Calibration Essentials Ebook PDF

Calibration
Essentials
Introduction
The proper calibration of instruments is always one of the most essential, yet historically one of
the most overlooked processes in today’s plants and factories. Even today, with rapidly evolving
technology and tools to enhance production and efficiency, the need to calibrate and ensure
consistent, reliable measurement is always there.
With innovative new tools and equipment seemingly released every day, how do you know that
you’re taking the most efficient path towards your facilities calibrated, automated production? To
help you find that certainty, the calibration experts at Beamex have teamed with the International
Society of Automation (ISA) to provide this in-depth guide to calibration automation, complete
with all the information you need to ensure a fully calibrated and reliable facility.
The informative new eBook, Calibration Essentials, comes with 60 detailed pages, covering every-
thing you need to know about today’s calibration processes including:
• A comprehensive big picture guide on how to manage a facility-wide calibration program

for Industrial Automation and Control Systems.
• Informative overviews of calibration considerations, such as tolerance errors, and calibration

uncertainty, as well as practice scenarios and solutions to manage them.
• An in-depth look at some of the new smart instrumentation and WirelessHART instruments
and how to effectively calibrate them
• A technical discussion on the pros and cons of an individual instrument calibration strategy
versus a loop calibration strategy.
• Detailed guidelines to ensure facility and employee safety and security, as well as
compliance with standards, when conducting calibration tasks
• And much more
With this eBook, from Beamex and ISA, you now have the resource
you need to ensure that your facility is safely and efficiently
getting the most out of your instrumentation. This road-
map to calibration has tools for workers at every
level of your facility to standardize your
effort and facilitate an advanced, auto-
mated production environment.
Contents
Management of a Calibration Program for Industrial

Page 5
Automation and Control Systems
ISA Recommended Practice
Calibration Out of Tolerance

Page 25
By: Heikki Laurila
Calibration Uncertainty Page 30

By: Heikki Laurila
Calibration Intervals
ISA Publication: Calibration: A Technician’s Guide Page 37
By: Mike Cable
Criteria for Instrument Selection in Relation to the

Measurement Requirements Page 38
ISA Publication: Calibration Handbook of Measuring Instruments
By: Alessandro Brunelli
Calibrating WirelessHART Transmitters Page 42

By: Heikki Laurila
Calibration Essentials 3
Calibrating Smart Instruments Page 46
By: Heikki Laurila
Loop Calibration vs. Individual Instrument Calibration

By: Mike Cable
New Methods for Calibrating Loops Page 51

By: Ned Espy
Safety Considerations
by Mike Cable
Data Integrity Page 58

By: Heikki Laurila
Definition of Terms and Acronyms

ISA Recommended Practice: Management of a Calibration Page 62
Program for Industrial Automation and Control Systems
Management of a Calibration Program for
Industrial Automation and Control Systems
Introduction
Purpose acquired during these assessments not only aids in

the establishment of future calibration intervals, but
The purpose of this recommended practice is to is also critical in the allocation of capital and oper-
provide the basic framework for developing and ational resources. Clearly defined policy and proce-
maintaining a consistent calibration program for dures support the efforts of maintenance planners to
industrial automation and control systems (IACS), schedule adequate labor and equipment for calibra-
including instrumentation used in safety instrument- tion both during and between facility outages. Fol-
ed systems. This recommended practice provides lowing established calibration procedures and using
guidance on methodologies for calibration of an IACS correct equipment reduces the likelihood of human
by considering the accuracy of each loop required by errors due to incorrect practices; avoids acting on
the process and, if necessary, adjusting a loop com- incorrect information; ensures the desired results
ponent(s) to achieve the desired loop or component of the calibration efforts; and promotes the correct
accuracy. operation of an IACS.
Accurate, reliable, and repeatable operation of loops In an IACS, more hardware faults occur in the mea-
in an IACS is vital to maintaining the safety and reli- suring instrumentation, and control valve compo-
ability of a facility. Where repeatable measurements nents than in the control components of the IACS.
take place in the monitoring or control of a facility, A calibration program can aid in early detection of
the measuring instruments and test equipment these failures. Inadequate calibration and main-
must have repeatable outputs. A well-considered tenance of an IACS may increase the likelihood of
calibration program, correctly implemented and system problems, including:
maintained, can directly contribute to the assurance
of the desired operation of the IACS for the facility. A • Inaccuracy of measurements and control;
calibration program establishes periodic assessments • System not responding correctly or as desired;
of loop/component performance over time. Data • Reduced awareness of instrument performance
and actual need for calibration and maintenance;
and NOTE 3 The result of a calibration may be record-
• Potential for reporting of incorrect environmen ed in a document, sometimes called a “calibration
tal or other data. certificate” or a “calibration report.”
However, targeted reduction, or increase, in calibra-

tion activities based upon the assessment of data
derived from an instrument asset management cali- According to standards ISO 9001:2008 7.6 and ANSI/
bration program can both reduce maintenance costs NCSL Z540.3-2006, calibration is a comparison of
and improve reliability and safety. the device being tested against a traceable reference
instrument (calibrator) and documentation of this
comparison. Although calibration does not formally
Various definitions of the term calibration can be
include any adjustments, in practice, adjustments are
found:
possible and often included in the calibration pro-
cess.
The formal definition of calibration by the Interna-
tional Bureau of Weights and Measures is: “Opera- In the absence of a calibration program, mainte-
tion that, under specified conditions, in a first step, nance practices for line, vessel or equipment-mount-
establishes a relation between the quantity values ed devices such as pressure, temperature and level
with measurement uncertainties provided by mea- indicators, unfortunately, may occur only when the
surement standards and corresponding indications error in reading becomes large enough to be obvious
with associated measurement uncertainties (of the to the operator or technician. Maintenance person-
calibrated instrument or secondary standard) and, nel routinely make decisions based on these devices.
in a second step, uses this information to establish a A faulty indication on such a device could lead to the
release of energy or other unsafe action. A well-con-
relation for obtaining a measurement result from an
sidered calibration program that periodically mea-
indication.” (from International vocabulary of me-
sures actual loop accuracy can provide confidence in
trology - Basic and general concepts and associated the indications and drive the calibration intervals for
terms (VIM) (JCGM 200:2012, 3rd edition)). these devices.
A definition found in the NIST Handbook 150:2001 Instrumentation based on newer technology, e.g.,
is: “1.5.8 Calibration: Set of operations that establish, “smart” devices, is more accurate than older technol-
under specified conditions, the relationship between ogy devices and more stable, requiring less frequent
values of quantities indicated by a measuring instru- calibration monitoring. A calibration program should
ment or measuring system, or values represented by also accommodate these device types appropriately,
a material measure or a reference material, and the such as calibration check frequency and accuracy of
corresponding values realized by standards. calibration equipment required.
Instrumentation worth purchasing and installing is

worth calibrating and maintaining. All instrumenta-
tion, including the highest quality devices, will drift
NOTE 1 The result of a calibration permits either over time and provide less accurate measurements.
the assignment of values measured to the indica- Calibration corrects for that drift. Thus, it is critical
tions or the determination of corrections with
that all instruments are calibrated on their appropri-
ate intervals.
respect to indications.
Companies striving to maintain a safe working
environment while ensuring the reliability of their
NOTE 2 A calibration may also determine other
facilities may use calibration as a means of verifying
metrological properties such as the effect of in- the functionality and accuracy of their equipment.
fluence quantities.
Like other aspects of maintenance, there are many
things to consider when establishing a company
calibration program. Certainly, this is the case with 1.2.2 BPCS and SIS functionality
the calibration of monitoring and control loops in This document does not include any consideration
a facility IACS. This document presents a recom- for how instrumentation and control signals are
mended approach to developing, implementing and handled within the BPCS or SIS other than including
maintaining a calibration program that is intended to indication of the signal in loop accuracy calculations.
increase accuracy and reliability of an IACS, decrease
maintenance costs, and improve quality control.
1.2.3 Control valve and other final control devices
More importantly, this approach is also intended to
calibration
increase safety of operation as the result of increased
accuracy and reliability of the instrumentation. This document does not cover the essential mainte-
nance of the mechanical aspects of these and other
This approach to calibration has proven successful mechanical devices.
when companies have adhered to the concepts set
forth in these guidelines, enabling those companies 1.2.4 Regulatory requirements
to realize the full benefits from a standardized ap- Regulatory requirements related to loop/instrument
proach to calibration. accuracies are not included in this recommended
practice.
The intended audience for this document is any
company or industry that utilizes instrumentation in 1.2.5 SIS instrument calibration
the monitoring and control of a process or facility.
The calibration of monitoring and control loops that
are part of safety instrumented systems is included in
Organization
the scope of this recommended practice. The docu-
This recommended practice is organized to provide mentation and management of these instruments as
recommendations on: part of safety instrumented systems is excluded. For
more information, see ISA-84.00.01 (IEC 61511 Mod)
• Establishing a calibration program; and all its parts.
• Calibration program activities; and
• Calibration program management. 1.2.6 Instrument criticality
This document mentions criticality as it applies to a
1 Scope calibration program. This document does not modify
any guidelines for establishing criticality provided in
1.1 General applicability ISA-TR91.00.02.
The recommended practice detailed in this docu-
ment defines a baseline definition and model of a 1.2.7 Control loop performance
quality management system that can be utilized to The performance of final control elements and the
implement and maintain a calibration program for tuning of control loops are excluded from the scope
industrial automation and control systems. It is ap- of this recommended practice.
plicable to all IACSs.
1.2.8 Hazardous areas
1.2 Inclusions and exclusions Instruments can be located in hazardous locations.
These instruments are to be installed, operated
1.2.1 Manufacturer specific calibration procedures and maintained in accordance with requirements
This document does not provide or recommend for those hazardous locations. The precautions for
manufacturer-specific calibration procedures for working on these instruments are outside the scope
specific instruments as these are established by the of this document. Refer to other codes and standards
instrument manufacturer and are outside the scope addressing electrical equipment in hazardous loca-
of this document. tions for applicable requirements, such as ANSI/ISA-
12.01.01-2013, Definitions and Information Pertain-
ing to Electrical Equipment in Hazardous (Classified) which in turn is a fundamental part of the plant’s/
Locations. facility’s maintenance. As such the management and
scheduling of maintenance personnel, and other
2 References resources such as test equipment, to perform cali-
bration activities is usually executed via work orders
ISA-TR91.00.02-2003, Criticality Classification Guide- from the plant’s/facility’s maintenance management
line for Instrumentation system (computerized or otherwise). An overarching
reliability or asset management program in turn may
ANSI/ISA-5.1-2009, Instrumentation Symbols and
guide the plant’s/facility’s maintenance management
Identification
system. Consequently, consideration should be given
ISA-5.4-1991, Instrument Loop Diagrams to how the calibration program interacts with other
programs and systems in the plant/facility.
ANSI/ISA-12.01.01-2013, Definitions and Informa-
tion Pertaining to Electrical Equipment in Hazardous
Where instrument and systems performance certi-
(Classified) Locations
fication is mandated by regulating agencies, federal,
ISA-84.00.01 Parts 1-3 (IEC 61511 Mod), Functional state or local, the owners should ensure that techni-
Safety: Safety Instrumented Systems for the Process cians involved in the calibration process are familiar
Industry Sector with those mandatory requirements. Examples in the
NIST Handbook 150 Checklist:2001, National Volun- USA are EPA, NRC or CEMs permits for emissions
tary Laboratory Accreditation Program (NVLAP) and discharge to water bodies.
Liptak, Bela G., Editor. Instrument Engineers’ Hand-

3.2 Calibration program concepts
book, Process Measurement and Analysis, Fourth
Edition, International Society of Automation, 2003 Monitoring and control instrumentation currently
used in an IACS range from pneumatics to “smart”
Ultimate Calibration, 2nd Edition, Beamex, 2009
(microprocessor-based) digital electronics. These de-
Cable, Mike. Calibration: A Technician’s Guide, Inter- vices are as varied as the processes they monitor. De-
national Society of Automation, 2005 vices with moving parts require regular maintenance.
ISO 9001:2015, Quality management systems, Clause These mechanical devices are much more suscepti-
“Control of monitoring and measuring equipment”, ble to mechanical performance issues (e.g., binding
International Organization for Standardization and dragging due to environmental contamination)
than are other measurement and control instrumen-
ANSI/NCSL Z540.3-2006 (R2013), Requirements for tation. Analog electronic instrumentation is subject
the Calibration of Measuring and Test Equipment, to drift of settings and output. Digital instrumenta-
NCSL International tion has multitudes of parameter settings that must
ANSI/NCSL RP-1-2010, Establishment and Adjust- be set correctly to achieve desired operation. Devices
ment of Calibration Intervals, NCSL International not operating to their manufacturer’s specifications
and/or not correctly applied or configured for the
JCGM 200:2012, 3rd edition. International vocabulary specific application can result in operational issues,
of metrology - Basic and general concepts and associ- such as off-spec quality, productivity issues, and safe-
ated terms (VIM) ty issues. In addition, there is always the instrument
3 Establishing a calibration program

failure. All of this results in a need for the loops that
are important to safety, quality and correct operation
3.1 General
of the facility to be periodically calibrated.
A calibration management program includes de-
fined procedures and tasks for the calibration of Understanding and adhering to the following guide-
all relevant instruments. Ongoing calibration is lines, explained in the sections that follow, is required
a fundamental part of instrument maintenance, to achieve the full benefits of the recommended
approach to a calibration program set forth in this
document.
3.3 Calibration program

A calibration program for an IACS formalizes a
methodology to periodically verify the performance
accuracy of the loops/components in that IACS and,
when necessary, make adjustments to those compo-
nents to bring them within their manufacturer-rated
accuracy and/or the loop within its required perfor-
mance tolerance.
Each user company/facility is encouraged to estab-

lish a calibration program specific to its needs. This
recommended practice discusses the essential fea-
tures of a calibration program and provides guidance
on how to establish such a program.
importance of this concept is that, instead of only
This proposed approach to a calibration program for looking at the performance of an individual device
an IACS takes into consideration all known loop mea- (e.g., field instrument), it looks at the performance
surement errors and establishes loop performance of all of the components (sensor, transmitter, wiring,
tolerances based on the specific process require- input card, etc.) that must work together to measure
ments. Successful implementation of this approach a process parameter. Also, this concept establishes a
requires management’s commitment to make this a tolerance requirement for each loop that is based on
living process. process needs rather than simply assuming man-
ufacturer ratings. The accuracy of a measurement
containing multiple components (a.k.a. a loop) (as
3.4 Device calibration
a simple example, a sensor, transmitter, and input
Device calibration focuses on one particular device card) is unknown unless it is calibrated as a system.
within a loop and ensures that this discrete compo-
nent within the loop has been compared and ad- For example, a temperature transmitter could be
justed, if necessary, to a reference standard. This is calibrated as a single device but when installed with
typically done before the component or instrument the sensor, wiring, and indicator, the loop indication
is installed, typically by the manufacturer, and is used might not be within the desired tolerance due to
as an initial benchmark to ensure accuracy and then issues with other components of the loop. The key
is periodically checked to ensure device accuracy. parameter is whether the entire loop is providing a
The period, or calibration interval, may be set by the measurement within the tolerance needed for correct
authority having jurisdiction, the plant maintenance operation of the process. Each component in a loop
program, the manufacturer’s recommendation, or has a rated accuracy. The inaccuracy or error of all
by another agreed upon method. When this results components in the loop results in the total loop inac-
in multiple periods for a device, the shortest period curacy greater than that of any one component.
should be used.
3.6 Competency
3.5 Loop calibration Trained personnel are essential to the success of a
A key concept of this recommended practice is estab- calibration program. This includes understanding the
lishing required loop tolerance and evaluating, when- process, regulatory requirements, calibration tools,
ever practical, calibration performance of the entire calibration procedures, and instrument/systems
loop, which includes multiple components. The training required by the manufacturer.
4 Calibration program planning 4.2 Loops and components in the calibration
program
4.1 General
For a meaningful and efficient calibration program
Components and systems require periodic cali- to be conducted, it is essential that complete IACS
bration to confirm that they are operating safely, documentation exists from the original design and
efficiently, and in compliance with regulations. To construction of the plant/facility and that this docu-
ensure such calibrations occur as necessary, compa- mentation has been maintained in a current state. As
nies should have a structured calibration program, a a minimum, this documentation should consist of:
formalized methodology, that applies to all compo- 1.) an instrument index/list with all instrument and
nents and systems covered by the program, consist- control devices identified with loop and component
ing of standards, procedures and tasks for verifying functional tagging (e.g., ANSI/ISA-5.1); 2) loop dia-
performance accuracy and adjusting or replacing grams, e.g., per ISA-5.4, or other instrumentation wir-
deficient/defective components/systems as neces- ing interconnect drawings showing all components
sary to ensure operation occurs in accordance with and wiring, including control panels and instrument
manufacturer-rated accuracy parameters and/or junction boxes; and 3) instrument location plans for
required loop performance tolerances. all tagged devices, control panels and junction boxes.
Additionally, piping and instrumentation (or similar)
Critical steps in the process of developing a calibra- drawings (P&ID’s), and electrical area classification
tion program include: drawings would be beneficial.
a) establishing a loop/instrument tagging system; 4.3 Component and loop criticality determination
b) developing a comprehensive list of loops and If it does not already exist, establish a loop criticali-
instrumentation equipment requiring periodic ty classification system for loops in the facility. The
calibration; criticality of each loop is one of the driving factors
c) establishing tolerance requirements for each loop; for establishing calibration intervals. ISA-TR91.00.02
provides a structured technique for establishing crit-
d) establishing a loop criticality system and criticali
icality classifications. Classification examples from
ty ratings for each of those loops;
that document include safety, environmental, and
e) establishing loop calibration verification intervals asset protection.
(frequencies) based on the assigned criticality and
manufacturer recommended interval or user Using the criticality classification system established,
experience; the applicable criticality classification for each loop
and component in the calibration program should be
f) establishing calibration equipment requirements,
identified and documented.
based on loop tolerance and instrument accuracy
requirements;
4.4 Establishing required loop tolerance
g) establishing loop/instrument calibration record
forms and calibration record-keeping systems; Each loop should be evaluated against and calibrat-
ed to a specified tolerance, which is the permissible
h) establishing necessary calibration procedures;
deviation from the actual value, otherwise referred to
i) establishing staff qualifications and training re as the loop tolerance.
quirements;
The tolerance for each loop may be established in
j) performing calibration checks at required intervals
either of two ways: (1) by establishing the required
and documenting results; and
accuracy of the loop necessary to meet safety, qual-
k) establishing a calibration program verification ity, and/or production requirements of the process,
methodology. or (2) by establishing the theoretical tolerance of the
loop based on the rated accuracies of all of its com- d) facilitates focusing calibration efforts to the areas
ponents. In both methods, the theoretical tolerance that provide the most benefits or have the greatest
of the loop should be calculated to demonstrate needs;
whether the loop is capable of meeting the required
e) provides management with a measure for auditing
accuracy (in method 1 or to establish the target toler-
loop performance and calibration work (staff and
ance for the loop (in method 2). Also, the rangeability
equipment performance); and
of the specific application must be considered. The
wider the rangeability (measurement span) required f) clarifies loop performance expectations for opera
(to accommodate expected/unexpected load vari- tions and maintenance, including regulatory re
ances), the more inaccurate the measurement is quirements where applicable.
likely to be.
4.5 Calculating theoretical loop tolerance
Establishing an acceptable tolerance for each loop: Theoretical loop tolerance is calculated by using the
Root-Sum-Square (RSS) method, which combines the
a) clearly defines the level of acceptable perfor rated accuracy of each component in the loop. Each
mance for each loop; of the listed effects on accuracy is squared and all
b) provides a defined measure to use in the periodic squared terms are added together. The square root of
check of loop performance; this sum provides a combined uncertainty in either
percentage or engineering units of the loop. Section
c) facilitates tracking of loop performance over time;
8 provides examples for calculating loop tolerance by
using the manufacturers’ data and the RSS method. propriate test equipment, unqualified calibration
Each component may have multiple possible sources personnel, changes to location environment, or other
of error (e.g., accuracy, temperature effect, stability, possible causes.
etc.). Thus, the equation below should include all
possible errors for all loop components. Shown below is an illustrative example of initial
calibration intervals being assigned based on some
Tolerance = (error 1)2+(error 2)2+(error 3)2+(error n)2 criticality ranking. The example is not intended to
be used directly or to assign intervals to criticality
Since it is unlikely that each of these components will rankings.
experience their maximum error at the exact same
time, this calculation provides a “typical” error that Table 4-1 – Example criticality ranking (CR) and ini-
might be expected from a combination of devices tial verification frequencies
each with its own error statement. If all of the compo-
nents were to experience their maximum error at the Critically
same moment, this would be expressed by the sum of Ranking System Frequency Notes
all of the maximum individual errors. CR #4 Safety instrumentation 12 months See note 1
CR #4 Boiler trip instrumentation 12 months

To perform this calculation correctly, all errors should
CR #4 Turbine trip instrumentation 12 months
be expressed relative to the actual loop/component
application measurement span and in the same man- CR #4 Anti-water induction
instrumentation 12 months
ner (e.g., percent of full scale, percent of reading).
Per regulatory
CR #4 Regulatory instrumentation
requirements
4.6 Calibration confirmation intervals
CR #4 Fire system instrumentation 12 months
The calibration confirmation interval for a particular CR #3 Boiler control instrumentation Outage
loop or component is a function of (a) criticality of Outage
CR #3 Turbine control instrumentation
the loop/component; (b) the performance history
Reliability monitoring system
of the loop/component; (c) the ruggedness/stability CR #3
instrumentation
Outage
of the component(s); and (d) the operating environ- CR #3 Ash system Instrumentation Outage
ment. Effective calibration programs base calibration
CR #3 Reclaim System instrumentation Outage
intervals on these factors. The user must determine
the applicable verification frequency for each loop/ CR #2 Boiler transmitters (non-control) 36 months or as observed
component in the calibration program. Historical CR #2 Boiler switches (non-control) 36 months or as observed
performance information and/or instrument asset CR #2 Turbine transmitters (non-control) 36 months or as observed
management diagnostics maintained by the user can
CR #2 Turbine switches (non-control) 36 months or as observed
be very helpful in this effort.
CR #2 Performance transmitters 36 months
Maintenance records, possibly analyzed using statis- CR #2 Performance thermocouples 36 months

tical data techniques, can be useful in determining as needed or
CR #1 Local gauges
whether or not to modify calibration confirmation available
intervals. By analyzing instrument drift over time, CR #1 Recorders

as needed or
available
calibration costs can be reduced and efficiency im-
as needed or
proved. If no other data is available, initial frequen- CR #1 Valve positioners (non-control)
available
cies may consider manufacturer’s recommendations
as needed or
and then be adjusted as calibration performance CR #1 Indicators
available
data is obtained.
Note 1 to entry: Intervals for SIS-related instruments

If a high calibration fail rate is observed, before
follow the testing interval required to meet the SIL for
changing the calibration interval, check for correct
the specific SIF, which may be shorter or longer than 12
application and specifications of the device, inap
months
4.7 Calibration equipment the TUR shall be equal to or greater than 4:1.”). Refer
to Example 6 below for an example calculation of
Equipment used to evaluate a loop’s performance
TUR. If multiple calibration equipment components
against its required tolerance or to calibrate a device
are used to perform the calibration of a loop/compo-
should be certified, accurate, corrected for local con-
nent, the aggregate uncertainties of the calibration
ditions and have adjustments controlled.
equipment components should be calculated using
the RSS method and that aggregate uncertainty be
Test equipment accuracy requirements are directly
four times greater than the loop/device tolerance.
related to required loop tolerances. Establishing un-
reasonably small loop tolerances will result in requir-
NOTE A communicator (e.g., HART, fieldbus) is not usually
ing more accurate test equipment, which in turn can
a traceable calibration standard – check the manufacturer’s
result in increased maintenance costs.
product documentation for intended use of any measurement
capability the communicator may have. A communicator is
Equipment includes calibration standards, which are
necessary to configure/trim/diagnose certain devices but a
traceable to national or international standards. In
traceable standard is also needed to perform a calibration.
cases where a national/international standard is not
available or practical, a reproducible standard should
Equipment should be corrected for the effects of local
be established and maintained via an independent
conditions (e.g., ambient temperature, atmospheric
entity (preferred) or in-house.
pressure).
Equipment should be identified with a unique identi-
Equipment should have a means to control access
fication number, part number, and range/accuracy.
to adjustments affecting performance on devices
Equipment should be certified, typically by the man-
after they have been confirmed or calibrated. These
ufacturer or a third party, on a regular, fixed period
should be sealed or otherwise safeguarded to prevent
(annually or otherwise as recommended by the man-
unauthorized changes. Seals or safeguards should
ufacturer) to be operating correctly and within all
be designed and implemented such that tampering
manufacturer’s specifications. A sticker should be af-
will be detected. The calibration process procedures
fixed to the equipment documenting the most recent
should include directions to install seals/safeguards
date of re-certification and when it is due for the next
and actions to be taken when seals or safeguards are
re-certification. The device’s calibration should be
found damaged, broken, bypassed or missing.
traceable to a national or other acceptable standard.
These recommendations are critical to making an
Equipment documentation should include a history
accurate evaluation of loop/component performance
of each certification and a traceability statement or
and to making any calibration adjustments to a
uncertainty budget.
component. Test equipment of insufficient accura-
cy may result in incorrect evaluation, unnecessary
Equipment should be of greater accuracy than the
activities, and even leaving a component performing
device being calibrated and traceable to a higher
less accurately.
standard. There are basically two ways to maintain
traceability during calibration – the use of an uncer-
4.8 Calibration record system
tainty budget (performing uncertainty calculations
for each measurement), or using a test uncertainty A calibration record system is necessary to provide
ratio (TUR) of ≥ 4:1. To maintain traceability, without documentation of when calibration activities have
using uncertainty budgets or calculations, ensure been undertaken on which pieces of equipment,
standards used for calibration are at least four times what the results were, and what (if any) corrective
(4:1) more accurate than the test equipment being action was taken. The system should include forms
calibrated. (ANSI/NCSL Z540.3- 2006 states: “Where to be used to record the calibration activities and the
calibrations provide for verification that measure- method for storage of all calibration records to facili-
ment quantities are within specified tolerances... tate future access. This system may be paper- or elec-
Where it is not practical to estimate this probability, tronic-based. The calibration service record (CSR)
process is repeated until the performance values
meet the tolerance criteria. These adjusted values
then become the “as left” value for that calibration
interval. When the next calibration interval comes
up, the calibration activity again starts with reading
and recording the “as found” performance, and the
cycle completes as before.
When documenting calibration information (as

found, as left), calculate and document the accuracy
of the observed indications. Ensure that the calculat-
ed accuracies are expressed in the same manner as
the required loop/device tolerance (e.g. engineering
units, percent of full scale, percent of reading, engi-
neering units) to enable direct comparison. Refer to
Clause 8 for an example.
Calibrations may be performed using 5 check points

(0, 25, 50, 75, and 100%) or 3 check points (0, 50, and
100%). Calibration should be checked and recorded
in both directions (inputs increasing from 0 to 100%
and inputs decreasing from 100 to 0%) to observe if
the loop/device exhibits any hysteresis.
All data should be recorded on a CSR and stored in a

centralized location for easy review and audit.
4.9 Calibration procedures

To promote consistent methodology and results,
provides essential information on a particular loop/ calibration procedures, or calibration work instruc-
component. This includes the calibration points with tions, should be established for each loop/compo-
the acceptable loop/component tolerances. Using nent. These procedures should detail the method,
the CSR allows staff to capture the calibration data, equipment and criteria to be used to check and
which highlights how much drift has occurred in the perform calibration and should be part of a formal
loop/component since the last calibration. An exam- procedures management program. Procedures
ple of a CSR is provided in the annex. covering on-line, in-situ, or in service calibration of
devices will require specifics regarding the process
Recording of calibration activity includes document- conditions, with
ing “As Found” and “As Left” readings. Calibration
programs serve the dual purpose of checking loops details about the interaction of the instrument being
for conformance to tolerance requirements, and if calibrated/tested and the process, interactions such
tolerance is not met, making adjustments to equip- as alarms, interlocks, control points, or other aspects
ment to bring it into tolerance. The term “as found” or impacts related to the online instrument being
is used to describe the performance (readings and calibrated/tested. In general, calibration procedures
errors) of a device when it is first tested during the should include the following:
calibration activity. If it is found to be within toler-
ance, no adjustment is made and the performance is a) description of the loop or component to be
recorded as the “as left”. However, if an adjustment calibrated;
is made, the equipment is then tested again. This b) overview of the tasks to be performed;
c) equipment and personnel requirements; and the completion of an internal certificate program or
d) steps to be followed for the calibration. through a recognized independent third party certifi-
cation program. A subject matter expert should then
Persons performing calibration on in-service loops/ evaluate an individual’s ability to perform the various
components, should meet with the appropriate oper- applicable specific types of calibration. The owner
ating personnel to discuss the work to be performed should keep a register of qualified individuals by
before work is begun. component type. It is important to understand that
someone may be fully qualified to calibrate one type
The latest version of the applicable calibration proce- of component and not another. In addition, certain
dures should be available and referred to during the regulatory requirements may establish minimum
process to ensure awareness of any recent changes. qualification pre-requisites for personnel performing
calibrations. Requirements for periodic training and
Annex A provides an example of a loop calibration re-certification to ensure continued competency of
procedure document. It is not intended to be a personnel should also be established.
specific recommendation on the calibration proce-
dure itself, as this will depend on the application, 4.11 Responsibilities
the device manufacturer’s recommendations, and
Preferably, a calibration program definition should be
owner-specific methodology.
an integral part of the initial design and build capital
expenditure phases of the plant/facility’s life cycle.
NOTE Some instruments require both sensor and output
Whether the calibration program is being created
adjustment. Refer to manufacturer’s documentation for full
for the initial facility build or created/rejuvenated
calibration instructions.
during the operation phase of the life cycle, given the
number of loops in a typical industrial facility, fully
4.10 Calibration personnel
implementing an automation and control system
To ensure the safety and reliability of the measure- calibration program is a major undertaking. It may
ment and control system, only competent personnel well be part of a broader scoped instrument asset
should be allowed to perform calibrations. Suggested management or reliability program. As with any
areas of knowledge include: significant program, for it to be successful, company
management will need to provide sufficient support
a) general maintenance skills to the creation and on-going execution of the pro-
• operations and process safety gram. To ensure consistency, a project team should
• electrical safety be created to manage the project of establishing the
• troubleshooting calibration program. The project will require support
b) instrument maintenance skills from the facility technicians, engineers and, most
• instrumentation maintenance and repairs, espe importantly, senior management.
cially with the specific instruments involved in
the calibration
• loop checking
• calibration
• calibration record-keeping
c) process philosophy, control philosophy, and con-
ditions of process service
• process philosophy and control philosophy
• process conditions (e.g., service, pressure, tem
perature, flow, level)
• alarms, cause and effect matrix
Confirming the general qualifications of the person-

nel involved in calibration can be achieved through
The management of the plant, facility, or automation accuracy cannot be achieved, the component should
system should be responsible for establishing and be repaired or replaced.
owning the program and ensuring compliance with
the IACS calibration program. This would include Calibration confirmation is not achieved until and
establishing a process for auditing compliance with, unless the fitness of the measuring equipment for the
and periodically updating, the program. intended use has been demonstrated and document-
All aspects of the calibration program should be ed. Calibration confirmation should include calibra-
completely documented for an effective calibration tion verification, any necessary adjustment or repair,
program, including, but not limited to: subsequent recalibration, and comparison with the
loop tolerance requirements for the intended use.
a) calibration procedures (required method);
Software used in the calibration processes and calcu-
b) calibration equipment (required accuracies,
lations of results should be documented, identified
re-certification frequency and procedures, personnel
and controlled to ensure suitability for continued
training, inventory);
use. Calibration software, and any revisions to it,
c) personnel training and certification; should be tested and/or validated prior to initial use,
d) personnel responsibilities and accountabilities; approved for use, and archived.
e) required loop tolerances and manufacturer-sup-

6 Calibration program management
plied device accuracies;
f) required forms; 6.1 Control
g) calibration record-keeping system; and An effective calibration program should include a
means to track calibration activity by loop/compo-
h) required calibration frequencies.
nent, define required interval, forecast calibrations
due, document calibrations, and trend results. Many
5 Loop and component performance third-party software applications exist that can help
verification automate a calibration program.
5.1 General 6.2 Assurance

Performance verifications should be made using the An effective calibration program should be peri-
“loop” calibration method whenever possible. In odically monitored for compliance. This includes
those cases where this is not possible, loop compo- verification that loop calibration checks are being
nent accuracies should be verified and used to calcu- performed at the required intervals; all calibration
late actual loop accuracy, which is then compared to activities are being documented, including all cal-
required loop tolerance. culations; and calibration equipment being used is
maintained and has current certification. Manage-
The verification portion of the calibration proce- ment should ensure that this periodic monitoring of
dure should define the steps necessary to check the the calibration program is consistently performed
performance of the loop. This should include a task and results documented.
list providing specific steps necessary to perform the
check, including equipment and methodology. 6.3 Improvement
If the calibration verification of a loop/component Calibration programs should be evaluated and up-
reveals that the loop/component is not within the dated on a periodic basis or when significant modifi-
required tolerance, a calibration adjustment may be cations or projects are completed to ensure the pro-
required to one or more loop components according gram is evergreen. Audits should not only consider if
to defined procedures. The calibration adjustment loops/components have been added or removed, but
procedure for that component defines the overall should also include a review of the criticality rating,
steps necessary to calibrate the specific device type. process conditions, accuracy capability and tolerance
A task list provides the specific steps necessary to cal- requirements for each loop/component.
ibrate the specific component. If the desired device
Calibration results should be trended and reviewed
periodically to determine if performance is accept-
able or if a modified frequency of calibration or
device technology should be considered in order to
ensure desired loop performance is maintained.
7 Examples
7.1 Example 1 – Calculating instrument total error

To illustrate calculation of total error of an instru-
ment, assume a differential pressure transmitter
application that has been tested/calibrated at a
temperature and pressure different from the applica-
tion temperature and pressure. Total error includes
transmitter error and any zero or span shift caused
by these temperature and pressure variations. The
following are assumptions for this example:
Transmitter accuracy (E) – +/- 0.2% of actual span (under atmospheric ambient conditions)
Maximum range – 0 - 750 inches H2O (0 - 186.82 kPa)
Span – 0 - 100 inches H2O (0 - 24.9 kPa)
Operating pressure – 1000 PSIG (6,894.78 kPa)
Operating temperature – within a 50 ºF (27.8 ºC) difference of the temperature at which it was calibrated
Temperature zero shift – (Tz) – for temperature variation of 100 ºF (55.5 ºC), assumed to be +/-0.5% of maximum
range
Temperature span shift – (Ts) – for temperature variation of 100 ºF (55.5 ºC), assumed to be +/-0.5% of actual reading
Pressure zero shift – (Pz) – for pressure of 2000 PSIG (13,789.5 kPa), assumed to be +/-0.25% of maximum range
Pressure span shift – (Ps) – for pressure of 2000 PSIG (13,789.5 kPa), assumed to be +/-0.5% of actual reading
Calculating each error component
E = +/-0.2%
Tz = 0.5(750 in./100 in.) (50 ºF/100 ºF) = +/-1.875%
= (0.5(186.82 kPa/24.9 kPa) (27.78 ºC/55.5 ºC) = +/-1.875%)
Ts = 0.5(50 ºF/100 ºF) = (0.5(27.78 ºC/55.5 ºC )) = +/-0.25%
Pz = 0.25(1000 PSIG/2000 PSIG) (750 in./100 in.) = 0.9375%
= (0.25(6,894.78 kPa/13,789.5 kPa) (186.82 kPa/24.9 kPa) = 0.9375%)
Ps = 0.5(1000 PSIG/2000 PSIG) = (0.5(6,894.78 kPa/13,789.5 kPa)) = +/-0.25%
Calculate the total error, using RSS method, at an assumed actual differential pressure reading of 100 inches H2O
(24.9 kPa)
Total Error = (E)2+(Tz)2+(Ts)2+(Pz)2+(Ps)2
Total Error = (0.2)2+(1.1875)2+(0.25)2+(0.9375)2+(0.25)2 = 1.57%
This is the total error associated with the transmitter for this example application. It is not the total loop measure-
ment error, which would include the errors from all other loop components. It is not the maximum possible error
associated with the transmitter as that would be the sum of all error components, assuming they all occurred
simultaneously.
NOTE This example is taken from the Instrument Engineer’s Handbook.
7.2 Example 2 – Calculating theoretical loop tolerance

To illustrate calculation of theoretical loop tolerance, assume a static pressure transmitter application that has
been tested/calibrated at a temperature and pressure different from the application temperature and pressure.
Total error includes transmitter error and any zero or span shift caused by these temperature and pressure varia-
tions. The following are assumptions for this example:
Maximum range – 0 - 3500 PSI (0 - 24,132 kPa)

Span – 0 - 2500 PSI (0 - 17,237 kPa)
Transmitter accuracy (E): ± 0.075% of full scale
Transmitter stability over 5 years (S): ± 0.125%
Temperature zero shift (Tz) for temperature variation of 100 ºF (55.5 ºC), assumed to be +/-0.036% of maximum
range
Temperature span shift (Ts) for temperature variation of 100 ºF (55.5 ºC), assumed to be +/-0.036% of actual read-
ing
Control system analog input card accuracy (C) ± 0.100%
E = ± 0.075%
S = ± 0.125%
Tz = 0.036%(3500 PSI/2500 PSI) (50 ºF /100 ºF) = +/-0.025%
= (0.036%(24,132 kPa/17,237 kPa) (27.78ºC/55.5ºC) = +/- 0.025%)
Ts = 0.036%(50 ºF /100 ºF) = +/-0.018%
= 0.036%(27.78 ºC/55.5 ºC) = +/-0.018%
C = 0.100%
Total Error = (E)2+(S)2+(Tz )2+(Ts )2+(C)2
Total Error = (.075)2+(0.125)2+(0.025)2+(0.018)2+(0.100)2
Total Error = ±0.18% of full scale or ±4.5 PSI (31 kPa)
Calculating the theoretical tolerance using the RSS method, the loop is within calibration if the full-scale reading
is between 2495.5 and 2504.5 PSI (17,206 and 17,268 kPa)
7.3 Example 3 – Pressure measurement theoretical loop tolerance – pressure switch
To illustrate calculation of theoretical loop tolerance, assume a static pressure switch application that has been test-
ed/calibrated at a temperature and pressure different from the current conditions. Total error includes switch error
and zero/span errors due to temperature/pressure variations. The following are assumptions for this example:
Range and span: 0 - 3000 PSI (0 - 20684 kPa)
Switch accuracy (E): ± 1.0% of full scale
Switch hysteresis (H): ± 0.2%
Switch repeatability (R): ± 0.07%
Transmitter stability (S): ± 0.5% full range per year non-cumulative
Temperature zero shift (Tz) for temperature variation of 50 ºF (27.8 ºC), assumed to be +/-0.04% of maximum range
Temperature span shift (Ts) for temperature variation of 50 ºF (27.8 ºC), assumed to be +/-0.04% of actual reading
E: ± 1.0%
H: ± 0.2%
R: ± 0.07%
S: ± 0.5%
Tz = 0.04%(3000 PSI/3000 PSI) (50 ºF/100 ºF) = +/-0.02%
= (0.04%(17236.89 kPa/17236.89 kPa) (27.8 ºC/55.5 ºC) = +/-0.020%)
Ts = (0.04% (50 ºF/100 ºF)
= (0.04% (27.78 ºC/55.5 ºC)) = +/-0.020%
Total Error = (E)2+(H)2+(R)2+(S)2+(Tz )2+(Ts )2
Total Error = (1.0)2+(0.2)2+(0.070)2+(0.05)2+(0.020)2+(0.020)2
Total Error = ±1.14% of full scale or ±34.2 PSI (235.8 kPa)
7.4 Example 4 – Pressure measurement theoretical tolerance – pressure gauge

To illustrate calculation of total error of a static pressure gauge application that has been tested/calibrated at a tem-
perature and pressure different from the current conditions. Total error includes gauge error and zero/span errors
due to temperature variations. The following are assumptions for this example:
Range and span: 0 - 3000 PSI (0 - 20684 kPa)

Operating temperature: 86 ºF (30.0 ºC)
Gauge accuracy (E): ± 2.0%
Temperature span shift (Ts): ± 0.4% of span per 18 ºF (10 ºC) of temperature change from a reference temperature of
68 ºF (20.0 ºC)
E = ± 2.0%
Ts = (0.04%*(86 ºF-68 ºF)/18=±0.04%
= (0.04%*(30 ºC-20 ºC)/10 ºC) = ± 0.04%
Total Error = (E)2+(Ts )2
Total Error = (2.0)2+(0.04)2
Total Error = ±2.0% of full scale or ±60 PSI (413.7 kPa)
These examples highlight the large differences in the

uncertainties from system to system and from device
to device. Far less accuracy is required to calibrate
a switch (combined uncertainties ±1.14%) than is 7.6 Example 6 – Calculating calibration TUR
required to calibrate loops on the inputs to a control
For this example, assume a device is being calibrated
system (combined uncertainties ±0.18%). It is im-
at 10 EU (engineering units). The required tolerance
portant to note that there are actually more identifi-
is +/- 1% of reading. The reference standard being
able factors to uncertainty in the systems. However,
used in the calibration has an uncer-
none of these factors is typically significant enough (USL – LSL)
tainty of 0.5 EU.
to make a recognizable increase in the uncertainty of (2u)
the loop.
TUR =
7.5 Example 5 – Calculating calibration accuracy
USL = Upper specification Limit = 10 + (1% x 10) = 11 EU
When documenting readings during calibration,
LSL= Lower Specification Limit = 10 – (1% x 10) = 9 EU
accuracy of the readings should be calculated and
documented. The calculated accuracy should be 11– 9 2
u=uncertainty = = = 0.5 EU
expressed in the same manner as the stated loop 2*0.5 1
tolerance or device accuracy (e.g., % of full scale, % of
TUR
reading).
The TUR must be calculated using identical param-

For this example, assume a pressure measurement
eters and units for both terms. Accuracy statements
loop with a required tolerance of +/-1% of reading
expressed in percentages or other ratios must be
and a measurement span of 0 - 20 PSIG.
converted to absolute values of the basic measure-
Test Input Reading Accuracy ment units. The uncertainty of a calibration standard
(PSIG) (% of reading) is not necessarily the same as its accuracy specifica-
(% span)
tion. The uncertainty in the equation must include
the uncertainties of all test equipment involved in the
0 0
measurement process used for the calibration.
25 4.95 (0.05/5) = 1%
In this example, the TUR is 2:1, which would not
50 9.92 achieve the recommended 4:1 ratio for traceability.
(0.08/10) = 0.8%
75 14.93 (0.07/15) = 0.47%
100 19.9 (0.1/20) = 0.5%
This loop would meet its required tolerance of +/- 1% of

reading.
Annex A – Example documentation
A.1 On-line testing procedure of differential pressure transmitter
Unit 2 Point C2 104, PDT-211
Description
The loop consists of the differential pressure transmitter (HART), operator interface computer (OIC), recorder
XR-211A-3, and the data acquisition system (DAS). This loop is indication only and uses the 4-20 mA signal for the
process variable (rather than the HART digital PV).
Preparation
1) Inform shift supervisor and Unit 2 control operator of the work to be performed.
2) Determine what, if any, affect this work will have on the other systems.
3) Review calibration procedures.
4) Conduct a “job briefing” with all personnel assigned to the work before the job begins.
5) When all tasks are complete, validate that the system has been correctly returned to service.
General
The tasks shall be performed by a fully trained I&C technician working with either an apprentice I&C technician, a
calibration technician, or control operator. Each step is initiated by the I&C technician. The calibration technician,
apprentice I&C technician, or CO will record data and track the procedure.
The following items are required to perform the work:
1) Reference standard, which must have a current certificate traceable to NIST or equivalent national standard
a) Pressure source (0-100 PSIG) (0 – 689.5 kPa)
b) Milliamp measurement (0-25 mA)
2) Other equipment
a) HART communicator
3) Calibration record sheet
4) Task list for calibrating the static pressure transmitter
Procedure:
1. ___ Proceed with the differential pressure calibration task list.

2. ___ Notify the shift supervisor and or control operator when the calibration process is complete and the
device is returned to service.
3. ___ Sign the completed calibration service record and file
Comments: ___________________________________________________________________
_____________________________________________________________________________
_____________________________________________________________________________
_____________________________________________________________________________
Worker Signatures: ___________________________ ___________________________

Date: ________________________
A.2 Differential pressure (DP) calibration task list (HART transmitter)
1) Obtain the calibration record sheet for the differential pressure transmitter being calibrated.
2) Locate and verify the identity of the transmitter.
3) Close primary high side isolation valve (process isolation). Close high side isolation valve (transmitter isola-
tion).
4) Open equalizing valve.
5) Close low side isolation valve (transmitter isolation). Close primary low side isolation valve (process isolation).
6) Vent high and low sides of the transmitter.
7) Purge transmitter with instrument air to remove water.
8) Connect HART communicator.
9) Connect pressure source (reference standard).
10) Record the operator interface computer reading in “as found” on the CSR at 0%, 50%, and 100% input range,
both increasing and decreasing.
11) Evaluate whether the readings are within the required loop/device tolerance.
12) If the readings are within the required tolerance, record those readings again under “as left” and proceed to
step 15.
13) Adjust the device as necessary using the HART communicator and the device manufacturer’s instructions to
bring the OIC readings with tolerance.
NOTE The method described here is for a loop calibration. If a device calibration is found to be necessary that
adjusts the device 4-20 mA output, the reference milliamp measurement must be used to measure device output,
the HART communicator is not a reference standard.
14) Repeat steps 10 and 11 until the readings are within the required tolerance, then record those readings under
“as adjusted” and “as left”.
15) Remove the HART communicator and milliamp measurement (if used).
16) Remove pressure source (reference standard).
17) Verify that the equalizing valve is open.
18) Open primary low isolation valve (process isolation). Open the low side isolation valve (transmitter isolation).
19) Bleed air from the transmitter and close vents.
20) Verify that there are no leaks.
21) Close equalizing valve.
22) Open high side isolation valve (transmitter isolation). Open the primary high side isolation valve (process
isolation).
23) Confirm that the operator interface computer reading reflects current process conditions.
A.3 Calibration service record example form
CALIBRATION CERTIFICATE Printed: 31.05.2013 15:02:51

Certificate Number: BX-11163-CAL Printed by: Ned Espy
TAG ID: TT-175
TAG INFORMATION ASSET

Tag Name: Dryer 175 Temp Asset ID: TT-175
Work Order Number: A123456 Serial Number: 171684
Location: 2nd Floor, Grid 12 - next to control panel Manufacturer Model: Rosemount 248
Plant Structure: Butte Power Plant /Unit 1 /1N (Boiler 1)/1NG (Primary and Rangeability: T/C, mV, RTD/Ohm Inputs
secondary air)/1NG10 (Primary air)/ Operating Temp.: -40 to 185 degF Humidity: 0 to 99% RH
FUNCTION CALIBRATION EVENT

Function: Temperature Transmitter (tt) Calibration Date: 5/31/2013 3:00:08 PM
Transfer Function: Linear Next Cal. Date: 11/30/2013
Range (I/O): 0 to 175 °F 4 to 20 mA Environment Temp.: 72 °F Humidity: 40 %
PROCEDURE CALIBRATORS
Due Date: 11/29/2013 Interval: 6 months Input Calibrator: MC6 s/n: 602057 Due Date: 3/5/2014
Reject if Error >: 0.5 % of span Adjust To <: % of span Input Module: TC-R-OUT/ R1 s/n: 61735 Due Date: 3/7/2014
Classification: Safety Output Calibrator: MC6 s/n: 602057 Due Date: 3/5/2014
Calibration Strategy: OSHA1910 Output Module: IN s/n: 21773 Due Date: 3/6/2014
1.000 As Found (1)

As Left (2)
Error Limit
Output Error [% of span]
0.500
0.000
-0.500
-1.000
0.00 43.75 87.50 131.25 175.00
Input [°F]
1. As Found FAILED 2. As Left PASSED

Max Error: -0.5888 % of span Max Error: 0.0400 % of span
Nominal Actual Nominal Actual Found Nominal Actual Nominal Actual Found
Input Input Output Output Error Input Input Output Output Error
[°F] [°F] [mA] [mA] [% of span] [°F] [°F] [mA] [mA] [% of span]
0.00 0.000 4.0 3.9995 -0.0031 0.00 0.000 4.0 4.0016 0.0100
43.75 43.750 8.000 7.9782 -0.1362 43.75 43.750 8.000 8.0000 0.0000
87.50 87.500 12.00 11.9545 -0.2844 87.50 87.500 12.00 11.9985 -0.0094
131.25 131.250 16.000 15.9327 -0.4206 131.25 131.250 16.000 15.9983 -0.0106
175.00 175.000 20.000 19.9058 -0.5888 175.00 175.000 20.000 20.0041 0.0256
131.25 131.250 16.000 15.9299 -0.4381 131.25 131.250 16.000 16.0011 0.0069
87.50 87.500 12.00 11.9557 -0.2769 87.50 87.500 12.00 12.0064 0.0400
43.75 43.750 8.000 7.9761 -0.1494 43.75 43.750 8.000 8.0019 0.0119
0.00 0.000 4.0 4.0023 0.0144 0.00 0.000 4.0 4.0014 0.0088
Calibration Note: For more information, call Beamex at 800.888.9892 (www.beamex.com)
Calibrated by: Peter Parker Approved: Clark Kent, Calibration Supervisor

5/31/2013 3:00:08 PM 5/31/2013 3:02:42 PM
Page: 1/ 1
Calibration
Out of Tolerance
What does it mean and what to do next?
By: Heikki Laurila
Sometimes we hear of a calibration being Out of What does “Out of Tolerance” (OoT) mean?
Tolerance (OoT). What does this mean in practice?
How do we know if it really happened, what the Let’s start by discussing what out of tolerance means.
impact is and what we should do next? These are the In summary, it means that during a calibration some
main topics discussed in this article. The focus of this calibration point(s) failed to meet the required tol-
article is mainly for the world of process industry, but erance level. This causes the result of the calibration
most of the principles are valid everywhere you run to be out of tolerance, or also referred to as a failed
into an OoT situation. For a process plant, it can be calibration. This definition sounds pretty simple, but
an actual process instrument that is out of tolerance, when we start to look deeper, it turns out to be more
or it can also be a calibration standard, or any other complicated. So let’s continue…
measuring device.
What does out of tolerance mean? In How was it found to be out of tolerance?
summary, it means that during a
calibration some calibration point(s) Assuming tolerance levels are set correctly, when
failed to meet the required tolerance someone says that during a calibration an instru-
level. ment failed to meet its tolerance level, the next log-
ical question is: are you sure? And to continue, how
What is the tolerance level used? sure are you?
For a calibration, we normally specify a tolerance lev- This leads to the question; which calibrator was used
el (maximum permitted error) that a given test result to make the calibration, and what is the total uncer-
is compared against. If we say that the instrument tainty of the calibration? If the situation is where the
failed to meet the required tolerance level, it becomes calibration was outsourced to an external calibration
an out of tolerance case and specific follow-up proce- laboratory, what are the calibration uncertainties of
dures may be initiated. that laboratory?
Whenever a tolerance limit is not met, the next log- For every measurement and calibration, the total
ical question should be; what is the tolerance level uncertainty of the calibration is critical. Whenever
that was used? Or maybe the question should be that you make the compliance statement that something
what tolerance level should be used? passes or fails a calibration, or that it is inside or out-
side of the tolerance, you must consider the uncer-
For a process instrument, it is very typical that the tainty of the calibration, in order to be able make a
tolerance level used is the manufacturer’s accuracy proper decision.
specification of the installed process instrument (as
specified by the manufacturer). It means that if you How critical is it?
buy 100 similar transmitters and install them into 100
different locations in your plant, all of these loca- What is the next step when we have found an instru-
tions will have the same tolerance level. However, ment to be out of tolerance and our tolerance levels
in most cases, many of these installation locations are valid and the calibration was done with appro-
have different criticality and therefore they should priate uncertainty? This means that we must admit
also have different tolerance levels. A critical process that this really in an out of tolerance case, and action
location needs to have a smaller tolerance level and needs to be taken. Before hitting the “Panic” button,
many times it is also calibrated more often. Likewise, it is important to see how critical the case is.
in a non-critical location, it is a waste of resources,
time and money to calibrate as often and to use the Before hitting the “Panic” button, it
same tolerance level. Personnel who have the best is important to see how critical the
knowledge of the actual requirements of the process case is.
in question, should be involved when deciding toler-
ance levels for given locations. For a process instrument, it is good practice to have
the calibration tolerance a bit tighter than the actual
If it is a matter for your calibrator or your reference process requirement. By doing this, even if the instru-
standards, then it is more common to use the manu- ment fails slightly against the tolerance in calibra-
facturer’s specification as the tolerance level (at least tion, it does not have a dramatic effect on the actual
to begin with). Comparison to the manufacturer’s process. Generally, if the results are just slightly out
specifications during re-calibration or certifica- of tolerance, this may not be critical for the measure-
tion will indicate how reliable the equipment is for ments that have been done with it. An analysis needs
performing field calibration. Calibrator tolerances to be done in order to see if a failure is critical or not.
should also consider local needs and be adjusted
accordingly over time. During a criticality analysis, you should analyze what
is the impact of this out of tolerance case. If it is a ibration laboratory. The
process instrument, what effect does this amount highest level reference
of error in the process measurement have to the standards/calibrators in
actual process and for process control? Also, what the plant should be cal-
effect does this have to the actual end product being ibrated by an accredited
produced? calibration laboratory, to
assure proper traceability
In case of a calibration standard or calibrator, what coming into the plant.
effect does the error in the calibrator have to all the
various measurements and calibrations that have The higher in the
been done with that calibrator? traceability chain
an out of tolerance
For an extremely critical process or safety measure- situation occurs,
ment, redundancy may be added with two or more the larger effect it
simultaneous measurement instruments being has on all the
installed. In this case, a failure in one instrument’s instruments below
capability to measure correctly does not cause a that instrument in
critical failure. the chain.
For some measurements, the process instrument can The accredited calibra-
be checked before each measurement is made with tion laboratories are
it. For example, a weighing scale can be checked with responsible for ensuring
reference weight(s) before making measurements, that their standards have
or on a regular daily basis. A calibrator can also be been calibrated proper-
cross-checked against another similar level reference ly, and are traceable to
instrument periodically (between the regular recali- national standards or the
brations). The calibration interval can also be adjust- equivalent.
ed according to the criticality of the measurements
made with a given instrument. Depending on the The higher in the traceability chain an out of toler-
application, if the cost of an out of tolerance calibra- ance situation occurs, the larger effect it has on all
tion is very high, that will also affect the calibration the instruments below that instrument in the chain.
interval and strategy. Imagine a situation where a National Laboratory
finds out that one of its reference standards is out
Where in the traceability chain did an OoT occur? of tolerance, and that reference has been used to cali-
brate customers’ reference standards/calibrators for
The criticality and the consequences of a failed cali- a long period. This means that all instruments of all
bration depend on where in the traceability chain it customers are affected! Of course that is a worst case
occurred. scenario, and not very likely, but it gives you some
perspective of the traceability chain.
A general rule for any calibration is that it needs to be
traceable to a higher level standard. The traceability Wherever in the traceability chain
chain goes to a national standard or the equivalent, the OoT happened, all instruments
and there are normally several steps in between. A below that are affected.
process transmitter is normally not sent to a na-
tional calibration center for calibration. Instead, But back to the process plant…typically there are
the process instrument is typically calibrated with limited levels of instruments in the traceability chain
a process calibrator that is calibrated with a higher within a plant. The vast majority of calibrated in-
level standard, either in-house or in an outside cal- struments are the actual process instruments, which
and the previous good calibration to see if there is
anything that would indicate when the instrument
drifted out of specification. For example, there could
be a sudden rise in reported issues in the process, or
in the case of a calibrator - a time period where more
failed calibrations began to appear. You may also an-
alyze the history of that instrument to see if there is
an indication of any typical drift for that instrument,
and possibly interpolate the data to find the most
likely moment when it went out of tolerance.
It can be difficult to determine the actual moment

when an instrument failed to meet its tolerance. It
may be the case there is no option but to assume that
are at the bottom of the traceability chain. There are
all calibrations done after the previous successful
typically process calibrators used to calibrate the pro-
calibration are affected and suspect to be failed.
cess instruments, as most of the process instruments
are not sent out for calibration. At some plants there
can be also higher level reference standards that are
Impact analysis - what are the consequences?
used in-house to calibrate the process calibrators. So,
Once we know that the out of tolerance really hap-
there can be a few levels and the impact is different
pened and we have analyzed how much it was and
in each level. Whatever the highest level instrument
have an idea when it had occurred, the next step is to
is inside the plant’s traceability chain, it needs to be
evaluate the impact. You need to find out where this
calibrated regularly by a traceable, preferably accred-
failed instrument has been used and what measure-
ited, calibration laboratory, in order to get the trace-
ments are suspect.
ability into the plant.
In the case of a process transmitter, it is obvious

It is good to remember that wherever in the trace-
where it has been used, but in case of portable
ability chain the OoT happened, all instruments
measuring equipment, or a portable calibrator, it is
below that are affected.
different situation. One powerful option for a calibra-
tion management program is a “reverse traceability”
When did it happen?
report. This kind of report should list all the calibra-
tions where a specific instrument has been used,
When an OoT is noticed during a calibration, note
over a certain time period. This report is most help-
this is not the moment when the instrument started
ful when you need to analyze, for example, where a
to measure incorrectly or went out of tolerance. But
portable calibrator has been used. If you do not have
when did it happen? It is important to determine
an automated reverse traceability report and need to
when it happened because any measurements done
manually go through calibration reports to see where
after that moment are suspect. In case of a critical
that certain calibrator was used, it may take many
process measurement that failed, any products pro-
man hours to complete. However you do it, it needs
duced after that moment are affected and may need
to be done.
to be recalled, in the worst case.
In the case of a process instrument being out of toler-

It is not an easy task to determine the moment when
ance, you need to have your process specialist analyze
the instrument went out of tolerance. By checking
what the impact of this failure is for the actual process
the previous calibration data, and assuming the
and to your end product. In best case scenario, if the
instrument was left in acceptable condition, you
effect to the process measurement was so small, it will
can start from there. However, if there are no records
not cause any significant damage. However, in the
between the previous good calibration and the new
worst case, if the analysis tells you that the effects to
fail calibration, you should take a look at everything
the process, and to the products being produced, are
done in between. You can study the measurement
so big that the products produced do not meet their
results and any relevant data in between the OoT
specifications, then costs can be huge.
You need to find out where this elsewhere. Quality standards also require that these
failed instrument has been used and processes are documented and that responsibilities
what measurements are suspect. are specified.
In many processes, the quality of the end product A root cause analysis is typically required by quality
cannot be simply tested in the final product, but the standards to find out what caused an OoT to occur.
process conditions must be correct during the man- A risk analysis, or generally, risk-based thinking, is
ufacturing process. If this example involves food/ something required by the modern quality system
medicine or the heat treatment process of critical standards. Continuous improvement is also a com-
aerospace/automobile parts, then you are obligated mon quality requirement to ensure that you contin-
to inform your clients/customers, or even withdraw uously improve your quality system and learn from
products form market. Product withdrawal is a any mistakes, so that problems do not happen again.
dramatic consequence; it will get you into the news,
it will be very expensive, and it will have a negative Many companies, especially in regulated industries,
effect to your company brand, reputation and stock are using some form of a “deviation management
value. software system” where all OoT calibration cases are
recorded in order to control and document the pro-
In the case of a process calibrator that fails to meet cess of handling these cases.
its tolerance, you will need to evaluate how much the
failure had effect to all the measurements made with
in since its last known good calibration. Many times,
the calibrator is significantly more accurate than the SUMMARY
process instruments calibrated by it, so there is some
safety margin. In the best case scenario, even if the Summarizing the key points in this paper, if you
calibrator failed recalibration, the failure can be so get an out of tolerance calibration, you need to
small that it does not have significant effect to the do the following:
calibrations that have been done with it. But in the
worst case, if all the calibration work that has been • Verify what tolerance level was used and that
done with that calibrator are suspect, then you need it is a correct level.
to analyze the effect for each process measurement
that has been calibrated. As previously mentioned, • Verify the uncertainty used in making any
this can be a really big task as you need to do the decisions that a measurement is out of toler-
analysis for all the process measurements being ance and that the uncertainty is appropriate.
affected.
• How critical is this out of tolerance observa-
tion?
Quality assurance considerations
• Where in the traceability chain did this occur?
You may have heard your quality professionals
talking about CaPa, being an abbreviation of Correc- • When did it occur?
tive Actions and Preventive Actions. This is some- • Make an impact analysis to find out what the
thing that is stipulated by most quality standards, consequences are.
such as the very common ISO/IEC 9001 quality
standard as well as ISO/IEC 17025 used in accred- • Make quality assurance considerations.
ited calibration laboratories. Corrective actions are
obviously the actions you take to correct the situa-
tion, while preventive actions are all the actions you RELATED ARTICLE
take to prevent the same situation from happening
again in the future. The effectiveness of corrective
and preventive actions is important to review. Also, How often should instruments be calibrated?
all other similar instances should be considered to https://hubs.ly/H09rXwL0
see if there is any possibility for similar occurrences
Calibration Uncertainty
For Non-mathematicians
By: Heikki Laurila
This paper discusses the basics

of uncertainty in measurement
and calibration. It is not made
for mathematicians or metrolo-
gy experts, but rather for those
of you who are planning and
making practical measurements
and calibrations in industrial
applications.
Being aware of the uncertainty

related to the measurement is
a very fundamental concept. You should not really More than likely, as a result, you will get three some-
make any measurements unless you are aware of the what different answers, such as:
related uncertainty. Generally speaking, it seems that
the awareness and interest of uncertainty is growing, • The first person says it is about 60 cm. He used
which is great. a 10 cm plastic ruler, measured the string once
and came to this conclusion.
The uncertainty of measurements can come from
various sources, such as the reference measurement • The second person says it is 70 cm. He used a
device used for making the measurement, from envi- three meter measuring tape and checked the
ronmental conditions, from the operator making the results a couple of times to make sure he was
measurements, and from many other sources. right.
Being aware of the uncertainty • The third person says it is 67.5 cm, with an
related to the measurement is a very uncertainty of ±0.5 cm. He used an accurate
fundamental concept. You should measuring tape and measured the string multiple
not really make any measurements times to get an average and standard deviation.
unless you are aware of the related Also, he tested how much the string stretches
uncertainty. when it is pulled and noticed that this had a
small effect on the result.
Classic “piece of string” example
Even this simplified example shows that there are
Let’s start with an example to illustrate the measure-
many things that affect the result of a measurement;
ment uncertainty in practice; the example is to give
the measurement tools that were used, the method/
the same piece of a string to three persons (one at
process that was used and the way the person did the
a time) and ask them to measure the length of that
job.
string. With no additional instructions given. They
So, the question you should be asking yourself is:
can all use their own tools and methods to measure
When calibration work is performed at your plant,
it.
which of these three examples sound the most famil- being aware of all the related uncertainties?
iar to you?
What kind of “rulers” are being used at your site and The first simple, yet good, practice is that when you
what are the measuring methods/processes? normally make a measurement/calibration once,
If you just measure something once without know- try instead to repeat the same measurement several
ing the related uncertainty, the result is not worth times. Most likely you will discover small differences
much. in the measurements between the repeats. But which
measurement is the correct one?
Very short terminology course
Without diving too deep into statistics, we can say
Let’s take a very brief consider the essential terms that it is not enough to measure only once. If you
related to this subject. repeat the same measurement several times, you can
find the average and the standard deviation of the
So, what is the uncertainty of measurement? We can measurement. So you will learn how much the re-
simply say that it is the “doubt” of our measurement, sults can differ between repeats. This means that you
meaning that it tells us how good our measurement can find out what is the normal difference between
is. Every measurement we make has some “doubt”, measurements.
and we should know how much this “doubt” is to
decide if the measurement is good enough for the It is suggested to make a measurement multiple
purpose. times, even up to ten times, for it to be statistically
It is good to remember that error is not the same enough reliable to calculate the standard deviation.
as uncertainty. When we compare our device to be These kind of uncertainty components, that you get
calibrated against the reference standard, the error is by calculating the standard deviation, are called the
the difference between these two measurements. But
the error does not have any meaning unless we know
the uncertainty of the measurement.
So, I would like to say that:
If you don’t know the uncertainty of the measure-
ment, don’t make the measurement at all!
Too often we have seen, for example, that when a
person is making an important temperature mea-
surement in his process with, say, ±1.0 °C acceptance
limit, and finds a maximum error of 0.5 °C, he is
happy and says it “passes” and accepts the result.
Although, after analyzing the calibration process,
he could find that the total uncertainty of his mea-
surement process is ±2.0 °C. Therefore, the way the
calibration was done was not good enough for this Diagram: Standard deviation
application.
But as long as he did not know/care about the un- A-type uncertainty.
certainty, he could claim that it was a good “pass- You may say: What??? - Always repeating the same
ing” calibration, although in reality, it failed. measurement ten times is just not possible in prac-
tice!
From making a single measurement to knowing Luckily you don’t always need to make ten repeats,
your standard deviation but you should still experiment with your measure-
ment process by sometimes making several repeats
So, what should you do to start the journey towards of the same measurement. This will tell you what the
typical deviation of that whole measurement process mation as the uncertainty of your reference standard
is and you can use this knowledge in the future as an in your calibrations.
uncertainty component related to that measurement,
even if you just make the measurement once during Other uncertainty sources
your normal calibration.
Imagine that you would perform a temperature In the previous section I suggested that you repeat
measurement/calibration multiple times and you the measurement several times. But how about if
would learn that there could be a ±0.2 °C difference you ask a few of your colleagues to repeat that same
between the repeats. Next time you would make the measurement? Do you all get the exact same results?
same measurement, even if you would then make Often there are some
it just once, you would be aware that there is this differences between the
±0.2 °C possible difference, so you could take it into different persons making
account and don’t let the measurement go too close the measurement. So, does
to the acceptance limit. it mean that the person
So if you keep calibrating similar kinds of instru- making the measurement
ments over and over again, it is often enough to make also have an effect to
the measurement just once and use the typical exper- uncertainty? – yes, it does.
imental standard deviation. Of course you need to do This is especially the case
your homework and make the measurements and the if the instructions are not
calculations to find out the typical standard deviation at an appropriate level.
of that instrument type and that calibration process. What if you make the
In summary, you should always be aware of the same test and this time
standard deviation of your calibration process – it is you use different kind of
one part of the total uncertainty. reference standards (cali-
brators) to make the mea-
Your reference standard (calibrator) and its surement? Again, most
traceability likely you will find differ-
ences. And if the reference
Often, one of the biggest sources of uncertainty standards have different
comes from the reference standard (or calibrator) levels of accuracy (uncer-
that you are using in your measurements/calibra- tainty) you may even see
tions. Naturally to start with, you should select a relatively big differences.
suitable reference standard for each measurement. Often the reference stan-
It is also important to remember that it is not enough dard (or calibrator) used
to use the manufacturer’s accuracy specification to make the measurement
for the reference standard and keep using that as can be one of the biggest sources of uncertainty!
the uncertainty of the reference standards for years. Different environmental conditions may add addi-
Instead you must have your reference standards cali- tional uncertainty in certain calibrations.
brated regularly in a calibration laboratory that has If you need to read some form of analog display
sufficient capabilities (uncertainty small enough) to (analog gauge, temperature meter), you have lim-
calibrate the standard and to make it traceable. Pay ited readability, i.e. you can only read it to certain
attention to the total uncertainty of the calibration accuracy and there is a possibility to read it incor-
that the laboratory documented for your reference rectly (wrong viewing angle) which ads uncertainty.
standard. Also, you should follow the stability of your In case of digital readouts, the resolution (number of
reference standards between its regular calibrations. decimals) is always limited, which causes uncertainty
After some time, you will learn the true uncertainty of (you can only read to the last decimal).
your reference standard and you can use that infor- There are different technical aspects in the calibra-
tion process, applications and quantities that create anyhow easier to understand this with a relatively
additional uncertainties. For example in temperature simple formula:
calibration, it is imperative to wait long enough for
the temperature to stabilize and to ensure proper Total uncertainty =
probe immersion into temperature block; in flow
calibration you need to ensure a stabile flow; in pres-
sure calibration you must avoid any leaks and have
a stabile pressure, etc. Generally, any fluctuations or
changes in the variable to be measured will cause
additional uncertainty.
There are also some random variables that throw in
some additional spices to the soup. Where each “u” is one independent uncertainty
Also, you can use the experimental standard devi- component.
ation mentioned earlier as one uncertainty compo-
nent. Coverage factor/confidence level
So we can shortly summarize these additional
sources of uncertainty: When uncertainty is determined, it is typically
• Device under test multiplied with a coverage factor (k). Most often the
• Reference standard (calibrator) combined uncertainty is multiplied with 2 (k=2 or 2
• Method/process for making the measurements/ sigma). This multiplication is done in order to have
calibrations greater confidence level of the result. When the cov-
• Environmental conditions erage factor of 2 is used, it equals a confidence level
• The person(s) making the measurements of 95%. This is done because we are dealing with
• Additional uncertainty components depending statistical data and according normal (Gaussian)
on the quantity being measured/calibrated distribution 95% of the results are within the 2 sigma
range. So in practice, using the 2 sigma, 95% of the
All of these above listed uncertainty components are results will be within the given uncertainty budget.
referred Different sigma values give the following confidence
as the Type B uncertainty. levels:
• 1 sigma (k=1) = 68% confidence level
Adding uncertainties together => combined (68% of the results are within)
uncertainty • 2 sigma (k=2) = 95% confidence level
• 3 sigma (k=3) = 99.7% confidence level
The type A (standard deviation) is something you
can calculate, but often some of the various type B Normal (Gaussian) distribution
uncertainties needed to be estimated. Once standard
deviation is calculated and the various Type B
uncertainties are estimated, it is time to add them
together. Before that you need to make sure that all
uncertainties are in the same quantity/unit. Also, the
uncertainties should be having the same coverage
factor/confidence level.
When you add together uncertainty components
that are independent from each other, don’t just sum
them all together, that would make a too pessimistic
(worst-case) result. Instead, add the components
together using the root sum of the squares method.
That means, square each component, then sum
them together and finally take a square root of the
total sum. Although I said no formulas, maybe it is
When you add different uncertainty components We can interpret these different above cases as fol-
together, make sure they are all the same 1 sigma lowing:
values before adding them.
• Case 1: This is pretty clearly within the tolerance
Expanded uncertainty limits, even when uncertainty is taken into account.
So we can state this as a good “Pass” result.
Before the combined uncertainty component is • Case 4: This is also pretty clear case. The result is
published, you need to multiply the result with the outside of the tolerance limits, even when uncertain
selected sigma value in order to get the required ty is taken into account. So we can state this being
confidence level. After you have done the multiplica- a bad or “Fail” result.
tion, what you get is called expanded uncertainty, i.e. • Case 2 and Case 3: These cases are a bit more dif
uncertainty with certain confidence level included. ficult to judge. Sure it seems that in case 2 the result
is within the tolerance while in case 3 it is outside,
How to express uncertainty in results or calibra- especially if you don’t care about the uncertainty.
tion certificate But taking the uncertainty into account, we can’t
really say this with confidence.
In your calibration results, you should express the
uncertainty as ± value and also mention the coverage There are regulations (for example; ILAC G8:1996
factor/confidence level. For example, you can say - Guidelines on Assessment and Reporting of Com-
that the temperature is: 20.5°C pliance with Specification; EURACHEM / CITAC
with uncertainty ±0.1 °C (k=2). Guide: Use of uncertainty information in compliance
assessment, First Edition 2007) for how to state the
Compliance statement – pass or fail compliance of calibration. These guides suggest
stating a result as passed only when the error added
Most often the calibration of an instrument includes with uncertainty is less than the acceptance limit.
an acceptance criteria, i.e. there are limits within Also, they suggest to state failed only when the error
which the result is considered being passed and added (or subtracted) with the uncertainty is bigger
outside of which it is considered being failed. There than the acceptance limit. When the result is closer to
are different interpretations if/how the uncertain- the acceptance limit than half of the uncertainty, it is
ty should be taken into account when deciding for suggested to be called an “undefined” situation, i.e.
Pass/Fail. Let’s use some examples to study differ- you should not state pass or fail.
ent cases. In the below picture, the diamond shape We have seen many people interpreting the uncer-
illustrates the measurement result and the line above tainty and pass/fail decision in many different ways
and below indicates the total uncertainty for that over the years. In practice, the uncertainty is most
measurement. often not taken into account in the pass/fail decision,
but it is anyway very important to be aware of the
uncertainty, when making the decision.
Uncertainty examples
In the graphics below, there are some examples of

what different uncertainties can mean in practice.
The cases 1 and 2 have the same measurement result,
so without uncertainty we would consider these be-
ing the same level measurements. But when the un-
certainty is taken into account, we can see that case 1
is really terrible because the uncertainty is simply too
large to be used for this measurement with the given
tolerance limits.
Looking at case 3 and 4 it seems that the case 3 is
better, but with uncertainty we can see that it is not
good enough for a pass statement, while the case 4 is.
TUR / TAR ratio vs. uncertainty calculation
Again, I want to point out that we need to know the The TUR (test uncertainty ratio), or TAR (test ac-
uncertainty before we can judge a measurement re- curacy ratio), is often mentioned in various pub-
sult. Without the uncertainty calculation, the above lications. In short, this means that if you want to
cases 1 and 2 look similar, although with uncertainty calibrate a 1% instrument and you want to have 4:1
taken into account they are very different. ratio, your test equipment should be 4 times more
accurate, i.e. having 0.25% accuracy, or better. Some
A real-life example publications suggest that having a TUR/TAR ratio
large enough, there is no need to worry about uncer-
Below is a real-life example where the same RTD tainty estimation/calculation. The quite commonly
temperature transmitter has been calibrated using used ratio is 4:1. Some guides/publications do also
two different calibrators. This graphics were pro- have recommendations for the ratio.
duced using Beamex CMX calibration management Most often the ratio is used as in the above example,
software. You can easily see that in the first case, i.e. just to compare the specifications of the DUT
the results is very good and also the green vertical (device under test) and the manufacturer’s specifica-
uncertainty line is very short indicating a very small tions of the reference standard. But in that scenario
uncertainty. In the second case you can see that the you only consider the reference standard (test equip-
result is a little bit worse, but the uncertainty of that ment, calibrator) specifications and you neglect all
calibrator is much worse. other related uncertainties. While this may be “good
Well, needless to say, that the first case is done with enough“ for some, calibrations, this system does
a Beamex calibrator… ;-) not take some of the biggest uncertainty sources
Anyhow, when you see the uncertainty graphically it into account. So it is highly recommended to make
is very easy to notice the significance of it. the uncertainty evaluation/calculation of the whole
calibration process.
We also get asked quite regularly: “How many times
more accurate should the calibrator be, compared
SUMMARY
to the device to be calibrated?”. While some sugges-
tions could be given, there isn’t really a correct an- I hope this paper helped to give some practical
swer to that question. Instead you should be aware understanding of the uncertainty subject.
of the total uncertainty of your calibrations. And of
To very shortly summarize the key take-outs of
course, it should reflect to your needs!
some of the main topics:
• Be sure to distinguish “error” and “uncer-

tainty”
SOME USEFUL RELATED RESOURCES
• Experiment by making multiple repeats of
measurements to gain knowledge of the
• EA-4/02 - Evaluation of the Uncertainty of typical deviation
Measurement in Calibration
• Use appropriate reference standards (cal-
ibrators) and make sure they have a valid
• ILAC G8:1996 - Guidelines on Assessment traceability to national standards and that
and Reporting of Compliance with Specifi-
the uncertainty of the calibration is known
cation
and suitable for your applications
• EURACHEM / CITAC Guide: Use of uncer- • Consider if the effect of the environmental
tainty information in compliance assess- conditions have a significant effect to the
ment, First Edition 2007 uncertainty of your measurements
• ISO/IEC 17025:2005 - General requirements • Be aware of the readability and display res-
for the competence of testing and calibration olution of any indicating devices
laboratories
• Study the specific important factors of the
quantities you are calibrating
• ISO 9001:2015 - Quality management sys-
tems -- Requirements • Familiarize yourself with the “root sum of
the squares” method to add independent
• ISO 10012:2003 - Measurement manage- uncertainties together
ment systems-- Requirements for measure-
ment processes and measuring equipment • Be aware of the coverage factor / confi-
dence level / expanded uncertainty, of the
uncertainty components
• JCGM 101:2008 - Evaluation of measure-
ment data Guide to the expression of uncer- • Instead, or in addition to the TUR/TAR ra-
tainty in measurement tio, strive to be more aware of all the related
uncertainties
• Pay attention to the total uncertainty of the

RELATED WEBINAR calibration process before making pass/fail
decisions
Calibration uncertainty and why technicians If you have any comments or questions, and I
need to understand it hope you do, we are very happy to hear from you!
Contact us, www.beamex.com or [email protected]
https://hubs.ly/H09rYqg0
•• • ••
• • • • • •• • • • ••
Calibration Intervals ••
••• • • • • • • ••
By: Mike Cable
Probably the #1 question asked at calibration sem-

inars is “How do I determine the initial calibration
intervals?” The answer to this question is
difficult at first for someone new to developing a
calibration program. It ends up being pretty simple.
Initially we try to use a variety of resources
which include:
• Manufacturer recommendation
• National Conference of Standards Laboratories
Recommended Practice RP-1
• Past experience
• Intervals of similar existing instruments
In reality, it is a combination of all the above, but

mostly past experience. As an example, in my experi- Don’t be alarmed if you calibrate more or less often
ence, electronic transmitters have a calibration inter- than described above. It simply means you’re willing
val of 6 months and analog gauges have an interval of to take more or less risk based on the process and
a year. Many manufacturers’ specifications contain a quality standards at your facility or you have more
6-month stability specification. This stability specifi- history to base your calibration intervals on. Of
cation, in effect, only guarantees the accuracy spec- course, not all instruments fit into the same category.
ification for 6 months. Also, electronic transmitters Some instruments, particularly analytical instrumen-
are typically installed in applications that are “more tation, are calibrated more frequently, even to the
important” to the process. Even though these instru- point that the user performs a calibration check prior
ments are more reliable than analog gauges and fail to each use. On the other hand, some instruments
calibration less often, we check the calibration on a may have an interval of two years or more.
more frequent basis. This means we set our calibra-
tion intervals based on how much risk we are willing Calibration intervals may be adjusted over time.
to take. If we wanted an almost 100% assurance Once several calibrations have been performed, the
that our instruments were within calibration toler- calibration history of the device may be used to ad-
ance, we’d have to check the calibration almost every just the calibration interval. If the as-found calibra-
day. Obviously, that would be impracticable. So we tion data of a particular instrument does not meet
assume some risk that every once in a while a cali- the calibration tolerance, the calibration interval may
bration is not going to pass. Of course, our managers be shortened. If several instruments with a particular
and quality department don’t want to hear that, but manufacturer/model number are always well within
it happens and we need to educate other disciplines the calibration tolerance, the interval is increased.
that it does happen.
Criteria for Instrument Selection in Relation to the
Measurement Requirements
1 Introduction equate or it has excessive tolerance or insufficient
stability
The organization, in congruence with what is defined, • The likelihood of the onset of such nonconformity
documented, and regulated, must meet: is not large enough to be immediately detected
and eliminated, if not before delivery to the cus
• The quality objectives for the product tomer and or use
• The contractual requirements of the customer • The relationship between the amplitude of the
• The regulatory requirements of the market tolerance allowed on the product (process or
service) and the measuring equipment uncertain
It must identify and define which variables or charac- ty is too small (e.g., < 3)
teristics of the product (process or service) should be • The characteristic measured in a stage of produc
considered critical. For example, a variable, character- tion is mandatory, in that it could compromise its
istic, or measure can be considered critical when: application characteristics (e.g., for a manufac
tured mechanical device)
• The consequences of nonconformity of the • The characteristic measured in the testing phase
measured characteristic are serious and costly is “determining,” it could impair its metrological
(e.g., when it concerns the areas of security, characteristics (e.g., for a measuring instrument)
health, and the environment)
• The likelihood of nonconformity is not negligible, Once the organization defines variables, charac-
because the capacity of the process is not ad- teristics, and critical measures, it should design the
measurement processes suitable for the control least the following
of the product, process, or service, to prevent any elements:
nonconformity, in accordance with the requirements
prescribed in 7.2 of ISO 10012. • Manufacturer
• Type and description
Therefore, it must specify, document, and identify • Series or serial number
the relative measurement process, at least the follow- • Measuring range
ing elements: • Measuring resolution
• Measurement accuracy
• The type of product • Measurement uncertainty
• The measurement method • Functional operating conditions
• The measuring equipment • Environmental operating conditions
• The environmental conditions of measurement • Stability or possible eventual drift
• The ability and qualifications of the operating • Sensitivity to any influential quantities
personnel
Further, before being inserted into the measurement
In addition, for each measurement process, it must process, it must be submitted to the calibration pro-
identify and quantify the following characteristics cess and metrological confirmation provided for in
described in point point 7.1 of ISO 10012.
7.2.2 of ISO 10012:
3 Reference Equipment Selection
• Measuring range
• Measurement uncertainty Calibration of the measurement apparatus (also
• Maximum Permissible Error called instrument) may be made externally or inter-
• Stability, repeatability, and reproducibility nally within the organization with reference equip-
• Any other specific feature of interest ment (also called standard) having a measurement
uncertainty possibly at least three times lower than
Finally, remember that each individual process must the presumed uncertainty of the instrument to be
be validated before use and should be checked reg- calibrated.
ularly and continuously during operation to prevent
nonconformity over time. The relationship between the measurement uncer-
tainty between the instrument to be calibrated and
2 Measurement Equipment Selection the reference standard (also called test uncertainty
ratio or TUR) should also be possibly greater than
The equipment for measuring relative to the de- 3, but 3 could be sufficient given the treatment of
signed measurement processes should
be first chosen, then calibrated, and
then confirmed before being inserted
in the measurement process, according
to the requirements prescribed in point
7.1 of ISO 10012.
The initial selection must be made

according to the metrological charac-
teristics required for the above mea-
surement process, described in the in-
troduction, while the initial calibration
can be done directly from the supplier,
externally by a qualified laboratory, or
even internally at the company’s labo-
ratory, but it must always be traceable
to the International System (SI). In any
case, the measuring equipment must
be specified and documented for at
uncertainties. In fact, since the treatment of squared Therefore, see the recommended uncertainty ratios
uncertainties, the single contribution due to the of table 9-1, except in the following situations:
standard is approximately 1/10 compared to the con-
tribution of any instrument errors, and therefore, it • Cases where there is a technical normative or
increases the result of the instrument measurement regulation that specifies different minimum
uncertainty < 5% (see point 10). requirements
• Cases where it is not possible to observe the
On the other hand, more ratios between the mea- adequacy of the minimum acceptable ratio of 3,
surement uncertainties of the instrument and the as it works with measuring instruments with
standard certainly improve the resulting uncertainty. uncertainty slightly lower than the reference
They surely also increase the purchase and main- standards (this, of course, will increase the result
tenance costs of the reference standard, and of its ing uncertainty)
calibration, especially if this is carried out externally
to the organization.
Table 9-1. Adequacy of the Ratio of Instrument Uncertainty to be Calibrated and Standard
Intrument uncertainty
Adequacy
Standard uncertainty
Table 9-1 may also be the reference for determining the adequacy of the relationship between the tolerance band
of the product (process or service) and the measurement uncertainty (U) of the instrument used for the relative
measurement (point 9.4).
Example of Choice of Measuring Instrument in Compliance with Product Requirements
As an example, a typical exemplification of overexposed methods and criteria is illustrated in figure 9-1, concern-
ing the selection, calibration, and metrological confirmation of measuring instruments, used in a typical produc-
tion process.
Increasing measurement uncertainty
Product
U U U U
Reference
Value
Measured Variable
Ambiguity Ambiguity
zone zone
Not conform Conformity zone Not conform
Tolerance Band Tolerance Band
Product Tolerance Amplitude
Figure 9-1. Example of Typical Control of a Product (Process or Service)
Consider as an example the one provided in Annex A of ISO 10012:
For a critical operation in a process reactor, it is required that the pressure remains between 200 and 250 kPa. That
is then considered a critical variable to be controlled by means of a relative measurement process.
The design of the measurement process that satisfies this critical requirement of the customer can be developed
in the following phases:
1. Set a process measuring range more widely between 150 and 300 kPa.
2. Establish a Maximum Permissible Error of the manometer equal to 2 kPa.
3. Choose a suitable manometer accuracy class of 0.5% with a range of 0–400 kPa.
4. Calibrate the manometer with a reference standard with uncertainty of 0.3 kPa.
5. Metrologically confirm the manometer prior to use in the measurement process.
Obviously, the metrological confirmation of the manometer is “passed” if the maximum error found in the cali-
bration is less than or equal to the Maximum Permissible Error of 2 kPa!
This previous example can be schematized in the following way in relation to the critical requirements illustrated
in figure 9-1:
Input elements:
• Product Reference Value (PRV) @ 225 kPa

• Product Tolerance Amplitude (PTA) 200 – 250 kPa
• Product Tolerance Band (PTB) ± 25 kPa
• Customer Metrological Requirements (CMR) Err. ≤ 2 kPa
• Measuring Equipment Metrological Characteristics (MEMC) Err. ≤ 2 kPa
• Reference Equipment Metrological Characteristics (REMC) Unc. ≤ 0.3 kPa
Output element:
• Metrological confirmation passed if: MEMC ≤ CMR

Some considerations for the example:
If one compares the specific ratio of the Product Tolerance Band in respect to the Maximum Permissible Error,
or PTB/ CMR, and the specific ratio of the instrument Maximum Permissible Error with respect to the reference
Maximum Permissible Uncertainty, or MEMC/REMC, with typical ratios presented in table 9-1, one can get the
conclusions highlighted in the last column of the following prospectus:
Relations between the input elements:
Typical Annotation
PTB / CMR = 25 / 2 = 12.5 3 – 10 Excessive
MEMC / REMC = 2 / 0.3 = 6.6 3 – 10 Normal
Obviously, relations between the input elements considered excessive over 10, may still be considered practicable
if they do not involve unreasonable additional costs for purchase and management of the measurement. This is
because they improve the product and control of the measuring process, and thus minimize the eventual zones
of ambiguity corresponding to the measurements of the amplitude limits of the product tolerance (see figure 9-1).
Calibrating
WirelessHART
Transmitters
By: Heikki Laurila
WirelessHART transmitters are becoming more wired HART transmitter is compatible with analog
popular. What are they and how do they differ from control systems.
wired HART transmitters? Why do the WirelessHART
transmitters need to be calibrated and how is the HART is a digital communication
calibration done? protocol that enables
communication with a field device.
A very brief history of HART
WirelessHART
The HART (Highway Addressable Remote Transduc-
er) protocol was developed in the mid-1980s by Rose- WirelessHART was approved and ratified by the HCF
mount Inc. for use with a range of smart measuring Board of Directors, and introduced to the market in
instruments. Originally proprietary, the protocol September 2007, becoming the first officially released
was soon introduced for free use, and in 1990 the industrial wireless communication standard. The
HART User Group was formed. In 1993, the regis- WirelessHART network uses IEEE 802.15.4 compati-
tered trademark and all rights in the protocol were ble radios operating in the 2.4GHz radio band. Each
transferred to the HART Communication Foundation device in the mesh network can serve as a router
(HCF). The protocol remains open and free for all to for messages from other devices. The WirelessHART
use without royalties (Source: HCF). HART is a digital transmitter does not have an analog mA signal. It
communication protocol that enables only has the digital signal which is available wireless-
communication with a field device. The commu- ly, or through a screw terminal.
nication allows you to read and write settings, read Since the transmitter is wireless, power cannot be fed
measurement results, receive diagnostic data, etc. via cables; instead, the transmitter needs a battery
for power. The battery life and communication speed
Wired HART signal are inversely proportional. Sometimes wireless trans-
mitters can be programmed not to send a wireless
The wired HART Protocol uses Frequency Shift Keyed signal very often which lengthens the lifespan of
(FSK) digital communication signal superimposed the batteries. The communication speed can also
on top of the 4-20mA standard analog signal. The be increased if necessary. It is possible to use Wire-
lessHART even on a control circuit. In practice, the logical calibration. Configuring parameters of a HART
WirelessHART transmitters are usually used in mon- transmitter with a communicator is not metrological
itoring applications, which tend to change slowly, as calibration and it does not assure accuracy. For a real
well as in applications that are difficult to wire. metrological calibration, a traceable reference stan-
Any existing wired HART transmitter can also be dard (calibrator) is always needed.
made wireless by adding the wireless adapter avail-
able from many instrument manufacturers. If the How to calibrate a wired HART transmitter
control system is analog, reading only the mA signal,
an additional WirelessHART host system can be built It is good to remember that a HART transmitter has
to process all of the additional information available two different outputs that can be used and calibrated:
in the HART devices. This can include information the analog mA output and the digital HART output. In
that is not available via the analog control system, most cases, customers still use the analog output.
for example, advanced diagnostics and predictive To calibrate the analog output, generate or measure
maintenance. the transmitter input and at the same time measure
the transmitter output. A dual function calibrator
HART status and the future able to handle transmitter input and output at the
same time is needed, or alternatively two separate
Over 30 million HART devices are installed and in single-function calibrators; for example, if someone
service worldwide. The wired HART technology is the wants to generate a pressure input and measure it
most widely used field communication protocol for accurately with a calibrator and at the same time
intelligent process instrumentation. The HART share measure the analog mA output with an mA meter.
equals nearly half of the installed base of intelligent The calibration process changes slightly if you want to
transmitters. Various studies estimate growth for calibrate the digital HART output. Obviously, it is still
HART in the future as well. The new WirelessHART needed to generate/measure the transmitter input
standard seems to be a new booster for the HART pro- the same way as for an analog transmitter, using a
tocol. Data from studies predicts exponential growth calibrator. To see what the transmitter digital HART
for WirelessHart over the next 10 years. output is, some kind of HART communicator with the
ability to show the digital HART signal is needed. A
What is meant by “calibration” HART transmitter can have several digital variables
depending on the transmitter type. In the case of ana-
According to international standards, calibration is a log or digital output, you would progress through the
comparison of the device being tested against a trace- range of the transmitter at a few points and record the
able reference instrument (calibrator) and documen- input and output signals to document the calibration.
tation of this comparison. Although calibration does
not formally include any adjustments, in practice,
adjustments are possible and often included in the
calibration process.
What is meant by “configuration”
Configuration of a HART transmitter means changing

the transmitter settings and parameters. The con-
figuration is typically done with a HART communi-
cator or with configuration software. It is important
to remember that although a communicator can be
used for configuration, it cannot be used for metro-
How to calibrate a WirelessHART transmitter with a HART communicator that is able to read the
transmitter via the screw terminals. Since the Wire-
Firstly, it is good to remember that, although the lessHART transmitters are made according to the
WirelessHART transmitter has a different output than HART7 standard protocol, a communicator able to
the wired HART transmitter, the WirelessHART trans- support the HART7 standard is needed. If there is a
mitter also needs to be calibrated. As the calibration separate calibrator for the input and communicator
verifies the transmitter accuracy, i.e. the relationship for the output, the readings will need to be manually
between the physical input and transmitter output, written down and the calibration documented. How-
the need for calibration does not change, whether ever, if there is a calibrator and communicator built
wireless or wired, digital or analog. into one device, the input and output can be handled
The input of a WirelessHART transmitter needs
to be generated (or measured) the same way as the Wired and wirelessHART transmitter
analog or wired HART transmitter, using a refer-
ence standard or a calibrator. The output of the
transmitter needs to be read at the same time. A
WirelessHART transmitter does not have any analog
output; it only has a digital output. The digital output
can be read in two different ways.
One way is to read the output signal wirelessly, but
the wireless signal can be very slow. Depending on
the transmitter configuration, it may be transmitting
All of the WirelessHART transmitters

also have screw terminals allowing
a wired connection with the
transmitter.
its output only once per minute. In any case, the
wireless signal is not really suitable for calibration.
For example, in the case of a pressure transmitter
calibration, there may be small leaks in the pressure
connections or hoses, causing the input to change
rather frequently. If the output is read very seldom,
there could be a significant uncertainty and error simultaneously with the same device. If the device
between the saved calibration input and output data. also has a documenting feature, the calibration
Also, if there is any need to trim (adjust) the transmit- can be automatically documented without paper.
ter, or make any other configurations, these cannot If a wired HART transmitter needs to be trimmed,
be done wirelessly. the sensor section (A/D conversion), as well as the
All of the WirelessHART transmitters also have analog (D/A conversion) section, will also need to be
screw terminals allowing a wired connection with the trimmed. In the case of the WirelessHART transmit-
transmitter. While being connected via the screw ter- ter, there is no analog section, so it is enough to trim
minals, the digital output can be read quickly enough the sensor section.
for calibration purposes and any configuration or
methods, such as trimming methods, are accessible. Why calibrate
Therefore, the WirelessHART transmitter should be
calibrated with a wired connection to the transmit- A modern transmitter is advertised as being smart
ter’s screw terminals. and very accurate. Sometimes people may say that
The input can be generated or measured with a there is no need for calibration at all because the
reference calibrator. The output needs to be read transmitters are so “smart.” Why should smart trans-
mitters be calibrated then?
First of all, changing of the output protocol of a
Example
transmitter does not change the fundamental need for
calibration. Let’s take an example of calibrating an Emerson
There are numerous reasons to calibrate instruments 648 WirelessHART temperature transmitter. The
initially and periodically. The main reasons are: transmitter is configured for RTD measurement
with sensor type Pt100 (Alpha385). Disconnect
• Even the best instruments do drift with time,
the RTD sensor and connect the MC6 to simulate
especially when used in demanding processing
the RTD sensor. Connect the MC6’s HART termi-
conditions.
nal to the transmitter’s screw terminals and con-
• Regulatory requirements, such as quality systems,
figure the MC6 to read the Primary Variable (PV)
safety systems, environmental systems, standards,
of the transmitter, which is the digital output. The
etc.
range to be calibrated is 0 °C to 100 °C (32 °F to
• Economic reasons: any measurement has direct
212 °F). Configure the MC6 to progress the input
economic effects.
signal from 0 to 100 °C (32 °F to 212 °F) in steps of
• Safety reasons: employee safety as well as custom
25%, stepping up and down. Then configure the
er/patient safety.
MC6 to wait 10 seconds in each step to allow the
• To achieve high and consistent product quality and
transmitter to stabilize. Of course, the transmit-
to optimize processes.
ters damping should be taken into account when
• Environmental reasons.
deciding the calibration delay. In completing
these steps, we have programmed the maximum
The Beamex MC6 field calibrator and commu-
error tolerance to 0.5% of the full scale.
nicator
When the connections are complete, calibra-
tion can begin. The calibration will go through
The new Beamex MC6 is a device that combines a field
the required input steps fully automatically, stop-
communicator and an extremely accurate multifunc-
ping for the delay, and then going on to the next
tional process calibrator.
step. Once the calibration is completed, a dialog
With the Beamex MC6, the smart transmitter’s input
will appear stating whether the calibration was
can be generated/ measured at the same time the
successful or not (Pass/Fail). Next, save the cali-
digital output is read. Thus, they can be done simul-
bration into the MC6’s memory. Later on, upload
taneously and the results can be automatically stored
the calibration results to calibration manage-
into the MC6 memory for later viewing or uploading to
ment software to be saved in the database and
calibration software.
print a calibration certificate if necessary.
For configuration of the smart transmitters, the MC6
If the As-Found calibration failed, or you want
includes a field communicator for HART, WirelessHART,
to trim the transmitter, you can use MC6 HART
FOUNDATION Fieldbus H1 and Profibus PA protocols.
communication. While trimming, it is possible to
All required electronics are built-in, including power
simultaneously simulate the required input with
supply and required impedances for the protocols.
the MC6, so no other device is needed. Once the
The Beamex MC6 can therefore be used both as a com-
calibration is completed, run another automat-
municator for configuration and as a calibrator to cali-
ic calibration procedure to perform an As-Left
brate smart instruments with the supported protocols.
calibration.
While a normal HART communicator can be used to
configure and read the HART digital output, it alone
cannot be used to calibrate or trim transmitters. You
will need an additional calibrator for that purpose, RELATED VIDEO
which leads to a situation where you need two separate
devices, which lack the automatic calibration proce-
dure and documentation. Therefore, a device such as How to calibrate HART pressure transmitters
the Beamex MC6, is superior for calibration of wired or https://hubs.ly/H09rYYv0
wireless HART transmitters.
Configuring and Calibrating
Smart Instruments
By: Heikki Laurila
So called “smart” instruments are ever more popular In any case, for the engineers who need to configure
in the process industry. The vast majority of deliv- and calibrate the transmitter, the digital communi-
ered instruments today are smart instruments. These cation protocol is the biggest difference compared to
new smart instruments bring new challenges to the conventional transmitters. Engineers can no longer
calibration and configuration processes. But what are simply measure the output analog signal, but they
these smart instruments and what is the best way to need to have the possibility to communicate with the
configure and calibrate them? transmitter and read the digital signal. That brings a
Beamex has recently introduced a new revolutionary whole new challenge - how can the digital output be
tool, the Beamex MC6 –Advanced Field Communica- read?
tor and Calibrator, that will help to overcome these Thinking of the opposite of a smart transmitter, i.e. a
challenges. non-smart transmitter, would be a transmitter with a
purely analog (or even pneumatic) output signal.
What is a “Smart” transmitter?
Smart transmitter protocols
A process transmitter is a device that senses a physi-
cal parameter (pressure, temperature, etc.) and gen- There are various digital protocols that exist among
erates an output signal proportional to the measured transmitters considered smart. Some are proprietary
input. The term “smart” is more of a marketing term protocols of a certain manufacturer, but these seem
than a technical definition. There is no standardized to be fading out in popularity and favor is being given
technical definition for what smart really means in to protocols based on Open Standards because of the
practice. interoperability that they enable.
Generally, in order for a transmitter to be called
smart, it will utilize a microprocessor and should Most of the protocols are based on open standards.
also have a digital communication protocol that can The most common transmitter protocol today is the
be used for reading the transmitter’s measurement HART (Highway Addressable Remote Transducer)
values and for configuring various settings in the protocol. A HART transmitter contains both a con-
transmitter. A microprocessor-based smart trans- ventional analog mA signal and a digital signal super-
mitter has a memory that can perform calculations, imposed on top of the analog signal. Since it also has
produce diagnostics, etc. Furthermore, a modern the analog signal, it is compatible with conventional
smart transmitter typically outperforms an older type installations. Recently the HART protocol seems to be
of conventional transmitter regarding measurement getting more boosts from the newest WirelessHART
accuracy and stability. protocol.
The fieldbuses, such as FOUNDATION Fieldbus and
Profibus, contain only a digital output, no analog sig-
nal. FOUNDATION Fieldbus and Profibus are gaining
a larger foothold on the process transmitter markets.
This article will discuss “smart” transmitters, includ-

ing HART, WirelessHART, FOUNDATION Fieldbus
and Profibus PA protocols.
Configuration
One important feature of a smart transmitter is that

it can be configured via the digital protocol. Config-
uration of a smart transmitter refers to the setting of
the transmitter parameters. These parameters may
include engineering unit, sensor type, etc. The con-
figuration needs to be done via the communication
protocol. So in order to do the configuration, you
will need to use some form of configuration device,
typically also called a communicator, to support the
selected protocol.
It is crucial to remember that although a commu-
nicator can be used for configuration, it is not a
reference standard and therefore cannot be used for
metrological calibration. Configuring the parameters
of a smart transmitter with a communicator is not in
itself a metrological calibration (although it may be
part of an Adjustment/Trim task) and it does not as-
sure accuracy. For a real metrological calibration, by
definition a traceable reference standard (calibrator)
is always needed.
time, or alternatively two separate single-function
Calibration of a smart transmitter calibrators.
But how can a smart transmitter, with output being
According to international standards, calibration is a a digital protocol signal, be calibrated? Obviously the
comparison of the device under test against a trace- transmitter input still needs to be generated/mea-
able reference instrument (calibrator) and docu- sured the same way as with a conventional transmit-
menting the comparison. Although the calibration ter, i.e. by using a calibrator. However, to see what the
formally does not include any adjustments, potential transmitter output is, you will need some device or
adjustments are often included when the calibration software able to read and interpret the digital proto-
process is performed. If the calibration is done with a col. The calibration may, therefore, be a very chal-
documenting calibrator, it will automatically docu- lenging task; several types of devices may be needed
ment the calibration results. and several people to do the job. Sometimes it is very
To calibrate a conventional, analog transmitter, you difficult or even impossible to find a suitable device,
can generate or measure the transmitter input and especially a mobile one, which can read the digital
at the same time measure the transmitter output. output.
In this case calibration is quite easy and straight Wired HART (as opposed to WirelessHART) is a
forward; you need a dual-function calibrator able hybrid protocol that includes digital communication
to process transmitter input and output at the same superimposed on a conventional analog 4-20mA out-
About Beamex MC6
Beamex® MC6 is an advanced, high-accuracy field

calibrator and communicator. It offers calibration
capabilities for pressure, temperature and various
electrical signals. The MC6 also contains a full field-
bus communicator for HART, FOUNDATION Field-
bus and Profibus PA instruments.
The usability and ease-of-use are among the main

features of the MC6. It has a large 5.7” color touch-
screen with a multilingual user interface. The robust
IP65-rated dust-and water-proof casing, ergonomic
design and light weight make it an ideal measure-
ment device for field use in various industries, such
put signal. The 4-20mA output signal of a wired HART
as the pharmaceutical, energy, oil and gas, food and
transmitter is calibrated the same way as a conven-
beverage, service as well as the petrochemical and
tional transmitter. However, to do any configuration
chemical industries.
or trimming, or to read the digital output signal (if it
is used), a HART communicator is needed.
The MC6 is one device with five different operational
modes, which means that it is fast and easy to use,
and you can carry less equipment in the field. The
The solution
operation modes are: Meter, Calibrator, Document-
ing Calibrator, Data Logger and Fieldbus Commu-
The new Beamex MC6 is a device combining a full
nicator. In addition, the MC6 communicates with
field communicator and an extremely accurate
Beamex® CMX Calibration Software, enabling fully
multifunctional process calibrator. With the Beamex
automated and paperless calibration and documen-
MC6, the smart transmitter’s input can be generat-
tation.
ed/ measured at the same time as reading the digital
output. The results can be automatically stored into
In conclusion, the MC6 is more than a calibrator.
the memory of the MC6 or uploaded to calibration
software.
When it comes to configuration of the smart trans-
mitters, the MC6 includes a full field communicator
for HART, WirelessHART, FOUNDATION Fieldbus H1
and Profibus PA protocols. All required electronics
are built-in, including power supply and required
impedances for the protocols.
The Beamex MC6 can be used both as a communi-
cator for the configuration and as a calibrator for the
calibration of smart instruments with the supported
protocols. The MC6 supports all of the protocol com-
mands according to the transmitter’s Device Descrip-
tion file. Any additional communicator is therefore
not needed.
There are some other “smart” process calibrators on
the market with limited support for different proto-
cols, typically only for one protocol (mostly HART)
and offering very limited support. In practice, a sepa-
rate communicator is needed in any case.
WHY CALIBRATE?
A modern transmitter is advertised as being

smart and extremely accurate and sometimes
sales people tell you they don’t need to be calibrat-
ed at all because they are so “smart”. So why would
you calibrate them?
First of all, the output protocol of a transmitter
does not change the fundamental need for cali-
bration. There are numerous reasons to calibrate
instruments initially and periodically. A short
summary of the main reasons include:
• Even the best instruments and sensors drift

over time, especially when used in demanding
process conditions.
• Regulatory requirements, such as quality sys-

tems, safety systems, environmental systems,
standards, etc.
• Economical reasons – any measurement hav-

ing direct economical effect.
• Safety reasons- employee safety as well as cus-

tomer/patient safety.
• To achieve high and consistent product quality

and to optimize processes.
• Environmental reasons.
RELATED WEBINAR
Configuring and calibrating smart instruments

hubs.ly/H09rZ0t0
Loop Calibration
vs.
Individual Instrument Calibration
By: Mike Cable
An individual instrument calibration is a cal- Individual Calibration

ibration performed only on one instrument.
The input and output are disconnected. A Cons Pros
known source is applied to the input, and the
output is measured at various data points 1. Entire loop is not verified 1. Correct instrument will be
throughout the calibration range. The instru- within tolerance adjusted
ment is adjusted, if necessary, and calibration 2. Mistakes on re-connect 2. More compatible with
is checked. multifunction calibrators
3. Less efficient use of time
to do one calibration for
A loop calibration is performed from the each loop instrument as
opposed to one calibra
sensor to all loop indications with all the
tion for the loop
loop components connected. For example, a
temperature sensor connected to a tempera-
ture transmitter would be inserted in a tem-
perature bath/block. (Note: Either the bath/ Loop Calibration
block would be calibrated or a temperature
standard would be used in the bath/block for Pros Cons
traceability.) The temperature of the bath/
block would be adjusted to each data point 1. Entire loop, including 1. Wrong instrument may be
sensor, is verified within adjusted to bring the loop
required to perform the calibration. All local within calibration
tolerance
and remote indications would be recorded. It
is also recommended to record the transmit- 2. Mistakes on re-connect 2. Not as compatible with
minimized multifunction calibrators
ter output. If all indications and transmitter used for “paperless” data
output are within tolerance, the loop is within 3. More efficient use of time collection
to do one calibration for
tolerance. If any loop component is not with-
loop as opposed to one
in tolerance, then a calibration is performed calibration for each loop
on that instrument. Do not adjust a transmit- instrument
ter to correct a remote indication.
New Methods for Calibrating Loops
By: Ned Espy
Instrument technicians are following practices that effectively calibrate advanced instrumentation, like
were set up many years ago and it is not uncommon multivariable and smart/digital instruments. With
to hear, “this is the way we have always done it.” the complexity of instrumentation, there is more
Measurement technology continues to improve and pressure than ever on the calibration technician.
is becoming more accurate. Technicians with 30+ years’ experience at a single
plant are retiring and cannot easily be replaced by a
The typical approach to calibration has been to reg- younger technician or be properly outsourced.
ularly test instrumentation that influence effective
control, safe operation, quality or other relevant cri- The idea of a loop can mean
teria. In most cases, scheduling is conservative and different things to different people
methods at a particular site have slowly evolved over due to their work background and/
time. Instrument technicians are following practices or industry.
that were set up many years ago and it is not uncom-
mon to hear, “this is the way we have always done it.” Documentation requirements are becoming much
Measurement technology continues to improve and more common for improved quality, environmental
is becoming more accurate. It is also becoming more monitoring, and for adhering to government reg-
complex– why test a fieldbus transmitter with the ulations. Calibration software, like Beamex CMX
same approach as a pneumatic transmitter? Perform- calibration management software, is often required
ing the standard five-point, up-down test with an to store and analyze detailed data as well as to create
error of less than 1% or 2% of span does not always calibration certificates and reports. All of these fac-
apply to today’s more sophisticated applications. tors should cause scrutiny and evaluation of current
In general, calibration tasks require special skills practices. Simpler and more efficient test methods
and an investment in test equipment. Sophisticat- need to be considered to ensure proper plant opera-
ed, highly accurate and multifunctional calibration tion.
equipment, such as the Beamex MC6 advanced While not a new concept, there are advanced cali-
field communicator and calibrator, are required to bration techniques based on loop testing. In some
cases, it is best practice to perform individual in-
strument calibration to achieve maximum accuracy
(e.g. custody transfer metering). However, there are
viable methods where a loop can be tested end-to-
end and if readings are within acceptable tolerances,
there is no need to break into the loop for individual
instrument testing. To be effective, a common-sense
approach is required with the goal to minimize
downtime, maximize technician efficiency while
ensuring reliable control and maintaining a safe work
environment.
What is a loop?
A closed loop test is the best practice; if an accu-
rate temperature is made for the control process,
The idea of a loop can mean different things to
it does not matter how the individual instruments
different people due to their work background and/
are performing. The DCS/PLC value is what is used
or industry. In practice, a loop is simply a group of
to make any control changes, alarms, notifications,
instruments that in combination make a single mea-
etc. However, if the loop measurement has a signif-
surement or effect a control action in a process plant.
icant error, then the error of each instrument in the
A typical temperature example would be a tempera-
loop should be checked and corrected one by one in
ture element (RTD or T/C) that in turn is connected
order to bring the final measurement back into good
to a transmitter, which is connected in a series to a
operation.
local indicator and finally a control system input card
In some cases, it is not possible to make a closed
(DCS or PLC). The signal is then displayed on one or
loop test. In the example loop, it may be extremely
more control panels and the measurement is ulti-
difficult or expensive to remove the probe from the
mately used to control the process. When evaluating
process or the probe cannot be inserted into a tem-
a loop for testing, an important distinction to make is
perature block/bath. If this is the situation, then an
whether a closed loop test should be performed or an
open loop test can be performed where the tempera-
open loop test?
ture element is disconnected from the transmitter
A closed loop is an end-to-end test; in the tem-
and a temperature calibrator is used to simulate a
perature loop example (figure 1), the temperature
signal into the transmitter. As in the closed loop test,
element would need to be removed from the pro-
the final displayed measurement would be compared
cess and placed in a temperature block, such as the
to the simulated temperature and the error inter-
Beamex temperature blocks, or temperature bath in
preted, etc. While the loop is open, it would be good
order to simulate the process temperature. The final
to check the installed temperature element; perhaps
displayed measurement would be compared to the
a single-point test could be done by temporarily
simulated temperature and the error interpreted.
inserting a certified probe/thermometer into the pro-
FIGURE 1 – EXAMPLE TEMPERATURE LOOP cess and comparing that measurement against the
element’s output when connected to a calibrator.
Analysis of loop error
Error limits can be somewhat difficult to determine

and many mistakes are made when it comes to set-
ting error limits. One common judgment is to base
process measurement tolerance on a manufacturer’s
specification. Some manufacturers are better than
others, but the marketing department may have as
much to say about an accuracy specification as an ditional error to the loop measurement. A common
R&D engineer. Furthermore, accuracy statements practice to compute loop error is to utilize a statis-
are generally an “off-the-shelf” value that does not tical average or a root mean- square (RMS) calcula-
include such things as long term stability (typically tion. With regard to the temperature loop example,
a significant error component), repeatability, tem- assume the RTD sensor is rated ±0.5 °F, the transmit-
perature effects and more. Sensor and transmitter ter is ±0.10% span (span = 50 to 250 °F) and the DCS
accuracy should be a consideration of what the input card is ±0.25% span (span = 50 to 250 °F). The
process measurement tolerance should be, not the loop error could be evaluated as follows:
final value.
The best method is to have a collaborative discus-
sion between the control engineer, quality engineer
and/ or the safety engineer with the instrument engi- The most conservative approach would be to simply
neer in setting a realistic and practical tolerance. It is sum up the errors (0.5 + 0.2 + 0.5 or ±1.2 °F). The final
extremely important to keep in mind that the tighter decision should also consider the criticality of the
the tolerance, potentially, the more expensive it will measurement along with evaluation of the impact
be to not only make the measurement, but to main- the error will have on the process and/or the risks
tain the measurement. The balance falls somewhere involved. The discussion should not end here. The
between the required tolerances to create efficient control engineer will strive for the lowest number
control, the best quality and maintain the highest possible (±0.75 °F), but there are other factors. An
safety versus minimizing downtime, maximizing
technician efficiency and/ or utilizing optimum test
equipment. In practice, it is common to see ±1% of
span (or ±2% or even ±5%). However, this does not
easily apply to flow measurements (typically a per-
cent of reading or rate) or analytical instruments (pH
or ppm, for example).
One good way to look at error is

to think in terms of the loop’s input
engineering units.
One good way to look at error is to think in terms
of the loop’s input engineering units. As regards the
temperature loop example (figure 1), the discussion
should focus on what minimum temperature error
creates the highest operating efficiency without com-
promising quality or safety and can be realistically
measured by the calibration/test equipment. One
other complication for loop error is that a given loop
is no more accurate than the least accurate compo-
nent contributing to the measurement. Today’s trans-
mitters are extremely accurate and provide excellent
performance. However, temperature sensors are typ-
ically not nearly as accurate and, depending on the
process, can exhibit significant drift. If a typical RTD
is rated to ±0.5 °F, a control engineer cannot expect
better than ±0.5 °F to control the process. In reality,
even though the transmitter and DCS analog-to-dig-
ital conversion can be significantly more accurate, it
must be recognized that these components add ad-
evaluation of the test equipment is required. The Depending on the process, more frequent testing
typical temperature block has an accuracy anywhere may be required, but in any case, management will
from 0.3 °F to 1.0 °F, and it is good practice to have a have a high level of confidence that accurate mea-
4:1 ratio of test equipment versus process measure- surements are being made. With repeatable work
ment. To make a proper temperature simulation, a methods, technicians will recognize common issues
reference probe (RPRT or SPRT, reference or second- and there should be efficiency gains. If calibrations
ary primary resistance thermometers) along with an are being documented, analysis of test cycles can be
accurate PRT meter, such as a Beamex MC6 with the analyzed and most likely intervals can be extended or
optional RPRT probe, would both need to be utilized at least optimized. The need for troubleshooting and
to achieve an improved measurement error of 0.1°F emergency repairs will always be required, but the
to 0.2 °F. This could impose a significant investment loop cycle should be reset whenever such an event
in test equipment, depending on the industry, and occurs. This methodical approach effectively provides
it should be noted this will require a higher cost of contact to every instrument in the plant while mini-
maintenance for the more accurate test equipment. mizing disturbances to the loop integrity and deliver-
For example, what if the quality engineer reports that ing the very best measurements to the control system.
an error of ±5 °F is all that is needed to make a good
product? Why impose an unnecessary burden on the
instrumentation department? If the control engineer Multivariable loop test example
has no objection (along with input from reliability,
safety, etc.), a practical approach would be to set a Flow measurements can be very demanding and of-
loop tolerance of ±2.0 °F, assuming the temperature ten require very tight performance tolerances. In the
block has an accuracy of ±0.5 °F over the range of 50 case of natural gas or steam measurements, a small
to 250 °F. While not as accurate as the instrumen- error can amount to significant errors in billing, thus
tation in the loop, it is better than 2:1 for what is creating extra scrutiny by management. A common
required to make a quality product and allows the example of orifice measurement is to compensate
calibration technician to utilize a simple combina- the differential pressure measurement by factoring
tion of equipment. in the process temperature and static pressure. These
While this is just one scenario, it is good practice to three measurements can be processed by the DCS to
determine the “weakest link” in the loop and not set make an accurate flow calculation. However, there
an unrealistic performance tolerance. When look- are now DP flow meters (aka, “multivariable”) with an
ing at fuel costs or process efficiencies, this type of integrated process RTD and static pressure measure-
analysis could easily justify a larger investment in test ment that provide a compensated flow measurement
equipment along with frequent testing if the cost/ output; the flow calculation is built into the smart
risk of error is high. With good judgment, striking a transmitter.
balance and avoiding unreasonable testing requests, If the three measurements are independently pro-
manufacturing objectives can be met. cessed by the control system, typical test procedures
apply, but a loop test should be done to verify the ac-
curacy of the compensated flow reading. While multi-
LOOP TESTING EXAMPLES
FIGURE 2 – EXAMPLE OF A MULTIVARIABLE LOOP
Temperature loop test example
Should a process plant have hundreds of tempera-

ture loops like the example (figure 1), there are good
benefits with loop testing. While it takes time to make
a test with a temperature block, the calibration tech-
nician can effectively check 2, 3 or more instruments
that make up the loop. With this type of approach,
it may make sense to invest in more rugged and/or
more accurate probes in order to minimize failures.
variable meters appear to be complex, a loop test can non-linear table can be entered into the Beamex
be set up to quickly verify that the meter is correctly CMX software for a specific tag and downloaded into
measuring the flow to a desired per cent of reading the MC6 for testing. Additionally, the three tests can
accuracy by identifying the measurement compo- be performed with the process variables versus each
nents. As an example, consider a steam application: HART value that is used in the compensated output
calculation. The only additional test tool required
would be a Beamex temperature block.
The loop test should simply be a 5-point check of
inH2O vs. lbs/hr at 0%, 50%, 100%, 50% and 0%. If
all of the measurements fall within a 1% reading, the
technician can pack up his tools and move on to the
next instrument. If the loop test result is marginal or
a failure, then 3 tests of the DP pressure versus HART,
RTD temperature versus HART and static pressure
versus HART will need to be performed and adjusted
as needed. Upon completion of the three variables
that plug into the flow calculation, a quick check of
the 4–20 mA output should be done as well. Assum-
ing one or more of the inputs require adjustment, a
final As Left loop test of the improved flow output
will document indicate that the meter is in good op-
erating condition and make documentation of it. It is
For this example, a non-linear test should be set up a time-saver to focus on the non-linear input vs. flow
where the expected lbs/ hr output is calculated for output for a multivariable loop and this will result in
specific pressure in put test points assuming a con- a much simpler maintenance task for the instrument
stant, typical 450 ºF temperature and a static pres- technician.
sure of 14.735 psi since the low side of the transmitter
is vented to atmosphere for testing. Consulting with
the control engineer, expected measurements may Other loop examples
look like this:
A pressure loop can be easily checked by applying a
pressure to the input transmitter and comparing it
to the DCS or final control reading. This can be done
very quickly and can be much more effective than
merely testing the transmitter. Any batch control
loop should be evaluated for loop testing with the
goal to make work more efficient for the technician
while verifying that control measurements are as
accurate as possible.
This same technique should be considered for
control valve testing where an mA input into the I/P
is compared to an mA output (feedback). This would
also apply to smart control valve positioners using a
communicator to step the valve and monitor the dig-
ital feedback. By making 10% test points, a quick test
on a valve will verify that it is operating correctly. In
most cases, the test should pass and the technician
The Beamex MC6 offers very unique features for can make a quick round of testing of critical control
testing multivariable transmitters. The proceeding valves.
An overlooked component of a flow loop is the
primary element (orifice plates, annubars or averag-
ing pitot tubes). These are critical for a proper flow
measurement and while they cannot be calibrated,
they should be inspected for damage or wear.
Another critical area where loop testing should be
considered is safety instrumented systems (SIS).
When the process is down, it is common to follow a
script of testing procedures that can include cali-
bration of single instruments. However, whenever
possible, consider checking an entire loop where the
integrity of a critical measurement can be verified,
especially for temperature (utilizing a block/bath)
or pressure measurements. Also, it may be possible
to perform quick and simple tests on a SIS while testing. With the Beamex Integrated Calibration Solu-
the process is up and running to ensure systems are tion, flexible tests can be designed to meet a variety
operating properly. of applications. The Beamex solution delivers the
highest level of automation while providing detailed
Conclusion documentation and electronic reporting.
By approaching the task of calibration with a fresh
In many, many process plants, calibration is per- look, there are plenty of opportunities to “do more
formed by simply checking the transmitter. It takes with less” and effectively make contact with every
time to use a temperature block/bath, but consider instrument in the plant more efficiently using loop
how important it is to test all the devices that make calibration strategies. Logical and careful planning of
up a given measurement. Transmitters are not the loop testing strategies will result in improved control
only devices that drift. Temperature probes drift due performance without compromising quality, reliabili-
to thermal stress/shock and vibration or physical ty or safety of plant operations.
damage. DCS/ PLC input cards drift as much or more
than transmitters. If loops are not being tested, how
can a good measurement be made? Without good RELATED WEBINAR
measurements, how can optimum control, safety,
reliability and quality be ensured?
How to avoid the most common mistakes in
As instrumentation and automation evolve, so
field calibration
should the methods for calibrating instrumentation.
Loop testing is not a new concept, but it is underuti- https://hubs.ly/H09rZ390
lized as an effective strategy for instrumentation
Safety
Considerations
By: Mike Cable
Obviously, performing calibrations safely is very • Include specific safety considerations in each
important. One lapse on safety could cost you or calibration procedure. For example, if we know
your co-worker your lives. Even if it’s not a life lost, there is a tank that does not have a thermowell
minor injuries caused by unsafe work practices are installed for the resistance temperature detector
preventable. Safety of the product is also of concern (RTD), highlight this fact. Better yet, if possible,
when performing calibrations in a manufacturing get a thermowell installed.
environment. There are many resources available
from ISA on the topic of safety, including Chapter 1 • Keep the shop and work areas clean and free of
of Troubleshooting: A Technician’s Guide, by William trip hazards.
L. Mostia. Here are a few things we, as calibration
technicians, can do to improve safety in our day-to- • Work with a partner or at least make sure some-
day work activities. one knows where you are working at all times.
• Some instruments are always installed at diffi-

cult places to reach. If it’s possible to install some
permanent platform, have it done. Otherwise use
safety harnesses, ladders, and lifts properly.
• Technicians may be exposed to lethal electri-

cal voltages. Know what the high voltage areas
are, de-energize electrical circuits that are not
required, and use proper electrical safety prac-
tices (insulated floor mat, rubber electrical safety
gloves, roped off area, safety man outside the area
with a rope tied around you).
• Ensure electrical power cords are properly insulat-

ed. Ensure equipment is properly grounded.
Data Integrity in Calibration Processes
Calibration in the pharmaceutical industry
By: Heikki Laurila
Although this article primarily focuses on data integ- Data integrity is the maintenance of,
rity requirements for the pharmaceutical industry, it and the assurance of the accuracy
is a topic that impacts many other industries as well. and consistency of the data over its
Similar requirements are already being implement- entire life-cycle.
ed into other sectors, such as food and beverage and
power generation
It’s all about trust
As a concept, data integrity is by no means a new
one, it has been around for several decades. Anyhow, Often, when we buy an everyday product, we can
in this article, we look at the data integrity more quickly see if the product is operating properly, or
from the calibration process point of view, and focus if it is faulty. For example, if you buy a new TV and
mainly on the pharmaceutical and regulated indus- turn it on, you can quickly see if it is working or not.
try. At first, we look at data integrity generally: what it But with different products it is not so easy to see if
is, why it is important and what a breach could cause. you have a proper product. This is especially the case
The ALCOA plus concept is also briefly discussed. with medicines. When you pick up a medicine, how
I remember in the early 90’s when we had pharma- do know that it is a product working properly accord-
ceutical customers auditing us prior to a calibration ing to design specifications? In most cases you can’t
software purchase, and data integrity was already tell that, so it is all about trust – you must be able to
then one of the normal topics discussed during such trust that the medicine you take is a proper one.
a supplier audit. So, it is not a new topic.
What is Data Integrity?
Data integrity is fundamental in a pharmaceutical

quality system ensuring that products are of the
required quality.
In every process, there is a lot of data produced.
Data integrity is the maintenance of, and the assur-
ance of the accuracy and consistency of the data over
its entire life-cycle. It is a critical aspect to the design,
implementation and usage of any system which
stores, processes, or retrieves data. The term Data In-
tegrity is pretty widely used and has different mean-
ings in different contexts. The term itself is outdated
and was initially used in computing. The integrity of
the data collected and recorded by pharmaceutical on comment rounds. Both will naturally have effect
manufacturers is critical to ensuring that high quality with the pharmaceutical industry. Sure, previously
and safe products are produced. To ensure the integ- there has been guidance for the good manufactur-
rity of data, it should be protected from accidental or ing practice (CGMP), such as 21 CFR parts (210, 211,
intentional modifications, falsification and deletion. and 212), discussing data integrity related issues, but
With many processes in the process industry, you these mentioned new updates will raise the focus.
cannot just simply test the final product to see if it is One additional reason why more focus has been
a proper one. Instead you must assure that the con- put to data integrity is the increase of the use of mo-
ditions during the process are correct in order for it bile devices in calibration processes. This includes
to produce the correct product. These critical condi- applications used in tablets and mobile phones. It
tions must naturally be recorded and maintained to also includes the increase of the use of documenting
assure they were correct. This is certainly the case in calibrators, which automatically store the calibra-
many processes in a pharmaceutical plant. tion results in their memory during a calibration and
transfer this data to calibration software. Since the
use of automated documenting calibrators will im-
Why is data integrity important at the moment? prove the business case of a calibration system, they
are being more widely used.
Data integrity has recently risen to an even more To learn more on what a documenting calibrator
important topic than before. is and how it benefits the calibration process, please
Data integrity related violations have led to several check the blog post: What is a documenting calibra-
regulatory actions such as warning letters and import tor and how do you benefit from using one?
alerts. Actually, a large number of the recent warning As results of all these, data integrity is getting more
letters issued by FDA are somehow related to data and more acute.
integrity.
As international regulatory agencies have more
focus on data integrity, the FDA, WHOA and MHRA Impacts of breach of data integrity
auditors have been trained to better recognize data
integrity issues. The impact of breach of data integrity can be looked
MHRA (Medicines & Healthcare products Regula- as the impact to customer and impact to the phar-
tory Agency in UK) has recently released new guide maceutical company.
“GMP Data Integrity Definitions and Guidance for For the customer, the impact can be that the medi-
Industry” (March 2015). There is a deadline set for cine does not have the required effect, patient safety
pharmaceutical companies to comply at the end can be compromised and in a worst case it can cause
of 2017. Also, FDA has released “Data Integrity and even loss of lives.
Compliance With CGMP - Guidance for Industry” For the pharmaceutical company, the impact can
(April 2016). This is still in draft mode but has been be a warning letter from FDA, bans of license to
produce, negative reputation, loss of customer con- The ALCOA plus ads a few attributes to the list:
fidence, reduction of market share, and reduction of • Complete
share price. • Consistent
• Enduring
• Available
Accidental or Intentional
A brief description of these attributes are included in
A breach of data integrity may be accidental or the following table:
intentional. Often, there are computerized systems
involved to handle the data and the users may not
be aware of any issues in such systems. Certainly,
the majority of data integrity issues are accidental
and non-intentional. Anyhow, in looking at some of
the FDA warning letters, it indicates that in the very
worst cases there have been intentional falsifying of
records.
Main steps towards better data integrity
Many pharmaceutical companies seem to agree that

the main steps towards better data integrity are:
• Better education and communication

• Detection and mitigation of risks
• Focus on technology and IT systems
• Governance of data integrity
Validation is also something that is a must for any

computerized system in the pharmaceutical industry.
It is good to remember that ANSI defines systems as:
people, machines, and the methods organized to per-
form specific functions. So, it is not only the comput-
er system that needs to be validated.
ALCOA and ALCOA plus
The acronym ALCOA has been around since the

1990’s being used by regulated industries as a frame-
work for ensuring data integrity. It is the key to good
documentation practice (GDP). ALCOA relates to
data, whether paper or electronic, and is defined by What could cause data integrity issues?
FDA guidance as:
• Attributable Some practical and general things that could cause
• Legible data integrity issues in any systems are: lack of
• Contemporaneous training, user privileges, poor or shared passwords,
• Original control of a computerized system, incomplete data
• Accurate entry, and lack of audit data records for changes and
modifications.
Conclusion
Although the data integrity concept has existed for a

long time, it has recently risen to be more acute due
to the use of mobile tools and added focus of regula-
tory agencies. In the end, data integrity is common
sense - to ensure the integrity of data throughout
its life cycle - in practice with various systems and
tools being used, it gets more complicated. Since
the impacts of the breach of data integrity can be
enormous, it is something that needs to be a high
priority.
The first trap to avoid for consumers – fraud

drugs
Although not really a data integrity issue for the USEFUL REFERENCES
industry, this is an important point for consumers.
People are buying more from the internet nowadays
and you can also buy medicine from internet, but • 21 CFR Part 11, Electronic Records; Electronic
unfortunately you don’t always get what you order. A Signatures:
huge amount of medicine bought online are fraudu- www.fda.gov/RegulatoryInformation/Guidances/
lent. Sometimes the packaging is obviously inappro- ucm125067.htm
priate, so it becomes apparent that the medication
is a counterfeit. Unfortunately, that is not always the • MHRA GMP Data Integrity Definitions and
case and people do, at times, consume fraudulent Guidance for Industry, March 2015:
medicine. It is clear that the fraudulent medication www.gov.uk/government/uploads/system/up-
does not provide the expected cure, but it is also a loads/attachment_data/file/412735/Data_integri-
big risk for our safety and at its worse, it may be even ty_definitions_and_guidance_v2.pdf
lethal.
• Data Integrity and Compliance with CGMP
Guidance for Industry DRAFT GUIDANCE,
New regulation for product packaging to avoid April 2016:
frauds www.fda.gov/downloads/drugs/guidance-
complianceregulatoryinformation/guidance-
To better control these fake drugs, the European sucm495891.pdf
Medicines Agency (EMA) has recently introduced
a new regulation that will require all prescription • FDA warning letters are public and can be
drug makers in all (but three) EU (European Union) found here:
countries to incorporate new safety features on their www.fda.gov/ICECI/EnforcementActions/Warn-
product packaging by February 2019. The regulation, ingLetters/default.htm
which is part of a broader effort to combat falsified
medicines in the EU, will require drug makers to add • European Medicines Agency (EMA), recent
a unique identifier and an anti-tampering device to regulation for product packaging:
the packaging of most centrally authorized products. www.raps.org/Regulatory-Focus/
This naturally adds another burden and cost for the News/2016/02/09/24281/EU-Regulation-Re-
drug manufacturers to build the systems to sup- quires-New-Safety-features-on-Drug-Packaging-
port this, but this will certainly be beneficial for the by-2019/
customers. Although this specific regulation is for the
European Union area, it will have a global effect.
Definition of
Terms and Acronyms
Terms

accuracy error
as part of the calibration activity, the degree of con- the difference between an indicated value and the
formity of a device’s output to its actual input value, actual value expressed as either
typically expressed in terms of the measured variable,
for example percent of full scale value or percent of a) percent of full scale – difference between the
actual reading; usually measured as an inaccuracy indicated value and the actual value, expressed
and expressed as accuracy as a percentage of the device’s full scale (mini-
mum-to-span) range of values or
Note 1 to entry: The difference between these expres-
sions can be great. The only way to compare accuracy b) percent of reading – difference between the
expressed in different ways is to calculate the total error indicated value and the actual value, expressed
at certain points in engineering units. as a percentage of the actual value

adjustment hysteresis
as part of a calibration activity, the act of adjusting a the deviation in output at any point within the
device to meet a known standard instrument’s sensing range, when first approaching
this point with increasing input values, and then
with decreasing input values (in other words, when a
calibration device produces a different output at the same input
procedure (3.1.18) of checking or adjusting (by com- point based on whether the input is increasing or
parison with a reference standard) the accuracy of a decreasing)
measuring instrument
[SOURCE: ISO 15378, modified – “process” was re-
placed with “procedure”] industrial automation and control system
collection of personnel, hardware, software and pol-
calibration work instructions icies involved in the operation of the industrial pro-
step by step instructions for performing a calibration cess and that can affect or influence its safe, secure
on a specific device or loop and reliable operation [ANSI/ISA-62443-3-3]
Note 1 to entry: This system includes sensors and
final elements and may be either a BPCS or an SIS or
criticality a combination of the two.
criticality ranking or classification of items or events
based upon relative significance, importance, or
severity. instrument asset management
coordinated work processes of an organization to
ensure the intended capability of assets is available
device
a piece of instrument hardware designed to perform
a specific action or function loop
combination of two or more components (devices or
control functions arranged so that signals pass from
device calibration one to another for the purpose of measurement, indi-
a calibration performed on only one device cation, or control of a process variable
loop accuracy test uncertainty ratio
the degree of conformity of a loop’s measured or ratio of the span of the tolerance of a measurement
controlled variable indicated value to the variable’s quantity subject to calibration to twice the 95%
actual value expanded uncertainty of the measurement process
used for calibration

loop calibration
SOURCE: ANSI/NCSL Z540.3-2006]
a calibration performed on a loop, consisting of two
or more devices
(USL – LSL)
TUR=
loop tolerance (ku)
the permissible limit of variation, from a known stan-
dard, in the loop indicated process measurement USL = Upper Specification Limit
LSL= Lower Specification Limit
management program u=uncertainty
activity that manages a group of related projects and/ k = 2 approximates 95% of coverage
or work processes in a way that provides benefits
and control not available by managing each activity tolerance
individually and independently permissible limit or limits of variation in a measure-
ment with respect to the actual value
precision
the repeatability of a measurement over time traceability
the property of the results of a measurement or the
procedure value of a standard whereby it can be related to a
sequence of tasks with a defined beginning and end stated reference, usually national or international
that is intended to accomplish a specific objective standards, through an unbroken chain of calibrations
all having stated uncertainties

rangeability Note 1 to entry: For example in the US, this is the
measurement range over which the error statement National Institute of Standards and Technology or ILAC
is guaranteed recognized laboratories.

repeatability verification
the variation in outputs/indications of a device/ as part of a calibration activity, the act of checking
loop under the same conditions, and over a period of the accuracy of a loop or loop component to de-
time. Often expressed in the form of standard devia- termine whether it is performing within required
tion. Also see precision. tolerances

task work process
a single piece of work that needs to be done and does set of interrelated or interacting procedure(s) (3.1.18)
not have interacting elements requiring manage- which transforms inputs into outputs
ment
[SOURCE: ISO 9000, ISO 55000, modified – “activi-
[SOURCE: IEC 62304, modified – “a” was deleted,
ties” was replaced with “procedures”]
“and does not have interacting elements requiring
management” was added]

Acronyms
ANSI: American National Standards Institute
BPCS: Basic Process Control System
CEM: Continuous Emission Monitoring
CSR: Calibration Service Record
EPA: Environmental Protection Agency
HART: Highway Addressable Remote Transducer
IACS: Industrial Automation and Control System
ILAC: International Laboratory Accreditation Cooperation
ISA: International Society of Automation
ISO: International Organization for Standardization
LRV: Lower Range Value
NCSL: National Conference of Standards Laboratories
NIST: National Institute of Standards and Technology
NRC: Nuclear Regulatory Commission
RP: Recommended Practice
RSS: Root Sum Square
SIS: Safety Instrumented System
TUR: Test Uncertainty Ratio
URV: Upper Range Value

Calibration Essentials Ebook PDF

Uploaded by

Copyright:

Available Formats

Calibration Essentials Ebook PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Calibration Essentials Ebook PDF

Uploaded by

Copyright:

Available Formats

What are some of the topics covered in the eBook?

What are some of the topics covered in the eBook?

What are the benefits of calibrating at the loop level versus the individual instrument level?

What are the benefits of calibrating at the loop level versus the individual instrument level?

Calibration

• A comprehensive big picture guide on how to manage a facility-wide calibration program

• Informative overviews of calibration considerations, such as tolerance errors, and calibration

• And much more

Management of a Calibration Program for Industrial

Calibration Out of Tolerance

Calibration Uncertainty Page 30

Criteria for Instrument Selection in Relation to the

Calibrating WirelessHART Transmitters Page 42

Loop Calibration vs. Individual Instrument Calibration

New Methods for Calibrating Loops Page 51

Data Integrity Page 58

Definition of Terms and Acronyms

Purpose acquired during these assessments not only aids in

However, targeted reduction, or increase, in calibra-

Instrumentation worth purchasing and installing is

Liptak, Bela G., Editor. Instrument Engineers’ Hand-

3 Establishing a calibration program

3.3 Calibration program

Each user company/facility is encouraged to estab-

CR #4 Boiler trip instrumentation 12 months

Maintenance records, possibly analyzed using statis- CR #2 Performance thermocouples 36 months

intervals. By analyzing instrument drift over time, CR #1 Recorders

Note 1 to entry: Intervals for SIS-related instruments

When documenting calibration information (as

Calibrations may be performed using 5 check points

All data should be recorded on a CSR and stored in a

4.9 Calibration procedures

Confirming the general qualifications of the person-

e) required loop tolerances and manufacturer-sup-

5.1 General 6.2 Assurance

7.1 Example 1 – Calculating instrument total error

Total Error = (0.2)2+(1.1875)2+(0.25)2+(0.9375)2+(0.25)2 = 1.57%

NOTE This example is taken from the Instrument Engineer’s Handbook.

7.2 Example 2 – Calculating theoretical loop tolerance

Maximum range – 0 - 3500 PSI (0 - 24,132 kPa)

Total Error = (E)2+(S)2+(Tz )2+(Ts )2+(C)2

Total Error = (.075)2+(0.125)2+(0.025)2+(0.018)2+(0.100)2

Total Error = ±0.18% of full scale or ±4.5 PSI (31 kPa)

Total Error = (E)2+(H)2+(R)2+(S)2+(Tz )2+(Ts )2

Total Error = (1.0)2+(0.2)2+(0.070)2+(0.05)2+(0.020)2+(0.020)2

Total Error = ±1.14% of full scale or ±34.2 PSI (235.8 kPa)

7.4 Example 4 – Pressure measurement theoretical tolerance – pressure gauge

Range and span: 0 - 3000 PSI (0 - 20684 kPa)

Total Error = (E)2+(Ts )2

Total Error = (2.0)2+(0.04)2

Total Error = ±2.0% of full scale or ±60 PSI (413.7 kPa)

These examples highlight the large differences in the

The TUR must be calculated using identical param-

75 14.93 (0.07/15) = 0.47%

100 19.9 (0.1/20) = 0.5%

This loop would meet its required tolerance of +/- 1% of

A.1 On-line testing procedure of differential pressure transmitter

Unit 2 Point C2 104, PDT-211

The following items are required to perform the work:

Worker Signatures: _ _