OPTOTUTORIAL: SNAP PAC PID

A Supplement to the SNAP PAC Learning Center

Form 1641-120824—August 2012

43044 Business Park Drive • Temecula • CA 92590-3614

Phone: 800-321-OPTO (6786) or 951-695-3000
Fax: 800-832-OPTO (6786) or 951-695-2712
[Link]

Product Support Services

800-TEK-OPTO (835-6786) or 951-695-3080
Fax: 951-695-3017
Email: support@[Link]
Web: [Link]
 OptoTutorial: SNAP PAC PID
Form 1641-120824—August 2012
Copyright © 2003–2012 Opto 22.
All rights reserved.
Printed in the United States of America.
The information in this manual has been checked carefully and is believed to be accurate; however, Opto 22 assumes no
responsibility for possible inaccuracies or omissions. Specifications are subject to change without notice.
Opto 22 warrants all of its products to be free from defects in material or workmanship for 30 months from the
manufacturing date code. This warranty is limited to the original cost of the unit only and does not cover installation, labor,
or any other contingent costs. Opto 22 I/O modules and solid-state relays with date codes of 1/96 or later are guaranteed
for life. This lifetime warranty excludes reed relay, SNAP serial communication modules, SNAP PID modules, and modules
that contain mechanical contacts or switches. Opto 22 does not warrant any product, components, or parts not
manufactured by Opto 22; for these items, the warranty from the original manufacturer applies. These products include,
but are not limited to, OptoTerminal-G70, OptoTerminal-G75, and Sony Ericsson GT-48; see the product data sheet for
specific warranty information. Refer to Opto 22 form number 1042 for complete warranty information.

Wired+Wireless controllers and brains and N-TRON wireless access points are licensed under one or more of the following
patents: U.S. Patent No(s). 5282222, RE37802, 6963617; Canadian Patent No. 2064975; European Patent No. 1142245;
French Patent No. 1142245; British Patent No. 1142245; Japanese Patent No. 2002535925A; German Patent No. 60011224.
Opto 22 FactoryFloor, Optomux, and Pamux are registered trademarks of Opto 22. Generation 4, ioControl, ioDisplay,
ioManager, ioProject, ioUtilities, mistic, Nvio, [Link] Web Portal, OptoConnect, OptoControl, OptoDataLink, OptoDisplay,
OptoEMU, OptoEMU Sensor, OptoEMU Server, OptoOPCServer, OptoScript, OptoServer, OptoTerminal, OptoUtilities, PAC
Control, PAC Display, PAC Manager, PAC Project, SNAP Ethernet I/O, SNAP I/O, SNAP OEM I/O, SNAP PAC System, SNAP
Simple I/O, SNAP Ultimate I/O, and Wired+Wireless are trademarks of Opto 22.
ActiveX, JScript, Microsoft, MS-DOS, VBScript, Visual Basic, Visual C++, Windows, and Windows Vista are either registered
trademarks or trademarks of Microsoft Corporation in the United States and other countries. Linux is a registered
trademark of Linus Torvalds. Unicenter is a registered trademark of Computer Associates International, Inc. ARCNET is a
registered trademark of Datapoint Corporation. Modbus is a registered trademark of Schneider Electric. Wiegand is a
registered trademark of Sensor Engineering Corporation. Nokia, Nokia M2M Platform, Nokia M2M Gateway Software, and
Nokia 31 GSM Connectivity Terminal are trademarks or registered trademarks of Nokia Corporation. Sony is a trademark of
Sony Corporation. Ericsson is a trademark of Telefonaktiebolaget LM Ericsson. CompactLogix, MicroLogix, SLC, and RSLogix
are trademarks of Rockwell Automation. Allen-Bradley and ControlLogix are a registered trademarks of Rockwell
Automation. CIP and EtherNet/IP are trademarks of ODVA.
All other brand or product names are trademarks or registered trademarks of their respective companies or organizations.

ii OptoTutorial: SNAP PAC PID

Table of Contents

Chapter 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
How to Use This Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Skills You’ll Learn in this Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Lesson 1: Basic PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
Lesson 2: Understanding Proportional, Integral, and Derivative . . . . . . . . . . . . . . . . . . .2
What You Need. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
SNAP PAC Learning Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
Purchasing a SNAP PAC Learning Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Downloading Sample Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Creating a Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
Ready for the Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
For Those Who Can’t Wait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9

Chapter 2: Basic PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11

Skills You Will Learn. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
What is PID? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Where PID Resides in a SNAP PAC System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
A Simple PID System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Plotting the Simple System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
System Dead Time and Scan Interval (Scan Rate) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Activity 1: Configure System and Observe System Dead Time . . . . . . . . . . . . . . . . . . . . . . 19
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Configure PID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Determine Scan Interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
For Those Who Can’t Wait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Chapter 3: Understanding Proportional, Integral, and Derivative . . . . . . . . .33

Skills You Will Learn. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

OptoTutorial: SNAP PAC PID i i

 Concepts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
For Those Who Can’t Wait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Proportional, Integral, and Derivative Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Understanding the Proportional Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Proportion Constant: Positive versus Negative Gain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Input and Output Capabilities Need to Be Well Matched . . . . . . . . . . . . . . . . . . . . . . . 37
“Proportion” versus “Gain” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Proportional Control Doesn’t Fully Correct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Understanding Integral . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Integral Windup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Proportional and Integral Work for Many Loops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Understanding Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Considerations for Choosing Derivative Tuning Parameters . . . . . . . . . . . . . . . . . . . . 50
Activity 2: Tune a PID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Preparation: Configure PID Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Determine Ambient Temperature and Setpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Tune Gain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Measure the Steady State Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Apply an Integral Tuning Constant to Correct Steady-State Error . . . . . . . . . . . . . . . 58
Test Gain and Integral Constants Against a Setpoint Change . . . . . . . . . . . . . . . . . . . 60
Refine Integral Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Tune Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Configuring a PID Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Features of the PID Inspect Window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
PID Debug Mode—Plot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Features of the Plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Moving a Plot’s Scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Displaying Cursors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Changing the Cursor Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Using Delta-X and Delta-Y Cursors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Resetting Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
PID Debug Mode—I/O Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
PID Debug Mode—Misc. Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Debug PID—IVAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Debug PID—Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
PID Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Defining a New Control Engine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

ii OptoTutorial: SNAP PAC PID

 Chapter i

1: Introduction

How to Use This Tutorial

This tutorial shows how to configure and tune a PID loop on a SNAP PAC rack-mounted
controller or brain, using the SNAP PAC Learning Center. The introduction describes the
hardware, firmware, and software needed to complete the tutorial.
Each lesson has two sections: Concepts and Activity.
• The Concepts section provides general background information; you do not need a
SNAP PAC Learning Center to benefit from this section. If you have little experience
with PID or need a refresher, you’ll find the Concepts sections especially helpful.
• The Activity section provides step-by-step instructions for using PAC Control with a
SNAP PAC Learning Center and is independent of the Concepts section. In other words,
you do not need to study the Concepts before following the Activities.
Also, this tutorial assumes that you may at any point be eager to depart from the
instructional sequence and explore the PID features. Sections entitled “For Those Who Can’t
Wait” provide some guidance.

Skills You’ll Learn in this Tutorial

Lesson 1: Basic PID Control

• Configuring input, output, and setpoint
• Defining valid range of input
• Clamping output
• Configuring scan interval
• Observing a PID
• Changing tuning parameters in real time
• Adjusting views of a graph: changing resolution of X and Y axes
• Determining system dead time and scan interval

OptoTutorial: SNAP PAC PID 11

WHAT YOU NEED

Lesson 2: Understanding Proportional, Integral, and

Derivative
• Setting gain constants for heating and cooling systems
• Understanding proportional control
• Understanding the integral calculation and the I constant
• Understanding the Derivative calculation and the D constant
• Tuning a PID loop

What You Need

Skills—A basic understanding of PAC Control (provided by the SNAP PAC Learning Center
User’s Guide or by the PAC Control User’s Guide).
Hardware—SNAP PAC™ Learning Center
Software—PAC Project Basic or PAC Project Pro
Configuration file—The sample PAC Control Tag database [Link] is provided with
the Learning Center or in a separate downloadable zip file (OptoTutorial 1641: PID with PAC
Control). The .otg file provides the point configuration for the heater and temperature
probe used in this tutorial.

SNAP PAC Learning Center

Because this tutorial is a supplement to the SNAP PAC Learning Center, this tutorial assumes
the following:
• The Learning Center is assembled according to the directions in the SNAP PAC Learning
Center Guide (Opto 22 form #1638).
• Communication is established between your computer and the SNAP PAC.
The SNAP PAC Learning Center provides a combination temperature probe and heater, used
in this tutorial to demonstrate PID control over a heating system.

2 OptoTutorial: SNAP PAC PID

 CHAPTER 1: INTRODUCTION

NOTE: This probe is referred to as a “temperature sensor” in the SNAP PAC Learning Center Guide

Figure 1-1: SNAP PAC Learning Center and its PID simulator

The PID simulator consists of a temperature

probe that connects to analog input point 12
and a heating element that connects to
analog output point 09.

To use this tutorial, make sure your PID simulator is connected and that all the analog points
and wiring are connected as described in the SNAP PAC Learning Center Guide (form 1638).

Purchasing a SNAP PAC Learning Center

If you need to purchase a SNAP PAC Learning Center, you can do so through our website.
Follow this link or search for part number SNAP-PACLC. Or call Opto 22 Sales at
800-321-6786.

Downloading Sample Files

A zip file with everything you need (the complete tutorial and sample tag database) is
available at [Link]
The sample tag database contains the point configuration used in this tutorial. (The tutorial
does not use all the points of the SNAP PAC Learning Center, and the unused points are not
configured.)

Getting Started

Creating a Strategy
You will need a strategy to configure and tune a PID loop; but once it’s configured, the PID
loop runs independently of a strategy. In other words, you can run a PID loop without any
flowchart logic in your strategy.
1. From the Windows Start menu, select Programs > Opto 22 > PAC Project > PAC Control
Basic

OptoTutorial: SNAP PAC PID 33

GETTING STARTED

2. Create a new control strategy.

a. Choose File > NewStrategy or click the New Strategy button.
b. Navigate to the directory where you stored your sample tag database, for example
C:\Program Files\Opto22\PAC Project\SNAP PAC Learning Center\PID Tutorial

c. In the Create New Strategy dialog box, type PID Tutorial and click Open.

4 OptoTutorial: SNAP PAC PID

 CHAPTER 1: INTRODUCTION

3. Associate your Control Engine.

a. Right-click the
Control Engines
folder.
b. From the pop-up
menu, click Add.
c. In the Configure
Control Engines
dialog box, click
Add.
d. The Control
Engine you
defined for your
SNAP PAC
Learning Center
should be
displayed in the
Select Control
Engine dialog box
(if not, see
“Defining a New
Control Engine”
on page 80).
e. Select your
Control Engine and click OK.
f. Click OK to close the Configure Control Engines dialog box.

4. Import the tag database.

a. Right-click the I/O Units folder.
b. Select Import.

OptoTutorial: SNAP PAC PID 55

GETTING STARTED

c. Navigate to the directory where you stored your sample tag database. For example,
c:\Program Files\Opto22\SNAP PAC Learning Center

d. Select [Link] and click Open.

5. Edit the I/O unit.
a. Expand the I/O Units folder.
b. Under the I/O Units folder, double-click SNAP_PAC_LearningCtr.

6 OptoTutorial: SNAP PAC PID

 CHAPTER 1: INTRODUCTION

c. Change the IP address to the IP address of your SNAP PAC.

NOTE: Alternatively, you can use the Loopback IP address ([Link]). This directs PAC Control
to use the same IP address that is configured for the Control Engine. (The loopback I/O
address is explained in Lesson 2 of the SNAP PAC Learning Center Guide).

Ready for the Tutorial

You are ready for the activities when you have the following configured:

Control engine

Learning Center (Import configuration

from [Link]; change IP address.)
Module 02, Point 09, analog output point (controls
a heating element in the PID simulator)
Module 03, Point 12, analog input point (reports
the temperature of the PID simulator)

Module 04, Point 16, analog input point (connects to the

Learning Center potentiometer. This lets you optionally
control the setpoint with the potentiometer.)

OptoTutorial: SNAP PAC PID 77

GETTING STARTED

It is essential that points 09 and 12 are configured. The configuration for these points is
shown below:
Point 09, Heater

Point 12, Temperature

8 OptoTutorial: SNAP PAC PID

 CHAPTER 1: INTRODUCTION

For Those Who Can’t Wait

The goal of this tutorial is to present a thorough, step-by-step explanation of the concepts
of PID and features provided in PAC Control. For users who prefer to experiment first, here is
a summary of the features that can be configured in PAC Control.
A SNAP PAC rack-mounted controller or brain provides 96 PID loops. Each PID loop can be
configured with unique settings for the following:
• Input variable (process variable)
• Setpoint
• Output (controller output)
• PID algorithm
• Constants for proportional, integral, and derivative
• PID scan time
• Feed forward gain
• Out-of-range input (upper and lower)
• Upper and lower clamps on controller output
• Minimum and maximum changes in output
• Scaling of input and output
Essential configuration for all PID loops:
• Scan rate
• Input (process variable)
• Output (controller output)
• Setpoint
• PID algorithm
• Input low and high range
• Positive or negative gain value. (For the Learning Center’s heating system, use a
negative gain. For more discussion on gain, see “Proportion Constant: Positive versus
Negative Gain.” on page 36.)
Optional configuration parameters:
• Output upper and lower clamps
• Output minimum change
• Output maximum change
• Output for assumed failure
• Feed forward gain
• Calculate square root of input

OptoTutorial: SNAP PAC PID 99

GETTING STARTED

Once the basic configuration settings are made, the following must be done:
• The strategy must be downloaded to the SNAP PAC.
• The strategy must be run to save the PID configuration to the I/O unit.
• The Mode must be set to Auto. (Auto is the default setting.)
• Communication with PAC Control must be Enabled. (Enabled is the default setting.)

10 OptoTutorial: SNAP PAC PID

 Chapter 1

2: Basic PID Control

Skills You Will Learn

In PAC Control Configure mode:
• Configuring input, output, and setpoint
• Defining valid range of input
• Clamping output
• Configuring scan interval
In PAC Control Debug mode:
• Observing a PID
• Changing tuning parameters in real time
• Adjusting views of a graph: changing resolution of X and Y axes
• Determining system dead time and scan interval

Concepts

What is PID?
A proportional integral derivative (PID) control system monitors a process variable,
compares the process variable to a setpoint, and calculates an output to correct any
difference (error) between the setpoint and process variable. The mathematical formulas
that do this vary, but all PID systems share fundamental concepts:
• They evaluate a process variable against its setpoint.
• They control an output to reduce the difference between the process variable and
setpoint.
• The output is the result of proportional, integral, and derivative calculations.
• The effects of proportional, integral, and derivative calculations are modified by
user-determined P, I, and D constants.

Optotutorial: SNAP PAC PID 11

11
CONCEPTS

• The P, I, and D constants need to be adjusted (or tuned) for each system.

Where PID Resides in a SNAP PAC System

In the SNAP PAC system, the PID algorithm runs independently of PAC Control, so the PAC
Control strategy doesn’t need to monitor every variable and configuration setting
associated with the PID algorithm.

PAC Control can change configuration PID loops run independently

settings and start and stop PIDs. of PAC Control.

In the SNAP PAC R, both PAC Control and PID run on the same hardware, but the two parts
remain independent.

12 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

A Simple PID System

A tank, thermostat, and heater is a system with all the elements required for PID control:
Process Variable (PV)
The examples Temperature Controller
throughout will use
sample values Setpoint
comparable to what you
would get with the PID PV = 70
simulator probe
PV should be: 80
(supplied with the SNAP
A thermostat
PAC Learning Center) at Heater output should be provides the
room temperature. increased/ decreased
desired setting for
Controlling temperature
tank temperature
is one type of PID
system. Others may Controller
control pressure, mixture Output
ratios, cooling, flow rate, Output Device
and so on. Heat Source

Figure 2-1: A Basic PID System

Temperature is the process variable monitored by the PID loop used with this tutorial. The
process variable is also referred to as the PID input, or just input. The PID’s output is used to
control a device that will have an effect on the process variable, in this case a heater. A PID
loop’s output is often referred to as the “Controller Output.” In this context, the controller
output is referring to the PID and not to PAC Control’s control engine.

In this tank heating system, the process variable (PV) is the temperature of the tank. The
setpoint is supplied by a thermostat. The controller compares the setpoint to the process
variable and decides the appropriate output to achieve the setpoint. But the system
becomes complex because of the following:
• The output of the PID loop must be scaled to the output device. For example, to turn
up the heat, the controller must convert a difference in temperature to a setting on a
gas valve controlling a heater. Increasing the temperature by 10 degrees may mean
opening the gas valve by 10%, 5%, or 50%—this varies according to the device.
• The difference between the setpoint and the process variable is called the error. This
error is changing in time, so the controller output must also be changing to avoid
overshooting the setpoint.
• A PID algorithm not only maintains a process variable at a setpoint; it must also react to
disturbances, which are external factors affecting the stability of the system. Typical
disturbances are a change in load (for example, filling the tank) or a change in setpoint.
• Equipment and safety limitations may restrict how quickly you can change a setting in
the output device or what conditions determine when the output device can be
turned on or off.
• The range of valid input for the process variable should be defined so the PID can
properly react to or ignore out-of-range inputs.

Optotutorial: SNAP PAC PID 13

13
CONCEPTS

The diagram below shows more elements of this system.

Process Variable (PV): PID Controller

Tank temperature,
reported by temperature Calculate Error: Setpoint
sensor (Error = Process Variable – Setpoint)

70° F 70
Input to PID
80 80 degrees

10 Setpoint may be a setting

within the PID controller, read
from an instrument or operator
interface, or the output from
another PID loop.

Output to Use Error The error is the basis of the controller

heater to Calculate output. The PID controller
Heat Controller manipulates this variable with
Output
proportional, integral, and derivative
calculations.
Output Variable: Controller Output
Percentage that gas valve
is open, which in turn Other factors in calculation of controller output are not
controls the heater shown, such as scaling of the output variable, clamps to
its range, and minimum or maximum on and off times.

Figure 2-2: Details of a Basic PID System

NOTE: PAC Control calculates error by subtracting the Setpoint from the Process Variable. This
calculation creates a negative error when the Process Variable is below the setpoint. This mode
is used for “pump-up” control, such as maintaining level, pressure, flow, and heating.

14 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

Plotting the Simple System

In this tank heating system, the process variable and setpoint are on the same scale, while
the controller output is on a different scale. If the values of the process variable, setpoint,
and output are plotted in time, they would form the graphs shown below.

Process Variable and Setpoint

150° F
Controller output corrects
Stable system and stabilizes system
Setpoint

Process
Variable Process variable
System disturbance,
restored to setpoint
e.g., tank is filled
0° F
Time

Output Variable
Gas valve 80% open

100%
Output
0%
Time
Gas valve 20% open

Figure 2-3: Plots of the Process Variable, Setpoint, and Output

In these graphs, the Y-axis represents the values of the process variable, setpoint, and
controller output. The X-axis represents time. The X-axis is shown with demarcations for
units of time. Though the process variable is changing continuously, the controller only
knows its value when the process variable is scanned. The vertical lines represent the scan
interval, which is an important component in configuring a PID loop.

Optotutorial: SNAP PAC PID 15

15
CONCEPTS

The chart below shows the step changes in controller output.

Process Variable, Setpoint, and Controller Output

Error at time
of scan

Setpoint

Process Variable
(Input)

Time

Output
(Correction) Controller output at time of scan

Time
Scan
interval

Figure 2-4: Actual Values of Process Variable, Setpoint, and Output over Time

NOTE: The input change is truly continuous, as it is shown here. The output,
however, is a step shape. At each scan, the controller sets an output value that
continues unchanged until the next scan.

Figure 2-4 shows how the configured scan interval affects the controller output. With
shorter scan intervals, the output graph resembles a continuous curve, with frequent
controller adjustment and therefore better control of the process variable. Setting an
appropriate scan interval is critical to successful PID tuning.
To summarize, Figures 2-1 through 2-4 show the following concepts relevant to PID
configuration and tuning:
• Setpoint, process variable, output, and scan interval are the key components of the
system.
• Setpoint can be from an input device or any variable accessible to the controller.
• The output device can be a physical device that effects change, or it can be used as the
input or setpoint to another PID. In this example, the output controls a heater.
• The setpoint and process variable are on the same scale.
• The output device (heater) may be on the same scale as the process variable and the
setpoint, but often it is not.

16 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

• Scan interval is the interval at which the controller evaluates the process variable
against the setpoint and calculates the output.

System Dead Time and Scan Interval (Scan Rate)

All PID calculations occur at a time interval you specify by choosing a scan rate when you
configure the PID. This makes scan rate a very important setting. The value you use for scan
rate will vary with each system, and measuring the system dead time can help you assess
the value of your scan interval.
System dead time is the delay between a change in output and its measurable effect on a
system. This delay occurs as a result of inertia in the system and is often a necessary part of
a properly designed system. For example, the designer of this tank heating system would
not place the temperature sensor too close to the heater, as readings would show
temperature values not typical of the whole tank; rather, the temperature sensor would
reside where its reading indicates the average temperature. This design means that time
must pass between a change in the output and seeing the effect of this change in the
temperature sensor.

System dead time is the delay

between when a change in output
occurs and when it has a measurable
effect on the system.
Most systems have a lag due to
inertia. In the tank example, the heat
affects the bottom first, and
convection (or an agitator, not part
of this example) slowly distributes
the heat, eventually registering on
the temperature sensor.

Figure 2-5: System Dead Time in a Tank Heating System

Optotutorial: SNAP PAC PID 17

17
CONCEPTS

Plotting a disturbance in a system reveals the system dead time:

System Dead Time

System dead time: Delay between a
disturbance and measurement of the
disturbance’s effect on the process variable

System is stable

Controller
Output
Change in process variable observed

Change in output (disturbance)

Process
Variable

Figure 2-5: Plot of System Dead Time

System dead time is easily observed by changing the controller output in a stable system.

A scan interval should be shorter than one-third of the system dead time, measured by
changing the output in a stable system. This formula is not absolute, as any system may
have qualities that determine the scan interval.

System Dead Time and Scan Interval

System dead time provides a boundary for maximum scan rate value.

Change in
Output
System dead time

Stable system
Process
Variable

A typical scan interval should be shorter

than 1/3 the system dead time.

Controller
Output

Figure 2-6: Testing Maximum System Interval against the System

Dead Time

18 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

For example, equipment limitations may require longer intervals. In many systems,
however, the scan interval is faster than the time required to observe a change in the
process variable, which can cause the controller to overcorrect. Any overcorrection can be
removed with proper tuning of the PID, which is explained in Lesson 2.
In Activity 1, you will configure a PID loop controlling a heating system and observe the lag
in this system. This activity will provide practice with using the PID in manual mode and
with manipulating its plot.

Activity 1: Configure System and Observe System Dead Time

Preparation
1. Start PAC Control.
2. Open your strategy.

Configure PID
You will configure only the basic components needed to observe system dead time.
Configuring additional features may interfere with determining the system dead time. For
example, you do not want your PID loop to exercise any control, so you will want to be in
Manual mode. Rather, you just want to plot the input (process variable), setpoint, and
output (controller output).
1. Select Mode > Configure.
2. Open PID Loop configuration.
a. In the strategy tree, right-click the PIDs folder.

Optotutorial: SNAP PAC PID 19

19
ACTIVITY 1: CONFIGURE SYSTEM AND OBSERVE SYSTEM DEAD TIME

b. From the pop-up menu, choose Add.

c. Select the first loop (which should be selected by default) and click Add.
3. Enter PID configuration settings.
a. Type or select the following configuration information (some of these will be
selected as defaults):
– Name: Control_Tank_Temperature
– Input: Select I/O Point from the first list and Temperature from the second list
– Low Range: 0
– High Range: 100
– Setpoint: Select Host from the first list and type 0 in the initial value field
– Output: Select I/O Point from the first list and Heater from the second list
– Lower Clamp: 0
– Upper Clamp: 100
– Mode: Manual
– Scan Rate: 1

20 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

Name your PID loop.

Select point used as the

process variable.

Low and high range:

define valid input

Setpoint: Host allows you to control

setpoint from PAC Control
Select point used as
output device.

The heater is scaled to percent, so 0

and 100 prevent the PID from
calculating outputs that are outside of
this logical range.
Mode: Manual prevents the PID from
making output changes so you observe
the system dead time.

Scan rate: Use 1 as a starting point. This

number determines how often the
input is scanned and the controller
output is updated.

b. Click OK to close the Add PID Loop dialog box.

c. Click Close to close the Configure PID Loop dialog box.
Notice that your PID loop is now listed under the PIDs folder:

Optotutorial: SNAP PAC PID 21

21
ACTIVITY 1: CONFIGURE SYSTEM AND OBSERVE SYSTEM DEAD TIME

Determine Scan Interval

1. Download and run the strategy.
a. Choose Configure > Debug.
b. Acknowledge any download messages.
c. When in Debug mode, choose Debug > Run.
Running the strategy sends the PID configuration to the I/O configuration.

Control Engine and I/O Unit

Control Engine I/O Unit

The control engine is the brain's onboard The I/O unit is the part of the brain devoted to
capability for controlling one or more I/O management of the I/O modules. This part can
units using a PAC Control strategy. be configured in PAC Control or in PAC Manager
(and then imported into PAC Control).
Control logic
Active I/O unit configuration

Flowcharts and variables

I/O unit configuration

The I/O unit
configuration
imported or created
in PAC Control is
loaded to the I/O unit Debug Run
when the strategy is
run.

2. Open the PID viewer.

22 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

3. In the PIDs folder, double-click Control_Tank_Temperature.

Your PID should be operating as follows:

– Your PID is in Manual mode, so it reads the input and setpoint but makes no
changes to the output.
– Your output is 0 (blue line).
– Your setpoint is 0 (yellow line).
– Your input (red line) shows the room temperature in Fahrenheit.
– Graphs for input, output, and setpoint are being plotted against time.
4. Set PID output to 30 percent.
a. In the Output field, type 30.
b. Click Apply.

Optotutorial: SNAP PAC PID 23

23
ACTIVITY 1: CONFIGURE SYSTEM AND OBSERVE SYSTEM DEAD TIME

Notice change in both the PID Input and PID Output graphs.

Enter new value for

Output here and
click Apply.

Gaps in a plot indicate that scanning

was interrupted, in this case, when
the Output field was edited.

It takes a few minutes for your system to stabilize. Viewing the input graph at a finer
resolution helps you know when the system has stabilized.

24 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

5. Change view of input axis.

a. Click and drag the input axis so that the plot of the input is at the bottom of the
graph:

Click here to drag

the axis up or down.
Click and drag the input axis so that the
input plot is at the bottom of the visible
graph. This point will be centered when
you change the input axis.

Optotutorial: SNAP PAC PID 25

25
ACTIVITY 1: CONFIGURE SYSTEM AND OBSERVE SYSTEM DEAD TIME

b. Click Input Axis and select View 1% Span.

26 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

6. Change the time axis.

a. Click Time Axis and select View 1 Minute Span.

b. If your graph has stopped tracking, which can happen when focus is on any of the
tuning fields, choose Reset Scale Tracking from the Time Axis menu (or click the
graph).
Until you are familiar with the PID plot, it is recommended that you avoid using the
shortest settings (10 seconds or 1 second) for the time axis. After you’ve observed a
change in the input, you can zoom in on the graph, which is described later.

Optotutorial: SNAP PAC PID 27

27
ACTIVITY 1: CONFIGURE SYSTEM AND OBSERVE SYSTEM DEAD TIME

7. Wait for the system to stabilize.

A stable system will exhibit little change in the input value, which is shown numerically
and graphically.

Note: If your temperature doesn’t stabilize, a likely cause is a varying room temperature. Make
sure your PID simulator probe is not too near to a heating or cooling outlet.

Observe the graph of the input field. The system is stable when the input value does
not vary significantly (some drift in the input value can be expected). Stabilization may
take several minutes, depending on the type of system.
8. Under the Time Axis menu, choose Reset.
a. Click the Time Axis button.
b. Click Reset Scale Tracking.
This is a precautionary step, as changing settings on the plot can cause the trend to
stop plotting data. Resetting the time axis ensures that you are viewing the real-time
values.
NOTE: Be prepared to do the following step quickly. In Step 9 you change the output. A
change in the input will occur within a couple of seconds. As soon as you see a change in
input, you will click and drag the time axis (X-Axis), which will stop the plot so you can closely
examine the system dead time.

9. Observe system dead time with a 20 percent increase in output.

a. Type 50 in the Output field.
b. Click Apply.

28 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

A change in the input will occur quickly:

The input axis (upper

white arrow added to
this picture) begins to
respond to the change
in output (lower white
arrow).

c. Click and drag the Time Axis (X-Axis) so that the plot shows the change in output
and the change in input.
10. Display a Delta X cursor.
a. From the Data menu, select Cursor.

b. Right-click the cursor and choose Style > Delta X.

Optotutorial: SNAP PAC PID 29

29
ACTIVITY 1: CONFIGURE SYSTEM AND OBSERVE SYSTEM DEAD TIME

The Delta X cursor displays the time difference between the two vertical bars.

11. Measure the System Dead Time.

The time between the change in output and the change in input represents the system
dead time.
a. To set the vertical measurement bars, click a vertical line and drag it left or right.
b. Place the first bar after the output change. Place the second bar before the change
in input.

30 Optotutorial: SNAP PAC PID

 CHAPTER 2: BASIC PID CONTROL

In this example, the system

dead time was shown to NOTE: With a scan interval of 1 second you may conclude
be about 1.8 seconds. Your that your measurement of the system dead time is
accurate to the nearest second. However, the PID viewer
results may vary, but a
plots the current values of the Process Variable and
typical system dead time Setpoint, regardless of the scan time. Therefore, your
for the PID simulator is measurement of system dead time is accurate.
around 2 seconds,
suggesting that 1 second In most applications, you can set the scan rate as fast as
is too long of a scan possible and use this test of system dead time only to
interval, but 0.4 should be better understand the reaction time of your system. The
adequate. SNAP PAC will easily handle shorter scan intervals, and for
many applications you might begin tuning by setting the
12. Type a new system scan rate to 0.1.
interval.
a. In the Scan Rate field, type 0.4
b. Click Apply.
13. Save Tuning.
a. Click the Save Tuning
button.
b. Click Yes.

14. Return to Configure mode.

a. Click Close to close the View PID Loop (Scanning) dialog box.
b. Choose Mode > Configure.

For Those Who Can’t Wait

At this point you may be very eager to complete the essential PID configuration and see a
working PID.
Lesson 2 discusses gain, integral, and derivative settings. In that activity, you will observe
and tune a working PID loop. All of this follows a thorough discussion of P, I, and D
calculations along with an explanation of the different PID algorithms available.
If you wish to experiment on your own now and see your PID loop, you will need to change
the settings as described below.

Optotutorial: SNAP PAC PID 31

31
ACTIVITY 1: CONFIGURE SYSTEM AND OBSERVE SYSTEM DEAD TIME

1. In Configure mode, open the PID loop configuration.

a. In the Strategy Tree, right-click PID ControlTankTemperature.
b. Select Modify.
2. Type or select the following configuration information.
a. Setpoint: Change
from Host to I/O
Point – Setpoint.
This setting will get
the setpoint from
the potentiometer.
b. Mode: Change
Manual to Auto.
c. Gain: Change 1 to
–10
In lesson 2, you will
see why it is critical
for this system to
have a negative
number for the
gain constant.
d. Integral: Change 0
to 0.5

3. Download your strategy.

4. Start your strategy.
5. Open the PID Inspect dialog box.
It is recommended that you observe your loop’s response to setpoint changes a few
degrees above room temperature.
When finished, return to Configure mode.

32 Optotutorial: SNAP PAC PID

 Chapter 2

3: Understanding Proportional,
Integral, and Derivative

Skills You Will Learn

• Setting gain constants for heating and cooling systems
• Understanding proportional control
• Understanding the integral calculation and the I constant
• Understanding the derivative calculation and the D constant
• Tuning a PID loop

Concepts

For Those Who Can’t Wait

This lesson provides a thorough discussion of proportion, integral, and derivative
calculations as they are deployed in PAC Control’s algorithms. Though the Concepts section
provides useful background to Activity 2, you may wish to begin the activity after reading
the concepts on proportional control and then resume reading the Concepts section when
the Activity begins to tune the integral.

Proportional, Integral, and Derivative Calculations

We intuitively use PID when making corrections to driving, cooking, walking, and so on. The
trick is getting machinery to do what humans do: examine the system, apply corrective
input, watch, compensate, and anticipate.
A PID algorithm is designed to do this with numbers. It calculates a difference between the
current and desired values and uses this difference, called error, to apply a corrective
response. The error is examined over time with the help of the integral term, which corrects
any tendency of the system to stabilize at the wrong value. The rate of change is analyzed
by the derivative calculation, which tries to dampen a tendency to overcorrect.

OptoTutorial: SNAP PAC PID 33

33
CONCEPTS

PID Controller
Process Calculate Error
Variable (PV)

(Input)
Setpoint

Controller output comprises

proportional, integral, and
Proportional
derivative calculations. The
Term
influence of each part of the
Integral calculation depends in part
Output Term on the values you assign to
Heat constants during a process
called tuning.
Derivative
Output Term
Device
Controller Output

Figure 3-7: Heating System with Gain, Integral, and Derivative

The PID algorithm uses the process variable and setpoint to make gain, integral, and derivative
calculations. These separate calculations are referred to as the Proportional Term, Integral Term, and
Derivative Term. Each of these terms reacts to error or error over time. The extent to which each term
influences the controller output is determined by the P, I, and D tuning constants.

A successful PID loop requires well-chosen constants for gain, integral, and sometimes
derivative (many loops work with P and I constants only). Choosing your constants is part
understanding their role in the PID algorithm (explained in this tutorial) and part trial and
error, a process referred to as tuning.
The SNAP PAC offers five PID algorithms:
• Velocity Type B (the default type)
• Velocity - Type C (available only for PAC-R, EB, and SB with firmware version 8.5e and
higher)
• ISA
• Parallel
• Interacting
The offering of five algorithms illustrates that there is no one formula for PID control. Each
algorithm has unique qualities, but the general concepts of proportional, integral, and
derivative, and tuning constants, is similar across all. The following discussion describes
generalizations of the effect of P, I and D components of the calculation. Specific attributes
of the calculation will be explained as well.

34 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

Understanding the Proportional Calculation

In proportional
control, the output Proportional Control
signal is a Error signal
proportional (Deviation from
setpoint) Output
response to the response
error signal. This Setpoint (heater)
proportional
response is
achieved by the Temperature (Input)
gain constant.
Figure 3-8 shows In proportional control, the output signal is a proportional (or
multiplied) response to the error signal. Here is shown a gain
proportional of 1. Moving the blue triangle left or right is analogous to
control as a using a higher or lower gain setting.
teeter-totter, where
the size of the error Figure 3-8: Proportional Control
has a direct effect on the
controller output.

Imagine the black line swinging

Gain in Proportional Control
up or down depending on the
value of the process variable. With Controller
a gain of 1, the error and controller Setpoint
Error output
output are equal.
In this analogy, increasing the
gain moves the fulcrum point (the Gain > 1
triangle ) left, where a small
Controller
error can produce a large change output
Setpoint
in the controller output.
Decreasing the gain moves the
fulcrum point to the right, where
Gain =1
the error produces less change in
controller output.
Error Controller
All of these gain settings are output
Setpoint
examples of proportional
responses to the error.
0 < Gain < 1

Figure 3-9: Effect of Various Gain Settings

OptoTutorial: SNAP PAC PID 35

35
CONCEPTS

Proportion Constant: Positive versus Negative Gain

The gain constant used in the PID algorithms can be a positive or negative number. When
the process variable drops below setpoint in a heating system, the error is a negative
number. This negative number must be converted to a positive number that increases the
heater’s output. Using a negative gain constant achieves this. The reverse is required for
cooling: exceeding the setpoint should result in a controller output that increases the
output of a cooling device. Figure 3-10 compares the effects of positive and negative gain
constants in heating and cooling systems.

Proportional Control: Positive and Negative P Constants

A proportional response to the error is achieved
by multiplying the error by a constant.

E = Error
P = Proportion constant
P constant can be positive, negative,
whole number, or fraction. CO = Controller output

Heating System and Negative P Constant

Output
response
Setpoint (Heater)

Input: Temperature

When using proportional control in a heating system, a

negative constant achieves the desired controller output.

Cooling System and Positive P Constant

Output
response
(Fan)

Setpoint

Input: Temperature

When using proportional control in a cooling system, a

positive constant achieves the desired controller output.

Figure 3-10: Heating and Cooling, Positive and Negative Gain

36 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

Input and Output Capabilities Need to Be Well Matched

There are several layers of design that achieve the appropriate balance between the output
device and the process variable being controlled:
• The physical equipment—In the tank example, the heater must have the capacity to
change the tank temperature within the time required by the application.
• The selection and scaling of the module—All PID calculations reduce to ranges that
the input and output modules are capable of attaining. If you have an output device
that responds to values of 0 to 10 volts, then a module of this range offers the best
resolution.
• The valid range options in PID configuration—One of the uses of these ranges is to
calculate a value called span, which is used in all the PID algorithms provided in PAC
Control. Span is a ratio of the output range to the input range and has the effect of
correlating the input and output scales.

“Proportion” versus “Gain”

Up to this point our discussion of proportional control has described a hypothetical use of a
numerical constant to achieve a proportional response to the error. The actual
implementation of a proportional constant varies with each PID algorithm. For example, in
the Parallel algorithm, the P-term includes the result of the gain tuning constant multiplied
with the error. The other algorithms, however, use the gain not only to multiply the error,
but to also multiply the integral term, the derivative term, or both. In this use, the
proportion constant affects the gain of the P, I, and D terms. This is why in actual use, the
term gain is used instead of proportional constant.

Span is used in all the algorithms to compensate for

differences in scale between the input and the output.

Gain multiplies P, I, and

D terms.

Gain multiplies P, I, and D

terms.
Gain multiplies P term.

Gain multiplies P
and I terms.
Figure 3-1: Portions of the PID Algorithms Showing the Use of the
Gain-Tuning Constant
NOTE: These are parts of the algorithms selected to emphasize the role of the
gain-tuning constant. In these parts of the algorithms, Velocity B and Velocity C are the
same. For the complete algorithms, see “PID Algorithms.” on page 79.

OptoTutorial: SNAP PAC PID 37

37
CONCEPTS

Proportional Control Doesn’t Fully Correct

There are many control devices that use gain only. For example, a float on a tank valve can
be used to control the opening of an intake valve. As the float rises, it slowly closes the valve.
In this mechanical control system, where there is no change in setpoint, a proportional
correction can be adequate. However, most systems that use PID control will not be
adequately controlled using a proportional correction only. For example:
• Some systems will overshoot the setpoint, creating an undesirable oscillation between
overcorrection and undercorrection. Such a system is unstable, inadequate for
changing load or setpoint, and may cause excessive wear on the equipment.
• In systems where a proportional correction is stable, the process variable often
stabilizes at the wrong value. This is called steady-state error or offset. To correct for this,
you need to look at error over time.
Figure 3-11 depicts the proportional-only correction in one system with an arbitrary gain
constant of –2, chosen to keep the arithmetic simple. The tank is exposed to a 50° F breeze.
In calculation 1 (see 1 in Figure 3-11), a 50° F input (process variable) generates an error of
–20, which is multiplied by –2 to create a controller output of 40. All of the controller output
goes toward raising the process variable to the setpoint.
In calculation 2 ( 2 ), the tank temperature has been raised to 55° F, so the error is –15. This
generates a controller output of 30. However, as the tank is now above the ambient
temperature, part of the energy intended to correct the error is being used to compensate
for heat lost to the environment.
As the process variable continues to rise ( 3 and 4 ), the output decreases while more of its
energy is used to maintain against heat loss and less goes to raising the temperature to the
setpoint. Eventually, all of the controller output goes to just maintaining the tank against
heat loss.
In this example, an error of -5 maintains the temperature at 65 degrees, which in turn
produces a controller output of 10. The output of 10 produces an error of –5, which
perpetuates the current state.

38 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

Proportional-only control Process variable

usually has a steady-state Setpoint
error, also called offset.
Error Setpoint = 70
Cool breeze
50° F

If process
variable is: Error is: Controller output is:
65 –5 –2 10

60 –10 –2 20

55 –15 –2 30

Heat 50 –20 –2 40

Energy required to maintain Controller output used

temperature against heat loss. to raise temperature.
Figure 3-11: Proportional Control and Steady State Error
This is a simplified example of what can happen in a system with one gain constant. The top of the gray box
restates how the error is calculated. A gain constant of –2 is arbitrarily chosen to keep the arithmetic simple. The
circled numbers (1–4) show the controller calculations at different values for the process variable. The hatched
( ) and solid ( )bars in the controller output box show how much of the controller output is used to
correct the error.

So in this example, the controller output matches the heat needed to maintain the
temperature's current status. There is no guarantee that this equilibrium occurs at the
setpoint. If it did, it would do so for one load, one setpoint, and one gain setting. If setpoint,
load, or both change, an offset would still occur.

OptoTutorial: SNAP PAC PID 39

39
CONCEPTS

A plot of the system shown in Figure 3-11 appears in Figure 3-12.

Steady State Error

Setpoint Steady-state
error, or offset

Process Correcting an offset

Variable requires comparing the
effect of corrections
over time, or the
accumulated error
(Process Variable and Controller Output are on different scales)

Controller
Output

Figure 3-12: Steady-State Error in Proportional-Only Controller Output

Proportional-only correction demonstrates a need for a calculation that can consider the effect of an
output over time.

Understanding Integral
In some systems it may be possible to increase the gain to reduce the offset error, but
usually it is better to use an integral term. For example, if you are using the Learning Center’s
PID probe at room temperature with a setpoint 5 degrees above room temperature, you
may see little performance improvement among gain settings of -10, -20, and -30.
However, by giving the I tuning constant a positive value, such as 0.5, you enable the
calculation that generates an I term.
The I term affects the controller output by increasing or decreasing the value that would be
produced by gain alone. (The I constant doesn’t need to be high, sometimes just enough to
get the integral factored into the calculation.)

40 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

Generally, the I term modifies the controller output as shown in Figure 3-13.

Integral Calculation Correcting Steady-State Error

Cumulative error at
time of scan
Setpoint Steady-state
error, or offset

Process Correcting an offset

Variable requires comparing the
effect of two
corrections over time

(Process Variable and Controller Output are on different scales)

Controller
Output

Output if P acted alone

Figure 3-13: Integral Correcting Steady-State Error

The I Term examines the cumulative error and either increases the controller output or decreases it,
depending on whether the error is positive or negative. The influence of the P term will vary according to the
value of the integral tuning constant.

Comparing a proportional calculation to an integral calculation reveals the measurement of

integral to be more complicated mathematically. Proportional correction is multiplication of
the error by a gain. Integral correction, however, requires the measurement of error over
time.

OptoTutorial: SNAP PAC PID 41

41
CONCEPTS

Integral Calculation Correcting Steady-State Error

Setpoint

Process True Mathematical

Variable Integral

Mathematical
Integral
The integral is the area between the setpoint
and the process variable.

Figure 3-14: True Mathematical Integral

Calculating a true integral requires measuring the area between the graphs of the setpoint and the process
variable. PID algorithms simplify this measurement, because calculating a true integral would take a lot of
processing power and might be no more effective than approximation.

Each of the available algorithms calculates an I Term, which determines the extent to which
error over time (accumulated error) influences the controller output. Also, each of these
calculations approximates the integral measurement. For example, the Velocity algorithms
estimate integral by multiplying the scan interval by the error.

The Velocity Algorithms’ Simplified Integral Calculation

Current integral is the scan
interval multiplied by the
Scan interval current error.
Setpoint

Error
Process
Variable

Current
Integral

Figure 3-15: Integral Measurement in the Velocity Algorithms

The approximation is due to the scan interval: an infinitely fast scan interval would produce
a perfect integral, but such a scanning rate is not practical. In addition, the influence of the I
term must be tunable.

42 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

The influence of the I Term in the Velocity B and C algorithms is calculated with this
equation:
Term I = Tune I * System Interval * Err.

Term I is 0 until a positive Tune I constant is provided. (NOTE: Tune I must be a positive
number.)
The Interacting, ISA, and Parallel algorithms estimate integral by measuring the
accumulated error from the current and previous scans.

Integral Measurement in the Interacting, ISA, and Parallel Algorithms

Setpoint

Process
Variable Error
Integral = 36 + 3 = 38
Current
Integral Integral = 29 + 7 = 36

Integral = 20 + 9 = 29

Integral = 10 + 10 = 20

Integral = 0 + 10 = 10

Figure 3-16: Integral Measurement in the ISA, Interacting, and Parallel

Algorithms

The influence of the I term in the Interacting, ISA, and Parallel algorithms is calculated in
these two equations:
Integral = Integral + Error
Term I = Tune I * System Interval * Integral
Term I is 0 until a Tune I constant is provided. Tune I must be a positive number.
With typical scan rates of 0.1 seconds, none of these algorithms measures the true
mathematical integral. Nor do they need to, because the Tune I constant determines how
effective Term I is in the particular application. This setting is not determined
mathematically, but empirically, by watching the effect of various settings in your PID
system.

Integral Windup
The Integral component of a PID calculation looks at the error over time. This allows the
Integral to efficiently correct for offset. However, there are times when the cumulative error
is not relevant to the necessary controller output. In our tank-heating scenario, imagine that

OptoTutorial: SNAP PAC PID 43

43
CONCEPTS

the heater failed overnight and was not repaired until the following morning. After the
repair, the heater runs at 100 percent for a significant length of time. Adding the error
accumulated during the night makes no difference to the output —it’s still at 100 percent
with or without the integral.

Integral Windup

Setpoint
Integral accumulated
Process error
Variable When the system
returns to normal, the
historical error is
Current irrelevant and should be
Integral The historical error at this point discarded.
has no beneficial contribution to
a controller output that is
already at maximum.

Controller
Output

Figure 3-17: Integral Windup

Integral windup refers to an integral measurement that exceeds a practical limit within the
PID calculation. Typically, anytime the controller output exceeds the range of the output
device, the accumulated error is irrelevant or impairs effective control when the system
returns to normal.

44 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

The Velocity algorithms measure integral by multiplying the current error by the scan
interval. Because this calculation only considers the current error, the Velocity algorithms do
not suffer from integral windup.

The Velocity Algorithms’ Simplified Integral Calculation Current integral is the scan
interval multiplied by the
Scan interval current error.
Setpoint

Process Error
Variable

Current
Integral
Previous integral
calculations aren't
considered in the
current calculation

Figure 3-18: Velocity Algorithms and Integral Windup

In the ISA, Interacting, and Parallel algorithms, the integral measurement is an accumulation
of error. These algorithms have an override function, called anti-integral windup, that is
invoked when the controller output exceeds the range of the output device. (The range of
the output device is set by the PID configuration settings Input–Low Range and Input–High
Range.) The anti-integral windup feature does the following:
• Disables the derivative calculation (if the D tuning constant is not 0) while the output is
out of range.
• Recalculates an integral value to moderate change in output when control is restored
to the PID algorithm. This creates a soft landing; in contrast, the true integral or a zero
integral may cause an abrupt change.
• Maintains a designated output setting (if configured to do so in the output options for
when input is out of range).
• Monitors the process variable in order to relinquish control to the PID algorithm when
the process variable has returned to the valid range.

OptoTutorial: SNAP PAC PID 45

45
CONCEPTS

When anti-integral windup is in effect for the ISA, Parallel, and Interacting algorithms, you
are not able to see the regular PID calculations until the output returns within the range and
control is restored to the algorithm.

Figure 3-19: PAC Control Debug, PID Loop Viewer

Anti-integral windup is activated in this PID heating loop in which the setpoint was
lowered below the process variable. Anti-integral windup asserts control when the
controller output is either over or under the range of the output device.

46 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

Proportional and Integral Work for Many Loops

In many PID calculations, a combination of proportional and integral control is sufficient. In
such applications, the effect of the gain and integral tuning constants follows the patterns
shown in Figure 20.

0.5 x Integral 2 x Integral

(I tune is too low) Optimal Integral (I tune is too high)

2 x Gain
Gain too 1 2 3
high

Optimal
Gain 4 5 6

0.5 x Gain
(Gain too 7 8 9
low)

Setpoint Process Variable (Input)

Figure 3-20: Proportional and Integral Tuning Guidelines

NOTE: This diagram is provided courtesy of Carl Wadsworth of APCO Inc.

Figure 3-20 represents nine tests in which a stable system is disturbed by a change in
setpoint. The columns and rows represent the same disturbance but with different gain and
integral tuning constants. Top to bottom rows represent too much gain, optimum gain, and
too little gain. When your system resembles one of the graphs in the second row, you then
try various integral settings. Left to right columns represent too little integral, optimum
integral, and too much integral.
In most systems, best results come from first tuning gain and then integral. The optimal
gain and integral tuning depends on your application.
If your system requires better response to changing loads or better dampening of change,
then apply a derivative constant in addition.

OptoTutorial: SNAP PAC PID 47

47
CONCEPTS

Understanding Derivative
A derivative term is based on the rate of change in the process variable. The measurement
of derivative produces a term that increases or decreases the controller output based on the
speed at which the process variable is approaching or moving away from setpoint.
Mathematically, the derivative is a line that is tangential to the graph of the process variable
at a point in time. By analyzing the slope of the line, which way it points and how steep it is,
the derivative term adjusts the controller output.

Mathematical Derivative

Derivative becomes more influential to

Process the controller output.
Variable

Setpoint

Derivative
contribution will try
to reduce any
over-correction.

Derivative contribution
becomes less influential.

Figure 3-21: Influence of the Derivative on Controller Output

Like integral, a true mathematical derivative may require more processing power than the
PID calculation needs, so the measurement is simplified, and tuning parameters help
correct for the estimate (as well as adjust the algorithm to the unique needs of a real
system).

48 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

In the ISA, Parallel and Interacting algorithms, the final D term is based on the slope
multiplied by a ratio of the D tuning parameter to the scan interval.

Derivative Measurement in ISA, Parallel, and Interacting Algorithms

Scan interval

Setpoint

Process
Variable
Derivative measurement calculates the
slope of a line between the current and
previous process variables.

Previous process
variable Current process
variable

Slope of line

The D constant will either increase or decrease the influence of the slope,
depending on 1) whether Tune D is greater or less than the system
interval, and 2) whether the scan interval is greater than 1.

Figure 3-22: Derivative Measurement in ISA, Parallel, and Interact-

ing Algorithms

OptoTutorial: SNAP PAC PID 49

49
CONCEPTS

In the Velocity algorithms, the final D term is based on the difference between the current
slope and the previous slope multiplied by a ratio of the D tuning parameter to the scan
interval. This is similar to the other algorithms, but the last two slopes are considered in the
calculation of the D term.

Derivative Measurement in Velocity Algorithms

The Velocity algorithms calculate the difference between the last two slopes.

Process
Variable Previous slope
PV_2
Current slope

PV_1

Current Process Variable (PV)

Scan Interval
Current slope minus
previous slope

Difference between
slopes

The D constant will either increase or decrease the influence of the slope,
depending on 1) whether Tune D is greater or less than the system
interval, and 2) whether the scan interval is greater than 1.

Figure 3-23: Derivative Measurement in Velocity Algorithms

Considerations for Choosing Derivative Tuning Parameters

When tuning the D constant, these considerations apply:
• Many loops work adequately without a D term. Use a D term to reduce overshooting of
the setpoint or in applications that have a long system dead time.
• When choosing a D tuning value, be aware of your scan interval. The D term may differ
in magnitude depending on whether your D tuning constant is smaller or larger than
your scan interval.

50 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

• Try D tune values close to your scan interval, and observe the effects of higher and
lower values.

Activity 2: Tune a PID

There is no absolute way to tune a PID loop, as systems vary. In many applications, safety
has to be considered first and precautions made to prevent a process from becoming
unstable or equipment from being damaged. Each system will have its own safety
requirements, which you must determine. In addition, PID loops from different
manufacturers will behave differently.
All specific tuning parameters cited in this lesson apply to the Learning Center’s PID
simulator, which is safe when used properly.

Preparation: Configure PID Loop

In Activity 1, you configured the minimum features needed to monitor the system dead
time. Then you modified the scan rate so that it was one-third the system dead time.
In this activity, you will tune the PID loop on the SNAP PAC Learning Center. Before
proceeding, make sure your configuration settings are set as described below.
1. In Configure mode, open the PID loop configuration.
a. In the Strategy Tree, right-click the PID Control_Tank_Temperature.
b. Select Modify.
2. Provide PID configuration.
a. Enter the following information:
Input: I/O Point – Temperature
Input low range: 0
Input high range: 100
Setpoint: Host – 0 (Initial value)
Output: I/O Point
Output lower clamp: 0
Output upper clamp: 100
Algorithm: ISA
Although you can use any of the algorithms, a non-Velocity algorithm is chosen for
this activity to demonstrate steady-state error. Unlike the ISA, Parallel and
Interacting algorithms, the Velocity algorithms require an initial integral constant
(in addition to gain), which may prevent the steady-state error that this activity
intends to demonstrate.
Mode: Manual
Scan Rate: 0.4

OptoTutorial: SNAP PAC PID 51

51
ACTIVITY 2: TUNE A PID

Gain: –5
A final gain constant will be determined by tuning, but before you can tune your
PID, your gain constant must be either a positive or negative according to the type
of system you have. PAC Control’s PID algorithms calculate error by subtracting
setpoint from the input (PV – Input), therefore producing a negative error when
heat needs to be applied. To convert the error to the appropriate output, gain must
be a negative value.
Tune I: 0
Tune D: 0
b. Click OK to close
the Edit PID Loop
dialog box.

3. Download and run your strategy.

Determine Ambient Temperature and Setpoint

You will tune the gain by observing the response to a setpoint increase of 5° F above
ambient temperature. Before doing this, you need to determine the ambient temperature
and arrange your graph so that you can easily see the effect of the setpoint change on the
input.
1. In Debug mode, double-click the PID icon on the Strategy Tree.
The View PID Loop dialog box appears with the setpoint, process variable (input), and
controller output plotted in real time. (As your loop should be in manual mode,
disregard the setpoint and output values at this time.)
2. In the Output field, type 0 and click Apply.
3. Wait for your temperature value to stabilize.

52 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

The following steps are based on an ambient temperature of 75° F degrees. Hopefully,
your ambient temperature isn’t too far from this.
4. Change your setpoint.
a. In the Setpoint
field, type a
value that is 5°
F warmer than
ambient
temperature.
For example, if
your ambient
temperature is
75, type 80.
b. Click Apply.

5. Adjust the span of the input axis to 10 percent.

a. Click Input Axis.
b. Choose View 10% Span.

OptoTutorial: SNAP PAC PID 53

53
ACTIVITY 2: TUNE A PID

6. Center the input axis on the setpoint.

Click and drag the
input access scale
up or down until
the setpoint is
centered vertically.
Make sure the
setpoint and the
input graphs are
displayed.

NOTE: You can also choose Center on Setpoint from the Input Axis menu, but you may still
need to drag the scale up or down to show both graphs.

Tune Gain
In tuning the gain, you should be prepared to recognize the following:
• Process instability
• Steady-state error
• Expected workload of your output device
An unstable PID loop will oscillate above and below setpoint, with each oscillation
diverging farther from setpoint. This is unlikely to happen with the Learning Center’s PID
simulator, as the potting of the temperature probe and heat element creates a slow heat
loss and the heating element is not very powerful. With a real application, you must be
careful to reduce the chance of creating an unstable PID loop by starting with low gain
settings and gradually increasing. (In a real application, you may also clamp output and
configure safe settings for when input is out of range.)
You will start with low gain settings, just as in a real application, but you will be evaluating
whether your gain settings produce what you think is a reasonable response from the
controller. For example, if a 5-degree error produces an output of 5 percent, this is less than
what is expected with our heater. It is safe to push our heater to 80 or 100 percent to
respond to our 5-degree load change.
You can raise the gain significantly and never attain the desired control. You need to
recognize when to use an I constant without overshooting the gain.
After you find gain settings that allow the controller to run the range of the output device,
you will look for a tendency for the input (process variable, temperature) to stabilize at the
wrong value. This is steady-state error and an indication that an integral term is needed.

54 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

1. In the Mode field, select Auto and click Apply.

Observe the response of the controller. Typically, you’ll notice that the controller output
drops before this setpoint is reached, and therefore a gain of –5 is too low.

The white arrows (added

to the screen capture)
show the effect of a gain
of –5: an error of 4
produces an output that
is 20 percent of what the
heater is capable of. Your
result may be different,
but clearly the output is
not sufficient, as the
temperature is too low.

2. Try various increases in gain.

a. Type –10 in the Gain field.
b. Click Apply.
Observe the input and the output levels. Output is still low—approximately 40
percent.
c. Type -30 in the Gain field.
d. Click Apply.

OptoTutorial: SNAP PAC PID 55

55
ACTIVITY 2: TUNE A PID

At a gain of –30, you should be able to raise the output to 80 percent or higher. In this
example, the output went out of range briefly before dropping to 35 percent.

The white arrows

(added to the screen
capture) show the
effects of gain settings
of –5, –10, and –30.
The gain of –30 results
in a steady state error
of about 1 degree.

A gain of –30 drives the output across the full range of the heater, and the process is
showing a steady-state error. At this point, you can test the gain by trying the 5-degree
load change with gain settings between –10 and –30.
3. Change the setpoint to 70 and wait for your system to stabilize at the ambient
temperature.
This step assumes your ambient temperature is around 75 degrees. If your temperature
is different, choose any setting below ambient temperature that sets the output to 0.
You may need to wait a few minutes for your input to reach ambient temperature.
4. Test a gain of –20.
a. Make sure you wait until the input is stable at room temperature.
b. In the Gain field, type –20.
This value is chosen because it is halfway between –10 and –30, and therefore it
will narrow the range of gain values.
c. In the Setpoint field, type a value 5° F above ambient temperature, for example, 80.
d. Click Apply to apply both changes.
5. Watch the input response and the output.

56 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

Wait for the input to stabilize.

In the example shown here, the gain of –20 resulted in an initial controller output of about
80 percent, which quickly diminished. If you compare this graph to the graph in Step 2 (on
page 55), you’ll see that gains of –20 and –30 have a similar result: both drive the input to
within 2° F of setpoint. Since a gain of –30 is not more effective than a gain of –20, it’s
reasonable to suspect that we have reached the useful limit of gain. (NOTE: Because of
differences in your environment and equipment, you may achieve this point with a different
gain setting.)

Measure the Steady State Error

1. Click Data and choose Cursor to display the cursor.
2. Right-click the data cursor displayed in the PID plot and choose Style > Delta Y.
3. Drag measurement bars to setpoint and to input.
a. Click and drag the upper red cursor to the setpoint graph.

OptoTutorial: SNAP PAC PID 57

57
ACTIVITY 2: TUNE A PID

b. Click and drag the lower red cursor to the input graph.

In this example, the steady state error is about 1.5° F below setpoint, and a gain of -20 yields
an output around 80 percent of the range of the output device. At this point it is reasonable
to deploy an integral term to correct for the steady-state error.

Apply an Integral Tuning Constant to Correct Steady-State

Error
In the ISA algorithm, the gain will multiply the integral (and derivative) terms.

58 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

In the Integral field, type 0.1 and click Apply.

NOTE: In this activity, all
screens showing a PID
loop’s response to
setpoint changes
represent actual tests.
This test was done when
the ambient temperature
dropped to 73° F, and so a
setpoint of 78 was used.

In this example, an integral constant of 0.1 corrected the offset and continued to maintain
the input close to setpoint with gradual changes in the output:

If your results are different, try higher or lower integral values, for example, 0.05 or 0.2. Also
be aware that reducing the offset within a degree is acceptable, especially if your ambient
temperature is fluctuating. You do not need to get perfect results, as the final tuning
involves testing the tuning parameters against load changes.
Two of three objectives are being met with these tuning parameters:
• The setpoint is being maintained.

OptoTutorial: SNAP PAC PID 59

59
ACTIVITY 2: TUNE A PID

• The output of the heater is steady.

The third goal achievable with this equipment is effective response to setpoint changes,
which can be tuned using P and I only.

Test Gain and Integral Constants Against a Setpoint Change

1. Lower setpoint to a minimum of 5 degrees below ambient temperature.
a. In the Setpoint field, type 70 (or any value low enough to change the output to 0).
b. Click Apply.
2. Wait for the input to stabilize at ambient temperature.
3. Change the setpoint to 5 degrees above ambient temperature.
a. Type 80 or any value that is approximately 5° F above ambient temperature.
b. Click Apply.
4. Compare the graph to the gain and integral matrix on page 61.
You can compare your graph to the sample graphs in a variety of ways:
a. Review the graph on screen.
b. Print the graph. Move the Time Axis so you can see the change in setpoint, and
then click Data > Print.
c. Save the graph as a file to view in a graphics program. Move the Time Axis so you
can see the change in setpoint, and then choose Data > Save As. After saving your
file, open it in your graphics program.

60 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

0.5 x Integral 2 x Integral

(I tune is too low) Optimal Integral (I tune is too high)

2 x Gain
(Gain too 1 2 3
high)

Optimal
Gain 4 5 6

0.5 x Gain
(Gain too 7 8 9
low)

Setpoint Process Variable (Input)

In this example, the graph most closely resembles graphs 6 and 9, suggesting that the gain
may too low, but integral is definitely too high.
Your results may vary, but you can apply the same method to your PID loop. Identify the
curve that best matches your results, and experiment with changes to gain and integral. It is
recommended that you change only one tuning constant in each test.
5. The following steps will test if higher gain produces a graph more like graph 6. Then,
depending on the results, lower integral will be tested.
6. Repeat load change test with more gain.
a. In the Gain field, type –30
b. In the Setpoint field, type a low setpoint, such as 70
c. Click Apply.
d. Wait for your input to stabilize at ambient temperature.
e. Change the setpoint to create a 5° F load change; for example, if your ambient
temperature is 74° F, change setpoint to 79° F.

OptoTutorial: SNAP PAC PID 61

61
ACTIVITY 2: TUNE A PID

The following graphic shows the response to a 5° F load change with the tuning constants
shown.

At first glance, the difference in response seems

negligible, but the increased gain has made a Aggressive vs. Conservative
detectable improvement. The loop that exceeds The terms aggressive and
the setpoint is narrower, indicating a more conservative are extremely
aggressive response from the controller. If you subjective and depend on the
achieved similar results, you may now wish to process. Aggressive tuning can be
repeat the test with variations on the integral thought of as tuning where the
constant. (Use the Delta-Y cursor to measure speed of reaching the setpoint is
overshoots.) the primary concern, and
If you are not achieving similar results, try the overshoot is tolerated to gain fast
following: response. Conservative tuning
• Repeat the test with higher or lower gain can be thought of as tuning
settings according to which graph shown on required in a system where
page 47 is closest to your results. overshoot is not tolerated, and
speed of response is sacrificed to
• Or you can return the gain to a previous prevent overshoot.
setting and try modifying the integral instead.
You can proceed to the next step, keeping
your gain setting.

62 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

Refine Integral Tuning

1. Reduce the integral.
It is a judgment to say whether the last graph resembles chart 3, 6, or 9. The way to use
the charts is to compare the change between your graphs and determine if a change
resembles the trends between charts 1–9. In the example shown, the shorter duration
of the oscillation above setpoint suggests chart 9 changing to chart 6. On this
assumption, you would reduce the integral constant.
a. In the Integral field, type 0.05.
b. In the Setpoint field, type a low setpoint, such as 70.
c. Click Apply.
d. Allow the input to stabilize at ambient temperature.
e. Change the setpoint to 5° F above ambient temperature and click Apply.
f. Observe your PID loop.

In many applications, the minor fluctuation around the setpoint is acceptable, and
these applications use gain and integral only. In some applications, the fluctuations at
the setpoint indicate that the gain is too high (too much gain makes a system
unstable) or that a derivative constant is required.
At this point, your PID loop should be well tuned using just the gain and integral
constants. You may wish to try setpoint changes with this integral setting but with
different gain values. For example, compare the graph above with the graph below.
The major difference between these is the gain setting. The gain –30 (above) achieved

OptoTutorial: SNAP PAC PID 63

63
ACTIVITY 2: TUNE A PID

a better response to the load change. A gain setting of –20 resulted in a slower
response.

2. Click Save Tuning to save your tuning parameters.

Tune Derivative
Derivative is not necessary for this loop, and it is likely that you will not improve the
loop’s performance with a D tuning constant. But to examine the effect of derivative
constants, do the following:
1. Make sure your PID setpoint is at 5° F above ambient temperature.
(It should be from previous steps.)
2. Set the view the Input Axis to 1 percent of span.
This will allow you to see minor changes affected by the derivative term.
3. Use the Delta-Y Cursor to measure error.
4. Apply D tuning constant.
a. In the Tune D field, type 0.1 and click Apply.
b. Observe the graph for a minute.
5. Try various settings from 0.1 to 10.
For the purpose of examining the derivative term, you may wish to return your Gain
and Integral tuning constants to previous, less-effective settings and then compare

64 OptoTutorial: SNAP PAC PID

 CHAPTER 3: UNDERSTANDING PROPORTIONAL, INTEGRAL, AND DERIVATIVE

these with variations of derivative. For example, a derivative of 10 makes a noticeable

difference in keeping the input near the setpoint.

Many PID systems are effectively controlled with Gain and Integral constants only and no Derivative
constant. In this example, the Gain and Integral settings are maintaining the temperature at 0.07 from
setpoint. To demonstrate the effect of the derivative constant, the resolution of the Input axis was
increased to show a 1 percent span. At this resolution, the plot reveals changes of 0.01 degrees F.

The left side of the plot shows the effect of Gain at -30, Integral at 0.1, and no Derivative constant. The
arrow shows when a Derivative constant of 10 was applied. The right side of the plot shows how the
Derivative constant is keeping the input closer to setpoint. However, the biggest difference is the
spikes in the output. Factors to consider are whether it’s more important to have precise adherence to
setpoint or to have steady output. Spikes in the output can cause mechanical stress on your process or
accelerate device failure times.

Follow-up
This tutorial has provided a basic introduction to the concepts behind PID control and some
hands-on practice using PAC Control’s PID features. With the basic skills taught in this
tutorial, you should be able to explore many other features on your own. Here are a few
areas you may wish to start with:
• Repeat the tuning process with each of the four PID algorithms. To select a different
algorithm, return to Configure mode.
• Try Min. Change and Max. Change settings to simulate output devices that need their
operation restricted.
• Experiment with Feed Forward constants.

OptoTutorial: SNAP PAC PID 65

65
ACTIVITY 2: TUNE A PID

• View the real-time calculations during Debug mode.

• Simulate the effect of cascading loops, where one loop controls the setpoint of
another. You can configure the setpoint to the Learning Center’s potentiometer to
simulate a frequently changing setpoint.

66 OptoTutorial: SNAP PAC PID

 Appendix A

Reference

Configuring a PID Loop

A PID loop is configured in PAC Control’s Configure mode.

• Name: Type a unique, descriptive name for the PID.

• Description: A description of the PID is optional.

SNAP PAC Learning Center User’s Guide 67

67
CONFIGURING A PID LOOP

• Input: The process variable being monitored by the PID loop. Input options are I/O
Point, Host, or PID Output.
– Choose I/O Point if this PID will get its process variable from an I/O point, and then
choose an I/O point from the drop-down list of configured points (or type a point
name to configure a new point on the same I/O unit).
– Choose Host if this PID will get its process variable from PAC Control, for example,
from a PID or I/O point on another I/O unit, or from a process variable that must be
manipulated before use in the PID loop. Type an initial value for the Input.
– Choose PID if this PID is getting its process variable from the output of another PID
on this brain (used in cascading control loops). Choose a PID from the drop-down
list of configured PID loops on this brain.
• Square Root: If you chose I/O Point or PID for step C, you can check this box if you wish
the error to be calculated based on the square root of the process variable. (This is
applicable to flow control systems where volumetric flow is proportional to the square
root of a signal from a flow transducer.)
• Low Range and High Range: Set the valid range of the process variable by typing a
number for the low range and the high range. This range tells the PID what is valid
input and is also used to create a Span term that aligns the input and output scales;
therefore this range affects the tuning of the PID loop. (Options for responding to
out-of-range input are described under Output options for when input is out of range.)
• Setpoint: Choose the source for the setpoint: I/O point, Host, or PID Output.
– Choose I/O Point if you wish to control the setpoint using a device such as a
potentiometer, and then choose an I/O point from the drop-down list of
configured points (or type a point name to configure a new point).
– Choose Host if you wish to control setpoint using PAC Control or PAC Display™, for
example, from a PID or I/O point on another I/O unit, or from a process variable that
must be manipulated before use in the PID loop. Type an initial value for the Input.
– Choose PID Output if this PID is part of a cascading control system in which
another PID loop on the brain will control the setpoint of this PID Loop.
• Output: Choose the destination for the controller output: I/O Point or Host.
– Choose I/O Point to use an output point to correct the process variable (input),
and then choose an I/O point from the drop-down list of configured points (or type
a point name to configure a new point).
– Choose Host to use the output for controlling the setpoint or input to another PID.
The output can be preset or changed at any time through PAC Control by first
switching to Manual mode. For example, if you have determined that the output
should start at 40 percent whenever the system is activated, switch to Manual mode
and then set the PID output to this value. Then switch back to Auto mode.
• Lower Clamp and Upper Clamp: Set the valid range of the output by typing a
number for the Lower and Upper Clamp. This range tells the PID what is valid output

68 Optotutorial: SNAP PAC PID

 APPENDIX A: REFERENCE

and is also used to create a Span term that aligns the input and output scales; therefore
this range affects the tuning of the PID loop.
Setting clamps is particularly important if the device controlled by the output signal
has “dead areas” at either end. For example, say the output is scaled 0-10. It is
connected to a valve that begins to open at 1.25 and is “effectively” fully open at 5.75
(even though it may only be 70 percent open). Set Lower Clamp to 1.2 (valve closed)
and Upper Clamp to 5.75 (valve effectively fully open). This prevents reset windup,
potentially resulting in dramatically improved control when the output value has
reached either limit and has to suddenly reverse direction.
The upper and lower clamp settings prevent the output from exceeding a desirable
range. Minimally, these should correspond to the range of an output point (for
example, the heater is scaled to 0 to 100 percent and so the lower and upper clamps
are 0 and 100).
These settings can also be used to ensure that the output device doesn’t shut off (for
example, keeping a heater or circulation pump running regardless of the PID output) or
that the output never reaches a destructively high setting (for example, keeping a
burner or motor below maximum).
• Minimum Change and Maximum Change: (Optional) The minimum and maximum
change allow you to do two things:
– Minimum Change allows you to prevent an output device from turning on until
there is a minimum desirable workload. For example, you may not want the heater
to turn on to correct a 1 degree error, only to quickly turn off again. A value of 0
disables Minimum change.
– Maximum Change allows you to prevent undesirable effects from too drastic a
change in output. For example, you could use this to moderate the increase in a
pump’s output to prevent pipe breakage. A value of 0 disables Maximum Change.
The Minimum Change and Maximum Change are for each scan period, and therefore
these settings may be a factor in choosing your scan rate.
• Output options for when input is out of range: In the event that the input variable
goes out of range, you can instruct the PID to 1) freeze output at the current value, 2)
switch to Manual mode and stay there until PAC Control logic or an operator
intervenes, 3) set output at a desired minimum or maximum, or both 2 and 3 (set the
output and go to Manual mode).
– For 1, leave these options unchecked.
– For 2 only, check Switch to manual mode when input goes out of range.
– For 3 only, check Force output when input is out of range and type output values for
when the input is below or above the range. (NOTE: This action is independent of 2,
meaning the output can be forced and the PID can be switched to Manual mode;
however, when the PID is being operated in Manual mode, this feature is not
active.)
– For 2 and 3, check both options and type out-of-range values.

Optotutorial: SNAP PAC PID 69

69
CONFIGURING A PID LOOP

• Algorithm: Select a PID algorithm: Velocity B, Velocity C, ISA, Parallel, Interacting.

• Mode: Auto activates the PID. Manual is for debugging and requires you to control the
PID output.
• Scan Rate (Scan Interval): The scan interval is the first and one of the most important
parameters to set. This setting specifies how often (in seconds) the PID loop will scan
the input and calculate an output. The minimum value is 0.001 (1 ms). Factors to
consider are 1) system dead time and 2) tasks on the brain competing for processing
power. Special equipment limitations may require a long scan interval, but generally,
scan interval should be shorter than one-third the duration of the system dead time.
• Gain: Gain enables the controller to implement a proportional response to the error.
The Proportional Gain (P) acts directly on the change in error since the last output
calculation. Type a positive or negative number for gain. Heating systems require a
negative number and cooling systems a positive number. Gain is usually refined during
the tuning process. Higher gain results in increased output change. Too much gain
results in output oscillation. Too little gain results in slow performance.
Depending on the algorithm, gain may also be used as a multiplier on the integral
term, the derivative term, or both.
• Integral: Integral corrects steady-state error common with proportional control. A
positive integral value is required. Integral that is too low results in undershooting the
setpoint. Integral that is too high results in overshoot, oscillation, or both.
By multiplying the integral by the scan time, the PID loop will provide the same integral
action even if the scan interval is changed. For example, if the integral part of the PID
loop moves the output 1 percent every second on a scan interval of 1 second, then if
the scan interval is changed to 0.1 second, it will still result in a change of output of 1
percent every second, but it will be implemented as 0.1 percent every 0.1 second.
• Derivative: The derivative contributes an appropriate value to the output that is
consistent with where the input will be at the next scan if it continues at its current rate
of change.
Derivative is used in feed forward applications and in systems where the loop dead
time is long. Most applications that do not have long dead times and long time
constants are not significantly helped by derivative, and it can be counterproductive.
Most PID loops in industry are run as PI loops.
To disable derivative, set its tuning constant to zero.
• Feed Forward Initial and Feed Forward Gain are constants that are multiplied and
added to the controller output.

70 Optotutorial: SNAP PAC PID

 APPENDIX A: REFERENCE

Features of the PID Inspect Window

The Inspect mode provides tools for tuning PID loops:
• Real-time plots of input, setpoint, and output
• Real-time editing of tuning parameters:
– Input
– Setpoint
– Output
– Gain
– Integral
– Derivative
– Scan rate
– Mode
– Minimum and Maximum output change
– Output high and low values
• Real-time views of the PID algorithm’s calculations

PID Debug Mode—Plot

Typically, tuning a PID requires manipulating the P, I, and D constants while observing the
effects of these settings on a real-time plot of the input, output, and setpoint. With the Plot
view, you can:
• Adjust setpoint and output to observe system dead time. (NOTE: the Setpoint must be
configured to Host in order to freely change its value in Debug mode.)
• Adjust Gain, Integral, and Derivative constants.
• Adjust the view of the Input Axis, Output Axis, and Time Axis.
• Save and print the plot as it appears on screen.
• Display cursors that report precise numerical values.

Optotutorial: SNAP PAC PID 71

71
PID DEBUG MODE—PLOT

Features of the Plot

Option for enabling or
Mode option for activating (Auto) disabling communication
PID tuning and deactivating (Manual) the PID. between the I/O unit and the
Input, Setpoint, and control engine.
Output. These fields can This feature also
be edited when set to reflects whether
Host. Output can also be the strategy is
edited when Mode is set running.
to Manual.

Setpoint and Input plot.

Adjust resolution using
Input Axis menu below
the plot. Click and drag
to move the focus.

Output plot. Adjust

resolution using the
Output Axis menu
below the plot. Click
and drag to move the
focus.

Time axis. Adjust

resolution using the
Time Axis menu below
the plot. Click and drag
left or right to move
the focus.
Any changes to settings must be Save tuning to update the PID configuration with the
applied to take effect. current settings.

Add Watch allows you to select the following PID values for inclusion in a Watch
Window: input, setpoint, output, gain, integral constant, derivative constant,
system interval and mode.

72 Optotutorial: SNAP PAC PID

 APPENDIX A: REFERENCE

Moving a Plot’s Scale

You can click and drag the

Input, Output, and Time
axes.
White arrows show the
direction to drag the scale
and are not part of the
plot.

Displaying Cursors
Choose Cursor under the Data menu to display numeral data. Cursors allow you to drag
vertical or horizontal lines to identify Input, Setpoint, Output, or Time values. Once
displayed you can drag the red line to the desired place on the plot.

The X-Y Cursor displays the

time and value of the
selected Y axis (for example,
Input).
Change the selected Y axis
to Setpoint, Output, or Input
by right-clicking the cursor
and choosing Channel from
the pop-up menu.

Optotutorial: SNAP PAC PID 73

73
PID DEBUG MODE—PLOT

Changing the Cursor Settings

You can right-click the cursor information box to change the cursor.

Selecting the data

represented in the cursor

Selecting the cursor function

Using Delta-X and Delta-Y Cursors

The Delta X cursor displays the difference in time between two vertical red lines. Each red
line can be dragged to a different position, as shown by the arrows (which are not part of
the plot).

74 Optotutorial: SNAP PAC PID

 APPENDIX A: REFERENCE

Resetting Tracking
When you manipulate the time axis of the cursors, the plot remains as a fixed point. To
restore the view of the current values, choose Reset Scale Tracking under the Time Axis
menu.

Optotutorial: SNAP PAC PID 75

75
PID DEBUG MODE—I/O DETAILS

PID Debug Mode—I/O Details

The I/O Details tab displays

PID configuration
information.
Changes to the editable
values are saved to the PID’s
configuration using the Save
Tuning button.
The remaining information
can be changed using
Configure mode.

PID Debug Mode—Misc. Details

For applications requiring

Feed Forward control, the
initial values of Feed Forward
and Feed Forward Gain can be
tuned and saved.
NOTE: For all four PID
algorithms, these two values
are multiplied and then added
to the output; therefore, a
value of 0 in either field results
in no change to the output.
These values can also be
changed using PAC Control’s
Set Feed Forward and Set Feed
Forward Gain commands.

76 Optotutorial: SNAP PAC PID

 APPENDIX A: REFERENCE

Debug PID—IVAL
The IVAL tab shows the difference between the PID XVALs and IVALs:
• The XVAL, or external value, is the “real” or hardware value as seen by the I/O unit. This
value is external to the control engine.
• The IVAL, or internal value, is a logical or software copy of the XVAL that is in the control
engine. The IVAL may or may not be current, since it is updated to match the XVAL only
when a strategy in the control engine reads from or writes to a PID. In the example
below, the Setpoint is configured to Host, so it has an IVAL.

PID values of I/O unit

PAC Control’s representation of

PID values. These won’t have
values until Host is selected
(Input, Setpoint, Output) or PAC
Control invokes PID-related
commands.

This is a view of the PID IVALs while a strategy is running. The strategy reads the Input
value (PAC Control command Get PID Input), which updates all of the PID IVALs.

Optotutorial: SNAP PAC PID 77

77
DEBUG PID—ALGORITHM

Debug PID—Algorithm

During configuration, you

select one of five
algorithms.
The algorithms and
real-time calculations are
shown during Debug
mode.

78 Optotutorial: SNAP PAC PID

 APPENDIX A: REFERENCE

PID Algorithms
OutputHigh and OutputLow are from upper
and lower output clamps.
InputHigh and Inputlow are from input high
range and low range.

Optotutorial: SNAP PAC PID 79

79
DEFINING A NEW CONTROL ENGINE

Defining a New Control Engine

This tutorial assumes that you have completed the SNAP PAC Learning Center lessons and
therefore have created a control engine. A control engine resides on the PC workstation and
is a reference for PAC Control to communicate with the PAC R1 brain. If you are working on a
different computer, you will need to define a new control engine.
1. In the strategy tree, right-click Control Engines and click Add.
2. In the Configure Control Engines dialog box, click Add.
3. In the Select Control Engine dialog box, click Add to open the Control Engine
Configuration dialog box.

4. In the Control Engine Name field, type any name, for example SNAP PAC R
5. Type the IP address of your SNAP PAC, for example [Link]
6. Click OK.
You can now associate your Control Engine.

80 Optotutorial: SNAP PAC PID