Lesson 9

Types of Process Controllers

Lesson Objectives
At the completion of this lesson, you will be able to understand the different types of
process controllers and how they are applied to control an industrial plant process.

Feedback Control Loop

In the process industries, control systems were put to practical use long before the theory
of their operation or methods of analyzing their performance were available. Processes
and control systems were determined by intuition and accumulated experience. This
approach, unscientific as it was, was successful.
Tolerance for the intuitive approach is diminishing today, but it is still a valid way to
obtain knowledge of the basics of the subject, so we’ll start by considering a simple
process and the way a human operator might handle its control.

Figure 9.1 Process Heater Example

Suppose there is a process such as shown in Figure 9.1. A source of liquid flows into the
tank at a varying flow rate. There is a need to heat this liquid to a certain temperature. To
do this there is available hot oil from another part of the plant that flows through coils in
the tank to heat the liquid. By controlling the flow of hot oil we can obtain the desired
temperature. The temperature in the tank is measured and read out on a recording
thermometer mounted within view of the valve on the hot oil line.

6/00 Process Control for Operators 5725013.02 197

Honeywell
Feedback Control Loop Lesson 9 - Types of Process Controllers

The operator has been told to keep the temperature at a DESIRED VALUE, or set point
of 300°. He compares the reading on the thermometer with this mental target and decides
what he must do with the valve to try to bring the temperature of the liquid to the desired
value. In reading the actual value of the temperature and mentally comparing it with the
desired value the operator is providing feedback.
If we draw a block diagram of this system, including the operator which shows the flow
of information between components, it looks like this:

Figure 9.2 Diagram of System with Operator

This is not a very detailed block diagram, but we can see here an interesting feature of
this type of control -information flows in a loop. The operator, to be sure, is a part of the
loop, so it is not automatic control, but it is closed loop control, or feedback control,
because the results are measured, compared to the desired results, and the valve is
manipulated accordingly.

198 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Feedback Control Loop

If we wish to supplant the operator with an automatic system, we will need a device that
will compare the measured variable with the desired value, or SET POINT, and initiate
control action according to what this comparison shows. A control loop with such a
device is shown in Figure 9.3.

Figure 9.3 Diagram of System with Automatic Controller

Every new technical field starts out with a language barrier of sorts. Words perfectly
familiar in one field suddenly take on a new meaning in another. So it becomes necessary
from time to time to undertake the somewhat boring task of establishing definitions.
We’ll start by analyzing the components of the control loop in Figure 9.4, which is a more
hardware-oriented version of Figure 9.3.

Figure 9.4 Feedback Control Loop

6/00 Process Control for Operators 5725013.02 199

Honeywell
Feedback Control Loop Lesson 9 - Types of Process Controllers

1. Error Generator - detects the difference between SET POINT AND PROCESS
VARIABLE. Can be mechanical, pneumatic or electric. In the loop shown the set
point value is entered manually.
(Note: In the actual hardware sometimes the error generator is incorporated as part of
the controller and sometimes it is a separate unit.) The difference between set point
and process variable is called the ERROR or DEVIATION.
2. Controller - basically a logic machine that changes its output in accordance with the
error signal it receives from the error generator. The manner in which it changes its
output is built into it through a system of logic called its modes of control
information the controller needs from the error generator:
a) Polarity (sign) of the error
b) Size of the error
c) Rate of change of the error
3. Final Control Element – the element which directly changes the value of the
manipulated variable.
Examples:
a) Pneumatic diaphragm value
b) Motorized valve
c) Contactor
d) Silicon Controlled Rectifier (SCR)
e) Rheostat
4. Process - the manipulation of raw materials to derive a more valuable product.
5. Primary Element or Sensor – a device that converts the process variable energy
into a measurable form.
(PROCESS VARIABLE = The variable being measured and controlled.)
Examples:
a) bourdon tube
b) orifice plate
c) thermocouple
d) float
e) Filled thermal system

200 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Feedback Control Loop

6. Display (optional) - shows value of process variable and set point, or sometimes set
point and deviation.
7. Transmitter (optional) - changes value of process variable into a standard signal for
transmitting over distances.
The notion of FEEDBACK is crucial to automatic control. Unless the results of the
control manipulations can be compared against objectives, there is no way to arrive at a
logical control strategy. When feedback exists, the system is said to be operating in a
CLOSED LOOP fashion.
(Sometimes, however, control systems operate on information not directly obtained from
the process variable. For example, if the liquid level in a tank were to be controlled it
would be feasible to do so just by adjusting the output flow in proportion to the inlet
flow. There is no closed loop with respect to liquid level: just a relationship between a
flow and a valve position which indirectly, but not infallibly, establishes a level. When a
quantity is being controlled indirectly in this fashion, it is considered to be OPEN LOOP
control.)

6/00 Process Control for Operators 5725013.02 201

Honeywell
Characteristics of Controllers Lesson 9 - Types of Process Controllers

Characteristics of Controllers
Mode of Control
In our section covering the feedback control loop we defined a controller as a logic
machine that changes its output in accordance with the error signal it receives from the
error generator. The change in output from the controller repositions the final control
element, which, finally, adjusts the flow of energy into the process.
The manner in which the energy flow is adjusted in response to the error signal has been
built into the controller through a system of logic called its Mode of Control.

ON-OFF Control
The simplest mode of control is ON-OFF, in which the final control element has but two
positions: fully open (ON) and fully closed (OFF). If the process in Figure 9.5, in which a
liquid is being heated by hot oil flowing through coils, drops below the Set Point of
300°F, ON-OFF control would open the valve all the way. This admits hot oil faster than
is necessary to keep the liquid at 300°F. and as a result the liquid temperature will rise
above 300°F. When it does. ON-OFF control will close the valve all the way and stop the
flow of hot oil. With no oil flowing, the temperature of the liquid will drop, and when it
drops below 300°F the control will open the valve and the cycle will be repeated. This
cycling is an ever present feature of ON-OFF control, because the only two energy input
levels are too much and too little.

Figure 9.5 Process Heater

The appeal of ON-OFF control lies in its simplicity. A vast amount of ON-OFF control is
in use. Its principle limitation is the inevitable cycling it causes in the process.

202 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Characteristics of Controllers

Graphically, ON-OFF control can be illustrated by Figure 9.6. Sometimes, as shown,

there is a neutral zone around the set point. This is often intentional, to keep the
components from wearing due to opening and closing too often.

Figure 9.6 On-Off Control Action

6/00 Process Control for Operators 5725013.02 203

Honeywell
Proportional Control Lesson 9 - Types of Process Controllers

Proportional Control
Since the cycling found in processes using ON-OFF control is due to the rather violent
excursions between All and Nothing, could we not eliminate the cycling by maintaining a
steady flow of hot oil that was just sufficient to hold the temperature of the liquid at
300°F? For each rate of flow of liquid in and out of the tank there must be some ideal
amount of hot oil flow that will accomplish this. This suggests two modifications in our
control mode.
We must:
1. Establish some steady flow value for the hot oil that will tend to hold the temperature
at the set point, and
2. Once this flow value has been established, let any error that develops cause an
increase or decrease in hot oil flow.
This establishes the concept of Proportional Control. Corrective action is now
proportional to the amount of deviation between process variable and set point.
But now we need a different kind of control valve on our process. It must be capable of
being positioned to any degree of opening from fully closed to fully open. This will
generally be either a pneumatic diaphragm or electric motor operated valve.
Graphically, proportional control can be illustrated by Figure 9.7. The amount of control
action (valve change) for a given error can be quite variable, but in Figure 9.7 it is shown
as one to one. The valve would move 1% of its travel for a 1% change in error.

Figure 9.7 Proportional Control Action

204 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Proportional Control

This would be described mathematically as follows:

V = K(E) + M
where
E = error
K = proportional gain
M = a constant which is the position of the valve when the error is zero
The proportional gain, or just GAIN, is a measure of how sensitive the valve change will
be to a given error. Historically, this proportionality between error and valve action has
gone under various names, such as throttling range and gain, but mostly it has been called
Proportional Band, or just PB, which is the expression stating the percent change in error
required to move the valve full travel.
On the basis of the previous definition, look at the graphs, Figure 9.7, Figure 9.8, Figure
9.9, and determine the proportional band in each.

Figure 9.8 Narrow Proportional Band

6/00 Process Control for Operators 5725013.02 205

Honeywell
Proportional Control Lesson 9 - Types of Process Controllers

Figure 9.9 Wide Proportional Band

Proportional Band can be related to Gain as follows:
Gain, K = 100% ÷ PB
Using this relationship, compute the Gains in Figure 9.8 and Figure 9.9.
The more modern way of looking at this mode of control is to think in terms of gain (K),
but in the field you will still hear it called proportional band (PB).
The M factor has to be that valve position which supplies just the right flow of hot oil to
keep the temperature at the set point. It is often called the manual reset term.
A controller designed to provide proportional control must have two adjustments, one for
the K and one for the M (manual reset).
There is, however, a rather serious limitation to proportional control. If there are frequent
load changes to the process, it will hardly ever keep the process variable at the set point.
The reason is that there is only one valve position for each value of the process variable.
But if there are load changes like a change in flow rate of liquid such that more hot oil
than before is needed to maintain the 300°F, this controller has no way of providing it
except through the manual reset adjustment.

206 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Proportional Plus Integral Control

Proportional Plus Integral Control

A proportional controller does not change the position of the valve enough to keep the
process variable at the set point when a load change occurs, as has been explained above.
Integral (sometimes called Reset) action will sense that an error, or offset is present after
proportional action has taken place and continue to change the valve position further in
an attempt to eliminate the error completely. Controllers with integral action will move
the valve at a speed proportional to the size of the error present.
The Proportional plus Integral modes are illustrated in Figure 9.10. Assume a step change
in Set Point at a point in time, as shown. First, there is an immediate change in valve
position equal to K(E) due to the Proportional Mode. At the same time, the Integral
Mode, sensing there is an error, begins to move the valve at a rate proportional to the size
of that error.

Figure 9.10 Proportional Plus Integral Control Action

6/00 Process Control for Operators 5725013.02 207

Honeywell
Proportional Plus Integral Control Lesson 9 - Types of Process Controllers

Since the illustration pictures a constant error, the valve rate will be constant. It will be
seen that after an interval of time. T1, a change in valve position equal to the original
proportional change has taken place. T, is called the Integral Time. An adjustment that is
made to an integral controller determines the slope of the integral response portion of the
graph. The dotted lines show other settings of the integral adjustment. When time is used
to express integral action it is called the Integral Time. This Integral Time is expressed in
Minutes per Repeat, abbreviated MPR. This term refers to the number of “Minutes per
Repeat” that the integral action is repeating the valve change produced by the
proportional control alone. The larger the number the slower the integral action and the
smaller the number the faster the integral action.
Quite commonly you may hear, its reciprocal used, which is called Reset. Reset is
expressed in “Repeats per Minute” abbreviated RPM. This term refers to the number of
times per minute that the reset action is repeating the valve change produced by
proportional control alone.

208 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Proportional Plus Integral Control

Proportional Plus Derivative Control

It seems reasonable that a process with a rapidly changing error would benefit from
additional control action. The derivative (sometimes called rate), mode of control does
just that. The movement of the valve is proportional to the rate of change of error or
process variable. This additional correction exists only while the error is changing. It
disappears when the error stops changing even though the error is not zero.
Graphically, this action is illustrated in Figure 9.11.

Figure 9.11 Proportional Plus Rate Control Action

It can be seen on this graph that the valve position change with derivative action exceeds
that which it would have been with proportional action alone. It can also be seen that on a
ramping error, the valve reaches any given position at an earlier time than it would have
with proportional action alone. This difference in time is the Rate Time, or Derivative
Time, Tp.
Controllers with this mode of control are designed with an adjustment on the amount of
derivative action. This adjustment is made in terms of rate time. Longer rate times
increase the amount of rate action.

6/00 Process Control for Operators 5725013.02 209

Honeywell
Proportional Plus Integral Control Lesson 9 - Types of Process Controllers

It is also of interest to consider derivative action in response to a step change in error, as

is illustrated in Figure 9.12.

Figure 9.12 Rate Action Response to a Step Change

Changes in rate time here show up in the length of time for the decay of the valve
position to the position it would have assumed with proportional action alone.

210 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Proportional Plus Integral Control

Proportional Plus Integral Plus Derivative

Finally, the full three-mode controller is achieved by combining the three modes
simultaneously, Thus the valve position will be determined by adding the effects of the
three modes.
The graph in Figure 9.14 illustrates how an increase in the flow of the liquid to be heated
will be responded to by the various control modes in terms of the process variable, and
liquid output temperature.
Graphically, its response to a step change in error is:

Figure 9.13 Three Mode Control Action

Figure 9.14 Response to Increase in Liquid Flow

6/00 Process Control for Operators 5725013.02 211

Honeywell
Proportional Plus Integral Control Lesson 9 - Types of Process Controllers

Controller Selection
The selection of the proper modes of control to use on a particular process is a complex
matter, but there are a few general statements that can be made for guidance.

ON-OFF
On-Off is popular because of its simplicity. In general it functions satisfactorily if the
process has a large capacitance and minimum dead time. It will accommodate load
changes to some extent, but such changes should not be rapid or large. Cycling at the new
load will have a different average value depending on the direction of the load change.
In industry, ON-OFF control is ideally suited, for example, to the control of temperature
in a cooking kettle where the only load changes are due to changes in ambient
temperature. The capacitance is large and the load changes are small.

Proportional
Proportional control reduces cycling below that of ON-OFF control. It does a particularly
good job when process capacitance is large and dead time small. These characteristics
promote stability and allow the use of a narrow proportional band, which gives faster
corrective action and less offset.
When the process has these favorable characteristics, proportional control can even make
moderate load changes tolerable. When the proportional band must be made wider,
however, even a small load change leads to offset.

Proportional Plus Integral (P+I)

The primary advantage of proportional plus integral is that it will eliminate offset with
load changes. It can be used even when process capacitance is small and load changes are
large. The main limitation of a proportional-plus-integral controller is its inability to
prevent overshoots due to integral accumulation. When integral action responds to a large
enough error, or one that exists for a long time, by putting the valve into saturation (either
fully open or fully closed), it is subsequently unable to change the direction of the valve
motion until the error changes sign, that is, until the process variable crosses the set point.
This is usually too late to prevent overshooting of the process variable. It is a problem
found particularly in the start up of processes, but any large or rapid load change may
cause it.

212 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Proportional Plus Integral Control

Proportional Plus Integral Plus Derivative (PID)

Derivative action can be very useful in minimizing overshooting of the process variable
when the controller is trying to compensate for large or rapid load changes. It is also
useful in preventing overshoot of the process variable in the start-up of batch processes.
On very slow moving processes derivative will have minimal affect. On noisy processes,
such as flow, derivative will amplify the noise and result in continual over correction. It
has been most widely used for temperature control, and least on pressure or flow
applications. In recent years, however, its use has been more widespread across all
control applications.
Each mode of control is applicable to processes having certain combinations of
characteristics. The simplest mode of control that will do the job is the best to use. Table
9.1 summarizes the guidelines for selection of control modes from various combinations
of process characteristics.
Table 9.1 Control Mode Selection Guidelines
Mode of Control Process Reaction Dead Time Load Changes
Rate
On-Off Slow Slight Small and slow
Proportional Slow or moderate Small or moderate Small, infrequent
Proportional + Integral Slow or moderate Moderate Small, faster
Proportional + Derivative Fast Small or moderate Slow, but any size,
frequent
Proportional + Integral + Fast Fast Fast
Derivative

Controller Tuning
Correct controller tuning is essential to good control. Tuning a controller means finding
the optimum settings for gain, integral, and derivative. Good control may have different
meanings for different processes.
At the outset we should realize that no corrective action is ever possible unless some
deviation or error exists. The feedback principle establishes this generalization. Since it is
not possible to achieve zero deviation control, can we state how closely we have
approached it? Many criteria have been advanced in an attempt to judge control
performance. It is impossible to say that any one of them is the absolute standard. Each
process will have its particular requirements.

6/00 Process Control for Operators 5725013.02 213

Honeywell
Proportional Plus Integral Control Lesson 9 - Types of Process Controllers

Essentially, however, most criteria strive to arrive at a compromise with three basic
goals:
• Minimum deviation following a disturbance.
• Minimum time interval before return to the set point.
• Minimum offset due to changes in operating conditions.
There are also diverse methods for finding the controller settings that will satisfy these
goals. Some are more precise than others, but all result in approximations that provide a
good starting point from which fine-tuning can be done. We will consider here two
different methods. The two require different amounts of information about the process.
As you would suspect, the more process information that can be incorporated, the more
precise will be the results. Both of these methods are based on conducting a few simple
tests in the field. They can be validated against theory.

214 Process Control for Operators 5725013.02 6/00

Honeywell
Lesson 9 - Types of Process Controllers Lesson Summary

Lesson Summary
In this lesson you have learned the different types of basic Process Controllers found in
control systems.
In the next lesson you will learn the different types of Multi-Loop Controllers and how
they are applied.

6/00 Process Control for Operators 5725013.02 215

Honeywell
Lesson Summary Lesson 9 - Types of Process Controllers

216 Process Control for Operators 5725013.02 6/00

Honeywell