ISA-TR75 25 02 - 2000 - Control valve response measurement from step inputs

Prévia do material em texto

Inputs
ISA–TR75.25.02–2000
T E C H N I C A L R E P O R T
ISA The Instrumentation,
Systems, and
Automation Society 
–
TM
Control Valve Response
Measurement from Step
Approved 30 December 2000
Copyright  2000 by ISA–The Instrumentation, Systems, and Automation Society. All rights 
reserved. Not for resale. Printed in the United States of America. No part of this publication may be 
reproduced, stored in a retrieval system, or transmitted, in any form or by any means (electronic, 
mechanical, photocopying, recording, or otherwise), without the prior written permission of the 
Publisher.
ISA
67 Alexander Drive
P.O. Box 12277
Research Triangle Park, North Carolina 27709
USA
ISA–TR75.25.02–2000
Control Valve Response Measurement from Step Inputs
ISBN: 1-55617-743-7
— 3 — ISA–TR75.25.02–2000
Preface
This preface, as well as all footnotes and annexes, is included for information purposes and is not part of 
ISA–TR75.25.02–2000.
The standards referenced within this document may contain provisions which, through reference in this 
text, constitute requirements of this document. At the time of publication, the editions indicated were valid. 
All standards are subject to revision, and parties to agreements based on this document are encouraged to 
investigate the possibility of applying the most recent editions of the standards indicated within this 
document. Members of IEC and ISO maintain registers of currently valid International Standards. ANSI 
maintains registers of currently valid U.S. National Standards. 
This document has been prepared as part of the service of ISA–The Instrumentation, Systems, and 
Automation Society, toward a goal of uniformity in the field of instrumentation. To be of real value, this 
document should not be static but should be subject to periodic review. Toward this end, the Society 
welcomes all comments and criticisms and asks that they be addressed to the Secretary, Standards and 
Practices Board; ISA; 67 Alexander Drive; P. O. Box 12277; Research Triangle Park, NC 27709; 
Telephone (919) 549-8411; Fax (919) 549-8288; E-mail: standards@isa.org.
The ISA Standards and Practices Department is aware of the growing need for attention to the metric 
system of units in general, and the International System of Units (SI) in particular, in the preparation of 
instrumentation standards. The Department is further aware of the benefits to USA users of ISA standards 
of incorporating suitable references to the SI (and the metric system) in their business and professional 
dealings with other countries. Toward this end, this Department will endeavor to introduce SI-acceptable 
metric units in all new and revised standards, recommended practices, and technical reports to the 
greatest extent possible. Standard for Use of the International System of Units (SI): The Modern Metric 
System, published by the American Society for Testing & Materials as IEEE/ASTM SI 10-97, and future 
revisions, will be the reference guide for definitions, symbols, abbreviations, and conversion factors.
It is the policy of ISA to encourage and welcome the participation of all concerned individuals and interests 
in the development of ISA standards, recommended practices, and technical reports. Participation in the 
ISA standards-making process by an individual in no way constitutes endorsement by the employer of that 
individual, of ISA, or of any of the standards, recommended practices, and technical reports that ISA 
develops.
CAUTION — ISA ADHERES TO THE POLICY OF THE AMERICAN NATIONAL STANDARDS 
INSTITUTE WITH REGARD TO PATENTS. IF ISA IS INFORMED OF AN EXISTING PATENT THAT IS 
REQUIRED FOR USE OF THE STANDARD, IT WILL REQUIRE THE OWNER OF THE PATENT TO 
EITHER GRANT A ROYALTY-FREE LICENSE FOR USE OF THE PATENT BY USERS COMPLYING 
WITH THE STANDARD OR A LICENSE ON REASONABLE TERMS AND CONDITIONS THAT ARE 
FREE FROM UNFAIR DISCRIMINATION. 
EVEN IF ISA IS UNAWARE OF ANY PATENT COVERING THIS STANDARD, THE USER IS 
CAUTIONED THAT IMPLEMENTATION OF THE STANDARD MAY REQUIRE USE OF TECHNIQUES, 
PROCESSES, OR MATERIALS COVERED BY PATENT RIGHTS. ISA TAKES NO POSITION ON THE 
EXISTENCE OR VALIDITY OF ANY PATENT RIGHTS THAT MAY BE INVOLVED IN IMPLEMENTING 
THE STANDARD. ISA IS NOT RESPONSIBLE FOR IDENTIFYING ALL PATENTS THAT MAY 
REQUIRE A LICENSE BEFORE IMPLEMENTATION OF THE STANDARD OR FOR INVESTIGATING 
THE VALIDITY OR SCOPE OF ANY PATENTS BROUGHT TO ITS ATTENTION. THE USER SHOULD 
CAREFULLY INVESTIGATE RELEVANT PATENTS BEFORE USING THE STANDARD FOR THE 
USER’S INTENDED APPLICATION.
ISA–TR75.25.02–2000 — 4 —
HOWEVER, ISA ASKS THAT ANYONE REVIEWING THIS STANDARD WHO IS AWARE OF ANY 
PATENTS THAT MAY IMPACT IMPLEMENTATION OF THE STANDARD NOTIFY THE ISA 
STANDARDS AND PRACTICES DEPARTMENT OF THE PATENT AND ITS OWNER.
ADDITIONALLY, THE USE OF THIS STANDARD MAY INVOLVE HAZARDOUS MATERIALS, 
OPERATIONS OR EQUIPMENT. THE STANDARD CANNOT ANTICIPATE ALL POSSIBLE 
APPLICATIONS OR ADDRESS ALL POSSIBLE SAFETY ISSUES ASSOCIATED WITH USE IN 
HAZARDOUS CONDITIONS. THE USER OF THIS STANDARD MUST EXERCISE SOUND 
PROFESSIONAL JUDGMENT CONCERNING ITS USE AND APPLICABILITY UNDER THE USER’S 
PARTICULAR CIRCUMSTANCES. THE USER MUST ALSO CONSIDER THE APPLICABILITY OF ANY 
GOVERNMENTAL REGULATORY LIMITATIONS AND ESTABLISHED SAFETY AND HEALTH 
PRACTICES BEFORE IMPLEMENTING THIS STANDARD.
The following people served as members of ISA Subcommittee SP75.25:
NAME COMPANY
C. Langford, Chairman Cullen G. Langford, Inc.
W. Weidman, Managing Director Parsons Energy & Chemicals Group
J. Beall Eastman Chemical Co.
D. Bennett Sampson Controls, Inc.
W. Bialkowski Entech Control Engineering, Inc.
W. Black Cashco, Inc.
M. Boudreaux Exxon Mobil Chemical
S. Boyle Neles Automation
D. Buchanan Union Carbide Corporation
F. Cain Flowserve Corporation
N. Cammy UOP LLC
J. Jamison Bantrel Inc.
S. Kempf Harold Beck & Sons, Inc.
P. Maurath Procter & Gamble Company
R. McEver Bettis Corporation
N. McLeod Elf Atochem
G. McMillan Solutia, Inc.
J. Reed Norriseal
The following people served as members of ISA Committee SP75:
NAME COMPANY
D. Buchanan, Chairman Union Carbide Corporation
W. Weidman, Managing Director Parsons Energy & Chemicals Group
A. Abromaitis Red Valve Company, Inc.
H. Backinger J. F. Kraus & Company
G. Barb Retired
H. Baumann H B Services Partners LLC
H. Boger Masoneilan Dresser
G. Borden Consultant
S. Boyle Neles Automation
R. Brodin Fisher Controls International, Inc.
F. Cain Flowserve Corporation
C. Corson Fluor Daniel Inc.
M. Coughran Fisher Controls
A. Engels Praxair, Inc.
— 5 — ISA–TR75.25.02–2000
H. Fuller Valvcon Corporation
J. George Richards Industries
A. Glenn Flowserve Corp.
L. Griffith Consultant/Retired
B. Guinon Shell Chemical
F. Harthun Retired
B. Hatton DeZurik Division Unit
J. Jamison Bantrel, Inc.
R. Jeanes TXU Electric
J. Kersh M. W. Kellogg Company
C. Langford Cullen G. Langford, Inc.
A. Libke DeZurik Valve Company
R. Louviere Creole Engineering Sales Company
O. Lovett Consultant/Retired
A. McCauley Chagrin Valley Controls, Inc.
R. McEver Bettis Corporation
H. Miller Control Components, Inc.
T. Molloy CMES Inc.
L. Ormanoski Frick Company
J. Ozol Commonwealth Edison
W. Rahmeyer Utah State University
J. Reed Norriseal
K. Schoonover Con-Tek Valves, Inc.
A. Shea Copes-Vulcan, Inc.
E. Skovgaard Leslie Controls, Inc. 
H. Sonderegger Tyco Flow Control
R. Terhune Retired
This standard was approved for publication by the ISA Standards and Practices Board on 
30 December 2000:
NAME COMPANY
M. Zielinski, Vice President Fisher-Rosemount Systems, Inc.
D. Bishop Consultant
P. Brett Honeywell, Inc.
M. Cohen Senior Flexonics, Inc.
M. Coppler Ametek, Inc.
B. Dumortier Schneider Electric 
W. Holland Southern Company
A. Iverson Ivy Optiks
R. Jones Dow Chemical Co.
V. Maggioli Feltronics Corp.
T. McAvinew BatemanEngineering, Inc.
A. McCauley, Jr. Chagrin Valley Controls, Inc.
G. McFarland Westinghouse Process Control Inc.
D. Rapley Rapley Consulting Inc.
R. Reimer Rockwell Automation
J. Rennie Factory Mutual Research Corp.
H. Sasajima Yamatake Corp.
R. Webb Altran Corp.
W. Weidman Parsons Energy & Chemicals Group
J. Weiss EPRI
ISA–TR75.25.02–2000 — 6 —
J. Whetstone National Institute of Standards & Technology
M. Widmeyer EG&G Defense Materials
R. Wiegle CANUS Corp.
C. Williams Eastman Kodak Co.
G. Wood Graeme Wood Consulting
— 7 — ISA–TR75.25.02–2000
Contents
1 Purpose........................................................................................................................................... 9
2 Scope.............................................................................................................................................. 9
3 Definitions ....................................................................................................................................... 9
4 Control valve response ................................................................................................................. 14
4.1 Measurement of control valve response ............................................................................... 14
4.2 System response .................................................................................................................. 14
4.3 Test environments used to determine control valve response.............................................. 16
4.4 Size of input signal change – regions ................................................................................... 17
5 Maintenance and design issues affecting process control............................................................ 17
5.1 Stem seal .............................................................................................................................. 18
5.2 Valve seat shutoff.................................................................................................................. 18
5.3 Valve seat type...................................................................................................................... 19
5.4 Process fluid effects .............................................................................................................. 19
5.5 Mechanical tolerances .......................................................................................................... 19
5.6 Structural stiffness................................................................................................................. 19
5.7 Pneumatic positioner............................................................................................................. 19
5.8 Actuator size/type.................................................................................................................. 20
5.9 Electric and hydraulic actuators ............................................................................................ 20
5.10 Flow effects ......................................................................................................................... 20
5.11 Valve sizing and selection................................................................................................... 20
6 Process and control design issues................................................................................................ 20
6.1 Control loop process gain – range and variability ................................................................. 20
6.2 Over-sizing ............................................................................................................................ 22
6.3 Control valve inherent characteristic ..................................................................................... 22
6.4 Closed loop performance – control valve dynamic specification........................................... 23
6.5 Nonlinear regions .................................................................................................................. 24
7 Static behavior tests...................................................................................................................... 25
7.1 Important measures of static behavior .................................................................................. 25
7.2 Applications affected and classes of performance................................................................ 25
7.3 Testing considerations .......................................................................................................... 26
7.4 Data presentation.................................................................................................................. 28
7.5 Design and maintenance factors important to static behavior .............................................. 29
ISA–TR75.25.02–2000 — 8 —
8 Small amplitude and medium amplitude dynamic response tests (regions 2 and 3) .................... 29
8.1 Important measures for regions 2 and 3 ............................................................................... 30
8.2 Applications affected and classes of performance................................................................ 32
8.3 Testing considerations for small amplitude and medium amplitude dynamic response 
 (regions 2 and 3) ................................................................................................................... 32
8.4 Data presentation for regions 2 and 3................................................................................... 33
8.5 Design and maintenance factors at small amplitude and medium amplitude 
 (regions 2 and 3) ................................................................................................................... 33
8.6 Oscillatory response.............................................................................................................. 34
8.7 Performance near the closed position................................................................................... 34
9 Large amplitude dynamic response tests (region 4) ..................................................................... 35
9.1 Important measures in region 4 ............................................................................................ 35
9.2 Applications affected and classes of performance................................................................ 36
9.3 Testing considerations for large amplitude dynamic response ............................................. 36
9.4 Design and maintenance factors important at large amplitude............................................. 37
10 References.................................................................................................................................... 37
— 9 — ISA–TR75.25.02–2000
1 Purpose
This technical report describes the characteristic response of a control valve to step input signal changes. 
It considers the factors that affect this response, the impact of the response on the quality of process 
control, and the appropriate control valve specifications. In this document, a control valve is the complete 
control valve body, with actuator and any accessories required for normal operation assembled and ready 
for use. This document supports standard ANSI/ISA-75.25.01-2000, "Test Procedure for Control Valve 
Response Measurement from Step Inputs." See the standard for the test procedures.
Users and manufacturers have developed a better understanding of the effects of control valve response 
characteristics on process control. This document identifies and defines four regions of control valve 
response to step input changes of varying sizes. Existing standards do not include the definitions and 
methods to measure certain valvecharacteristics now understood to be important. This technical report 
provides guidance that can be used to relate the control valve performance to process control.
2 Scope
This technical report applies to throttling control valves in closed loop control applications. The concept 
has some application to open loop control applications. It does not address control valves used in on-off 
control service. The “control valve” in the context of this document includes the following components:
Valve: A valve is a device used for the control of fluid flow. It consists of a fluid containing valve body 
assembly, one or more ports between connection openings and a moveable closure member, which 
opens, restricts or closes the port(s) (see ANSI/ISA-75.05.01-2000, "Control Valve Terminology").
Actuator: An actuator is a device that supplies the force and causes the movement of the valve closure 
member. Commonly these are fluid or electrically powered (see ANSI/ISA-75.05.01-2000). Actuators often 
use air but other types use electric, hydraulic and electro-hydraulic power. 
Motion conversion mechanism: A mechanism installed between the valve and the power unit of the 
actuator to convert between linear and rotary motion where required. The conversion may be from linear 
actuator action to rotary valve operation or from rotary actuator action to linear valve operation.
Accessories: Additional devices used in the operation of the control valve. As described in 
ANSI/ISA-75.05.01-2000, typical examples include a positioner, transducer, signal booster relay, air set, 
snubber, etc.
3 Definitions
This document and ANSI/ISA-75.25.01-2000 make use of terms as defined in ANSI/ISA-51.1-1979 
(R1993) "Process Instrumentation Terminology", and some of the essential terms are repeated here for 
convenience. In the specific area of nonlinear dynamics, it was determined that some terms defined in 
ANSI/ISA-51.1-1979 (R1993) lacked the precision desired for these documents. Others were inconsistent 
with the terminology used in the nonlinear control literature. A common set of definitions is used in 
ANSI/ISA-75.25.01-2000 and this document. Those used only in this document are marked with an 
asterisk (*).
3.1 backlash:*
in process instrumentation, a relative movement between connected mechanical parts, resulting from 
looseness when motion is reversed [ANSI/ISA-51.1-1979 (R1993)]. Sometimes also referred to as slop, 
lost motion, or free play.
ISA–TR75.25.02–2000 — 10 —
Figure 1 — Dead band and resolution
3.2 closed loop time constant:*
the time constant of the closed loop response of a control loop, used in tuning methods such as Internal 
Model Control (IMC) and Lambda Tuning. The closed loop time constant is a measure of the performance 
of a control loop.
3.3 dead band:
the range through which an input signal may be varied, with reversal of direction, without initiating an 
observable change in output signal [ANSI/ISA-S51.1-1979 (R1993)]. In this technical report and in 
standard ANSI/ISA-75.25.01-2000 it is defined in percent of input span. Note that in some other literature 
this definition is used for dead zone. 
3.4 dead time (Td): 
the time after the initiation of an input change and before the start of the resulting observable response. 
3.5 dead zone:*
a zone of input for which no value of the output exists [ANSI/ISA-S51.1-1979 (R1993)].
3.6 dynamic response:
the time-dependent output signal change resulting from a defined time-dependent input signal change. 
Commonly used input signal changes include impulse, pulse, step, ramp, and sinusoid [McGraw-Hill 
"Dictionary of Scientific and Technical Terms", fifth edition, 1994]. Dynamic means that the control valve is 
Dynamics are not shown
b
a c d
Input
TimeA
m
pl
itu
de
Output
c ≤ dead band < d
a < resolution ≤ b
— 11 — ISA–TR75.25.02–2000
moving. Dynamic response can be measured without process loading in bench top tests with simulated or 
active loading in a flow laboratory or under normal process operating conditions. 
3.7 gain ratio GZ
 
/ GZ02:
the response gain GZ divided by the response gain GZ02 determined from the multi-step test performed with 
a step size of 2 percent. The ideal gain ratio equals 1.0 for tests about any nominal position.
GR = GZ/GZ02
3.8 hunting:*
an undesirable oscillation of appreciable magnitude, prolonged after external stimuli disappears 
[ANSI/ISA-S51.1-1979 (R1993)]. Hunting can have two forms: oscillations occurring near the stability limit 
of a linear system or the limit cycling tendency of a nonlinear system.
3.9 limit cycle:
an oscillation caused by the nonlinear behavior of a feedback system. These oscillations are of fixed 
amplitude and frequency, and can be sustained in a feedback loop even if the system input change is zero. 
In linear systems, an unstable oscillation grows theoretically to infinite amplitude, but nonlinear effects limit 
this growth [Van De Vegte, J., "Feedback Control Systems", 2nd ed, Prentice Hall, 1990, p14]. See also 
hunting [ANSI/ISA-S51.1-1979 (R1993)].
3.10 memory:*
in the context of small signal nonlinear dynamics, is that property of a nonlinearity which makes it sensitive 
to the current direction, and the history of the input signal. Memory requires the inclusion of direction 
arrows on those line segments of an X-Y plot that are directionally sensitive. 
3.11 nonlinear system:*
a nonlinear system is one whose response depends on the amplitude and the nature of the input signal, as 
well as the initial conditions of the system. As an example, a nonlinear system can change from being 
stable to unstable by changing the size of the input signal. 
When a nonlinear system is driven towards a setpoint by feed back control action, it is likely to develop a 
limit cycle. The amplitude and frequency of such limit cycles are a function of the nature of the 
nonlinearities which are present, and the effective gain of the feed back control action. As the gain of the 
feed back is increased, the frequency of the limit cycle is likely to increase. More aggressive gain 
increases may produce behavior such as bifurcation, frequency doubling and eventually chaotic behavior.
3.12 nonlinearity:*
there are many types of nonlinearities, although they can be generally grouped into two main groups: 
simple nonlinearities without memory and more complex nonlinearities with memory [Van De Vegte, 
above, Gibson, J. E. "Nonlinear Automatic Control" McGraw-Hill 1963]. Not the same as in ANSI/ISA-
S51.1-1979 (R1993), linearity: the closeness to which a curve approximates a straight line.
3.13 overshoot:
the amount by which a step response exceeds its final steady state value. Refer to figure 24 of 
ANSI/ISA-S51.1-1979 (R1993). Usually expressed as a percentage of the full change in steady state 
value.
3.14 position Z:
the position of the closure member relative to the seated position. In this technical report and in standard 
ANSI/ISA-75.25.01-2000 expressed as a percent of span.
ISA–TR75.25.02–2000 — 12 —
3.15 resolution:
smallest step increment of input signal in one direction for which movement of the output is observed. 
Resolution is expressed as percentage of input span. The term in this document means: the tendency of a 
control valve to move in finite steps in responding to step changes in input signal applied in the same 
direction. This happens when the control valve sticks in place, having stopped moving after the previous 
step change.
3.16 response:
the time history of a variable after a step change in the input. In this technical report and in standard 
ANSI/ISA-75.25.01-2000, the step response can be stem position, flow, or another process variable.
3.17 response flow coefficient CvR:*
apparent flow coefficient as determined by testing in an operating type environment.The data available in 
the operating environment may differ from the laboratory data required by valve sizing standards.
3.18 response gain GZ:
the ratio of the steady state magnitude of the process change ∆Z divided by the signal step ∆s that caused 
the change. One special reference response gain is defined as that calculated from the 2 percent step 
size response time test. This is designated as GZ02. 
GZ = DZ / Ds
GZ02 = DZ02 / Ds02
3.19 sampling interval Dts:
the time increment between sampled data points. It is the inverse of the sampling rate, f0. ∆tS = 1/f0. As 
used in this technical report and ANSI/ISA-75.25.01-2000, since more than one variable is being sampled, 
it is the time between the sets of sampled data. Ideally, all variables in one set are sampled at the same 
time. If data is recorded using analog equipment, the time constant for the recording equipment shall be 
less than or equal to the maximum allowed ∆t.
3.20 sampling rate f0:
the rate at which data samples are taken or the number of samples per unit time. See sampling interval.
3.21 shaft windup:*
in rotary valve systems, the tendency of the drive shaft to twist under load while the closure member is 
stuck at a given position. 
3.22 sliding friction Fr or Tr:
the force or torque required to maintain motion in either direction at a prescribed input signal ramp rate.
3.23 static:
means without motion or change [McGraw-Hill "Dictionary of Scientific and Technical Terms", fifth edition, 
1994]; readings are recorded after the device has come to rest. Static performance can be measured 
either without process loading (bench top tests), with simulated or active loading, or under process 
operating conditions. This kind of test is sometimes called a dynamic test [McGraw-Hill above], which may 
cause confusion. The static behavior characteristics identified as important to the control valve 
performance are the dead band, the resolution, and the valve travel gain.
— 13 — ISA–TR75.25.02–2000
3.24 steady state:
a condition of a dynamic system when it is at rest at a given value. In testing the responses of a dynamic 
system, step test methods are often used. The resulting system transitions from an initial steady state 
value to a new final steady state value.
3.25 step change:
a nearly instantaneous step change made to an input signal of a dynamic system with the intention of 
stimulating a step response of the dynamic system. Such a test is used to characterize the step response 
of the dynamic system. 
3.26 step change time ∆tsc:
the time between the start of a signal input step and when it reaches its maximum value.
3.27 step test:
the application of a step change to an input signal in order to test the step response dynamics.
3.28 step response time (T86):
the interval of time between initiation of an input signal step change and the moment that the response of a 
dynamic reaches 86.5% of its full steady state value. The step response time includes the dead time 
before the dynamic response.
3.29 step size ∆s:
the difference between the beginning and ending signal in a step change expressed as a percent of the 
signal span.
3.30 stiction (static friction):*
resistance to the start of motion, usually measured as the difference between the driving values required to 
overcome static friction upscale and downscale [ANSI/ISA-S51.1-1979 (R1993)].
3.31 stick/slip:*
a term that attempts to explain jerky or “sticky” motion by postulating that static friction differs substantially 
from sliding friction. However, friction is rarely directly measured, and “sticky” behavior can be caused by 
other physical effects (e.g., positioner behavior, at small amplitudes).
3.32 stick/slip cycle:*
a term that attempts to describe a limit cycle caused when the control valve “sticks” and suddenly “slips” 
during a change in input signal. It is the result of static friction combined with a positioner and actuator 
system that does not provide enough force to overcome friction at low positioner error values. 
3.33 time constant τ:
for first order dynamic systems, the interval of time between initiation of an input signal step change and 
the moment that a first order dynamic system reaches 63.2% of the full steady state change. The term is 
used in this technical report and ANSI/ISA-75.25.01-2000 to describe the dynamic characteristics of the 
analog measuring instruments.
3.34 valve travel gain:*
the change in closure member position divided by the change in input signal, both expressed in 
percentage of full span.
GX = DX / Ds
ISA–TR75.25.02–2000 — 14 —
3.35 valve system approximate time constant (τ’):*
the time constant of a first order response without dead time, which may fit the actual control valve step 
response reasonably well. The approximate time constant is defined to provide a basis for comparison of 
the valve with other time constants, such as the closed loop time constant for the control loop. A first order 
system reaches 86.5% of its final step response value in two time constants; the approximate time 
constant is considered to be one half of the step response time, T86. The use of the approximate time 
constant in no way implies that the response of the control valve is first order. The step response of the 
control valve is typically complex, having dead time initially, followed by potentially complex dynamics 
before the steady state is achieved. T86 includes the dead time in the initial part of the response, as well 
as the possibility of slower settling in the last portion of the response. Some valve positioner designs 
attempt to achieve a slow-down in the final part of the response in order to limit overshoot. τ’ attempts to 
produce a simple linear time constant approximation of the control valve dynamic response, which can be 
compared to the closed loop time constant of the control loop on the same basis in time constant units. 
Note that as the portion of T86 that is dead time increases, this approximation becomes less ideal.
3.36 velocity limiting:*
the maximum rate of change that a system can achieve due to its inherent physical limitations.
3.37 wait time ∆tw:*
the time spent after a step input change waiting for the response to come to the new steady state value. 
3.38 X-Y plot:*
a plot of the output excursions plotted against input excursions. Input-output plots are useful for defining 
the steady state characteristics of nonlinearities.
4 Control valve response
4.1 Measurement of control valve response
The control valve is an integral part of the process control loop. The other parts are the sensing devices, 
the controller, and the process under control. A certain level of quality of process control performance is 
required for every application. To achieve this, the control valve must have the appropriate response 
characteristics.
The control valve modifies a fluid flow in response to a signal from a process controller. This is 
accomplished by moving the closure member resulting in a change in the flow coefficient. This response 
has both static and dynamic behavior characteristics. The static behavior characteristics identified as 
important to the control valve performance are the dead band, the resolution, and the valve travel. The 
dynamic response characteristics of interest are the dead time, overshoot and the step response time. 
Control valve performance may be improved by reducing the dead band, improving the resolution, 
stabilizing the travel gain, and reducing both the dead time and the step response times and by minimizing 
overshoot. The effect of the dynamic properties is a function of the process time constants. The effect of 
the static properties is a function of the process static gain properties. 
4.2 System response
Several parameters can be used to describe the response of the control valve.These parameters are the 
flow coefficient, the value of the measured variable, and the stem position. They are discussed below in 
detail. 
— 15 — ISA–TR75.25.02–2000
4.2.1 Flow coefficient 
The fundamental parameter is the valve flow coefficient. The flow coefficient, or Cv, is calculated from the 
flow through the valve, the fluid density and the differential pressure across the valve [see ISA-75.01-1985 
(R1995), “Flow Equations for Sizing Control Valves” and ANSI/ISA-75.02-1996, “Control Valve Capacity 
Test Procedure”]. For in-process testing, a response flow coefficient CvR, is defined in this technical 
report. The response flow coefficient CvR is calculated from the flow, the differential pressure and the 
density. It is defined this way to allow the use of the available data and test environment where this differs 
from the standard definition of Cv. This coefficient may differ from catalog values. Changes in the closure 
member position will change the flow coefficient even under choked flow conditions. The response flow 
coefficient will test for the following uncertainties:
a) The control system output signal may not be the exact stem position. 
b) The stem position may not be the exact closure member position. 
c) The relation between the closure member position and the flow coefficient is not certain. 
4.2.2 Process response 
The process response provides information similar to that provided by using the flow coefficient. The data 
is the change in temperature or in composition or in flow rate. It may be distorted by any errors in the 
process measurement, changes in differential pressure across the valve, noise, and changes in the fluid 
density. The process measurement instruments may not provide the desired resolution or accuracy. Any 
errors and all the dynamics of the transmitter and the process can distort the data. Consider these 
uncertainties in reporting the response of the control valve.
If the process installation includes a flowmeter, or if it is a relatively “fast”, self-regulating process, then this 
task is made easier. If the process is not a relatively “fast”, self-regulating process, it may be difficult to use 
the process response to determine the performance of the control valve. For instance, for an integrating 
process (such as a level or pressure), the fluid flow rate can be calculated from the rate of change or from 
differences over time in the process variable. In some applications, the process information may provide 
only a limited set of performance parameters such as dead band and resolution. 
4.2.3 Stem position
Stem position measures the response of the physical parts of the control valve. Some control valve 
accessory systems include the ability to measure and report control valve stem position and other 
parameters. This method can eliminate the uncertainties of the dynamic response parameters of the 
process and process measuring instruments. It is necessary to correct for the response of the stem 
measurement instrument. The stem position measurement is well suited for determining the dynamic 
response characteristics (dead time, step response time, overshoot) of the control valve. It does not 
account for the differences between the stem movement and the closure member movement or between 
the closure member movement and the flow coefficient. This makes it less useful for measuring dead 
band and resolution. These differences can be caused by such things as backlash in the actuator system, 
or stem1 windup on rotary valves. However, if the correlation between the stem position and the flow 
coefficient is known for a particular valve or valve design, it can be combined with the response of the stem 
position to predict the response of the flow coefficient. 
______ 
1
 Stem position refers to rising stem valves. Shaft position refers to rotary valves. This document uses stem to mean either stem or 
shaft.
ISA–TR75.25.02–2000 — 16 —
4.3 Test environments used to determine control valve response
The control valve can be tested in three different environments. Each of these environments has tradeoffs 
in practicality and measurement uncertainties. The test environment will determine which parameter (flow 
coefficient, process response or stem position) can be used to measure the response characteristics of the 
control valve. The test environments are discussed further and are as follows:
a) Bench test without process flow ( e.g. plant instrument shop, laboratory, manufacturing site)
b) Laboratory test with flow, simulating a plant process
c) In-process test during plant operation
Table 1 — Control valve output measurements and test environments
Table 1 summarizes the control valve test environment and the resulting information. Some of these 
combinations may be eliminated by constraints mentioned later in this document.
Test Environment Parameter Recommended for Measurement 
(listed order of preference within each environment)
Static Performance
Measures - Resolution, dead 
band, and valve travel gain.
Dynamic Performance
Measures - Dead time, response 
time, and overshoot.
Bench test
Tested on a bench, in a plant, a 
lab, a shop, etc. with no flow 
(may be under applied pressure), 
with valve completely assembled, 
and with recommended packing 
load.
Stem position (use only if a 
laboratory or in process test is 
not available)
Stem position
Laboratory test
Tested in a laboratory flow loop 
with flow
Flow coefficient for dead band 
and resolution
Process response, where 
possible
Stem position
Flow coefficient
Process response
In-process test
Tested in the actual process 
application
Flow coefficient (if 
measurements available)
Process response, where 
possible
Stem position
Flow coefficient (if 
measurements available)
Process response
— 17 — ISA–TR75.25.02–2000
4.4 Size of input signal change – regions
The character of the control valve response usually changes with the size of the change in the input signal. 
For the purpose of this document, four regions are defined.
The four regions are defined only as an aid to understanding control valve response. They are not 
intended for catalog data or purchase specifications. The size of each region and the boundaries between 
the regions are determined completely by the specifications used. For specification guidance see 6.5.2.
4.4.1 Region 1
Region 1 is defined as small input steps which result in no measurable movement of the closure member 
within the specified wait time. 
4.4.2 Region 2
Region 2 is defined as input step changes which are large enough to result in some control valve response 
with each input signal change, but the response does not satisfy the requirements of the specified time and 
linearity. 
4.4.3 Region 3
Region 3 is defined as step changes which are large enough to result in flow coefficient changes which 
satisfy both the specified maximum response time and the specified maximum linearity.
4.4.4 Region 4 
Region 4 is defined as input steps larger than in region 3 where the specified magnitude response linearity 
is satisfied but the specified response time is exceeded. 
5 Maintenance and design issues affecting process control
Control valves consist of a valve body, an actuator, and often accessories such as a positioner to form a 
control valve system. The interrelationship of these parts can be complex. It is advisable for the specifier 
and the manufacturer to communicate and to understand the requirements and constraints before 
selecting the valve that will perform correctly in service and can be maintained properly. Design, 
installation, service, and maintenance will affect valve performance. The user site may also have 
maintenance and business requirements that must be satisfied. Ongoing operationalconditions and 
requirements may affect ongoing control valve performance. It is not possible to make an intelligent 
selection of a suitable control valve for a given application without adequate information. Process fluid 
characteristics must be known: pressures, temperatures, (both normal and maximum), flows, (minimum, 
normal, and maximum), fluid characteristics (density, viscosity, vapor pressure, etc.), required leakage 
class, temperatures, and corrosion issues, are typical application requirements. The capabilities of the 
local maintenance organization, support by the manufacturer, local needs and desires should be 
considered. Some requirements may be determined by regulations and laws. Good knowledge resources 
include catalogs, experience, and knowledgeable manufacturer representative. The required speed of 
response should be estimated based on anticipated process control requirements.
Installation and operation outside the design conditions may damage the valve. The only defense is 
knowledgeable maintenance and management support. Poorly trained mechanics, poor records, 
inadequate spare parts, and no time to do it right, can lead to sticky valves, broken stems, actuators that 
do not move, and positioners that do not function. The discussion below is not intended to dictate 
technology.
ISA–TR75.25.02–2000 — 18 —
5.1 Stem seal
5.1.1 Materials
Most common valve stem seals are based on PTFE (polytetrafluoroethelene) and die-formed graphite. 
Until recently, graphite was required for applications involving high temperature, high pressure, fire safety, 
or because of chemical incompatibility with PTFE. Some implementations of graphite have a coefficient of 
friction much higher than PTFE and a tendency to stick to the stem, which will degrade the control behavior 
of the control valve. 
Recent developments provide alternatives. These often offer better sealing and reduced friction and can 
tolerate a wide range of temperatures. New materials and new formulations are available, along with 
combinations of the old materials and various design details (see references 16 through 19).
The packing system is now an engineered component of the control valve. Factors to be considered 
include
a) temperature, pressure, and chemical compatibility;
b) sealing and fire safe requirements; 
c) friction with actuator sizing and process control requirements;
d) service life, maintenance, and total life costs; and
e) project costs.
5.1.2 Design
Limitations on fugitive emissions have encouraged manufacturers to develop innovative material 
selections, combinations of packing materials, and installation configurations, such as “live loading”. Live 
loading is the use of a spring, typically a Belleville washer, to maintain a constant force on the packing. 
The more nearly constant and optimum compression load on the packing may improve the life of the 
packing and minimize both leakage and friction. New configurations typically use innovative material 
combinations and are designed to minimize leakage and friction. Extensive testing on packing leakage 
and stem friction by the vendors provides the basis for their design and recommendations. Deviations 
from these recommendations may result in higher leakage and greater friction. Attention to the details is 
paramount. The manufacturers offer a variety of design packing details, such as the number of rings of 
packing, style, and materials and filling materials. The packing box design provides a number of choices 
and details.
There is a balance between stem seal leakage and packing loading. Over tightening can lead to excessive 
stem seal friction. It will destroy the packing and result in poor dynamic performance and stem leakage.
5.2 Valve seat shutoff
Undesired static and dynamic control valve characteristics which may develop near the shut off position 
can be avoided when control is normally, and preferably, with the closure member well away from the seat 
position. 
There are designs where the valve seat continues to contact the closure member beyond the initial 
opening. This creates friction and may degrade resolution and dead band depending on the actuator and 
positioner.
— 19 — ISA–TR75.25.02–2000
5.3 Valve seat type
Valves that have seats in contact with the flow closure element while in the control range include some ball 
and plug valves and some cage guided globe valves. The magnitude of these effects will depend on the 
seat style, the design details and the materials used. Friction affects the static performance parameters; 
resolution and dead band. Review the catalog data and drawings for information.
5.4 Process fluid effects
Viscous and sticky fluids such as resins will tend to resist stem movement and will increase the resistance 
to movement. Rotary valves are less affected by process fluid and packing friction.
5.5 Mechanical tolerances
Mechanical backlash in any of the connections between the closure member, the actuator, and the 
positioner, will increase dead band. That is, two steps in the same direction may have a different result 
than if the direction were reversed between the steps. 
5.6 Structural stiffness
Weak or flexing linkages between the valve, the actuator, and the positioner can increase dead band and 
resolution.
5.7 Pneumatic positioner 
The positioner air flow capacity limits, varying gain, linkage wear, and internal friction can add significant 
non-linearity to the control valve response time. Internal friction can degrade dead band and resolution. 
Interactions between the positioner and the actuator can create dynamic non-linearity. 
5.7.1 Tubing/fitting capacity
Restrictions in the flow path into and out of a pneumatic actuator will slow the response. These restrictions 
may be undersized tubing, damaged tubing, undersized fittings, solenoid valves orifice size, and manual 
valves. Restrictions will have a lesser effect on small changes in signal but will affect the time of response 
for large (>10%) signal changes.
5.7.2 Actuator volume
The volume of air that the positioner must supply and exhaust limits the speed of response. A greater 
actuator volume requires a larger change in the mass of air in the actuator and may delay the response 
and increase the dead time. Three volumes may be described. The stroke volume is the change in 
volume during the stroke. The dead volume is the total volume minus the stroke volume. The total volume 
is fixed by the design. The stroke volume will vary with stem position and pressure creating a dynamic 
non-linearity. In any calculations, consider that it is actually the mass of air in the actuator that creates the 
pressure.
5.7.3 Supply pressure and capacity
Inadequate air supply capacity and pressure will limit the dynamic performance of the control valve. 
Undersized and limiting piping or tubing, air filters and supply regulators limit the air capacity. Dirty filters 
and partially closed block valves will slow or prevent response. The dead time and the time constant will 
ISA–TR75.25.02–2000 — 20 —
be increased. The times for opening and for closing the valve will probably differ from each other. See 
ANSI/ISA-7.0.01-1996, "Quality Standard for Instrument Air."
5.7.4 Valve pneumatic accessories
Accessories such as volume boosters, quick-release valves, and solenoid valves will all affect the 
performance. A volume booster can improve speed of response. 
5.8 Actuator size/type
Selection of the actuator size requires accurate information on friction, and process pressures, 
temperatures, and fluid characteristics. Safety factors in sizing actuators must consider safety, and the 
quality of the available information. Actuators are sized based on the minimum air supply pressure 
available but the actuator design must also withstandthe maximum air supply pressure. To reduce dead 
time and to minimize the dead band, available stroking power must be greater than the minimum force 
required to move the valve stem. An inadequate or undersized actuator or a positioner with poor 
performance will result in poor response in both response time and magnitude. 
The selection of the type of actuator involves considerations of valve size, and the design details, air 
supply pressure, and manufacturer offerings. 
5.9 Electric and hydraulic actuators
The actuator discussion above applies to all types of actuators, pneumatic, hydraulic and electric, but 
especially to pneumatic. Hydraulic actuators are expected to provide very good performance and are 
typically used for larger valves and the more difficult applications. Some types of electric motor actuators 
will have good resolution and may have a very small dead band, but they may lack the required speed 
when used with larger valves. These statements are only broad generalities. The user must investigate 
the data and claims from the manufacturers.
5.10 Flow effects
Dynamic imbalance from the effects of flow on the control valve closure member can degrade repeatability 
and dynamic linearity. Choking will limit flow capacity and vibration can affect the positioner performance. 
A valve operated in the flow-to-close mode may show stem instability as the plug approaches the seat and 
the hydrodynamic plug forces increase rapidly. Some butterfly designs have a reversal in flow induced 
shaft torque depending on position. These forces will vary with flow rate and pressure drop.
5.11 Valve sizing and selection
The installed flow characteristic of the valve may not be the same as the inherent flow characteristic of the 
valve. See further discussion in 6.3. 
6 Process and control design issues 
6.1 Control loop process gain – range and variability
A fluid process is much easier to control if the control dynamics remain nearly constant over the full range 
of operating conditions. 
— 21 — ISA–TR75.25.02–2000
The key dynamic parameters include: process gain, process time constant, dead time, controller 
dynamics, sensor dynamics, and control valve static and dynamic properties. The values of these 
parameters and their changes over the operating range will determine how well the process can be 
controlled to achieve the following: 
a) Acceptable level of process performance during process operation
b) Low process variability
c) Final product of acceptable uniformity
d) Low manufacturing cost
e) Ability to meet manufacturing demand
f) High level of plant safety
g) High level of environmental compliance
h) Successful startups, product grade transitions, and shutdowns
i) Recovery from process upsets
The control loop is subject to setpoint changes and to upsets. It establishes new operating conditions to 
recover from load disturbances or to meet the new setpoint. Parameters such as the required time 
constant and minimum dead time are set by the process design and instrument selection. The process 
gain is central to the valve selection, process dynamics, and fluid transport system. The process gain is 
determined by the process dynamics, the control scheme and the fluid transport system. The fluid 
transport system characteristics are determined by the pump/compressor, piping, equipment, and the 
control valve. In all cases, the valve capacity and characteristics influence the process dynamic control. 
Control design strategy can compensate to a certain extent for physical equipment and piping design 
limitations. The task of the design process is to select the right sized control valve and suitable 
characteristics. This will help keep the installed control loop process gain in the acceptable range, and 
reduce the variation over the normal operating range of the process.
The process gain of a self-regulating process is the ratio of the process variable change (e.g., temperature, 
composition, pressure, or flow) compared to the change in the controller output that caused the change. 
The process gain of an integrating process is the ratio of the change in the rate of change of the process 
variable (e.g., level, gas pressure) compared to the change in the controller output that was made to cause 
the change.
The process gain may be determined on-line by carrying out a series of step changes in the controller 
output. It may also be predicted from sizing calculations, or from a dynamic simulation of the plant design. 
In flow control applications the process gain is influenced by the relative pressure drop taken across the 
control valve, as compared to the rest of the fluid transport system, and the span of the sensor, transducer, 
or transmitter.
For non-integrating (self regulating) processes (integrating processes tend to be tank levels and certain 
types of vessel pressures) such as flows and pressures, the process gain should be within the range of
0.5 to 2.0 (% of span process variable change) / (% output change)
ISA–TR75.25.02–2000 — 22 —
Example: 
In a step test, if the controller signal output changed 5%, and the flow signal changed by 7.2% of span, 
then process gain = 7.2% / 5% = 1.44.
If the process gain is higher than 2.0, then 
the valve is oversized; 
has the wrong flow characteristic;
the process dynamics are too sensitive (high gain); or
the fluid transport system is oversized (i.e., pump may be oversized).
If the process gain is higher than 2.0, the process may be difficult to control. The valve dead band and 
resolution are multiplied by the process gain, therefore a high process gain increases the effective process 
dead band and resolution. The minimum flow change (dead band) determines the resolution of the 
resulting control action. A process gain lower than 0.5, will result in small flow changes and require higher 
controller gain. This does not create a control problem, but it may indicate a limitation to achieve adequate 
capacity.
There is a tendency for dead time to be longer with small changes in controller output for pneumatic 
actuators. This will have a de-stabilizing effect on most control loops. Dead time is a common cause for 
control loop limit cycles in control loops where the step response time in this region is not significantly 
shorter than the normal closed loop time constant of the control loop.
It is desirable to keep the process gain as constant as possible over the operating range of the process. 
This will reduce the need to retune the controller with changes in operating conditions. Process control 
performance is a function of the total loop. 
6.2 Over-sizing
Control valves, pumps, and pipelines are very often over-sized during the design of a plant to provide for 
an easy increase in plant capacity. The result is that the control valve will operate nearly closed. This often 
results in extremely poor control performance. The whole range of plant operation is reduced to a very 
narrow controller output and valve travel range. A system with a centrifugal pump operating at a small 
fraction of the design basis flow will have a higher control valve pressure drop. The pump will operate in 
the high head, low flow condition. The pipe and equipment friction pressure drops will be lower. Cavitation 
and damage to valves and pumps may occur. The control valve is an integral part of the fluid system, and 
the sizing deserves careful consideration. 
In a flow control loop, if the dead band and resolution are 1% of travel, and the process gain is 1.0 (percent 
of flow change per percent input signal change), then the flow signal will exhibit a dead band and 
resolution equal to 1% of span. If the process gain is 5, the flow resolution will be 5 x 1% = 5%. This may 
be so coarse that good control is impossible. No controller tuning can hideor eliminate this problem. The 
performance of the loop can be improved by either reducing the control valve dead band and resolution, or 
by lowering the process gain. 
The pump impeller and the control valve trim should be selected for the present actual required capacity. 
As production increases are desired, the control valve trim and the pump impeller can be replaced. The 
economic justification is the ability or inability to manufacture a product of adequate uniformity with poor 
control. 
6.3 Control valve inherent characteristic
The inherent characteristic of a control valve is the relation between flow and valve stem position at a fixed 
differential pressure. The installed characteristic is the relation between flow and stem position in a real 
— 23 — ISA–TR75.25.02–2000
installation where the pressures and density may vary with the flow and time. It is likely that the installed 
flow characteristics will differ from the inherent one. The choice between the standard catalog trim 
characteristics of "linear", “equal percentage” and "quick opening" is made to reduce the range of process 
gain over the control range. If the process gain varies excessively it can be at least partially corrected 
through control design strategy, by compensating for valve position in the control system by nonlinear 
compensation, or by compensating the controller gain by gain scheduling. Or, even through the 
adjustment or modification of the characterizing cam in the positioner. It is advisable to minimize the 
variation in the process gain, or the effective loop gain, over the operating range of the process so that the 
net variation is no more than +/- 50%.
6.4 Closed loop performance – control valve dynamic specification
The speed of response of a control loop is determined by the objectives of the process control strategy, 
which in turn is set by the process manufacturing goals. The required speed of response sets the control 
loop performance, which can be measured in terms of bandwidth, cross-over frequency, phase margin 
(see reference 10), closed loop time constant, setpoint overshoot, resonant peak and other measures. 
Modern tuning methods, such as those based on design synthesis, Internal Model Control (IMC and 
Lambda Tuning (see references 13 through 15), are all based on the determination of control loop 
performance by specifying a desired closed loop time constant. The closed loop time constant will differ 
depending on the tuning method used. The required speed of response will vary widely from loop to loop.
Figure 2 — Example of step response test with lost motion
One example is a design requirement to control the level of a large capacity tank with a closed loop time 
constant of 30 minutes in order to use the surge capacity of this large tank to reduce upsets in the rest of 
the process. This level controller requires a valve with a small dead band to avoid cycling but does not 
require fast response. Another example is a time critical loop, such as pressure control of an 
incompressible liquid header supplying a number of critical users. This may require a closed loop time 
constant as short as one second. This high speed of response reduces the interaction between the 
various flow control loops that supply each user. 
6.4.1 Control valve speed of response – step response time T86, approximate time constant t'
The control loop contains many dynamic elements: the process, the control valve and the transmitter. It is 
theoretically possible to tune a control loop to be faster than its internal dynamic elements. This can only 
be accomplished in a stable and robust manner when the internal dynamics are well behaved, and have 
parameters that are constant with time. This is seldom true in a plant environment, and even less so when 
Time (sec)
0 50 100 150 200 250 300 350 400 450
(%) 
35
40
45
50
55
60
65
70
1% Steps 2% Steps0.5% Steps 5% Steps 10% Steps
Flow Rate
Actuator Position
Input Signal
4-Inch Segmented Ball Valve with Diaphragm Actuator & Positioner
ISA–TR75.25.02–2000 — 24 —
control valve dynamics are involved. As dynamic parameters vary, the control loop could become unstable 
and process variability will increase. For this reason, it is common practice to tune control loops to be 
slower than the open loop dynamic of the component elements. 
The control valve system will not limit the control loop response if its speed of response is 10 to 20% of the 
next slowest control loop component. Minimum step sizes for good valves may range from 0.2 to 2.0% for 
the required T86. 
6.4.2 Impact of dead time on loop stability
The step response time (T86) is made up of two components, the dead time (Td) and the remainder of the 
time. Control loop stability is especially sensitive to dead time; this is the most de-stabilizing of the time 
dependent dynamics for a control loop (see reference 15). Equally de-stabilizing is the tendency of the 
dead time to vary. Pneumatic actuators tend to exhibit dead time while the positioner transfers sufficient 
power air to the actuator to overcome friction and to move the valve closure member. This tendency also 
is often amplitude dependent and smaller step changes exhibit a much longer dead time than larger step 
changes. For simplicity, a single response time specification, T86, is used to represent the total step 
response time of the control valve. It remains desirable to minimize the dead time portion of T86.
6.5 Nonlinear regions
The control valve response is nonlinear. Subclause 4.4 defines valve characteristic responses for four 
ranges of input step change size. Region 3 is the normal operation step size. T86 has meaning only within 
this range. A very good valve may have a dead band of 0.1 to 1.0% and a step resolution of 0.05 to 0.5%.
6.5.1 Control valve minimum position
It is more difficult for most control valves to make accurate small signal step movements when the closure 
member approaches the closed position (see figure 8). Friction increases as the seat is contacted. The 
change in flow coefficient also becomes less predictable at small openings. It is normal to specify a 
minimum valve position for which the valve dynamic specification applies.
6.5.2 Control valve dynamic specification
Control valve dynamic specifications should be based on the control loop dynamic requirements. 
For example:
Desired control loop dynamics – closed loop time constant of 10 seconds.
Control valve step response time T86 is 40% of the closed loop time constant, or 4 seconds, hence the 
desired control valve approximate time constant is 2 seconds.
The valve system shall respond to step changes from 1% to 10%. This means that the control valve will 
respond within the T86 maximum specification for step changes ranging from a minimum of 1% to a 
maximum of 10%.
The control valve has to operate within the above specification down to a minimum of 10% open.
The example above serves to illustrate that a range of dynamic specifications is needed for control valve 
speed of response, which depend on the specific requirements of the control loop. There are applications 
where a very fast response is needed, with a t' of 0.2 seconds, and a T86 of 0.4 seconds. Other 
— 25 — ISA–TR75.25.02–2000
applications exist where the speed of response of the control valve is not an issue, and it may be 
acceptable to specify a t’ of as long as 10 seconds, and a T86 of 20 seconds. Such a specification may 
allow the selection of a less expensive valve actuator. A t’ in the 1 to 3 second range, and a T86 in the 2 
to 6 second range will suit the majority of control valve applications in most process plants. Larger valves 
and larger actuators tend to have longer time constants.
7 Static behavior tests
7.1 Important measures of static behaviorSee ANSI/ISA-75.25.01-2000 for the standard testing procedures.
Clause 5 of ANSI/ISA-51.1-1979 (R1993) gives test procedures for a variety of static performance 
parameters that measure differences in the input-output relationship. Many of these are derived from the 
full-scale bench top calibration cycle familiar to most instrument engineers [e.g., figure 30 of ANSI/ISA-
51.1-1979 (R1993)]: dead band, linearity, repeatability, and reproducibility. These bench tests are 
valuable and provide considerable information. They do not measure the effects of varying friction and 
hydraulic forces. The hysteresis and dead band cannot be separated from this test, hysteresis and linearity 
are usually dominated by smoothly varying errors that accumulate significantly only over large strokes, and 
repeatability and reproducibility are important only in open loop systems2.
For most closed loop applications the key static performance discrepancies occur with small amplitude 
changes when the input reverses direction or when it continues in one direction after coming to rest. These 
are identified as dead band and resolution3 respectively.
The in-process test is conducted with the control valve installed in the process, and usually with the 
process transmitter used to measure the result (see reference 4). Bench top tests (without process 
loading) measuring actuator position or stem position provide necessary but not sufficient data to predict 
in-process performance. Possible development of better instrumentation would allow reliable detection of 
the closure member position. This could be used during normal process operation; any erratic flow 
behavior caused by the valve trim might still be missed. 
Another possible solution is to use actuator position data for conditions in which the manufacturer can 
provide data showing, for typical operating conditions (typical temperature, pressure drop, and valve 
friction), that actuator motion causes distinct changes in flow coefficient.
As a compromise, laboratory tests, with flow simulating a plant process load, allow accurate 
measurements with reasonably realistic loading. Such tests allow predictions of results under normal 
operating conditions. 
7.2 Applications affected and classes of performance
In many control loops, the ability of the control valve to make small moves (<1%) accurately is more 
important than the ability to make large moves quickly. The ability of the control valve to make small 
moves allows the process to be properly regulated. Reducing the controller gain, to eliminate the cycling, 
______
2
 These conclusions hold for the majority of troublesome applications experienced by the authors. However, readers interested in 
these other static performance measures will find thorough discussion of test methods in ANSI/ISA-51.1-1979 (R1993) and 
ISA-75.13-1996.
3
 The meaning of dead band – measured when reversing direction – is explicit in ANSI/ISA-51.1-1979 (R1993). However, the present 
meaning of resolution – measured while continuing in one direction – is adopted here consistent with our physical understanding and 
for lack of a better term in ANSI/ISA-51.1-1979 (R1993).
ISA–TR75.25.02–2000 — 26 —
degrades the response of the control loop to load disturbances and setpoint changes. Control is further 
degraded when the dissatisfied operator switches to manual control.
The non-linearities that occur at small signal amplitudes are important because they tend to generate limit 
cycles. The magnitude of the non-linearities determines the amplitude of the cycle. Controller tuning 
affects only the period of the cycle, not the amplitude. The period can range from fractions of a minute to 
several minutes, creating upsets that may affect the entire operation. It is very common for flow and 
pressure control loops to cycle, especially when the valves have high process gains. Control loops in a 
process area interact, sometimes very strongly. A pressure controller for a liquid header that supplies 
several users is a good example. When such a loop limit cycles, the flow loops supplying liquid to each 
user will also cycle, thus destabilizing the whole process area.
Dead band and resolution are also important in control of composition, pH, some level control applications, 
and temperature. These would be considered relatively slow systems, or “lag-dominant” (see reference 6), 
hence the importance of static behavior over dynamic response. Difficulties in establishing the desired 
process gain also contribute to the importance of static behavior in these applications. Advanced Process 
Control systems such as Dynamic Matrix Control make use of complex plant models that assume well-
behaved control valves; these generally require dead band less than 1%. 
The “small valve in parallel with a large valve” scheme has been used for difficult pH applications (see 
reference 6). The large valve is periodically adjusted to keep the small valve near 50% open. This is 
called “position control”. The small valve provides the control resolution, the large valve provides the 
rangeability. The disadvantage is in the dynamics for large changes in demand. The position control loop 
is tuned with low gain and moderate reset, response is delayed, and an upset is caused by the change in 
the relatively less precise large valve.
7.3 Testing considerations
7.3.1 General considerations
The following factors should be considered when defining the test procedure:
a) Will a pre-programmed input sequence be used or will it be manually operated?
b) How abruptly should the input be changed? Step changes are simplest, but not necessarily the most 
realistic.
c) What should be the amplitude of changes?
d) What should be the wait time or duration of steady input after each small change is made? At least 
one wait period of several minutes should be specified to detect sustained oscillation, which would 
invalidate measurement of static behavior.
e) How many cycles of reversal should be conducted?
f) What will be the nominal positions for the small-amplitude tests?
There is no single test sequence that can be generally applied to all control valves in all four environments. 
An example test is shown in figure 2. Testing requires a wait time long enough to allow the valve to reach 
the final position. This time may be several minutes. The pressure gauges associated with the positioner 
will provide an indication of the end of positioner action. 
— 27 — ISA–TR75.25.02–2000
Figure 3 — Example of a test using a series of small steps
7.3.2 Tests without process load in a plant instrument shop or control valve manufacturing site
This environment provides valve data but not process control data because field conditions are not present 
and the flow is not measured. Some realism can be simulated by tightening the valve packing to a 
specified value as described in 6.3.4.2 of ISA-75.13-1996, "Method of Evaluating the Performance of 
Positioners with Analog Input Signals and Pneumatic Output". Detailed test sequences can be run at 
multiple nominal positions. However, other factors work to limit realism; some known factors with today’s 
designs are listed below. These factors may also affect in-process testing.
a) Friction in many valves changes during the first hundred cycles of operation. Non-live-loaded PTFE 
packing may experience a relaxation of stress after a few cycles of operation.
b) Ball valves with tight shutoff and seals that self-lap in service may experience much higher friction after 
only a few hours of operation in the plant, but later friction may decrease.
c) Graphite packing for high-temperature service may experience much higher friction at room 
temperature.
d) PTFE packing for moderately high temperatures may experience lower-than-realistic friction at room 
temperatures.e) Significant vibration of the process piping may change the results for an installed valve.
f) It is not certain that the valve will hold the flow coefficient steady when the closure member is 
stationary (see reference 7).
g) Testing at manufacturing sites normally will be limited by operational limitations.
7.3.3 Laboratory testing with flow, simulating a plant process
This environment provides more information on valve performance by applying process loading and using 
the process variable to measure the results, as required in some specifications (see reference 4). Prior to 
testing, stroke the valve sufficient cycles to break-in all sealing and guiding surfaces. 
5 8 . 0
5 8 . 5
5 9 . 0
5 9 . 5
6 0 . 0
6 0 . 5
6 1 . 0
6 1 . 5
6 2 . 0
T im e ( s e c o n d s )
9 0 0 1 0 0 0 1 1 0 0 1 2 0 0 1 3 0 0 1 4 0 0 1 5 0 0 1 6 0 0 1 7 0 0 1 8 0 0 1 9 0 0 2 0 0 0 2 1 0 0 2 2 0 0 2 3 0 0
5 6 . 0
5 6 . 5
5 7 . 0
5 7 . 5
5 8 . 0
5 8 . 5
5 9 . 0
5 9 . 5
6 0 . 0
A c t u a t o r
P o s i t io n
( % )
I n p u t
S ig n a l
( % ) n o r e s p o n s eu p o n r e v e r s a l
0 . 1 % ≤ D e a d B a n d < 0 . 3 %
0 . 1 % < R e s o lu t io n ≤ 0 . 3 %
n o r e s p o n s e w h i le c o n -
t in u in g in s a m e d i r e c t io n
V a lv e X / A c t u a t o r Y / P o s i t io n e r Z , p a c k in g p e r in s t r u c t io n s
T e s t e d a t n o m in a l 6 0 % o p e n , 6 0 0 g p m , 3 8 p s id
ISA–TR75.25.02–2000 — 28 —
Figure 4 — Step response test 0.5% steps 7-8% dead band 
7.3.4 In-process testing during plant operation.
If the process is self-regulating4 (or, for integrating processes, if the time derivative can be calculated to 
infer flow results) and if noise and process disturbances are relatively small, this environment provides 
useful data but only at the allowable operating conditions. For some valve styles, this is the only way to get 
a useful answer. The process should be allowed to operate for some time after the valve is installed, to 
allow break-in of all sealing and guiding surfaces to their asymptotic frictional state.
7.4 Data presentation
The simplest method of presenting the results is direct plotting of the time series as in figures 2, 3 and 4. 
The dynamic response for the small steps can be measured on the same graphs. Multiple cycles can be 
shown, and multiple output variables can be shown if available (e.g., stem position and flow). The test 
shown in figure 3 is very time-consuming. Figure 4 shows an extreme case of a small step size test with a 
large dead band, a valve with high friction, and no positioner. Using a pre-programmed series of steps, 
bounds on the dead band and resolution can be set, as shown in the figure, providing overshoot does not 
occur on the output. In contrast, the ANSI/ISA-51.1-1979 (R1993) method of measurement requires an 
operator to move the input slowly in the smallest possible increment until an output change is observed. 
The labels on figure 3 show data interpretation according to ANSI/ISA-51.1-1979 (R1993) as follows. 
Dead band is the range through which the input signal may be varied in one direction, after a reversal of 
direction, without initiating an observable change in output signal. After the second step in the up 
direction, some motion did occur (albeit small); therefore, the dead band is less than the sum of two steps. 
Following ANSI/ISA-51.1-1979 (R1993), largest dead band from all recorded reversals is reported and 
from all nominal positions if more than one position was tested.
______ 
4
 Defined in ANSI/ISA-51.1-1979 (R1993)
Input Signal
Stem Position
61.5
58.6
55.6
52.6
49.7
 0.0 90.2 180.4 270.7 360.9 451.1
— 29 — ISA–TR75.25.02–2000
Another method is to plot only the static results (after the control valve stops moving on each step) in the 
x-y domain. However, if overshoot occurs, which prevents measurement of dead band and resolution, this 
information will be lost in transforming from the time series to x-y plots. 
Either plotting method can be used to determine the actual cause of the measured dead band and 
resolution if further measurements are taken of at least one internal state of the control valve. For 
example, with pneumatic actuators a pressure measurement enables determination of friction. Measuring 
motion at multiple locations enables determination of backlash.
Labeling of graphs should state clearly what loading was applied to the control valve, and the condition of 
the valve (new versus worn). For bench top testing, if comparisons are to be made of two control valves, 
or altered conditions within one control valve, it is helpful to have an independent measure of the total 
friction.
7.5 Design and maintenance factors important to static behavior
Friction is usually the dominant factor in the static behavior. Actuator sizing, positioner design, drive-train 
design, and other factors are also important. These factors are discussed in clause 5.
8 Small amplitude and medium amplitude dynamic response tests 
(regions 2 and 3)
Dynamic response, in this document, means valve response versus time for changes in control valve input 
signal. Dynamic response can be measured either without process loading (bench top tests) or with active 
loading.
This clause covers input amplitudes below which velocity limiting, region 4 (defined in clause 4), occurs. 
Steps smaller than 10% commonly avoid velocity limiting. It is possible for velocity limiting to occur on 5% 
steps, or even 2% steps, if the positioner has a small air flow capacity relative to the actuator volume.
ISA–TR75.25.02–2000 — 30 —
Figure 5 — Example graph of time series tests showing step response times 
Because control valves are not linear in behavior, the dynamic response is strongly amplitude dependent. 
Usually, 1% steps will give response times very different from 5% steps. The response is often much faster 
for 5% steps than 1% steps5. Small-step dynamic response cannot be scaled down from 100% steps, 
since the two amplitude ranges are governed by completely different dynamics. Finally, results become 
inconsistent as the amplitude is reduced to approach the dead band and resolution limits; this is the 
definition of region 2.
8.1 Important measures for regions 2 and 3
Frequency response, using sine wave inputs, is a useful test described in standards ISA-75.13-1996 and 
ISA-26-1968. The controller output signals of most control loops are simulated better by sine waves than 
by square waves (even for digital controllers correctly applied and tuned). However, frequency response is 
generally too difficult for tests within the scope of SP75.25, especially tests under normal process 
operating conditions. The simplest and currently most popular test, step response, may be the only 
method feasible in all environments for which tests are desired.
______ 
5
 For an example of this trend, see reference 8.
Time (seconds)
158 159 160 161 162 163 164 165 166 167 168
65
66
67
68
Time (seconds)
178 179 180 181 182 183 184 185 186 187
64
65
66
67
A
B C
I/P Input Signal
CA
B
Stem
Position
(%)
2% Steps on 4-Inch Globe Valves with Standard Actuators and Positioners
Nominal Q = 600 gpm, ∆P = 38 psid, water T = 120oF
Packing maintained per instructions
— 31 — ISA–TR75.25.02–2000
The closed loop input signal to a control valve is usually of limited bandwidth; open loop step response is 
misleading for some applications. Particular areas of uncertainty are the measured dead time and 
overshoot. Dead time, Td, is commonly measured after a step input. Tests show that dead times can be 
much longer with the slowly changing controller