The Proximity Probes and LVDTs
natural frequency, the damping coefficient, and a scale
Proximity probes and LVDTs are similar: Both are limited
factor. Scale factor relates the output to an acceleration
to steady-state acceleration or low-frequency vibration
input and is linked to sensitivity. Together, the natural
measurement; however, the LVDT vibration sensor has
frequency and damping coefficient determine
a slightly higher natural frequency, meaning that it can
vibration-sensor accuracy. In a system consisting of a
handle/detect more vibration. The proximity probe is
spring and attached mass, if you were to pull the mass
simply a spring mass attached to a potentiometer wiper.
back away from equilibrium and release the mass, the
mass would vibrate forward (past the equilibrium) and
Variable Reluctance Vibration Sensor
backward until it came to rest. The friction that brings
the mass to rest is defined by the damping coefficient, A variable reluctance vibration sensor uses permanent
and the rate at which the mass vibrates forward and magnets and movement through coils to measure
backward is its natural frequency. motion and vibration. This special vibration sensor
registers output only when the mass it is measuring is in
Ceramic piezoelectric vibration sensors are the most
motion, which makes it particularly useful in earthquake
commonly used sensors because they are the most
shock studies and oil exploration to pick up vibrations
versatile. You can use these vibration sensors for shock
reflected from underground rock strata.
measurements (explosions and failure tests), high-
frequency measurements, and slower, low-frequency
vibration measurements. They offer a higher than
average natural frequency. However, this sensor has
outputs typically in the millivolt range and requires a
high-input impedance, low-noise detector to interpret
voltages from its piezoelectric crystal.

07
The Complete Guide to Building a Measurement System

Position and Displacement

Driving factors when selecting a position sensor are excitation, filtering, environment, and whether a line of sight or
a direct, physical connection is required to measure distance. There is not one universally preferred sensor type, as
with pressure or force. Because sensors have measured position for a long time, both preferences and applications
affect this decision.

Position Sensor Price Environment Accuracy Sensitivity Comparison

Only certain that target is nearby

Hall Effect Sensor Low Standard On or off On or off
when depressing sensor
Optical Encoders: Accuracy determined by number
Varies Standard Varies High
Linear and Rotary of counts per revolution
Required to be physically attached
Potentiometers Low Standard High High
to moving target

Linear and Rotary Known for tolerance Handles a high degree of power
Variable Differential of dirty industrial Requires signal conditioning
High High High
Transformers (LVDT) environments RVDTs typically operate over any
or (RVDT) and precision angular range of ±30 to 70 °C
Not good where high resolution
is required
Noncontacting Not good for use when a large
Tolerance of dirty gap exists between sensor and
Eddy-Current environments target (optical and laser sensors
Medium Medium Varies
Proximity Probe Not sensitive to are better)
material between Good when mounted on a
sensor and target reasonably stationary mechanical
structure to measure nearby
moving machinery
Line of sight to target required
for measurement
Reflective Light Good for use when large gap exists
Varies Standard Varies High
Proximity Sensor between sensor and target
Accuracy determined by quality
of sensor

TABLE 5
Common Position Sensors

Hall Effect Sensors Potentiometers

With Hall effect sensors, the presence of an object Potentiometers use a sliding contact to create an
is determined when that object depresses a button. adjustable voltage divider that measures position. While
It is either “on,” and the object is touching the button, potentiometers provide a slight drag to the system to
or “off,” and the target could be anywhere. Hall effect which they are physically connected, potentiometers
sensors have been used in keyboards and even in robot are cheap compared to other position sensors and can
boxing battle competitions to determine when a blow offer great accuracy.
was delivered. These sensors provides no scale as to
how far away an object is from the sensor when the
button is “off,” but they are effective for applications
that do not require highly detailed position information.

08
The Complete Guide to Building a Measurement System

Optical Encoders which can add to cost. Also, these sensors must be
accurately aligned inside heavy, expensive packaging
Another common position sensor is the optical encoder,
and contain wound coils that are expensive to
which can be either linear or rotary. These devices
manufacture. In addition to their cost, they are known
determine speed, direction, and position with fast, high
for their high precision.
accuracy. As the name suggests, optical encoders use
light to determine position. A series of striped bars
Eddy-Current Sensors
divide the distance to be measured by counts. The more
counts, the higher the accuracy. Some rotary optical Eddy-current sensors are moderately priced and
encoders with up to 30,000 counts offer tremendous use magnetic fields to determine position. They are
accuracy. Also, because of their fast response time, used less in applications that require highly detailed
they are ideal for many motion-control applications. positioning information or where large gaps exist
between the sensor and the target. These sensors are
Sensors with physical components that attach to a
better used on assembly lines when mounted on a
system, such as potentiometers, add a small amount
reasonably stationary mechanical structure to measure
of resistance to the movement of the system’s parts.
nearby moving machinery or products. For more
Encoders hardly produce any friction when they move
precise positioning information, use a light proximity
and are very lightweight. However, several important
sensor instead.
items add to their cost: They require seals to operate
within a harsh or dusty environment, and in high-
Reflective Light Proximity Sensors
accuracy applications, they need their own bearings to
avoid misalignment when incorporated into products. Reflective light proximity sensors use a beam’s travel
time to and from a reflective target to determine
LVDTs distance. They have a quick response time and are
excellent for applications in which large gaps exist
LVDTs and their rotary counterparts (RVDTs) use magnetic
between the sensor and target. These sensors require
induction to determine position. They are both effective
line of sight, and accuracy and quality directly affect
for industrial and aerospace applications because
their price.
of their robustness. Both require signal conditioning,

Pressure
High or low pressure is relative—like with heat. It can be “hot” in a room, but the temperature in that room is nothing
compared to the temperature on the surface of the sun. With pressure, the comparison makes the measurement.

Pressure Relative
Tire Example Comparison
Measurement Types

Absolute pressure = standard atmospheric pressure

Absolute Relative to 0 Pa, the pressure in a vacuum
+ gauge pressure
Gauge Reading from tire pressure gauge Relative to local atmospheric pressure
Typically negative value when relative to local
atmospheric pressure Relative to either absolute vacuum (0 Pa)
Vacuum
or local atmospheric pressure
Flat tire = 0 kPa on vacuum gauge
Differential pressure = pressure difference between two
Differential Relative to another pressurized container
different tires
Sealed pressure = gauge pressure + difference between
Sealed Relative to sea level pressure
local atmospheric pressure and sea level pressure

TABLE 6
Relative Pressure Measurements

09
The Complete Guide to Building a Measurement System

There are five common pressure measurement types: Sealed pressure measurements are differential
Absolute, gauge, vacuum, differential, and sealed. pressure measurements taken with a known
Consider the following example of measuring the comparison pressure. Typically this pressure is
pressure within a tire, and note how each major type sea level, but it could be any pressure, depending
is relative to a different reference pressure. on the application.

An absolute pressure measurement includes Each of these measurement types could alter your
the standard pressure from the weight of the pressure values, so you need to know which type of
atmosphere (101.325 kPa) and the additional measurement your sensors are acquiring.
pressure within the tire. The typical tire pressure is Bridge-based (strain gages), or piezoresistive, sensors
34 PSI, or about 234 kPa. The absolute pressure is are the most commonly used pressure sensors because
234 kPa plus 101.325 kPa, or 331.325 kPa. of their simple construction and durability. These
A gauge pressure measurement is relative to the characteristics lower cost and make them ideal for
local atmospheric pressure and is equal to 234 kPa higher channel systems.
or 34 PSI.
These common pressure sensors can be either
Vacuum pressure is relative to either an absolute
conditioned or nonconditioned. Typically, conditioned
vacuum or local atmospheric pressure. A flat
sensors are more expensive because they contain
tire could have the same pressure as the local
components for filtering and signal amplification,
atmosphere or 0 kPa (relative to atmospheric
as well as excitation leads and regular measurement
pressure). This same vacuum pressure measurement
circuitry. If you are working with nonconditioned
could equal 234 kPa (relative to an absolute vacuum).
pressure bridge-based sensors, your hardware needs
Differential pressure is just the difference between signal conditioning.
any two pressure levels. In the tire example, this
means the difference in pressure between two Check the sensor’s documentation so that you
tires. It also could mean the difference between know whether you need additional components
atmospheric pressure and the pressure inside a for amplification or filtering.
single tire.

Force
Load Cell Sensors Price Weight Range Accuracy Sensitivity Comparison

Used with tanks and platform scales

Beam Style Low 10–5k lb High Medium
Strain gages are exposed and require protection
Used with tanks, platform scales
S Beam Low 10–5k lb High Medium
Better sealing and protection than bending beam
Used for truck, tank, and hopper scales
Canister Medium Up to 500k lb Medium High Handles load movements
No horizontal load protection
All stainless steel
Pancake/Low Profile Low 5–500k lb Medium Medium Used with tanks, bins, and scales
No load movement allowed

Either 0–50k lb Loads must be centered

Button and Washer Low or 0–200 lb Low Medium
typically No load movement allowed

TABLE 7
Common Load Cell Sensors

10
The Complete Guide to Building a Measurement System

At one time, mechanical lever scales primarily measured The canister load cell can handle larger loads than both
force. Today, strain-gage-based load cells are the S- and beam-style load cells. It also can handle load
most common because they do not require as much movement easily and is highly sensitive; however, the
calibration and maintenance as scales do. sensor requires horizontal load protection.

Load cells can be either conditioned or nonconditioned. Pancake or low-profile load cells are designed in such
Typically, conditioned sensors are more expensive a way that they require absolutely no movement to
because they contain components for filtering, achieve an accurate reading. If your application has
signal amplification, and excitation leads, as well as time constraints or requires quick measurements,
regular measurement circuitry. If you are working consider using a canister load cell instead.
with nonconditioned bridge-based sensors, your
Button and washer load cells typically measure the weight
hardware needs signal conditioning. Check the sensor’s
of smaller objects (up to 200 lb). Like pancake or low-
documentation so that you know whether you need
profile load cells, the object being weighed must not be
additional components for amplification or filtering.
moving to obtain an accurate measurement. The load also
Beam-style load cells are useful when a linear force is must be centered on what is usually a small scale. The
expected, and they are typically used to weigh both benefit to these load cells is that they are inexpensive.
small and large items (10 lb up to 5k lb). They have an
Read more about common sensor types
average sensitivity but are highly accurate. These load
and terminology
cells offer simple construction and low cost.
Download the Engineer’s Guide to Accurate
The S beam load cell is similar to the beam style except
Sensor Measurements
for its design. Because of this design difference (the
load cell’s characteristic S shape), the sensor is effective
for high side-load rejection and measuring the weight
of a load that is not centered. This low-cost load cell’s
design is also simple.

11
The Complete Guide to Building a Measurement System

How to Choose the Right

DAQ Device: Specifications
Overview
With so many DAQ devices to choose from, it can be difficult to select the right one
for your application. By understanding the signal you are trying to capture, you can
build a list of requirements that ensures that the DAQ system you select has the
accuracy and precision you need.
This chapter outlines five questions that you should ask before selecting your
hardware, to ensure the DAQ system you consider can meet your sampling and
accuracy needs:

13 What types of signals do I need to measure or generate?

13 Do I need signal conditioning?

14 How fast do I need to acquire or generate signal samples?

15 What is the smallest change in the signal that I need to detect?

15 How much measurement error does my application allow?

12
The Complete Guide to Building a Measurement System

What types of signals do I need to measure or generate?

While some DAQ devices can perform a single function, such as measuring and generating analog voltage signals,
others are purely digital, while others perform multiple functions. A data acquisition application typically relies
on a multitude of signals, so it is important to understand which functionality you need for your application.

DAQ Device Functions

Analog inputs measure analog signals.
Analog outputs generate analog signals.
Digital inputs/outputs measure and generate digital signals.
Counter/timers count digital events or generate digital pulses/signals.

A DAQ device that performs multiple functions typically is called a multifunction I/O device. Both DAQ devices that
perform a single function and multifunction I/O devices have a fixed number of channels. Your application dictates
which device is the right fit; however, it is good practice to consider if you might need to scale the system in the future.
If so, it’s probably cost-efficient to select a device that offers more channels than you currently require. This is especially
true if you are using the DAQ device for a limited amount of time on one application and plan to move on to another
application, expecting to use the same device.

Another option is a modular platform that you can customize to your exact requirements. A modular system consists
of a chassis to control timing and synchronization along with a variety of I/O modules. An advantage of a modular
system is that you can select different modules that have unique purposes for more configurations. With this option,
you can find modules that perform one function more accurately than a multifunction device. Another advantage is
that you can select the number of slots for your chassis. While a chassis has a fixed number of slots, you can purchase
a chassis that has more slots than you need now to give you the ability to expand in the future.

Do I need signal conditioning?

A typical general-purpose DAQ device can measure or generate +/-5 V or +/-10 V. Some sensors generate signals too
difficult or dangerous to measure directly with this type of DAQ device. Most sensors require signal conditioning (such
as amplification or filtering) before a DAQ device can effectively and accurately measure the signal.

For example, thermocouples output signals in the mV range that require amplification to optimize the limits of the
analog-to-digital converters (ADCs). Additionally, thermocouple measurements benefit from lowpass filtering to
remove high-frequency noise. Signal conditioning provides a distinct advantage over a DAQ device alone because
it enhances both performance and measurement accuracy.

13
The Complete Guide to Building a Measurement System

Table 8 summarizes common signal conditioning for different types of sensors and measurements.

Bridge
Amplification Attenuation Isolation Filtering Excitation Linearization CJC
Completion

Thermocouple X X X X
Thermistor X X X X
RTD X X X X
Strain Gage X X X X X
Load,
Pressure,Torque X X X X
(mV/V, 4–20 mA)
Accelerometer X X X X
Microphone X X X X
Proximity Probe X X X X
LVDT/RVDT X X X X
High Voltage X X

TABLE 8
Sensors and Measurement Signal Conditioning

If your sensor is listed in Table 8, you should be Choosing a DAQ device with a sample rate at least 10X
conditioning your sensor signals. You can build the frequency of your signal will measure or generate
external signal conditioning or use a DAQ device with a more accurate representation of your signal.
built-in signal conditioning. Many devices include
For example, suppose you want to measure a sine
direct connectivity to common sensors for convenient
wave that has a frequency of 1 kHz. According to the
integration. For a more in-depth guide on sensor
Nyquist theorem, you must sample at 2 kHz at least,
signal conditioning, download the Engineer’s Guide
but you should ideally sample at 10 kHz to measure
to Accurate Sensor Measurements.
or generate a more accurate signal representation.
Figure 1 compares a 1 kHz sine wave measured at
How fast do I need to 2 kHz and 10 kHz.

acquire or generate Once you know the maximum frequency component of

the signal that you want to measure or generate, you can
signal samples? choose a DAQ device with the appropriate sampling rate.

One of the most important specifications of a DAQ

device is the sampling rate—the speed at which the
DAQ device’s ADC takes signal samples. Typical
sampling rates are either hardware- or software-timed
and can reach rates of up to 14 MS/s. The sampling
rate for your application depends on the maximum
frequency component of the signal that you are trying
to measure or generate.

The Nyquist theorem states that you can accurately

reconstruct a signal by sampling 2X the highest
frequency component of interest. However, in practice, FIGURE 1

you should sample at least 10X the maximum frequency 10 kHz versus 2 kHz Representation of a 1 kHz Sine Wave

to represent your signal’s shape.

14
The Complete Guide to Building a Measurement System

What is the smallest How much measurement

change in the signal error does my application
that I need to detect? allow?
The smallest detectable change in the signal determines Accuracy is defined as a measure of the capability of an
your DAQ device’s resolution. Resolution refers to the instrument to faithfully indicate the value of a measured
number of binary levels an ADC can use to represent signal. This term is not related to resolution; however,
a signal. To illustrate, imagine how a sine wave would accuracy can never be better than the resolution of
be represented if it were passed through an ADC with the instrument. How you specify the accuracy of your
different resolutions. Figure 2 compares a 3-bit ADC measurement depends on the type of measurement
and a 16-bit ADC. A 3-bit ADC can represent eight (23) device. An ideal instrument would always measure the
discrete voltage levels. A 16-bit ADC can represent true value with 100 percent certainty, but in the real
65,536 (216) discrete voltage levels. The representation world, instruments report a value with an uncertainty
of the sine wave with a 3-bit resolution looks more like specified by the manufacturer. The uncertainty can
a step function than a sine wave, whereas the 16-bit depend on many factors, such as system noise, gain
ADC provides a clean-looking sine wave. error, offset error, and nonlinearity.

Typical DAQ devices have voltage ranges of ±5 V or A common specification for a manufacturer’s
±10 V. Represented voltage levels are distributed uncertainty is absolute accuracy. This specification
evenly across a selected range to take advantage of the provides the worst-case error of a DAQ device at
full resolution. For example, a DAQ device with a ±10 V a specific range. An example calculation for an NI
range and 12 bits of resolution (212 or 4,096 evenly multifunction device’s absolute accuracy is:
distributed levels) can detect a 5 mV change, whereas
a device with 16 bits of resolution (216 or 65,536 evenly ABSOLUTE ACCURACY = ([READING * GAIN ERROR] +
[VOLTAGE RANGE * OFFSET ERROR] + NOISE UNCERTAINTY)
distributed levels) can detect a 300 µV change.
ABSOLUTE ACCURACY = 2.2 MV
Many application requirements are met with devices
that have 12, 16, or 18 bits of resolution. However, if you Note that an instrument’s accuracy depends not only on
are measuring sensors with small and large voltage the instrument, but on the type of signal it’s measuring.
ranges, you likely can benefit from the dynamic data If the signal being measured is noisy, the measurement’s
range available with 24-bit devices. The voltage range accuracy is adversely affected. There are many DAQ
and resolution required for your application are primary devices with varying degrees of accuracy and price
factors in selecting the right device. points. Some devices provide self-calibration, isolation,
and other circuitry to improve accuracy. While a basic
DAQ device may provide an absolute accuracy of more
than 100 mV, a higher-performance device with such
features may have an absolute accuracy of around 1 mV.

Once you understand your accuracy requirements, you

can choose a DAQ device with an absolute accuracy that
meets your application needs.

Download the Engineer’s Guide to the Digitization

of Analog Signals
FIGURE 2 Compare different DAQ hardware products for
Sine Wave 16-Bit Resolution versus 3-Bit Resolution
your application

15
The Complete Guide to Building a Measurement System

How to Choose the Right DAQ

Device: Bus and Form Factor
Overview
Choosing a DAQ device involves more than selecting the right specifications for quality
measurements. Your environment and equipment should shape your choice of DAQ
device. Which form factor and ruggedness you choose determines where you can
use your system—whether in a controlled lab or the unpredictable field. The bus you
select determines not only your DAQ system’s throughput and latency, but also your
measurement device’s portability.
This chapter examines the most common PC bus options and outlines the technical
considerations to keep in mind when choosing the right bus and form factor for your
measurement application:

17 What are my single-point I/O requirements?

18 How portable should this system be?

18 How far will my measurements be from my computer?

18 How rugged will this system need to be?

18 Do I need to synchronize multiple devices?

19 Common Bus Selection Guide

19 DAQ Buses for PC-Based Systems

QUICK REFERENCES

19 Common Bus Selection Guide

19 DAQ Buses for PC-Based Systems

16
The Complete Guide to Building a Measurement System

How much data will I be What are my single-point

streaming across this bus? I/O requirements?
All PC buses have a limit to the amount of data that can Applications that require single-point reads and writes
be transferred over a certain period of time. Known as often depend on I/O values being updated immediately
the bus bandwidth, it’s often specified in megabytes and consistently. Based on how bus architectures are
per second (MB/s). If dynamic waveform measurements implemented in both hardware and software, single-
are important in your application, be sure to consider point I/O requirements could be the determining
a bus with enough bandwidth. factor for the bus that you choose. For example, in a
proportional integral derivative (PID) control system,
Depending on the bus that you choose, the total
single-point I/O is hugely important, and bus latency
bandwidth can be shared among several devices or
can directly impact the maximum control loop speed.
dedicated to certain devices. The PCI bus, for example,
has a theoretical bandwidth of 132 MB/s that is shared Bus latency is I/O responsiveness. It is the time
among all PCI boards in the computer. Gigabit Ethernet delay between when a driver software function is
offers 125 MB/s shared across devices on a subnet or called and the actual hardware value of the I/O is
network. Buses that offer dedicated bandwidth, such updated. Depending on the bus you choose, this
as PCI Express and PXI Express, provide the maximum delay could range from less than a microsecond to
data throughput per device. a few milliseconds.

When taking waveform measurements, you need to Another important factor in single-point I/O
achieve a certain sampling rate and resolution based applications is determinism, which is a measure of how
on how fast your signal is changing. You can calculate consistently I/O can execute on time. Determinism is
the minimum required bandwidth by taking the number important for control applications because it directly
of bytes per sample (rounded up to the next byte), impacts control loop reliability, and many control
multiplied by the sampling speed, and then multiplied algorithms are designed with the expectation that
by the number of channels. the control loop always executes at a constant rate.
Any deviation from the expected rate makes the
For example, a 16-bit device (2 bytes) sampling at
overall control system less effective and less reliable.
4 MS/s on four channels would be:
Therefore, when implementing closed-loop control
2 BYTES 4 MS applications, you should avoid buses such as wireless,
* * 4 CHANNELS = 32 MB/S
S SEC Ethernet, or USB that are high-latency with poor
determinism. In general, internal buses such as PXI
Your bus bandwidth needs to be able to support Express or the FPGA in an NI CompactRIO chassis
the speed at which data is being acquired, and it is are better for low-latency single-point I/O applications
important to note that the actual system bandwidth will than external buses such as USB or wireless.
be lower than the theoretical bus limits. Actual observed
bandwidth depends on the number of devices in a Communication bus software implementation plays
system and any additional bus traffic from overhead. a large role in bus latency and determinism. Buses
If you need to stream a lot of high-resolution data from and software drivers that support an RTOS provide
several channels, bandwidth may be the most important the best determinism, and therefore give you the
consideration when choosing your DAQ bus. highest performance.

17
The Complete Guide to Building a Measurement System

How portable should this How rugged will this

system be? system need to be?
Portable computing brings new ways to innovate with More tests than ever are being taken outside of
PC-based data acquisition. Portability could easily be pristine lab environments—such as in the field, where
the primary reason to choose one bus over another. humidity, shock, and vibration are concerns, or in test
In-vehicle DAQ applications, for example, benefit from cells, where extreme temperatures and spray-downs
compact, easily transported hardware. are commonplace.

External buses such as USB and Ethernet are particularly Consider the environments into which you may need
good for portable DAQ systems because of their quick to take your DAQ system, and ensure that your DAQ
hardware installation and laptop compatibility. Bus- system can handle it. Look for operating temperature
powered USB devices which are powered via the USB and shock and vibration ratings: For extreme
port offer additional convenience because they do not environments, consider the ingress protection (IP)
require a separate power supply. rating, which tells you a device’s dust- and water-
resistance level. For example, NI FieldDAQ™ devices

How far will my

have an IP rating of IP65 and IP67, which means that
they are dust-tight, water-jet resistant, and water-

measurements be submersion resistant.

from my computer? Do I need to synchronize

The distance between measurements you need and
where the computer is located can drastically vary from
multiple devices?
application to application. To achieve the best signal Many measurement systems have complex
integrity and measurement accuracy, you should place synchronization needs, whether that’s synchronizing
your DAQ hardware as close to the signal source as hundreds of input channels or multiple types of
possible. This can be a challenge for large distributed instruments. A stimulus-response system, for example,
measurements like those used for structural health might require the output channels to share the same
monitoring or environmental monitoring. sample clocks and start triggers as the input channels
Running long cables across a bridge or factory floor is to correlate the I/O and better analyze the results.
costly and can result in noisy signals. One solution to DAQ devices on different buses provide different ways
this problem is to use a portable computing platform to accomplish this. Synchronization techniques are
to move the entire system closer to the signal source. often categorized into signal-based or time-based.
With wireless technology, the physical connection
between the computer and the measurement Signal-based techniques offer the tightest
hardware is removed altogether, and you can take synchronization between devices, but can introduce
distributed measurements and send the data back cabling complexity. Timing signals—including a clock
to a central location. signal and a trigger pulse—are shared directly, with
devices connected by a physical cable. These cables
can be external (for synchronizing two USB DAQ devices
together, for example), or internal (for example, the
synchronization cabling built into PXI chassis for
synchronizing modules together).

18
The Complete Guide to Building a Measurement System

The PXI platform, which includes PXI and PXI Express, An increasingly popular time-based synchronization
offers the tightest synchronization between multiple technology is time-sensitive networking (TSN).
devices. This open standard was designed specifically TSN is an open-source update to the IEEE Ethernet
for high-performance synchronization and triggering, standard designed to address measurement and
offering several ways to synchronize I/O modules control system needs. Some DAQ devices support
within the same chassis as well as synchronizing a subset of the TSN standard denoted as 802.1AS,
multiple chassis. which offers tight synchronization between devices
using a standard Ethernet cable. Leveraging 802.1AS,
Time-based synchronization techniques simplify
these devices support synchronization within <1 μs
synchronization between devices, providing accurate
over a distance of 100 meters. Devices that support
synchronization without additional hardware or cabling.
TSN, or a subset of the standard, include certain
In a time-based system, multiple devices on a network
NI CompactDAQ chassis, NI FieldDAQ, and certain
such as Ethernet individually set their clocks to a common
NI CompactRIO devices.
time source.

Common Bus Selection Guide

Based on the five questions previously outlined, Table 9 shows a selection guide for the most common DAQ buses.

Distributed
Bus Waveform Streaming Single-Point I/O Portability
Measurements

PCI 132 MB/s (shared) Best Good Good

PCI Express 250 MB/s (per lane) Best Good Good
PXI 132 MB/s (shared) Best Better Better
PXI Express 250 MB/s (per lane) Best Better Better
USB 60 MB/s Better Best Good

Ethernet 2.0 125 MB/s (shared) Good Best Best

TABLE 9
Bus-Selection Guide Based on Application Requirements with NI Products

DAQ Buses for PC-Based Systems

While you can choose from many buses and form
factors, this section focuses on the most common
buses for a PC-based system:

PCI and PCI Express

USB
PXI and PXI Express
Ethernet

In Figure 3, all of these buses are represented in this

collection of NI DAQ products, from internal plug-in
FIGURE 3
options to hot-swappable external buses.
These NI DAQ products use several buses and form factors,
including PXIe, PCIe, USB, and Ethernet.

19
The Complete Guide to Building a Measurement System

PCI and PCI Express

The PCI bus is an older, yet still commonly used,
internal computer bus. With a shared bandwidth
of 132 MB/s, PCI offers data streaming and data
transfer for single-point control applications.

PCI Express is an evolution of PCI, offering a new

level of PC performance. The single biggest benefit
of PCI Express architecture is the dedicated bus
bandwidth provided by independent data-transfer
FIGURE 4
lines. PCI Express uses independent data lanes that PCI Express X Series Multifunction DAQ
are each capable of data transfer of up to 250 MB/s,
meaning that a single device can reach several GB/s
of bandwidth.

There are several DAQ devices to choose from in

the PCI/PCIe form factor, with multiple choices of
resolutions, sample rates, and signal conditioning.

Compare NI PCI and PCI Express DAQ devices

USB
USB delivers an inexpensive and easy-to-use connection
between DAQ devices and PCs. USB 2.0 has a maximum
theoretical bandwidth of 60 MB/s, which is shared among
all devices connected to a single USB controller.

USB devices are inherently latent and nondeterministic.

This means that single-point data transfers may not
happen exactly when expected, and therefore USB is
not recommended for closed-loop control applications
such as PID.

On the other hand, the USB bus has several FIGURE 5

characteristics that make it easier to use than USB DAQ adds data acquisition to any computer with a USB port.
internal PC buses. USB devices are hot-swappable
and plug-and-play, meaning that the PC will detect
a newly connected device and, with the right driver,
automatically install it.

Compare NI USB DAQ devices

20
The Complete Guide to Building a Measurement System

PXI
PXI bridges the gap between desktop PC systems,
high-end VXI, and GPIB systems. The PXI Systems
Alliance, with more than 200 members, maintains this
open standard, and in 2006, passed the PXI Express
specification to deliver PCI Express data-transfer
technology to the PXI platform.

Based on CompactPCI, PXI incorporates instrumentation

extensions and more rigid system-level specifications
FIGURE 6
to ensure an open yet high-performance measurement The PXI platform is composed of chassis, controllers, and
and automation specification. Benefits of PXI-based DAQ I/O modules.
systems include rugged packaging that can withstand
often-harsh industrial conditions. Also, PXI systems
offer a modular architecture, which means that you can
fit several devices in the same space as a single stand-
alone instrument, and you can expand your system far
beyond the capacity of a desktop computer with a PCI
bus. One of the most important benefits PXI offers is
its integrated timing and triggering features. Without
any external connections, you can synchronize multiple
devices using the internal buses resident on the PXI
chassis backplane.
FIGURE 7
Compare NI PXI DAQ options
Supporting 100 m per segment and the ability to use existing
network infrastructure, Ethernet data acquisition can extend
Ethernet the reach of your measurement system.

Ethernet is the backbone of almost every corporate

network in the world and is, therefore, widely available.
As a DAQ bus, Ethernet is ideal for taking portable or
distributed measurements at distances beyond the
5 m length of a USB cable. A single Ethernet cable
can extend 100 m before needing a hub, switch, or
repeater. This distance, combined with a large install
base of networks in labs, offices, and manufacturing
facilities, makes Ethernet an ideal choice for distributing
measurements to remote locations.

Select Ethernet DAQ devices use the TSN (802.1AS)

IEEE standard for easy time-based synchronization
between devices using a regular Ethernet cable.
Some CompactDAQ and CompactRIO devices and
all FieldDAQ devices use this standard for easy
multidevice synchronization.

Compare NI Ethernet DAQ options

Learn more about DAQ buses and form factors

Discover more about NI DAQ devices

21
The Complete Guide to Building a Measurement System

How to Choose the

Right Computer
Overview
Once you have chosen your DAQ device, it’s important to select the right computer for
your application. The computer can be the most crucial part of your data acquisition
system. It provides flexibility over traditional boxed systems by housing the DAQ
device, running the software to control the device, analyzing the measurements, and
saving the results.
This chapter explores the questions you should ask to choose the right computer for
your application:

23 How much processing power do I need?

23 Do I need my computer to be portable?

23 How rugged does the computer need to be?

23 Do I need my computer to be modular?

24 Do I need an RTOS?

24 Computer Selection Guide

QUICK REFERENCE

24 Computer Selection Guide

22
The Complete Guide to Building a Measurement System

How much processing How rugged does the

power do I need? computer need to be?
Nearly every computer has three main components that Ruggedness can be crucial if you are monitoring your
affect data-management capabilities: The processor, application in an extreme environment. Operating
the RAM, and the hard drive. conditions—for example, operating and storage
temperature, relative humidity, and maximum operating
The processor is the part of the computer that interprets
and storage altitude—determine computer ruggedness.
and executes instructions—think of it as the brain.
Standard off-the-shelf PCs are not designed to withstand
Processors in most new computers have multiple
industrial-environment conditions.
cores, meaning that the computer can use multiple
independent processors, or “cores,” to read and Typical specifications are 50 °F to 95 °F (operating
execute program instructions. temperature), -13 °F to 113 °F (storage temperature),
10,000 feet (operating altitude), and 15,000 feet (storage
Computer processing power also consists of the RAM,
altitude); any computers featuring specifications greater
the hard drive size, and the processor speed. With
than these are considered rugged.
more RAM, you improve speed and can run more
applications simultaneously. More hard drive space When designing your system, consider your environment.
gives you the ability to store more data. If heavy vibration or temperature swings could result in
critical data loss, it may be worth investing in a rugged
Finally, faster processors translate to faster application
or industrial PC.
operation. In general, faster is better, but processor
speeds across brands may not be equivalent. If you
need to analyze or save the data you acquired from Do I need my computer
your application, processing power is a key feature
to consider for your computer. to be modular?
Computer modularity is a factor if you are considering
Do I need my computer future applications or working on multiple applications.

to be portable? Modularity describes the degree to which you can

separate and recombine a system’s components. You
can modify and adapt the system to meet your current
Portability is imperative if you move frequently between
needs and plan future expansion, as well as upgrade
applications or locations. For example, a portable
individual components, without having to buy a whole
computer is essential for taking measurements in
new system.
the field and returning to the lab to analyze the data.
Portability also is important if you need to monitor For example, with a modular tower PC featuring PCIe
applications in different locations. slots, you can install a new hard drive if you need more
space or install a DAQ device with a faster analog-to-
Two considerations when assessing portability are
digital converter if you need faster sampling. Laptops
product size and weight (acknowledging that lighter PCs
and tablets provide portability, but they are more
can reduce performance). If you need a powerhouse
integrated, which makes them harder to upgrade.
of a computer, but need to take remote data, consider
Modularity can be an important feature if you need
building a distributed DAQ system, extending your
to adapt your current application to future demands.
measurement system into the field over Ethernet while
keeping your computer safe in a control room or lab.

23
The Complete Guide to Building a Measurement System

Do I need an RTOS?
The OS is an important feature to consider when choosing a DAQ computer. By far, the most common general-
purpose OS is Windows, but DAQ and control applications can require a more specialized OS.

An RTOS can operate deterministically, so that applications execute according to precise timing requirements.
An RTOS is deterministic because it does not determine which process happens when; rather, you define process
order and timing, giving you more control over your application and the ability to execute at faster rates than with
a nondeterministic OS. This is especially relevant in control applications in which you need to prioritize critical tasks
above others.

Learn more about RTOSs

Computer Selection Guide

Based on the previous six questions, Table 10 shows a selection guide for the most common types of computers.

Industrial
PXI System Desktop Controller, Laptop Tablet
NI CompactRIO

Processing Power Best Best Better Better Good

OS Compatibility Best Best Better Best Good
Modularity Best Better Better Better Good
Ruggedness Better Better Best Good Good
Portability Better Good Best Best Best

TABLE 10
Computer Selection Features

24
The Complete Guide to Building a Measurement System

How to Choose the Right

Driver Software
Overview
While you might be tempted to overlook driver software when selecting a DAQ
device, the driver behind your DAQ device can be one of the most important factors
in development time and device performance.
The driver handles the communication layer between hardware devices and application
software, giving you access to both high-level functions for quick measurements and
low-level control for fine-tuning complex tasks.
This chapter answers the questions you need to keep in mind when evaluating a DAQ
device’s driver software:

26 Is the DAQ driver compatible with my OS?

26 How well does the driver integrate with my application software?

26 What documentation comes with the driver?

27 Does the driver include any setup or diagnostic utilities?

27 Is the driver scalable to other DAQ devices?

25
The Complete Guide to Building a Measurement System

Is the DAQ driver

In a perfect scenario, the provided driver natively
integrates with your application software. In this case,

compatible with my OS? the driver is rewritten for the native language. This
provides better performance and a more seamless
You can choose from OSs including Windows, macOS, experience because functions and documentation are
and Linux, which offer different advantages for different directly built into the application software.
tasks, operations, and deployments. Each of these OSs In some cases, a wrapper may not be provided for
also may have different versions, distributions, or designs your preferred language, or even at all, so you must
for specific processors. For example, Windows offers manually write your own wrapper to interface with your
versions for 32-bit and 64-bit processors, and open- application software. When evaluating a DAQ system,
source Linux OSs include hundreds of varieties. Each check to see what languages the driver can support,
type, release, and version of an OS functions differently and if possible, access the documentation for that
and may or may not be cross-compatible. language. Ideally, you can get full driver functionality
As a result, DAQ drivers generally do not support every from the driver in a wide variety of application software.
single OS type and version. Most DAQ drivers work with
Windows OS releases, as they are the most common.
NI’s DAQ driver, NI-DAQmx, supports most Windows
What documentation
and several Linux variants. comes with the driver?
If you use an alternative OS, remember to confirm
Drivers feature many forms of documentation, including
whether the DAQ device driver supports it before
user manuals, functions references, release notes,
choosing a DAQ device. You generally can find the
known issues, and example code. Having to navigate
OS and version support in the driver readme files.
through poor, incomplete, or muddled documentation
wastes time. When a driver’s programming interface
How well does the is poorly documented, you can spend an unnecessary
and frustrating amount of time running functionality
driver integrate with trial-and-error tests. Although trial and error can

my application software? be a great way to learn functions and syntax, you

need to be able to refer to the manual. Therefore,
having well-organized, thorough documentation is
There are varying degrees of driver integration with
extremely valuable.
application software. At the core of every driver
is a library (often a DLL). This library manages the The best driver software documentation is complete,
communication with DAQ hardware. Normally, the easy to navigate, and simple to follow. Ideally, it offers
library comes with documentation and wrappers for example code specific to your preferred programming
various programming languages. These wrappers languages and provides detailed and useful error
are thin layers of code that translate the library’s messages. By evaluating the driver software’s
functions into a compatible interface for a particular documentation ahead of time, you can save yourself
programming language. potential headaches in the future.

26
The Complete Guide to Building a Measurement System

Does the driver include Is the driver scalable

any setup or diagnostic to other DAQ devices?
utilities? You might not be able to anticipate which changes and
expansions your current DAQ system may need in the
In addition to documentation, setup and diagnostic future. You may need to upgrade the device to higher-
utilities can help you get your application up and performance specifications or incorporate additional
running quickly and diagnose problems. measurements. Some DAQ drivers are designed for a
single device, and others are designed to work with a
With test panels, you can test hardware functionality wide range of devices.
at the most basic level before designing the end
application. You can generate and measure raw Single-device drivers typically are more lightweight than
signals and troubleshoot the DAQ hardware drivers that work with a wide range of devices. While
independently of other software and programming these drivers initially do the job, adding a new device
factors that could insert an extra level of uncertainty. or replacing an existing one could require significant
programming to integrate the corresponding new
Calibration utilities walk you through the steps
driver. The driver’s programming interface might
to self-calibrate your device to ensure it
be structured differently, requiring significant
measures accurately.
code changes.
Sensor scaling wizards help you easily map raw
voltage values to engineering units without having On the other hand, drivers that support a wide range of
to program the math yourself. devices are more easily scaled to additional functionality
and new devices. The programming interface is consistent
Some drivers even include complete configuration
among all devices, so adding a new device is essentially
wizards that encapsulate all of these utilities, walk
a drop-in replacement and requires little to no changes
you through setting up your measurement task,
to your code. These drivers also may support other
and help you take your first measurement in your
features that make synchronizing and combining
application software.
measurements from multiple devices easier.
Overall, setup and diagnostic utilities are very useful Read more about NI-DAQmx Driver Software for
when getting started with your DAQ device or DAQ Applications
diagnosing problems. Not all DAQ drivers include these
utilities, though. Explore the NI-DAQmx Driver Manual for
compatibility information

27
The Complete Guide to Building a Measurement System

How to Choose the Right

DAQ Software
Overview
Software lies at the core of modern DAQ systems. Most of your time with a DAQ
system is spent in the software environment, so it’s imperative that you select a
software tool that fits the needs of your application today and easily scales as your
system matures.
DAQ software ranges from ready-to-run application software to a fully customizable
programming language. Take into account your measurement requirements, available
development time, and programming expertise to help you choose the right tool for
your DAQ system.
Consider these questions before choosing your application software:

29 What is the difference between 30 Is there a community to reach out

application software and to when I get stuck?
development environment?
31 What kind of analysis do I need
29 How much time do I have to build to perform?
my measurement system?
31 What kind of data visualization do I need?
29 How long will it take to learn
the software? 32 Can I integrate my own custom
or legacy IP?
30 Does my software have training
options to help me get started? 32 Software selection table

QUICK REFERENCE

32 Software selection table

28
The Complete Guide to Building a Measurement System

What is the difference

The following sections refer to both application software
and development environments as two valid options for

between application your measurement system.

software and development How much time do I have

environment? to build my measurement
DAQ software tools range from ready-to-run application
software (no programming required) to fully customizable
system?
development environments. You can use either to build When choosing software, it’s important to understand
a robust and flexible measurement system, but there your own time constraints. You need time to learn the
are tradeoffs to both. software, set up your system, and potentially debug
Application software prioritizes ease of use so that your measurement code. When evaluating the time
you can acquire and process data with minimal (or no) you need to build your measurement system, consider
training. You can use the software to set up hardware, the long-term, too—if you plan to scale this test in the
visualize measurement channels, log data, and more, future, the time you invest in training today may yield
using drop-down menus and prebuilt screens. greater returns as your system grows and changes.

Typical application software is workflow-based—

NI FlexLogger™ software, for example, is designed How long will it take
around data-logging applications, including all
configuration, visualization, and events/alarming to learn the software?
you might need to record mixed measurements.
Ready-to-run application software tools are the easiest
You could encounter limitations when you require
and fastest to learn because they have abstracted
functionality outside of that workflow (for example,
user programming details, typically only requiring a
in postprocessing and reporting). In those cases, you
few details for setup. When deciding among ready-
may need to develop plug-ins or use another tool such
to-run software tools for your DAQ system, ensure
as NI DIAdem.
that the tool has the hardware support, processing
Development environments prioritize custom functionality capabilities, and analysis libraries necessary to meet
so that you can meet virtually any challenge in your your requirements. Also confirm that it offers proper
measurement system. Development environments resources—whether user manuals, in-product help
are extremely flexible because you can integrate information, online communities, or support forums—
DAQ drivers into the software and develop a custom to help you quickly learn the tool.
user interface (UI) and code to perform the exact
Application development environments often take
measurements or test routines that you need.
longer to learn, but the majority of that time is spent
A development environment such as NI LabVIEW can learning the language used within the environment
function as a data logger control system, postprocessor, to program your applications. If you can find an
report generator, and more. The trade-off for development application development environment that uses a
environments is that you need to spend time upfront language you are familiar with, you can reduce the time
to learn the programming language and develop the required to become a proficient programmer within
applications yourself. While this may seem like a large a new application development environment. Many
time commitment, modern-day development application development environments can integrate
environments provide a variety of tools to help you with—and even compile—several different languages
get started, reducing overall time spent. within a single framework.

29
The Complete Guide to Building a Measurement System

When evaluating application development environments the complex grammar and syntax rules that you must
that require you to learn a new language, consider those follow to successfully compile and run the code.
that direct your focus to your engineering problem, not
Graphical programming languages, such as the one
low-level programming-language details. For example,
offered in NI LabVIEW, are often easier to learn because
you might find that text-based languages, such as ANSI
the implementation is more intuitive and visually
C/C++, are often more challenging to learn because of all
consistent with the way an engineer thinks.

ANSI C Code LabVIEW Code

FIGURE 8
Sample Code

Does my software have Shipping examples—Good shipping-example sets

have enough code for the most common types of

training options to help DAQ applications. With these examples, you never
need to start from scratch.
me get started? You can save time by simply modifying the shipping
Also consider getting-started resources that come with examples to meet the needs of your system development.
the application software. These resources can help
you get up and running with a new software tool in
less time. Here are a few helpful resources for getting
Is there a community to
started with any software tool: reach out to when I get
Evaluation—A free evaluation gives you the stuck?
opportunity to test things out for yourself and
determine if the tool meets your application needs. The ecosystem that surrounds software is just as
important as the software tool itself. A healthy
Online curriculum—Online tutorials, videos, and
ecosystem provides a wealth of resources that make it
white papers are valuable when learning basic
easy for you to learn a new software tool and guide you
application software concepts.
with feedback as you develop your application. Before
Classroom instruction—An application software
you purchase, browse a community’s forums and
class is the perfect way to get up to speed and begin
determine how active it is and the kind of information
developing your DAQ system. Course price and level
being shared (code, discussions, tips, and tricks). You
of detail depends on the instructional setting. Often,
want a community that is heavy with activity and
you can find options ranging from free seminars
includes shared information that is closely aligned with
to formal classroom instruction, to instructor-led
the problems you are solving.
online courses.
Additionally, an application software’s ecosystem of
users often drives future development. Check to see

30
The Complete Guide to Building a Measurement System

if the organization behind an application software look for a tool offering functions related to your field
is responsive to its community’s needs and whether or application type.
the user base can provide input that guides future
Proper data-analysis tools contain more than 600 built-
software features.
in functions. While basic and complex math operations
are beneficial, you need functions specific to your
What kind of analysis area of interest. If your application deals with control,
look for proportional integral derivative (PID) control
do I need to perform? functions. If you have an optical character recognition
(OCR) application, make sure your tool contains
Raw data is not always the most useful way to those functions. Look for software products with an
communicate. Data transformations—removing signal ecosystem of add-ons to extend product functionality,
noise, compensating for environmental effects such as including third-party industry-expert analysis. This way,
temperature and humidity, and calibrating equipment you reduce time and cost associated with developing
error, for example—help turn raw data into useful these functions yourself.
data. Most engineering applications aim to produce
useful data, and that requires comprehensive signal If you need inline analysis, be sure that your application
processing for any analysis tool used in data acquisition. software includes built-in or expansion capabilities.
There are two main ways that you can analyze or If your needs involve offline analysis, your application
process data: software must be able to save data to a format that
your offline analysis package can consume.
Offline analysis, also known as postprocessing, takes
place after you’ve saved your data to file. Offline
analysis is a good fit for an application such as
What kind of data
FlexLogger software, which emphasizes logging data visualization do I need?
to be postprocessed after you run the test. See the
Data Management chapter for more details on Data visualization—from simply graphing an acquired
postprocessing and offline analysis. signal to correlating measurement data with video,
Inline analysis implies that data is analyzed in sound, or 3D model projection—is common in almost all
the same application in which you acquired it. measurement systems. Selecting the right visualization
If your application involves monitoring a signal technique could mean the difference between being
and changing testing variables based on incoming able to appropriately derive actionable information
data characteristics, inline analysis is the right from raw data and missing important insight.
choice. By measuring and analyzing certain signal As with analysis, visualization can happen inline or offline.
aspects, you can make the application adapt to The final chapter discusses offline visualization (typically
certain circumstances and incorporate execution considered report generation).
parameters—perhaps saving the data to disk in
case of alarm or increasing the sampling rate if the Inline reporting happens in the DAQ software so
incoming values exceed a threshold limit. To perform that you can follow data trends, view critical system
inline analysis, your application software must have information, and create an engaging UI. For example,
built-in signal analysis functions or the ability to you can display acquired data on a monitor so that
easily integrate external IP. a technician can see the signal being measured and
ensure proper connections.
Most data-analysis tool vendors produce a well-
If you run inline analysis with inline visualization, the
documented listing of their tool functions, which is
monitor may display a filtered version of the same
helpful if you know your specific signal-processing
signal. While this architecture gives you “instant
needs. But if you don’t know exactly what you need,
feedback,” since you can visualize acquired data in

31
The Complete Guide to Building a Measurement System

near-real time, it means that your chosen application Or, because application requirements change over
software must contain the required visualization tools. time, you’ve invested time and money creating analysis
routines or custom IP in older or alternative tools.
For visualization, most engineers require, at a minimum,
In these cases, look for a data analysis package that
basic charting and graphing capabilities. Luckily, almost
can incorporate these external analysis routines. Don’t
every data visualization tool on the market can make
reinvent the same functionality in the newer tool when
simple charts and graphs, and dedicated visualization
your existing algorithms are validated to work correctly.
tools offer robust additional capabilities that help you
learn more from your data. Whether you created your analysis routine in another
programming language, used a script in an older
It’s important to consider visualization scalability and
financial-analysis tool, or inherited some configuration
customizability. Out-of-the-box application software
file, confirm with your software vendor that you can
can include a wide variety of charts and graphs and
incorporate your legacy analysis routine in their data-
present multiple plots in a single graph. They might
analysis tool. If you can’t do so easily, you could spend
have indicators and visualization techniques, but may
precious time recreating your functionality in the new
not give you complete control over how you show your
tool. Modern data-analysis tools should be open to
data. Development environments, on the other hand,
using IP created in other environments.
offer extensive, in-depth customizability so that you can
control every facet of how you visualize your data. Be
sure to choose a development environment that makes Software Selection Table
GUI design easy—some text-based programming
languages make visualizing data challenging. Based on the questions and considerations covered
above, Table 11 highlights the differences between

Can I integrate my own application software and two types of development

environments: Graphical (as in LabVIEW) and text-
custom or legacy IP? based (as with Python or C++).

You might have a proprietary analysis algorithm that

simply can’t be purchased as add-on software.

Application Software Development Environment Development Environment

Selection Criteria
(No Programming) (Graphical Programming) (Text-Based Programming)

Example FlexLogger LabVIEW Python, C, C++

Ease of Use Best Better Good
Time to First Measurement/Test Best Better Good
Customizability Good Best Best
DAQ Driver Integration Best Best Good
Inline Analysis Tools Good Best Better
Inline Visualization Tools Best Best Better
External Hardware Integration Good Best Best
External IP Integration Good Best Best

TABLE 11
DAQ Software Selection Guide

Read about NI FlexLogger software, ready-to-use application software for data logging

Find out how NI LabVIEW, the development environment that accelerates engineering, can help you meet your
application and system needs

32
The Complete Guide to Building a Measurement System

How to Choose the Right

Data Management Software
Overview
You acquire data for a reason: To help you make decisions. While technology is
making data retention faster and richer, storing, processing, and sharing it remains
a real challenge.
Most DAQ systems collect data to analyze and ultimately present or share it in an
exchangeable report. You can choose from a wealth of data management tools, but
you should carefully consider the capabilities of your tool of choice to ensure that it
doesn’t become a bottleneck in your system.
This chapter outlines six questions to consider when choosing a reporting tool for
your application:

34 Can my data management software handle my data?

35 Does my data management software offer the analysis that I need?

35 Does my data management software offer the visualization that I need?

36 Can I use templates to simplify repetitive reports?

36 Can I automate reporting to save time?

36 Does my data management software export reports in the right format?

33
The Complete Guide to Building a Measurement System

Can my data management

DAQ applications collecting dozens of data channels
at megasample-per-second (MS/s) rates are also

software handle my data? commonplace. In these applications, data manipulation

and interaction happen on a signal (or channel) as a
We all have used spreadsheet programs to manipulate, whole. If you manipulate columns of individual cells, you
analyze, and share raw data with graphs and charts. But risk losing signal unity.
classic spreadsheet programs are designed for financial Manipulating entire columns at one time is cumbersome.
analysis. When you’re looking for a measurement Columns often contain descriptive information, such
system data-management tool, consider two major as a name or unit, in addition to raw numeric data. So
differences: First, the data-storage file format, and when you select a subset of the column (for example,
second, the volume of data. It’s imperative that a the range A3:A999), you introduce overhead and the
reporting tool not only load data from your chosen potential for inaccuracy or errors.
file format, but handle as much data as you need it to.

File Format
Traditional file types rarely meet all of the requirements
you need in a file format. For example, ASCII files are
exchangeable, but very large and slow to read and
write. Binary files read and write speeds can keep up
with high-speed hardware, but these files are difficult
to share.

The technical data management streaming (TDMS)

file format eliminates these challenges. TDMS files are
based on the TDM data model for saving well-organized
and documented test and measurement data.
FIGURE 9
With TDMS-formatted files, you do not have to redesign Microsoft Excel uses the cell as its fundamental building block. Even
simple data analysis must be applied to a cell and then repeated for
your application as your DAQ requirements increase.
all cells in a column (channel).
You simply extend the data model to meet your needs.
Because it was developed to meet the needs of all
engineers, TDMS is easy to use and offers high-speed Figure 9 shows how you can use Microsoft Excel to
streaming and exchangeability. perform a common engineering task: averaging five
temperature channels stored in columns to create a
Traditional financial analysis tools use the cell as their
resultant average channel. You must first implement
fundamental building block. Cells form rows and
the averaging calculation in the cell building block and
columns to make up a spreadsheet, an architecture
then copy it to or fill in all cells in the resultant column.
that is ideal for budgets and balance sheets. Simple,
single-point DAQ applications—for example, those that
collect one single data point an hour over the course
of a day—are often easily mapped to this architecture
because each individual data point holds more
importance when fewer data points are collected. Each
data point exists as a cell in a spreadsheet and must be
manipulated using this cell-based paradigm.

34
The Complete Guide to Building a Measurement System

Data Volume fitting are all examples of postprocessing tasks.

Postprocessing offers far greater data interactivity,
These days, common application data-streaming speeds
giving you the ability to truly explore both the raw data
often reach or exceed MS/s rates. An application that
and analysis results. Furthermore, it means that you
collects one single channel of data at 1 MS/s collects
don’t need to worry about analysis bottlenecking a live
a total of 1,000,000 data points in a one-second
acquisition, considering the amount of time that intense
acquisition. In a matter of minutes, billions of data
signal-processing algorithms can take when operating
points occupy gigabytes of hard drive space.
on large data sets.
Traditional data management tools attempt to open
When choosing data management software, consider
a large data file by loading every single data point
your post-processing needs. Check the data-management
into memory, which takes valuable time. Their cell-
software manual to confirm that it offers the appropriate
centric flexibility is ideal for business spreadsheets
math, plotting, and data-reduction functions for
where cell-level visibility is key, but it adds unnecessary
your system.
memory overhead for data sets with millions of values.
To avoid potential memory problems, traditional
reporting tools often impose a limit to the maximum Does my data management
number of data values that can be loaded for a given
column. This usually requires readjusting your storage software offer the
strategy by either choosing a new file format (possibly
having to rearchitect your application after the fact) or
visualization that I need?
segmenting data into many small files just so that your
When it comes to reporting, you probably require, at
reporting tools can open them.
a minimum, basic charting and graphing capabilities.
When designing a DAQ system, be sure that your data Luckily, almost every data management tool on the
management tool can handle your file format as well market can make simple charts and graphs. However,
as the volume of data you intend to acquire (leaving be certain that those charts can handle graphing the
flexibility for a change in requirements that might add volume of data you intend to plot, as many impose a
to the volume of data collected in the future). limit to data-point count.

If you think you might need to graph different curves

Does my data management with drastically different y-scales on the same chart,
your reporting tool needs to be able to distinguish
software offer the analysis between these scales. While many tools can, they also

that I need?
have a limiting maximum y-axis count.

In addition, think about whether your reporting needs

Inline analysis, or analysis performed as your data go beyond basic 2D graphing. For example, if you
is being acquired, is not always appropriate. If you need to represent data using polar plots, or if your
don’t need to make decisions as you acquire data, data would be best represented in a 3D graph, your
you may opt to perform offline analysis, also known reporting tool must support that.
as postprocessing. This involves saving acquired data
to disk for unlimited interaction and having your data
management software perform the analysis.

Because you perform this analysis after you acquire

the data, you are not limited by DAQ timing and
memory constraints, freeing up your processor for
more challenging tasks or trend analysis that requires
the entire dataset. Histograms, trending, and curve-

35
The Complete Guide to Building a Measurement System

Can I use templates to Can I automate reporting

simplify repetitive reports? to save time?
Many times, you need the same type of report for a Typically, a DAQ application uses one of two types
series of raw data files. For example, if you run the of reporting: infrequent or repetitive. Infrequent
same tests every week and have to report standardized reporting happens on an occasional basis, usually in
results, you reuse the same report layout across an interactive, customized way. Alternatively, repetitive
multiple data sets. Traditional reporting tools save the reporting is frequent and usually standardized, and
report display along with the raw data in a common often uses templates.
spreadsheet file, which makes it much harder to use a
If you have repetitive reporting needs, your data
particular report display for multiple data sets. Each
management software should be able to automate
data set contains its own report layout and formatting,
reporting. Even most traditional tools support macros
which means that if you need to make a modification to
or scripts that make this easier. Many have recording
the layout or formatting—for example, something as
modes so that you can interactively record scripts that
simple as changing the color of a curve—you must edit
automate lengthy evaluations or calculations.
every file to standardize on that change.

By creating templates, you can more easily create

custom reports to update with new data and results.
If you anticipate having to create the same report
multiple times across several data sets, you need a data
management tool that you can use to produce a report
template and apply it to different raw data files.

Does my data management software export reports in

the right format?
Reporting tools usually present the final output in an easily exchangeable format that you can email, print, or present,
regardless of the original raw data f ile format. With most reporting tools, you can export reports to several formats,
but be sure that your tool supports your most common format (whether that’s PDF, PowerPoint, imagery, or HTML).

Additionally, if you have immense reporting needs—for example, if your reports often span dozens of pages—ensure
that the reporting tool you choose can export your report in your desired format at your desired size. The last thing
you want to do is recreate all of your work at the very end of your system’s design simply because your reporting tool
can’t create reports at your required length.

Find out how NI DIAdem can help you manage and report your measurement data

36
The Complete Guide to Building a Measurement System

Next Steps
Let’s get started designing your measurement system. We’ll help you pick the right sensors to measure the physical
phenomena you’re interested in and select the proper DAQ device to read those signals to analyze, visualize, and
report out on your data.

FIGURE 10
Whatever your measurement challenge, NI has hardware and software to meet your needs.

Explore NI DAQ hardware and software

Consult an NI Sales Representative for your measurement system

Neither Emerson, Emerson Automation Solutions, nor any of their affiliated entities assumes
responsibility for the selection, use, or maintenance of any product. Responsibility for proper
selection, use, and maintenance of any product remains solely with the purchaser and
end user. Linkedin.com/company/niglobal/
NI, National Instruments, the National Instruments corporate logo, ni.com, NI FieldDAQ,
NI CompactDAQ, NI CompactRIO, NI DAQ, and NI LabVIEW are marks owned by one of the
companies in the Test & Measurement business unit of Emerson Electric Co. Emerson and the x.com/niglobal
Emerson logo are trademarks and service marks of Emerson Electric Co.
The registered trademark Linux® is used pursuant to a sublicense from LMI, the exclusive
licensee of Linus Torvalds, owner of the mark on a worldwide basis. All other marks are the Youtube.com/@NIGlobalYoutube
property of their respective owners.
The contents of this publication are presented for informational purposes only, and while
every effort has been made to ensure their accuracy, they are not to be construed as
warranties or guarantees, express or implied, regarding the products or services described Instagram.com/niglobal/
herein or their use or applicability. All sales are governed by our terms and conditions,
which are available upon request. We reserve the right to modify or improve the designs or
specifications of such products at any time without notice. Facebook.com/NationalInstruments
NI
11500 N Mopac Expwy
Austin, TX 78759-3504
© 2024 National Instruments. All rights reserved. 531748 ni.com