M3003 The Expression of Uncertainty and Confidence in Measurement

F
M3003
Edition 5 September 2022
The Expression of Uncertainty and Confidence in

Measurement
The Expression of Uncertainty and Confidence in Measurement
Contents
1 Introduction 3
2 Overview 4
3 In More Detail 10
4 Type A Evaluation of Standard Uncertainty 19
5 Type B Evaluation of Standard Uncertainty 21
6 Reporting of Results 22
7 Step by Step Procedure for Evaluation of Measurement Uncertainty 23
Appendix A Calibration and Measurement Capability 27
Appendix B Coverage Factor when there is a single dominant Type A input 28
Appendix C Coverage Factor when there is a single dominant Type B input 31
Appendix D Measurement Equations 38
Appendix E Some Sources of Error and Uncertainty in Electrical Calibrations 45
Appendix F Some Sources of Error and Uncertainty in Mass Calibrations 50
Appendix G Some Sources of Error and Uncertainty in Temperature Calibrations 52
Appendix H Some Sources of Error and Uncertainty in Dimensional Calibrations 54
Appendix J Some Sources of Error and Uncertainty in Pressure Calibrations using Dead Weight Testers 55
Appendix K Examples of Application 57
Appendix L Expression of Uncertainty for a Range of Values 77
Appendix M Assessment of Conformity with Specification 82
Appendix N Uncertainties for Test Results 83
Appendix P Comparing Independent Results using En ratio 86
Appendix Q Symbols 87
Appendix R References 89
w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

© United Kingdom Accreditation Service. UKAS copyright exists on all UKAS publications.
M3003 Edition 5 Page 2 of 90
1 Introduction
1.1 The general requirements that testing and calibration laboratories must meet if they wish to
demonstrate that they operate to a quality system, are technically competent and are able to
generate technically valid results, are contained within ISO/IEC 17025 [5]. This international
standard forms the basis for international laboratory accreditation and in cases of differences in
interpretation always remains the authoritative document. M3003 is not intended as a
prescriptive document and does not set out to introduce additional requirements to those in
ISO/IEC 17025 but instead aims to provide amplification and guidance on the current
requirements within the standard.
1.2 The purpose of these guidelines is to support policy on the evaluation and reporting of
measurement uncertainty for testing and calibration laboratories. Related topics, such as
evaluation of conformity with specifications, are addressed in UKAS LAB 48 [14]. Several
worked uncertainty examples are included in M3003 to illustrate how practical implementation
can be achieved. Further practical guidance for Medical Laboratories meeting requirements of
ISO 15189 [6] is provided in ISO TS 20914 [9].
1.3 The guidance in this document is based on information in the Guide to the Expression of
Uncertainty in Measurement, hereinafter referred to as the GUM [1]. M3003 is consistent with
the GUM suite of documents both in methodology and terminology. It does not, however, have
the same breadth of scope as the GUM, which also includes other methods of uncertainty
evaluation that may be more appropriate to a specific discipline, for example, the use of Monte
Carlo simulation [2].
1.4 M3003 is aimed both at the beginner and at those more experienced in the subject of
measurement uncertainty. In order to address the needs of an audience with a wide spectrum of
experience, the subject is introduced in relatively straightforward terms and describes the basic
concepts involved. Cross-references are made to a number of Appendices in which more
detailed information is presented for those seeking fuller guidance on the subject. For a more in-
depth understanding of measurement uncertainty, courses such as the practitioner course
offered by UKAS are recommended.
1.5 A number of changes have been made since the publication of M3003 Edition 4. These are too
numerous to list in detail. In general, they are:
• Editorial changes to address feedback and regular questions,
• Equation fonts have been converted throughout to improve presentation,
• Changes to improve clarity and consistency,
• Appendices A, B, C, D, K, L, M, have been significantly revised to address minor errors
and to improve clarity.
1.6 No further changes of any significance will be made to M3003 Edition 5 during its lifetime.
However, minor text modifications of an editorial nature may be made if the need is identified.
Any such changes will be listed below.
Date Details of amendment

2 Overview
2.1 In many aspects of everyday life, we are accustomed to the doubt that arises when estimating
how large or small things are. For example, if somebody asks, “what do you think the
temperature of this room is?” we might say, “it is about 23 degrees Celsius”. The use of the word
“about” implies that we know the room is not exactly 23 degrees but is somewhere near it. In
other words, we recognise that there is some doubt about the value of the temperature that we
have estimated.
2.2 We could, of course, be a bit more specific. We could say, “it is 23 degrees Celsius give or take
a couple of degrees”. The term “give or take” implies that there is still doubt about the estimate,
but now we are assigning limits to the extent of the doubt. We have given some quantitative
information about the doubt, or uncertainty, of our estimate.
2.3 It is also quite reasonable to assume that we may be more sure that our estimate is within, say, 5
degrees of the “true” room temperature than we are that the estimate is within 2 degrees. The
larger the uncertainty we assign, the more confident we are that it encompasses the “true” value.
Hence, for a given situation, the uncertainty is related to the level of confidence.
2.4 So far, our estimate of the room temperature has been based on a subjective evaluation. This is
not entirely a guess, as we may have experience of exposure to similar and known
environments. However, in order to make a more objective measurement it is necessary to make
use of a measuring instrument of some kind; in this case we can use a thermometer.
2.5 Even if we use a measuring instrument, there will still be some doubt, or uncertainty, about the
result. For example, we could ask:
“Is the thermometer accurate?”
“How well can I read it?”
“Is the reading changing?”
“I am holding the thermometer in my hand. Am I warming it up?”
“The relative humidity in the room can vary considerably. Will this affect my results?”
“Does it matter whereabouts in the room I take the measurement?”
All these factors, and possibly others, may contribute to the uncertainty of our measurement of
the room temperature.
2.6 In order to quantify the uncertainty of the room temperature measurement we will therefore have
to consider all the factors that could influence the result. We will have to make estimates of the
possible variations associated with these influences. Let us consider the questions posed above.
2.7 Is the thermometer accurate?
2.7.1 In order to find out, it will be necessary to compare it with a thermometer whose accuracy is
better known. This thermometer, in turn, will have to be compared with an even better
characterised one, and so on. This sequence leads to the concept of traceability of
measurements, whereby measurements at all levels can be traced back to agreed references.
In most cases, ISO/IEC 17025 [5] requires that measurements are traceable to SI units, which is

usually achieved by an unbroken chain of comparisons originating at a national metrology

institute.
In other words, we need a traceable calibration. This calibration itself will provide a source of
uncertainty, as the calibrating laboratory will assign a calibration uncertainty to the reported
values. When used in a subsequent evaluation of uncertainty, this is often referred to as an
imported uncertainty.
2.7.2 In terms of the thermometer accuracy, however, a traceable calibration is not the end of the
story. Measuring instruments change their characteristics as time goes by. Because they “drift”
regular recalibration is necessary. It is therefore important to evaluate the likely change since the
instrument was last calibrated.
If the instrument has a reliable history, it may be possible to predict what the measurement error
will be at a given time in the future, based on past results, and apply a correction to the reading.
This prediction will not be perfect and therefore an uncertainty on the corrected value will be
present. In other cases, there may be insufficient past data, or it may not indicate a reliable
trend, and a limit value may have to be assigned for the likely change since the last calibration.
This value can be estimated from examination of changes that occurred in the past. Evaluations
made using these methods yield the uncertainty due to secular stability, or changes with time, of
the instrument. This change with time is commonly known as drift.
2.7.3 There are other possible influences relating to the thermometer accuracy. For example, suppose
we have a traceable calibration, but only at 15 °C, 20 °C and 25 °C. What does this tell us about
its indication error at 23 °C?
In such cases we will have to make an estimate of the applicable calibration error, often by
interpolation between points where calibration data is available. The associated measurement
uncertainty might usually be interpolated in the same fashion, with some additional allowance for
uncertainty in the method of interpolation.
2.8 How well can I read it?
2.8.1 There will inevitably be a limit to which we can resolve the reading we observe on the
thermometer. If it is a liquid-in-glass thermometer, this limit will often be imposed by our ability to
interpolate between the scale graduations. If it is a thermometer with a digital readout, digital
rounding will define the limit.
2.8.2 For example, suppose the last digit of a digital thermometer is rounded so that its displayed
value changes in steps of 0.1 °C. The reading happens to be 23.4 °C.
The reading is a rounded representation of a larger series of values that the thermometer would
indicate if it had more digits available. In the case of a reading of 23.4 °C, this represents all
possible values in the range between 23.35 °C and 23.45 °C, which all round to 23.4 °C.
A reading of 23.4 °C therefore means that the value is somewhere between 23.35 °C and
23.45 °C. In other words, the 0.1 °C resolution of the display has caused a rounding error
somewhere between 0.05 °C and -0.05 °C (corresponding to plus or minus half of the display
resolution). As we have no way of knowing whereabouts in this range the value is located, we
have to assume the rounding error is zero with limits of ±0.05 °C. (Zero is the ‘expectation’ value
– it is the best estimate based upon the available information).
2.8.3 It can therefore be seen that there will always be an uncertainty of ± half of the change
represented by one increment of the last displayed digit. This rounding error does not only apply
to digital displays; it applies every time a number is recorded. If we write down a rounded result

of 123.456, we are imposing an identical effect by the fact that we have recorded this result to
three decimal places, and an error not exceeding 0.0005 will arise.
2.8.4 This source of uncertainty is frequently referred to as “resolution”, however it is more correctly
the numeric rounding caused by finite resolution.
2.9 Is the reading changing?
2.9.1 Yes, it probably is! Such changes may be due to variations in the room temperature itself,
variations in the performance of the thermometer and variations in other influence quantities,
such as the way we are holding the thermometer.
So, what can be done about this?
2.9.2 We could, of course, just record one reading and say that it is the measured temperature at a
given moment and under particular conditions. This would have little meaning, as we know that
the next reading, a few seconds later, could well be different. So, which is “correct”?
2.9.3 In practice, we will probably take an average of several measurements in order to obtain a more
representative value. In this way, we can “smooth out” the effect of short-term variations in the
thermometer indication. This average, or arithmetic mean, of a number of readings can often be
closer to the “true” value than any individual reading is.
2.9.4 However, we can only take a finite number of measurements. This means that we will never
obtain the “true” mean value that would be revealed if we could carry out an infinite (or very
large) number of measurements. There will be an unknown error, and therefore an uncertainty,
arising from the difference between our calculated mean value and the underlying “true” mean
value.
2.9.5 This uncertainty cannot be evaluated using methods like those we have already considered. Up
until now, we have looked for evidence, such as calibration uncertainty and secular stability, and
we have considered what happens with finite resolution by logical reasoning. The effects of
variation between readings cannot be evaluated like this, because there is no background
information available upon which to base our evaluation.
2.9.6 The only information we might have is a series of readings and a calculated average, or mean
value. We therefore need to use a statistical approach to determine how far our calculated mean
could be away from the “true” mean. These statistics are quite straightforward… the so-called
repeatability uncertainty is therefore estimated from the experimental standard deviation of the
mean, often referred to as simply the standard deviation of the mean.
NOTE: Standard deviation of the mean is also known as standard error.
2.9.7 It is often convenient to regard the calculation of the standard deviation of the mean as a two-
stage process. It can be performed easily by most scientific calculators or spreadsheet software.
2.9.8 First, we obtain the estimated repeatability standard deviation, 𝑠, e.g., using the values we
have measured. This facility is indicated on most calculators by the function key xσn-1. On some
calculators it is identified as s(x) or simply s.
In Microsoft Excel the STDEV.S cell function can be used.
2.9.9 The standard deviation of the mean is then obtained by dividing the estimate obtained in 2.9.8 by
the square root of the number of measurements that contributed to the mean value.

2.9.10 For example, suppose we record five consecutive readings with our thermometer. These are
23.0 °C, 23.4 °C, 23.1 °C, 23.6 °C and 22.9 °C, and we intend to report the mean 23.2 °C of these
five values.
2.9.11 We obtain an estimated standard deviation of 0.2915 °C.
2.9.12 Five measurements contributed to the mean value, so we divide 0.2915 °C by the square root
of 5, giving a repeatability estimate (standard deviation of the mean) equal to
0.2915 0.2915
= = 0.1304 ℃.
√5 2.236
2.9.13 Further information on the statistical processes used for evaluation of repeatability can be found
in Section 4.
2.10 I am holding the thermometer in my hand. Am I warming it up?
2.10.1 Quite possibly. There may be heat conduction from the hand to the temperature sensor. There
may be radiated heat from the body impinging on the sensor. These effects may or may not be
significant, but we will not know until an evaluation is performed. In this case, special
experiments may be required in order to determine the significance of the effect.
2.10.2 How could we do this? Some fairly basic methods come to mind. For example, we could set up
the thermometer in a temperature-stable environment and read it remotely, without the operator
nearby. We could then compare this result with that obtained when the operator is holding it in
the usual manner, or in a variety of manners. This would yield empirical data on the effects of
heat conduction and radiation. If such effects turn out to be significant, we could either improve
the method so that operator effects are eliminated, or we could include a contribution to
measurement uncertainty based on the results of the experiment.
2.10.3 Consideration of the measurement method reveals a number of important issues. The
measurement may not be independent of the operator and special consideration may have to be
given to operator effects (we may have to train the operator to use the equipment in a particular
way). Special experiments may be necessary to evaluate particular effects. Additionally, and
significantly, evaluation of uncertainty may reveal ways in which the method can be improved,
thus giving more reliable results.
2.11 The relative humidity in the room can vary considerably. Will this affect my results?
2.11.1 Maybe it will. If we are using a liquid in glass thermometer, it is difficult to see how the relative
humidity could significantly affect the expansion of the liquid. However, if we are using a digital
thermometer, it is possible that relative humidity could affect the electronics that amplify and
process the signal from the sensor. The sensor itself could also be affected by relative humidity.
2.11.2 As with other influences, we need means of evaluating any such effects. In this case, we could
expose the thermometer to an environment in which the temperature can be maintained at a
constant level, but the relative humidity can be varied… which would reveal how sensitive the
thermometer is to the quantity we are concerned about. Alternatively, we might rely upon
information published by the equipment manufacturer.
2.11.3 This question also raises a general point that is applicable to all measurements. Every
measurement we make has to be carried out in an environment of some kind; it is unavoidable.
So, we have to consider whether any particular aspect of the environment could have an effect
on the measured value and its uncertainty.
2.11.4 The significance of a particular aspect of the environment has to be considered in the light of the
specific measurement being made. For example, it is difficult to see how gravity could

significantly influence the reading on a digital thermometer. However, it certainly will affect the
results obtained on a precision weighing machine that might be right next to the thermometer!
2.11.5 The following environmental effects are amongst the most commonly encountered when
considering measurement uncertainty:
Temperature
Relative humidity
Barometric pressure
Electric or magnetic fields
Gravity
Electrical supplies to measuring equipment
Air movement
Vibration
Light and optical reflections
Furthermore, some of these influences may have little effect as long as they remain constant but
could affect measurement results when they start changing. Rate of change of temperature can
be particularly important.
2.11.6 It should be apparent by now that understanding of a measurement system is important in order
to identify and quantify the various uncertainties that can arise in a measurement situation.
Conversely, analysis of uncertainty can often yield a deeper understanding of the system and
reveal ways in which the measurement process can be improved, which leads on to the next
question…
2.12 Does it matter where in the room I make the measurement?
2.12.1 It depends on what we are trying to measure! Are we interested in the temperature at a specific
location, or in the average of the temperatures encountered at any location within the room, or
the average temperature at bench height?
2.12.2 There may be further, related questions. For example, do we require the temperature at a
particular time of day, or the average over a specific period of time?
2.12.3 Such questions have to be asked, and answered, in order that we can devise an appropriate
measurement method that gives us the information we require. Until we know the details of the
method, we are not in a position to evaluate the uncertainties that will arise from that method.
2.12.4 This question and those preceding it are important questions to ask. But the most important
question of all is one that should be asked before we even select a method and start our
uncertainty evaluation:
2.13 “What exactly is it that I am trying to measure?”
2.13.1 Until this question is answered, we are not in a position to carry out a proper evaluation of the
uncertainty. The particular quantity of interest (the quantity subject to measurement) is known as
the measurand. In order to evaluate the uncertainty in a measurement we must define the
measurand, otherwise we are not in a position to know how any particular influence quantity
affects the value we obtain for it.
2.13.2 The implication of this is that there has to be a defined relationship between the influence (input)
quantities and the measurand (output). This relationship is known as the measurement model.
This can be written as a measurement equation that describes how each influence quantity
affects the value assigned to the measurand. In effect, it is a mathematical description of the

measurement process. Further details about establishing a measurement equation can be found
in Appendix D. A proper analysis of this process also gives the answer to another important
question:
2.14 “Am I actually measuring the quantity that I thought I was measuring?”
2.14.1 Most measurement processes are such that the end result would be only an approximation to
the “true” value because of assumptions and approximations inherent in the chosen
measurement method. The model should recognise any such assumptions and uncertainties that
may arise from them should be accounted for in the analysis.
2.15 Summary
2.15.1 This section of M3003 has given an overview of uncertainty and some insights into how
uncertainties might arise. It has shown that we have to understand our measurement process
and the way in which various influences can affect the result. It has also shown that analysis of
uncertainty can have positive benefits in that it can reveal where enhancements can be made to
measurement methods, hence improving the reliability of measurement results.
2.15.2 The following sections of M3003 explore the issues identified in this overview in more detail.

3 In More Detail
3.1 The Overview section of M3003 has provided an introduction to the subject of uncertainty
evaluation and has explored a number of the issues involved. This section provides a slightly
more formal description of these processes, using terminology consistent with that in the GUM.
3.2 The International vocabulary of metrology (VIM) [4] defines a quantity (𝑄) as a property of a
phenomenon, body or substance to which a magnitude (expressed as a number and a
reference) can be assigned.
3.3 The purpose of a measurement is to assign a magnitude to the measurand; the quantity
intended to be measured. The assigned magnitude is considered to be the best estimate of the
value of the measurand. The uncertainty evaluation process will encompass a number of other
‘influence’ quantities that affect the result obtained for the measurand. These influence, or ‘input’,
quantities are often referred to as 𝑋 and the ‘output’ quantity, i.e., the measurand, is referred to
as 𝑌.
3.4 As there will usually be several influence quantities, they are differentiated from each other by
the subscript 𝑖. So, there will be several input quantities called 𝑋𝑖 , where 𝑖 represents integer
values from 1 to 𝑁, 𝑁 being the number of such quantities. In other words, there will be input
quantities of 𝑋1 , 𝑋2 , … , 𝑋𝑁 .
3.5 Each of these input quantities will have a corresponding value. For example, one quantity might
be the temperature of the environment – this will have a value, say 23 °C. A lower-case “𝑥”
represents the estimated values of the quantities. Hence the value of 𝑋1 will be 𝑥1 , that of 𝑋2 will
be 𝑥2 , and so on.
3.6 The purpose of the measurement is to determine the best estimate of the measurand, 𝑌. As for
the input quantities, the estimated value of the measurand is represented by the lower-case
letter, i.e., 𝑦. One of the first steps is to establish the mathematical relationship 𝑌 = 𝑓(𝑋𝑖 )
between the values of the input quantities, 𝑋𝑖 , and that of the measurand, 𝑌. This process is
examined in Appendix D.
3.7 The values 𝑥𝑖 of the input quantities 𝑋𝑖 will generally all have an associated uncertainty. This can
be expressed as 𝑢(𝑥𝑖 ), the standard uncertainty of 𝑥𝑖 . The process of ‘standardising’ the
available information about the uncertainty in 𝑥𝑖 is described shortly. The uncertainty 𝑢(𝑦)
associated with 𝑦 will involve a combination of the input uncertainties 𝑢(𝑥𝑖 ).
3.8 Some uncertainties, particularly those associated with the determination of repeatability, have to
be evaluated by statistical methods. Others have to be evaluated by examining other
information, such as data in calibration certificates, evaluation of long-term drift, consideration of
the effects of environment, etc.
3.9 The GUM [1] differentiates between statistical evaluations and those using other methods. It
categorises them into two types – Type A and Type B.
3.10 A Type A evaluation of uncertainty is carried out using statistical analysis of a series of
observations. Further details about Type A evaluations can be found in Section 4.
3.11 A Type B evaluation of uncertainty is carried out using methods other than statistical analysis of
a series of observations. Further details about Type B evaluations can be found in Section 5.
3.12 In paragraph 3.3.4 of the GUM it is stated that the purpose of the Type A and Type B classification
is to indicate the two different ways of evaluating uncertainty components, and the distinction

between Type A and Type B is for convenience in discussion only. Whether components of
uncertainty are classified as `random' or `systematic' in relation to a specific measurement process
or described as Type A or Type B depending on the method of evaluation, all components
regardless of classification are modelled by probability distributions, usually quantified by their
standard deviation or variance.
3.13 Therefore, any convention as to how uncertainty evaluations are classified does not affect the
estimation of the combined uncertainty (defined in 3.36). In this guide, when the terms `random'
and `systematic' are used they refer to the effects of uncertainty on a specific measurement
process. It is the usual case that random components require Type A evaluations and systematic
components require Type B evaluations, but there are exceptions.
3.14 For example, a random effect can produce a fluctuation in an instrument's indication, which is
both noise-like in appearance and significant in terms of uncertainty. But it may only be possible
to characterise it in terms of limits to the range of indicated values. This is not a common
situation but when it occurs a Type B evaluation of the uncertainty component will be required.
This is done by assigning limit values and an associated probability distribution, as in the case of
other Type B evaluations.
3.15 The input uncertainties, associated with the values 𝑥𝑖 of the influence quantities 𝑋𝑖 , arise in a
number of forms. Some may be characterised as limit values between which little is known about
the most likely place (within the limits) where the “true” value may lie. A good example of this is
the numeric rounding caused by finite resolution described in paragraph 2.8. In this example, it is
equally likely that the underlying value is anywhere within the defined limits of ± half of the
change represented by one increment of the last displayed digit. This concept is illustrated in
Figure 1.
3.16 𝑎 𝑎
probability density
𝑥𝑖 − 𝑎 𝑥𝑖 𝑥𝑖 + 𝑎
Figure 1
The expectation value 𝑥𝑖 lies in the centre of a distribution of possible values with a
half-width, or semi-range, of 𝑎.
3.17 In the resolution example, 𝑎 = 0.5 (for the resolution of 1.0).
3.18 As all underlying values are presumed equally likely, we can say that there is equal probability of
the value of 𝑥𝑖 being anywhere within the range 𝑥𝑖 − 𝑎 to 𝑥𝑖 + 𝑎, and zero probability of it being
outside these limits.
3.19 Thus, the uncertainty contribution associated with the value 𝑥𝑖 is characterised by a probability
density function (PDF), describing the range and relative likelihood of possible values of the
measurand.
By the GUM definition, the standard uncertainty 𝑢(𝑥𝑖 ) is equal to the standard deviation of the
corresponding PDF.

3.20 The probability distribution associated with an input quantity is a reflection of the available
knowledge about that particular quantity. In many cases, there will be insufficient information
available to justify choosing a more ‘informative’ distribution than a uniform, or rectangular,
probability distribution (as in Figure 1).
3.21 If more information is available, it may be possible to assign a different probability distribution to
the value of a particular input quantity. For example, a measurement may be taken as the
difference in readings on a digital scale – typically, the zero reading will be subtracted from a
reading taken further up the scale. If the sensitivity is constant, both of these readings might
have an associated rectangular distribution of identical size. If two identical rectangular
distributions, each of magnitude ±𝑎, are combined then the resulting distribution will be
triangular with a semi-range of ±2𝑎.
probability density
𝑥𝑖 − 2𝑎 𝑥𝑖 𝑥𝑖 + 2𝑎
Figure 2
Combination of two identical rectangular distributions, each with semi-
range limits of ±𝑎, yields a triangular distribution with a semi-range of ±2𝑎.
3.22 There are other possible distributions that may be assigned. For example, when making
measurements of radio-frequency power an uncertainty arises due to imperfect matching
between the source and the termination. The imperfect match usually involves an unknown
phase angle which means that a cosine function characterises the probability distribution for the
uncertainty. Harris and Warner [17] have shown that a symmetrical U-shaped probability
distribution arises from this effect.
probability density
𝑥𝑖 − 𝑎 𝑥𝑖 𝑥𝑖 + 𝑎
Figure 3
U-shaped distribution, associated with RF mismatch uncertainty. For this situation,
𝑥𝑖 is more likely to be close to one or other of the edges of the distribution.

3.23 An evaluation of the effects of (non) repeatability, performed by statistical methods, will usually
yield a Gaussian or normal distribution. Further details on this process can be found in Section 4.
3.24 When a number of distributions of whatever form are combined it can be shown that, apart from in
exceptional cases, the resulting probability distribution tends to the normal form in accordance with
the Central Limit Theorem.[16] The importance of this fact is that it makes it possible to use the
well-known properties of the normal distribution to assign a coverage probability to the likelihood
of the true value of the measurand being within a certain range of values, known as the coverage
interval.
Figure 4
The normal, or Gaussian, probability distribution, obtained when a number of
distributions, of any form, are combined and the conditions of the Central Limit
Theorem are met. In practice, if three or more distributions of similar magnitude
are present, they will usually combine to form a reasonable approximation to the
normal distribution.
The size of the distribution is described in terms of a standard deviation. The
shaded area bounds a region 1 standard deviation from the centre of the
distribution. This corresponds to approximately 68% of the area under the curve.
3.25 The exceptional case arises when one (or more) inputs to the combined uncertainty is dominant; in
this circumstance, to varying degrees the resulting distribution resembles that of the dominant
contribution(s).
NOTE 1: If the dominant contribution is normal, then clearly the resulting distribution will also be normal.
NOTE 2: The above statement and note may not be true when the measurement model is non-linear.
3.26 Whenever input uncertainties are expressed in terms of limit values (e.g., limits of a rectangular
distribution) rather than standard deviations, some processing is needed to ‘standardise’ them to
obtain 𝑢(𝑥𝑖 ), as described below.
3.27 When it is possible to assess only the upper and lower bounds of an error (as in the case of digital
rounding) a rectangular probability distribution should be assumed for the uncertainty associated
𝑎
with this error. Then, if 𝑎𝑖 is the semi-range limit, the standard uncertainty is given by 𝑢(𝑥𝑖 ) = 𝑖 .
√3
Table 1 gives the expressions for various situations.

Table 1 - Expression used to obtain the standard uncertainty for various probability distributions
Assumed Expression used to
probability obtain the standard Comment or example
distribution uncertainty
Rectangular 𝑎𝑖 A digital thermometer gives readings to one decimal place,
𝑢(𝑥𝑖 ) =
√3 that is they are expressed to within 0.1 °C. The numeric
rounding caused by finite resolution will have semi-range
limits of 0.05 °C. Thus, the corresponding standard
uncertainty will be
𝑎𝑖 0.05 ℃
𝑢(𝑥𝑖 ) = = = 0.029 ℃
√3 1.732
U-shaped 𝑎𝑖 A mismatch uncertainty associated with the calibration of an
𝑢(𝑥𝑖 ) =
√2 RF power sensor has been evaluated as having semi-range
limits of 1.3 %. Thus, the corresponding standard uncertainty
will be
𝑎𝑖 1.3 %
𝑢(𝑥𝑖 ) = = = 0.92 %
√2 1.414
Triangular 𝑎𝑖 A tensile testing machine is used in a testing laboratory
𝑢(𝑥𝑖 ) =
√6 where the air temperature can vary randomly but relatively
quickly and does not depart from the nominal value by more
than 3 °C. The machine has a large thermal mass and is
therefore most likely to be at the mean air temperature, with
no probability of being outside the 3 °C limits. It is reasonable
to assume a triangular distribution; therefore the standard
uncertainty for its temperature is
𝑎𝑖 3℃
𝑢(𝑥𝑖 ) = = = 1.2 ℃
√6 2.449
Normal 𝑠 A statistical evaluation of repeatability uncertainty is obtained
𝑢(𝑥𝑖 ) =
√𝑛 in terms of repeatability standard deviation 𝑠 and the number
(from of values 𝑛 contributing to the reported value.
repeatability
evaluation)
Normal 𝑈 A calibration certificate normally quotes an expanded
𝑢(𝑥𝑖 ) =
𝑘 uncertainty 𝑈 at a specified, high coverage probability. A
(from a coverage factor, 𝑘, will have been used to obtain this
calibration expanded uncertainty from the combination of standard
certificate) uncertainties. It is therefore necessary to divide the
expanded uncertainty by the same coverage factor to obtain
the standard uncertainty. (See 3.42)
Normal Tolerance Sometimes specifications are quoted at a given coverage
𝑢(𝑥𝑖 ) =
𝑘 probability (historically referred to as confidence level), e.g.,
(from a 95 % or 99 %. In such cases, a normal distribution might be
specification, assumed, and the tolerance limit is divided by the coverage
e.g. a factor 𝑘 for the stated coverage probability. (See 3.46)
manufacturer’s
For a coverage probability of 95 %, 𝑘 = 2 and for a coverage
specification)
probability of 99 %, 𝑘 = 2.58.
If a coverage probability is not stated, then a rectangular

distribution should be assumed.
3.28 The quantities 𝑋𝑖 that affect the measurand 𝑌 may not have a direct, one to one, relationship with
it. There may be a scaling factor, such as a multiplicative constant or different measurement
units, or 𝑌 may not vary linearly with 𝑋𝑖 (as in the relationship between area and radius of a
circle).

3.29 For example, a dimensional laboratory may use steel gauge blocks for calibration of measuring
tools. A significant influence quantity is temperature. Because the gauge blocks have a
significant temperature coefficient of expansion, there is an uncertainty that arises in their length
due to an uncertainty in temperature
3.30 In order to translate the temperature uncertainty into an uncertainty in length units, it is
necessary to know how sensitive the length of the gauge block is to temperature. In other words,
a sensitivity coefficient is required.
In this example, the steel used in the manufacture of gauge blocks has a temperature coefficient
of expansion of approximately +11.5 × 10−6 per °C, which provides the value for the sensitivity
coefficient.
3.31 The sensitivity coefficient associated with each input estimate 𝑥𝑖 is represented by 𝑐𝑖 . It is the
partial derivative of the model function 𝑓(𝑋𝑖 ) with respect to 𝑋𝑖 , evaluated at the input estimates
𝑥𝑖 . It is given by
𝜕𝑓 𝜕𝑦
𝑐𝑖 = | ≈
𝜕𝑋𝑖 𝑋 =𝑥 𝜕𝑥𝑖
𝑖 𝑖
In other words, it describes how the output estimate 𝑦 varies with a corresponding small change
in an input estimate 𝑥𝑖 .
3.32 If the functional relationship is not well known for a particular measurement system, or it cannot
easily be differentiated, the sensitivity coefficients can usually be obtained by the practical
approach of changing one of the input variables by a known amount, whilst keeping all other inputs
constant, and noting the change in the output estimate.
3.33 𝜕𝑦 Δ𝑦
In effect, this ‘numerical’ approach approximates the partial derivative by the quotient , where
𝜕𝑥𝑖 Δ𝑥𝑖
Δ𝑦 is the change in 𝑦 = 𝑓(𝑥𝑖 ) resulting from a change Δ𝑥𝑖 in 𝑥𝑖 . It is important to choose the
magnitude of the change Δ𝑥𝑖 around 𝑥𝑖 carefully. It should be balanced between being sufficiently
large to obtain adequate numerical accuracy in Δ𝑦 and sufficiently small to provide a
mathematically sound approximation to the partial derivative. The following example illustrates this
approach.
Example
The height ℎ of a flagpole is determined

by measuring the angle obtained when
observing the top of the pole at a
specified distance 𝑑. Thus ℎ = 𝑑 tan 𝛷
Both ℎ and 𝑑 are in units of length but

are related by tan Φ. In other words,
ℎ = 𝑓(𝑑, 𝛷) = 𝑑 tan 𝛷.
If the measured distance is 7.0 m and 𝛷

the measured angle is 37°, the estimated
height is
ℎ = 7.0 tan (37°) m = 5.275 m. 𝑑

3.34 If the maximum error in 𝑑 is, say, 0.1 m then the estimate of ℎ could be anywhere between
(7.0 − 0.1). tan(37°)m and (7.0 + 0.1). tan(37°)m, i.e., between 5.200 m and 5.350 m. So, a
change of ±0.1 m in the input quantity 𝑥𝑖 has resulted in a change of ±0.075 m in the output
0.075
estimate 𝑦. The sensitivity coefficient is therefore estimated to be 𝑐𝑑 = = 0.75.
0.1
3.35 Similar reasoning can be applied to the uncertainty in the angle 𝛷. If the maximum error in 𝛷 is
0.5°, then the estimate of ℎ could be anywhere between 7.0 tan(36.5°)m and 7.0 tan(37.5°)m,
i.e., between 5.179 m and 5.371 m. A change of ±0.5° in the input quantity 𝑥𝑖 has resulted in a
change of ±0.096 m in the output estimate 𝑦. The sensitivity coefficient is therefore estimated to
be
0.096 m
𝑐𝛷 = = 0.192 metre per degree.
0.5°
3.36 Once the standard uncertainties 𝑢(𝑥𝑖 ) and the sensitivity coefficients 𝑐𝑖 have been evaluated, the
uncertainties can be combined in order to give a single value of uncertainty to be associated with
the estimate 𝑦 of the measurand 𝑌. That value is known as the combined standard uncertainty
and is represented by the symbol 𝑢c (𝑦).
NOTE: The subscript “c” in 𝑢c (𝑦) is superfluous and can be omitted. It is retained here for consistency with the GUM [1].
3.37 The combined standard uncertainty is usually calculated from:
𝑁 𝑁
𝑢c (𝑦) = √∑ 𝑐𝑖2 𝑢2 (𝑥𝑖 ) = √∑ 𝑢𝑖2 (𝑦) (1)

𝑖=1 𝑖=1
where
𝑢𝑖 (𝑦) = |𝑐𝑖 |𝑢(𝑥𝑖 )
is the standard uncertainty corresponding to the 𝑖 th input quantity, expressed in terms of the
measurand.
NOTE: Equation (1) only applies when all 𝑥𝑖 are independent otherwise GUM equation 13 should be used.
3.38 In other words, the individual standard uncertainties, expressed in terms of the measurand, are
squared; these squared values are summed, and the square root is taken.
3.39 An example of this process is presented below, using the data from the measurement of the
flagpole described above. For the purposes of the example, it is assumed that the repeatability of
the process has been evaluated by making repeated measurements of the flagpole height, giving
an estimated standard deviation of the mean of 0.05 metres. See Section 4 for further details
about the evaluation of repeatability.
Note that there is no standardised format for presenting the content of an uncertainty budget and
many variations will be encountered in practice. In this table, to save on space, the standard
uncertainties 𝑢(𝑥𝑖 ) have not been separately evaluated and reported. Instead, all calculations
are performed in a single stage which is summarised in the final column.

Sensitivity
Source of Probability Standard uncertainty
Uncertainty Divisor coefficient,
uncertainty distribution 𝑢𝑖 (𝑦)
𝑐𝑖
Distance from 0.1 m
0.1 m Rectangular √3 0.75 0.75 = 0.0433 m
flagpole √3
0.5°
Angle 0.192 m/°
0.5° Rectangular √3 0.192 m/° √3
measurement
= 0.0554 m
Measurement 0.05 m
0.05 m Normal 1 1 1 = 0.05 m
Repeatability 1
Combined standard uncertainty 𝑢c (𝑦) = √0.04332 + 0.05542 + 0.052 = 0.0863 m
NOTE 1: The columns headed “Uncertainty” and “Probability distribution” represent the known information about the
corresponding input.
The term “Uncertainty” is used here in a general sense and might, as in the case of the first two terms, correspond to the
‘half-width’ for a range of possible values (e.g., for a range ±0.1 m, the half-width is 0.1 m). In the case of the final input
for this example (measurement repeatability) it represents a standard deviation.
The “Probability distribution” summarises the nature of the information known about the respective inputs and, in
association with the “Uncertainty” information determines the relevant “Divisor”. In this example the ‘Rectangular’
distributions reflect the lack of all information other than the limit values
The “Divisor” serves to standardize the information to establish the standardised input uncertainty 𝑢(𝑥𝑖 ).
NOTE 2: As is the case for all uncertainty evaluations, the combined standard uncertainty is a consequence of applying
GUM principles to a measurement model. In this example the model is ℎ = 𝑑 tan 𝛷 + δℎ𝑟 See Appendix D Measurement
Equations or [2] for a detailed explanation of processes for establishing a measurement model.
3.40 In accordance with the Central Limit Theorem, the PDF for 𝑦 is a normal distribution with
standard deviation equal to 𝑢c (𝑦), as illustrated in Figure 5.
𝑦
5.275 m
𝑦 − 0.0863 m 𝑦 + 0.0863 m
Figure 5
The measured value 𝑦 is at the centre of a normal distribution
with a standard deviation equal to 𝑢c (𝑦). The numerical
values relate to the example discussed above.
3.41 For a normal distribution,  one standard deviation encompasses 68.3 % of the area under the
curve. This means that there is about 68 % probability that the measurand lies within these
limits.

3.42 The GUM recognises the need for providing a coverage interval with a higher coverage probability
and achieves this by defining the coverage interval in terms of expanded uncertainty, 𝑈, which is
obtained by multiplying the combined standard uncertainty by a coverage factor. The coverage
factor is given the symbol 𝑘, thus the expanded uncertainty is given by
𝑈 = 𝑘 𝑢c (𝑦) (2)
Where necessary to avoid ambiguity, subscripts can be attached to both 𝑈 and 𝑘 to reflect
the corresponding coverage probability, e.g., 𝑈𝑝 , 𝑘𝑝 , 𝑈95% , 𝑘95% , …
3.44 In accordance with generally accepted international practice, it is recommended that a coverage
factor of 𝑘 = 2 is used to calculate the expanded uncertainty. This value of 𝑘 will give a
coverage probability of approximately 95 %, assuming a normal distribution.
NOTE: A coverage factor of 𝑘 = 2 actually provides a coverage probability of 95.45 % for a normal distribution. For
convenience this is approximated to 95 % (which actually corresponds to a coverage factor of 𝑘 = 1.96). However, the
difference is not generally found to be significant when model assumptions and the reliability of input quantities is taken
into consideration.
3.45 For example:

The measurement of the height of the flagpole had a combined standard uncertainty 𝑢c (𝑦) of
0.0863 m. Hence the expanded uncertainty
𝑈 = 𝑘 𝑢c (𝑦) = 2 × 0.0863 m = 0.17 m.
3.46 There may however be situations where a different coverage probability is required. For
example, in safety-critical situations a higher coverage probability may be more appropriate.
The table below gives the coverage factor necessary to obtain various levels of coverage for
a normal distribution.
Coverage probability Coverage factor

𝑝 𝑘
90 % 1.64
95 % 1.96
95.45 % 2.00
99 % 2.58
99.73 % 3.00
3.47 There may also be situations where a normal distribution cannot be assumed, and a different
coverage factor may be needed in order to obtain a coverage probability of approximately 95 %.
Such situations are described in Appendix B and Appendix C.

4 Type A Evaluation of Standard Uncertainty
4.1 If an uncertainty is evaluated by statistical analysis of a series of observations, it is known as a

Type A evaluation.
4.2 A Type A evaluation will normally be used to obtain a value for the repeatability uncertainty of a
measurement process. For some measurements, this ‘random’ component of uncertainty may not
be significant in relation to other contributions to uncertainty. It is nevertheless desirable for any
measurement process that the relative importance of random effects be established.
When there is a spread in a sample of measurement results, the arithmetic mean (average) of the
results should be calculated. If there are 𝑚 independent repeated values for a quantity, 𝑄 then the
mean value 𝑞̅ is given by:
𝑚
1 𝑞1 + 𝑞2 + ⋯ + 𝑞𝑚
𝑞̅ = ∑ 𝑞𝑗 = (3)
𝑚 𝑚
𝑗=1
4.3 The values obtained (𝑞𝑗 ) are considered to be a random, finite sample arising from a
measurement process whose underlying variability is characterised by a standard deviation .
It is instructive to ask - if we repeated the set of measurements… would we obtain the same
mean value? I.e., would we get the same value for 𝑞̅ ? This seems unlikely (except in the case
when the measurements are limited by poor resolution). In most cases, we would actually see a
distribution of values for 𝑞̅.
For samples of size 𝑛 the standard deviation of the distribution of these sample means is 𝜎⁄√𝑛,
known as the standard deviation of the mean (sometimes referred to as the standard error).
4.4 In practice however, it is not usually possible to obtain the value of 𝜎 and an estimate 𝑠 is instead
used, thus standard repeatability uncertainty
𝑠
𝑢rep = (4)
√𝑛
4.5 The dataset used to evaluate the estimate 𝑞̅ can be used to obtain an estimate 𝑠 for the standard
deviation .
𝑚
1 2
𝑠 = 𝑠(𝑞𝑗 ) = √ ∑(𝑞𝑗 − 𝑞̅ ) (5)
(𝑚 − 1)
𝑗=1

4.6 Example: Four measurements were made to estimate the value of a quantity 𝑞 and the
repeatability for the value. The results obtained were 3.42, 3.88, 2.99 and 3.17.
The mean value,
𝑚
1 3.42 + 3.88 + 2.99 + 3.17
𝑞̅ = ∑ 𝑞𝑗 = = 3.365
𝑚 4
𝑗=1
The estimated standard deviation,

𝑚
1 2
𝑠 = 𝑠(𝑞𝑗 ) = √ ∑(𝑞𝑗 − 𝑞̅ ) = 0.386
(𝑚 − 1)
𝑗=1
The estimated repeatability is therefore,

𝑠 0.386
𝑢rep = = = 0.193
√𝑛 √4
where in this case 𝑛 = 𝑚.
4.7 It may not always be practical or possible to repeat the measurement many times during a test or
a calibration. In these cases, a more reliable estimate of the standard deviation may sometimes
be obtainable from previously obtained data, based on a larger number of readings.
This approach must be treated with caution – it relies on the reliability of the previously obtained
data to represent the variation in the present measurements, i.e., it assumes that the underlying
standard deviation 𝜎 is the same in both cases. A previous estimate of standard deviation can
only be used if there has been no subsequent change in the measuring system or procedure that
could have an effect on the repeatability. If an apparently excessive spread in measurement
values is found, the cause should be investigated and resolved before proceeding further.
4.8 Example: Suppose that two measurements were made to estimate the value of a quantity 𝑥, i.e.,
𝑛 = 2.
However, the repeatability is to be estimated from 𝑚 = 20 previously obtained measurements,
with standard deviation
𝑠 = 0.247
The estimated repeatability is therefore,

𝑠 0.247
𝑢rep = = = 0.175
√𝑛 √2
NOTE: The degrees of freedom under such circumstances are 𝑚– 1, where 𝑚 is the number of measurements in the prior
evaluation. Indeed, this is the reason that a large number of readings in a prior evaluation can give a more reliable
estimate when only a few measurements can be made during the routine procedure. Degrees of freedom are discussed
further in Appendix B.

5 Type B Evaluation of Standard Uncertainty
5.1 If an uncertainty is evaluated by non-statistical analysis, it is known as a Type B evaluation.
5.2 Type B measurement uncertainties usually arise from fixed but unknowable (or poorly known)
measurement errors. In evaluating the components of uncertainty, it is necessary to consider and
include at least the following possible sources:
(a) For measuring instruments - the imported uncertainties associated with their calibration and any
drift or instability in their values or readings.
(b) The reported uncertainty assigned to reference materials and any drift or instability in their values.
(c) Effects arising from the use of ancillary equipment, including items such as connecting leads,
pipework, heaters etc., and any drift or instability in their values or readings.
(d) The equipment or item being measured, for example its resolution and any instability during the
measurement. It should be noted that the anticipated long-term performance of an item being
calibrated is not normally included in the uncertainty evaluation for that calibration.
(e) The operational procedure.
(f) The effects of environmental conditions on any or all of the above.
5.3 Having identified all the possible Type B components of uncertainty based as far as possible on
experimental data or on theoretical grounds, they should be characterised in terms of standard
uncertainties based on assigned probability distributions. The probability distribution of an
uncertainty obtained from a Type B evaluation can take a variety of forms, but it is generally
acceptable to assign a well-defined distribution for which the standard uncertainty can be
obtained from a simple calculation. These distributions and sample calculations are presented in
paragraphs 3.15 to 3.22 and in more detail elsewhere, e.g., JCGM-101 [2].
NOTE: It is a basic feature of the GUM framework that standard uncertainty is taken as the standard deviation of the
assigned probability distribution.
5.4 Whenever possible, corrections should be made for known errors revealed by calibration or other
sources. (It is not possible to make corrections for (random) repeatability errors)
The convention is that an error is given a positive sign if the measured value is greater than the
expected value. The correction for error therefore involves subtracting the error from the measured
value. On occasions, to simplify the measurement process, it may be convenient to treat such an
error, when it is small compared with other uncertainties (and when doing so has an insignificant
effect upon the overall evaluation) as if it were a systematic uncertainty of the uncorrected error
magnitude.
5.5 Measurement errors should not be confused with mistakes.

Common examples of mistakes are incorrectly applied corrections, transcription errors, and
faults in software designed to control or report on a measurement process. The effects of such
mistakes cannot readily be included in the evaluation of uncertainty and care is needed to avoid
them.

6 Reporting of Results
6.1 After the expanded uncertainty has been calculated, usually for a coverage probability of 95 %,
the estimated value of the measurand and associated expanded uncertainty should be reported
as
𝑦 ± 𝑈 and accompanied by the following statement:
6.2 "The reported expanded uncertainty is based on a standard uncertainty multiplied by a

coverage factor 𝑘 = 2, providing a coverage probability of approximately 95 %. The
uncertainty evaluation has been carried out in accordance with UKAS requirements".
6.3 In cases where the procedure of Appendix B has been followed the actual value of the coverage
factor should be substituted for 𝑘 = 2 and the following statement used:
6.4 "The reported expanded uncertainty is based on a standard uncertainty multiplied by a

coverage factor 𝑘 = 𝑋𝑋, which for a t-distribution with 𝑣eff = 𝑌𝑌 effective degrees of
freedom corresponds to a coverage probability of 95.45 %. The uncertainty evaluation
has been carried out in accordance with UKAS requirements".
6.5 In circumstances where a dominant non-Gaussian Type B contribution occurs, and the procedures
described in Appendix C have been followed the following statements should be used:
6.6 “The reported expanded uncertainty is based on a standard uncertainty obtained by

combining a dominant Type B uncertainty with other smaller uncertainties. The standard
uncertainty has been multiplied by a coverage factor 𝑘 = 𝑋𝑋 which, for this particular
combination, corresponds to a coverage probability of 95.45 %. The uncertainty
evaluation has been carried out in accordance with UKAS requirements.
For the purpose of further propagation, the measurement uncertainty can be imported
into subsequent uncertainty budgets in terms of:
1. a rectangular distribution with half width of 𝑎R = 𝑌𝑌, and
2. a normal distribution with a standard uncertainty 𝑢N = 𝑍𝑍”
6.7 If uncertainty is being reported as an analytical expression, refer also to Appendix L.
6.8 For the purposes of this guide "approximately" is interpreted as meaning sufficiently close that
any difference may be considered to be insignificant.
6.9 Uncertainties are usually expressed in units of the measurand or as relative values, for example as
μV
a percentage (%), parts per million (ppm), parts in 10𝑥 , , etc.
V
6.10 Measurement uncertainties should generally be reported to two decimal digits, as it is seldom
justified to report more. The numerical form of the measured value in the final statement should be
reported with the same number of decimal places as the measurement uncertainty.
6.11 Rounding should always be carried out at the end of the process in order to avoid the effects of
cumulative rounding errors.
6.12 In situations where the PDF describing inputs to an uncertainty evaluation are asymmetric, or
where the measurement model is non-linear, the resulting PDF for the measurand may also be
asymmetric. In such cases Monte Carlo Simulation (as described, for example, in JCGM-101 [2])
offers a more suitable approach to evaluation of measurement uncertainty and coverage intervals.

7 Step by Step Procedure for Evaluation of Measurement Uncertainty
The following is a step-by-step guide to the use of this guide for the treatment of uncertainties. The
left-hand column in the table gives the general case while the right-hand column indicates how this
relates to example K.4 in Appendix K. Although this example describes a calibration activity, the
process is quite general and applies to most measurement situations.
Example K.4: Calibration of a weight of nominal

General case
value 10 kg of OIML Class M1
7.1 Determine the mathematical relationship between It will be assumed that the unknown weight, 𝑊𝑋 , can be
values of the input quantities and that of the obtained from the following relationship:
measurand:
𝑊𝑋 = 𝑊𝑆 + δ𝐷𝑆 + δ𝐼𝑑 + δ𝐶 + ∆𝐴𝑏 + δ𝑊𝑟
𝑌 = 𝑓(𝑋1 , 𝑋2 , … 𝑋𝑁 )
See Appendix D for details.
7.2 Identify all corrections that have to be applied to It is not normal practice to apply corrections for this
the results of measurements of a quantity class of weight and the comparator has no
(measurand) for the stated conditions of measurable linearity error.
measurement.
Estimates for these values are therefore taken to be:
Drift of standard mass since last calibration δ𝐷𝑆 = 0
Effect of least significant digit resolution δ𝐼𝑑 = 0
Comparator linearity correction δ𝐶 = 0
Correction for air buoyancy ∆𝐴𝑏 = 0
Repeatability error δ𝑊𝑟 = 0
However, finite (i.e., non-zero) uncertainty

contributions have been determined for each of
these inputs.


General case
7.3 List components of uncertainty associated with

Type B inputs, including corrections and Source of uncertainty uncertainty Distribution*
uncorrected systematic errors treated as (mg)
uncertainties.
𝑊𝑆 Calibration of std. mass 30 N ( 𝑘 = 2)
Seek prior experimental work or theory as a basis
for assigning uncertainties and probability distribu- δ𝐷𝑆 Drift of standard mass 30 R
tions to the systematic components of uncertainty.
δ𝐼𝑑 Resolution effects 10 T
Calculate the standard uncertainty for each
component of uncertainty, obtained from Type B δ𝐶 Comparator linearity 3 R
evaluations, as in Table 1.
∆𝐴𝑏 Air buoyancy 10 R
For assumed rectangular distributions:
Then, in standardised form:
𝑎𝑖
𝑢(𝑥𝑖 ) =
√3 𝑢(𝑥1 ) = 𝑢(𝑊𝑆 ) =
30 mg
= 15 mg
2
For assumed triangular distributions:
30 mg
𝑢(𝑥2 ) = 𝑢(δ𝐷𝑆 ) = = 17.3 mg
√3
𝑎𝑖
𝑢(𝑥𝑖 ) =
√6 10 mg
𝑢(𝑥3 ) = 𝑢(δ𝐼𝑑 ) = = 4.08 mg
√6
For assumed normal distributions:
3 mg
𝑢(𝑥4 ) = 𝑢(δ𝐶) = = 1.73 mg
√3
𝑈
𝑢(𝑥𝑖 ) =
𝑘 10 mg
𝑢(𝑥5 ) = 𝑢(∆𝐴𝑏 ) = = 5.77 mg
√3
or consult [2] if the assumed probability distribution
is not covered in this publication. *see glossary for explanation of labels, N, R, T.
7.4 Use prior knowledge or make trial measurements From previous knowledge of the measurement process it
and calculations to determine if there is to be a is known that there is a significant random component of
random component of uncertainty that is uncertainty.
significant compared with the effect of the other
components of uncertainty.
7.5 If a random component of uncertainty is significant Three measurements were made of the difference
make repeated measurements to obtain the mean between the unknown weight and the standard weight,
from equation (3): from which the mean difference was calculated:
𝑚 0.015 + 0.025 + 0.020

1 𝑞1 + 𝑞2 + ⋯ + 𝑞𝑚 ̅𝑆 =
𝑊 = 0.020 g
𝑞̅ = ∑ 𝑞𝑗 = 3
𝑚 𝑚
𝑗=1


General case
7.6 If correlation is suspected use the guidance in None of the input quantities is considered to be
paragraph D.3 or consult other referenced correlated to any significant extent; therefore
documents. Equation (1) can be used to calculate the combined
standard uncertainty
7.7 Either calculate the standard deviation of the mean

A previous Type A evaluation had been made to
value from equations (4) and (5):
determine the repeatability of the comparison using
the same type of 10 kg weights. The standard
𝑚 deviation was determined from 𝑚 = 10
1 2
𝑠 = 𝑠(𝑞𝑗 ) = √ ∑(𝑞𝑗 − 𝑞̅ ) measurements using the conventional bracketing
(𝑚 − 1)
𝑗=1 technique and was calculated, to be
𝑠(δ𝑊𝑟 ) = 8.7 mg
𝑠 𝑠 (𝑞𝑗 ) Since the number of determinations taken when

𝑢rep = =
√𝑛 √𝑛 calibrating the unknown weight was 3 this is the value
of 𝑛 that is used to calculate the standard deviation of
or refer to the results of previous repeatability the mean using equation (4):
evaluations with an estimate 𝑠𝑝 based on a larger
𝑠(δ𝑊𝑟 ) 8.7
number of readings: 𝑢(𝑥6 ) = 𝑢rep = = = 5.0 mg
√𝑛 √3
𝑠 = 𝑠𝑝
𝑠 𝑠𝑝
𝑢rep = =
√𝑛 √𝑛
where 𝑚 is the number of readings used in the
evaluation of 𝑠 and 𝑛 is the number of readings that
contribute to the evaluation of the mean value.


General case
7.8 The units of all standard uncertainties are in terms of

Calculate the combined standard uncertainty for
uncorrelated input quantities using equation (1) if those of the measurand, i.e., milligrams, and the
functional relationship between the input quantities
absolute values are used:
and the measurand is a linear summation; therefore,
𝑁 𝑁 all the sensitivity coefficients are unity (𝑐𝑖 = 1).
𝑢c (𝑦) = √∑ 𝑐𝑖 2 𝑢2 (𝑥𝑖 ) = √∑ 𝑢𝑖 2 (𝑦)
𝑖=1 𝑖=1 Applying equation (1) gives
where 𝑐𝑖 is the partial derivative

𝑢c (𝑊𝑋 ) = √152 + 17.322 + 4.082 + 1.732 + 5.772 + 52
𝜕𝑓 𝜕𝑦
𝑐𝑖 = | ≈ ,
𝜕𝑋𝑖 𝑋 =𝑥 𝜕𝑥𝑖
𝑖 𝑖 𝑢c (𝑊𝑋 ) = 25 mg
or a known sensitivity coefficient.
Alternatively use equation (8) if the standard

uncertainties are all relative values and 𝑌 = 𝑓(𝑋𝑖 ) is a
pure product of terms
𝑁 2
𝑢c (𝑦) 𝑝𝑖 𝑢(𝑥𝑖 )
= √∑ [ ]
𝑦 |𝑥𝑖 |
𝑖=1
where 𝑝𝑖 are known positive or negative exponents in

the functional relationship.
7.11 If there is a dominant Type A contribution (i.e., a

𝑈95% = 2 × 24.55 mg = 49 mg
significant value obtained from a small number of
readings), use Appendix B to calculate a suitable
alternative value for coverage factor 𝑘 and use this A coverage probability of approximately 95 % is
value to calculate the expanded uncertainty. obtained with 𝑘 = 2. It was not necessary to use an
alternative value for 𝑘 as there are no dominant
Similarly, if there is a dominant Type B contribution, Type A or Type B contributions.
use Appendix C to calculate a suitable alternative
value for coverage factor 𝑘 and use this value to NOTE:
calculate the expanded uncertainty. The evaluation includes a single Type A component for
which the standard deviation, 𝑠(δ𝑊𝑟 ) is based upon 𝑚 = 10
Then, calculate the expanded uncertainty values, for which there are
𝜐rep = (𝑚 − 1) = 9 degrees of freedom.
𝑈𝑝 = 𝑘𝑝 𝑢c (𝑦) All other components are Type B, for which 𝜐 = , therefore
𝑢c 4 (𝑦)
𝜐eff = 𝑢rep 4 (𝑦)
= 5145, i.e., the Type A input is not dominant
(𝑚−1)
Also, the 𝑟-value for the largest Type B component is 0.99,

much less that a threshold value of 1.42 for dominant Type B
7.12 Report the result and the expanded uncertainty in

The measured value of the 10 kg weight is
accordance with Section 6.
10 000.025 g ± 0.049 g.
The reported expanded uncertainty is based on a

standard uncertainty multiplied by a coverage factor
𝑘 = 2, providing a coverage probability of
approximately 95 %. The uncertainty evaluation has
been carried out in accordance with UKAS
requirements.

Appendix A Calibration and Measurement Capability
A.1 The Calibration and Measurement Capability (CMC), defines the measurement capabilities,
ranges and boundaries of a calibration activity. In particular it defines the lowest measurement
uncertainty that can be achieved during a calibration under normal conditions.
A.2 For an accredited calibration laboratory, the CMC is described in the Schedule of Accreditation.
The associated measurement uncertainty relates to the calibration of real items for which the
laboratory has been accredited, using processes that were the subject of assessment.
A.3 The measurement uncertainty is calculated according to the procedures described in this guide,
and in the GUM [1], and is normally stated as an expanded uncertainty at a coverage probability
of 95 %, which usually requires a coverage factor of 𝑘 = 2.
A.4 An accredited laboratory is not permitted to quote a smaller uncertainty in certificates issued
under its accreditation but may report an equal or larger uncertainty if appropriate. For example, if
a particular item under calibration itself contributes significantly to the uncertainty (e.g., through
limited resolution or significant non-repeatability) then the uncertainty reported on a calibration
certificate will naturally be increased to account for such factors.
A.5 Refer to ILAC-P14 [11] and to UKAS LAB 45 [15] for further explanation of Calibration and
Measurement Capability.
NOTE: The term CMC also applies to the measurement capabilities of National Metrology Institutes that are published in
the BIPM key comparison database (KCDB) of the CIPM MRA.

Appendix B Coverage Factor when there is a single dominant Type A input
B.1 In some cases, it may not be practical to base a Type A evaluation on a large number of readings.
In these situations, if it is also the case that the Type A input is a significant part of the combined
uncertainty, this could result in the coverage probability being significantly less than 95 % if a
coverage factor of 𝑘 = 2 is used. In these situations, the value of 𝑘, or more precisely 𝑘𝑝 , where 𝑝
is the confidence probability, should be based on a t-distribution rather than a normal distribution.
This value of 𝑘𝑝 will give an expanded uncertainty, 𝑈𝑝 , that maintains the coverage probability at
approximately the required level 𝑝.
𝑦– 𝑘𝑢 𝑦 𝑦 + 𝑘𝑢
Figure 6
In Figure 6, the solid line depicts a normal distribution with standard deviation (standard
uncertainty) 𝑢.
A specified proportion 𝑝 of the values under the curve are encompassed by the interval
between 𝑦– 𝑘𝑢 and 𝑦 + 𝑘𝑢.
An example of the t-distribution is superimposed, using dashed lines. For the
t-distribution, a greater proportion of the values lies outside the interval 𝑦– 𝑘𝑢 to 𝑦 + 𝑘𝑢,
and a smaller proportion lies inside this region. An increased value of 𝑘 is therefore
required to restore the original coverage probability. This new coverage factor, 𝑘𝑝 , is
obtained by evaluating the effective degrees of freedom of 𝑢𝑐 (𝑦) and obtaining the
corresponding value 𝑡𝑝 , e.g., from a t-distribution table. The required coverage factor is
then 𝑘𝑝 = 𝑡𝑝
B.2 In order to obtain a value for 𝑘𝑝 it is necessary to obtain an estimate of the effective degrees of
freedom, 𝜐eff , of the combined standard uncertainty 𝑢c (𝑦). The GUM [1] recommends that the
Welch-Satterthwaite equation is used to calculate a value for 𝜐eff based on the degrees of freedom,
𝜐𝑖 , of the individual standard uncertainties 𝑢𝑖 (𝑦); where
𝑢c 4 (𝑦)
𝜐eff =
𝑢𝑖 4 (𝑦) (6)
∑𝑁
𝑖=1 𝜐𝑖
B.3 The degrees of freedom, 𝜐𝑖 , for contributions obtained from Type A evaluations are 𝑚 − 1, where
𝑚 is the number of values used to evaluate 𝑢𝑖 (𝑦).
B.4 It is often possible to take the degrees of freedom, 𝜐𝑖 , of Type B uncertainty contributions as being
infinite, that is, their value is known with a very high degree of reliability. If this is the case then the
calculation simplifies, as all the terms relating to the Type B uncertainties become zero. This case
is illustrated in the example in paragraph B.10.
B.5 However, it is possible for a Type B contribution to come from a calibration certificate, in the form of
an uncertainty based on a t-distribution (as is described in this Appendix) rather than a normal

distribution. This is an example of a Type B contribution that does not have infinite degrees of
freedom. In this eventuality the degrees of freedom will be as quoted on the calibration certificate.
B.6 Having obtained a value for 𝜐eff , the corresponding value from the t-distribution can be obtained,
either from tables (such as the table below), or by calculation.
B.7 Unless otherwise specified, the values corresponding to a coverage probability of 𝑝 = 95.45 %
should be used.
B.8 Normally 𝜐eff will not be an integer and, when using tabulated data, it will be necessary to
interpolate between the values given in the table. Linear interpolation will suffice for 𝜐eff > 3 in the
table provided below, higher-order interpolation should be used otherwise, or else the next lower
value of 𝜐eff may be used.
To calculate 𝑡𝑝 directly the Excel spreadsheet function T.INV.2T(1 − 𝑝, 𝜐eff )

can be used.
B.9 The required coverage factor is then 𝑘𝑝 = 𝑡𝑝 . This is the coverage factor required to calculate the
expanded uncertainty, 𝑈𝑝 , from 𝑈𝑝 = 𝑘𝑝 𝑢c (𝑦).
Unless otherwise specified, the coverage probability 𝑝 will usually be 95.45 %.
B.10 Example
B.10.1 In a measurement system a Type A evaluation, based on 4 observations, gave a value of 𝑢𝑖 (𝑦) of
3.5 units. There were 5 other contributions all based on Type B evaluations for each of which
infinite degrees of freedom had been assumed. The combined standard uncertainty, 𝑢c (𝑦), had a
value of 5.7 units.
Then, using the Welch-Satterthwaite equation:
5.74 5.74
𝜐eff = = × 3 = 21.1
3.54 3.54
+0+0+0+0+0
4−1
B.10.2 In the t-distribution table, the value of 𝜐eff for a coverage probability 𝑝 = 95.45 %, immediately
lower than 21.1 is 20. This gives a value for 𝑘𝑝 of 2.13 and this is the coverage factor that should
be used to calculate the expanded uncertainty…
The expanded uncertainty is 5.7 × 2.13 = 12.14 units.

Degrees of Values of 𝑡𝑝 (𝜐) from the t-distribution for 𝜐 degrees of freedom

freedom, 𝜐 𝑝 = 68.27 % 𝑝 = 90 % 𝑝 = 95 % 𝑝 = 95.45 % 𝑝 = 99 % 𝑝 = 99.73 %
1 1.84 6.31 12.71 13.97 63.66 235.78

2 1.32 2.92 4.30 4.53 9.92 19.21
3 1.20 2.35 3.18 3.31 5.84 9.22
4 1.14 2.13 2.78 2.87 4.60 6.62
5 1.11 2.02 2.57 2.65 4.03 5.51
6 1.09 1.94 2.45 2.52 3.71 4.90

7 1.08 1.89 2.36 2.43 3.50 4.53
8 1.07 1.86 2.31 2.37 3.36 4.28
9 1.06 1.83 2.26 2.32 3.25 4.09
10 1.05 1.81 2.23 2.28 3.17 3.96
11 1.05 1.80 2.20 2.25 3.11 3.85

12 1.04 1.78 2.18 2.23 3.05 3.76
13 1.04 1.77 2.16 2.21 3.01 3.69
14 1.04 1.76 2.14 2.20 2.98 3.64
15 1.03 1.75 2.13 2.18 2.95 3.59
16 1.03 1.75 2.12 2.17 2.92 3.54

17 1.03 1.74 2.11 2.16 2.90 3.51
18 1.03 1.73 2.10 2.15 2.88 3.48
19 1.03 1.73 2.09 2.14 2.86 3.45
20 1.03 1.72 2.09 2.13 2.85 3.42
25 1.02 1.71 2.06 2.11 2.79 3.33

30 1.01 1.70 2.04 2.09 2.75 3.27
35 1.01 1.70 2.03 2.07 2.72 3.23
40 1.01 1.68 2.02 2.06 2.70 3.20
45 1.01 1.68 2.01 2.06 2.69 3.18
50 1.01 1.68 2.01 2.05 2.68 3.16

100 1.005 1.660 1.984 2.025 2.626 3.077
∞ 1.000 1.645 1.960 2.000 2.576 3.000

Appendix C Coverage Factor when there is a single dominant Type B input
C.1 It is quite common in some measurement processes, particularly calibrations, for there to be a
component of uncertainty derived from a Type B evaluation that is dominant in magnitude
compared all other components. In these circumstances it can no longer be assumed that the
Central Limit Theorem applies (resulting in a normal distribution for the output, 𝑦.)
C.2 A commonly encountered example arises from the resolution of a digital indicating instrument. This
will have a rectangular distribution which, if the half-width 𝑎 is large, may dominate the shape of
the distribution for output 𝑦. Consequently, the properties of the normal distribution can no longer
be used to establish the coverage factor for a required coverage probability.
C.3 As calibration results are required to be presented in terms of a coverage interval, a different
approach has to be taken to obtain a suitable coverage factor. A method based upon the use of
tabulated values is described below.
C.4 The presence of a dominant rectangular input quantity can be detected by a ‘rule of thumb’…
for a 95.45 % coverage probability, it will ‘dominate’ (i.e. the error if using a coverage factor of
𝑘 = 2 will be more than 5 %) if its standard uncertainty 𝑢R (𝑦) is more than about 1.42 times the
combined standard uncertainty 𝑢N (𝑦) for the remaining 𝑁 − 1 inputs (assuming that these result
in a normal distribution).
𝑢R (𝑦) 𝑢N (𝑦)
I.e., a rectangular term ‘dominates’ when ≳ 1.42, or inversely ≲ 0.70
𝑢N(𝑦) 𝑢R (𝑦)
where
𝑎 𝑎R
𝑢R (𝑦) = 𝑐R =
√3 √3
which defines effective half width 𝑎R = 𝑐R 𝑎,
and (excluding the dominant term)
𝑁−1
𝑢N (𝑦) = √ ∑ 𝑢𝑖2 (𝑦)

𝑖=1
which as expected combine to give
𝑢𝑐 (𝑦) = √𝑢R2 (𝑦) + 𝑢N

2 (𝑦)
C.5 Example
A digital voltmeter is calibrated by observing its indication 𝑉ind when it measures an applied
reference voltage 𝑉ref . The measurement error 𝑉 is: 𝑉 = 𝑉ind − 𝑉ref .
The reference voltage is 𝑉ref = 1.00000 V; and the digital voltmeter reading is 𝑉ind = 1.001 V.
The expanded uncertainty (𝑘 = 2) of the applied voltage is 𝑈(𝑉ref ) = 0.00019 V.

The only other uncertainty of significance is due to the rounding of the voltmeter display, which
has a resolution of 0.001 V; therefore, there will be a possible rounding error of ±0.0005 V for
which a rectangular probability distribution is assumed.
The standard uncertainty associated with the potentially dominant rectangular term is
𝑎 0.0005 V
𝑢R (𝑦) = 𝑐𝑅 =1× = 0.000289 V
√3 √3
The standard uncertainty associated with the remaining terms (in this case there is just one
term) is
𝑁−1
0.00019 V
𝑢N (𝑦) = √ ∑ 𝑢𝑖2 (𝑦) = = 0.000095 V
2
𝑖=1
𝑢N(𝑦)
The value of the corresponding ratio is = 0.329
𝑢R (𝑦)
The value is less than 0.70 which indicates that the term dominates, and a normal distribution
cannot be assumed for 𝑦
C.6 A suitable coverage factor can be obtained from tables of factors derived from a convolution
between the dominant distribution and the distribution assumed for the combined remainder of
the terms (usually a normal distribution)
C.7 If a rectangular distribution and a normal distribution are convolved, the coverage factor 𝑘 for a
coverage probability of 95.45 % may be obtained from the following table:
𝑢N (𝑦) 𝑢N (𝑦) 𝑢N (𝑦)

𝑘95.45% 𝑘95.45% 𝑘95.45%
𝑢R (𝑦) 𝑢R (𝑦) 𝑢R (𝑦)
0.00 1.65 0.50 1.84 0.95 1.95
0.10 1.66 0.55 1.85 1.00 1.95
0.15 1.68 0.60 1.87 1.10 1.96
0.20 1.70 0.65 1.89 1.20 1.97
0.25 1.72 0.70 1.90 1.40 1.98
0.30 1.75 0.75 1.91 1.80 1.99
0.35 1.77 0.80 1.92 2.00 1.99
0.40 1.79 0.85 1.93 2.50 2.00
0.45 1.82 0.90 1.94 ∞ 2.00
C.8 For the example introduced above the coverage factor 𝑘 = 1.76.
The standard uncertainty for the example is
𝑢𝑐 (𝑦) = √𝑢R2 (𝑦) + 𝑢N

2 (𝑦)
= √0.000 2892 + 0.000 0952 V = 0.000 30 V
and the corresponding expanded uncertainty is
𝑈 = 𝑘 𝑢𝑐 (𝑦) = 1.76 × 0.000 304 V = 0.000 54 V

C.9 This approach provides a suitable coverage factor for establishing the coverage interval,
however for the purposes of propagation of uncertainties it is more helpful also to describe the
uncertainty in two parts - corresponding to the dominant distribution, and the distribution
representing the balance of the terms (usually taken as a normal distribution)
“The reported expanded uncertainty is based on a standard uncertainty obtained by

combining a dominant Type-B uncertainty with other smaller uncertainties. The standard
uncertainty has been multiplied by a coverage factor 𝑘 = 𝑋𝑋 which, for this particular
combination, corresponds to a coverage probability of 95.45 %.
1. a rectangular distribution with half width of 𝑎R = 𝑌𝑌, and
2. a normal distribution with a standard uncertainty 𝑢N = 𝑍𝑍”
C.10 The example described at C1.5 can be summarised and reported as follows
Symbol Source of uncertainty Value / V Probability distribution Divisor 𝑐𝑖 𝑢𝑖 (𝑉) / V

𝑽𝐫𝐞𝐟 Uncertainty of applied voltage 0.00019 Normal 2 1 0.000095
𝑉ind Digital rounding of indicator 0.0005 Rectangular √3 1 0.000289
Convolved
𝑢c (𝑉) Combined standard uncertainty 0.000095 0.000304
= 0.329
0.000289
Convolved
𝑈 Expanded uncertainty 0.000535
𝑘 = 1.76
Reported result:
For an applied voltage of 1.00000 V the voltmeter reading was 1.001 V
The measurement error was 𝑉 = 𝑉ind − 𝑉ref = 0.00100 V ± 0.00054 V
The reported expanded uncertainty is based on a standard uncertainty obtained by

combining a dominant Type B uncertainty with other smaller uncertainties. The standard
uncertainty has been multiplied by a coverage factor 𝑘 = 1.76 which, for this particular
combination, corresponds to a coverage probability of 95.45 %.
1. a rectangular distribution with half width of 𝑎R = 0.0005 volts,
2. a normal distribution with a standard uncertainty 𝑢N = 0.000095 volts”
The table below demonstrates how this uncertainty might be imported into a subsequent
uncertainty budget for the use of the voltmeter, seen in this partially completed example as the
entries in the first two rows:
Symbol Source of uncertainty Value / V Probability distribution Divisor 𝑐𝑖 𝑢𝑖 (𝑉) / V

Calibration error,
∆𝑽 0.0005 Rectangular √3
uncertainty term 1, 𝒂𝐑
Calibration error,
∆𝑉 0.000095 Normal 1
uncertainty term 2, 𝑢N
⋮ Other sources relating to use… ⋮ ⋮ ⋮
𝑉ind Rounding of indication in use 0.0005 Rectangular √3
⋮ ⋮ ⋮ ⋮

C.11 The use of a two-part expression such as this means that when both components are imported
into a subsequent uncertainty budget it is possible that 𝑎 will no longer be a dominant
component and a normal distribution can be assumed for the subsequent combined standard
uncertainty.
C.12 The same situation may be encountered with other distributions associated with Type B
uncertainties. An example is the U-shaped distribution associated with mismatch uncertainty in RF
and microwave systems. Similar reasoning applies here, and the suggested coverage probability
statement can be modified accordingly.

C.13 If a U-shaped distribution and a normal distribution are convolved, the coverage
factor 𝑘 for a coverage probability of 95.45 % may be obtained from the following
table:
𝑢N (𝑦) 𝑢N (𝑦) 𝑢N (𝑦)
𝑘95.45% 𝑘95.45% 𝑘95.45%
𝑢U−shape (𝑦) 𝑢U−shape (𝑦) 𝑢U−shape (𝑦)
0.00 1.41 0.50 1.77 0.95 1.93
0.10 1.47 0.55 1.80 1.00 1.93
0.15 1.51 0.60 1.82 1.10 1.95
0.20 1.55 0.65 1.84 1.20 1.96
0.25 1.60 0.70 1.86 1.40 1.97
0.30 1.64 0.75 1.88 1.80 1.99
0.35 1.67 0.80 1.89 2.00 1.99
0.40 1.71 0.85 1.90 2.50 2.00
0.45 1.74 0.90 1.92 ∞ 2.00
C.14 If a U-shaped distribution and a rectangular distribution are convolved, the

coverage factor 𝑘 for a coverage probability of 95.45 % may be obtained from the
following table:
𝑢R (𝑦) 𝑢R (𝑦) 𝑢R (𝑦)
𝑘95.45% 𝑘95.45% 𝑘95.45%
𝑢U−shape (𝑦) 𝑢U−shape (𝑦) 𝑢U−shape (𝑦)
0.00 1.41 0.45 1.75 3.0 1.80
0.10 1.48 0.50 1.78 4.0 1.75
0.15 1.53 0.60 1.82 5.0 1.72
0.20 1.57 0.70 1.86 6.0 1.70
0.25 1.62 0.80 1.88 7.5 1.68
0.30 1.66 0.90 1.89 10 1.66
0.35 1.69 1.0 1.90 20 1.65
0.40 1.73 2.0 1.86 ∞ 1.65
C.15 If two distributions of identical form, either rectangular or U-shaped, are

convolved, the coverage factor 𝑘 for a coverage probability of 95.45 % may be
obtained from the following table:
𝑢smaller (𝑦) 𝑘 for stated ratio 𝑘 for stated ratio
𝑢larger (𝑦) 2 Rectangular Distributions 2 U-shaped Distributions
0.00 1.65 1.41
0.05 1.65 1.44
0.10 1.66 1.49
0.15 1.69 1.53
0.20 1.71 1.58
0.25 1.74 1.62
0.30 1.77 1.66
0.35 1.79 1.69
0.40 1.82 1.72
0.45 1.84 1.75
0.50 1.86 1.77
0.60 1.89 1.81
0.70 1.91 1.83
0.80 1.92 1.85
0.90 1.93 1.86
1.00 1.93 1.86
C.16 Coverage factors for various other coverage probabilities are shown graphically
on the following pages.

2.9
2.8
2.7
2.6
2.5
2.4
2.3
coverage factor kp (r )
2.2
2.1
1.9
1.8 p=0.9
p=0.95
1.7
p=0.9545
1.6
p=0.975
1.5
p=0.99
1.4
p=0.995
1.3
0.0 0.5 1.0 1.5 2.0 2.5 3.0
uncertainty ratio, r
Coverage factors for the combination of normal and rectangular PDFs at several different coverage probabilities
𝑢N (𝑦)
for different uncertainty ratio: 𝑟 =
𝑢R (𝑦)

2.9
2.8
2.7
2.6
2.5
2.4
2.3
coverage factor kp (r )
2.2
2.1
1.9
1.8 p=0.9
p=0.95
1.7
p=0.9545
1.6
p=0.975
1.5
p=0.99
1.4
p=0.995
1.3
0.0 0.5 1.0 1.5 2.0 2.5 3.0
uncertainty ratio, r
Coverage factors for the combination of normal and U-shaped PDFs at several different coverage probabilities
𝑢N (𝑦)
for different uncertainty ratio: 𝑟 =
𝑢U−shape (𝑦)

Appendix D Measurement Equations
D.1 Establishing a Measurement Equation
D.1.1 A measurement model describes the relationship between an output quantity 𝑌 and input
quantities 𝑋𝑖 .
The ‘bottom-up’ approach of the GUM [1] law of propagation of uncertainties, as covered in
M3003, requires this to be a functional relationship, 𝑌 = 𝑓(𝑋1 , … , 𝑋𝑁 ), which also relates the
estimates of the input quantities 𝑥𝑖 to the estimate of the output, 𝑦 = 𝑓(𝑥1 , … , 𝑥𝑁 ). This is referred
to in this guide as a measurement equation.
A common approach to establishing a measurement equation [3] involves first defining a ‘basic
model’ which relates the output to the ‘physical’ input quantities. For example, this might be an
𝜋(𝑃1 −𝑃2 )𝑟 4
equation such as 𝑝 = 𝜌𝑔ℎ, or 𝑄 = , or it might be an equation defining a measurement
𝜂𝐿
error ∆𝐼 = 𝐼obs − 𝐼ref
The equation is then extended by the addition of ‘metrological’ terms representing possible
quantities that are not already part of the basic equation. These usually correspond to error
quantities for which an exact value is unknown. The best estimate of their value (in additive
terms) is zero, however each has a finite uncertainty.
D.1.2 For example, suppose that the voltage output of a transducer (the measurand) is directly
measured using a calibrated meter. A basic model could be written as
𝑉 = 𝑉ave + ∆𝑉cal
where
𝑉 is the estimate of the voltage output,
𝑉ave is the average of 𝑛 observations 𝑉𝑗=1 to 𝑛 (obtained under repeatability conditions of
measurement),
∆𝑉cal is an additive correction to the observed values, established from calibration of the
meter.
The estimate 𝑉 is further influenced by other factors, each of which corresponds to a poorly
known or unknowable measurement error δ𝑉𝑖 (each with a best estimate of zero value, but finite
uncertainty).
The measurement equation is therefore extended to account for these ‘metrological terms’ e.g.,
𝑉 = 𝑉ave + ∆𝑉cal + δ𝑉res + δ𝑉drift + δ𝑉T
where say
δ𝑉res is the error due to finite resolution of the observed values 𝑉𝑗
δ𝑉drift is the error due to drift in the correction ∆𝑉cal since the meter was last calibrated
δ𝑉T is the error due to possible temperature effects
D1.3 There is no single standard or ‘correct’ way to construct a measurement model. The process is
dictated by the nature of the information available and by the knowledge of the person
performing the evaluation, nevertheless all valid approaches should lead to a similar result

Similarly, there is no universally accepted standard for the choice of symbols to represent
different quantities in a measurement equation. Choices should be made primarily to ensure
clarity of meaning.
However, a useful convention is to use a δ symbol to represent any term whose best estimated
value is zero and to use a ∆ symbol to represent a difference, a measurement error or
correction.
D.1.4 In some cases, the measurement equation can be written as a pure product of terms (i.e., the
output quantity is obtained from only the multiplication or division of the input quantities)
𝑝 𝑝 𝑝
𝑦 = 𝑐𝑥1 1 𝑥2 2 … 𝑥𝑁𝑁 (7)
where the exponents 𝑝𝑖 are known positive or negative numbers
For example,
𝑄 = 𝑄ave × 𝑓cal × 𝑓res × 𝑓drift × 𝑓T
where say
𝑓cal is a multiplicative correction established from calibration of the meter
𝑓res is a factor representing error due to finite resolution of the observed values 𝑄𝑗 .
𝑓drift is a factor representing error due to drift in the correction 𝑓cal since the meter was last
calibrated
𝑓T is a factor representing error due to possible temperature effects
If the multiplicative factor is poorly known or unknowable, the best estimate of its value is one,
analogous to the value of zero attributed to poorly known or unknowable values in the additive
case.
D1.4.1 In this special case the uncertainty evaluation can be performed in terms of relative values (e.g., in
% terms, or in parts per million) and the relative standard uncertainty will then be given by,
𝑁 2
𝑢𝑐 (𝑦) 𝑝𝑖 𝑢(𝑥𝑖 )
= √∑ ( ) (8)
|𝑦| 𝑥𝑖
𝑖=1
Some examples are given below
2 2
𝑢(𝑃) 𝑢(𝐼) 𝑢(𝑉)
𝑃 = 𝑓(𝐼, 𝑉) = 𝐼𝑉, = √( ) +( )
|𝑃| 𝐼 𝑉
2 2
1 𝑢(𝐸) 𝑢(𝑚) 2. 𝑢(𝑣)
𝐸 = 𝑓(𝑚, 𝑣) = 𝑚𝑣 2 , = √( ) +( )
2
|𝐸| 𝑚 𝑣
2 2
1 𝑢(𝑉) 𝑢(𝑃) 𝑢(𝑍)
𝑉 = 𝑓(𝑃, 𝑍) = (𝑃𝑍)2 , = √( ) +( )
|𝑉| 2𝑃 2𝑍

Use of relative uncertainties can often simplify the calculations and is particularly helpful when the
input quantities and the uncertainties are already available in relative terms.
Equation (8) should not be used when the functional relationship includes any addition or
subtraction of quantities.
D.1.5 More generally there will be both additive and multiplicative terms in the measurement equation.
For example, a measurement model might be written as:
𝑄 = (𝑄ave × 𝑓cal ) + δ𝑄res + δ𝑄drift + δ𝑄T
where say
𝑄ave is the average of 𝑛 observations 𝑄𝑗=1 to 𝑛 (obtained under repeatability conditions of
measurement)
𝑓cal is a multiplicative correction established from calibration of the meter
δ𝑄res is the error due to finite resolution of the observed values 𝑄𝑗
δ𝑄drift is the error due to drift in the correction 𝑓cal since the meter was last calibrated
δ𝑄T is the error due to possible temperature effects
D.1.6 In all cases, to evaluate the measurement uncertainty, the contribution associated with each
‘input’ quantity in the measurement equation must be considered. For uncertainty budgets that
are represented in tabular format this will usually involve creating a separate line for each term
D.1.7 In situations where a suitable measurement equation cannot readily be established a ‘top-down’
approach is often adopted, resulting in a form of statistical model for the measurement. This
approach is not covered in M3003. The reader is referred to other guidance documents such as
EURACHEM/CITAG Guide CG4 [10] and ISO 21748 [8] for further information.
D.1.8 More detailed guidance on the development of measurement models of all types can be found in
JCGM GUM-6 [3].
D.2 Measurement repeatability in measurement equations
D.2.1 Measurement repeatability can be incorporated into a model in several different ways. The
choice will largely depend on how its value is to be estimated.
D.2.2 For example, the measurement equation
𝑉 = 𝑉ave + ∆𝑉cal + δ𝑉res + δ𝑉drift + δ𝑉T
might associate the measurement repeatability with the quantity 𝑉ave .
If the budget is presented in a table, it might appear as shown in the example below:

Source of Quantity Estimate Uncertainty PDF Divisor 𝑢(𝑥𝑖 ) 𝑐i 𝑢𝑖 (𝑉)

uncertainty 𝑋𝑖 𝑥𝑖 /volt /volt
/volt
Repeatability 𝑉ave 100.001 0.0051
Calibration ∆𝑉cal 0.001 0.0035
Resolution δ𝑉res 0 0.00005
Calibration drift δ𝑉drift 0 0.0001
Thermal effects δ𝑉T 0 0.00012
D.2.3 Alternatively, an additional error term, say δ𝑉rep might be used.
𝑉 = 𝑉ave + ∆𝑉cal + δ𝑉res + δ𝑉drift + δ𝑉T + δ𝑉rep
in which case 𝑉ave no longer features in the table in the example below, because all associated
errors are taken into account by the other quantities, in effect it is treated as a constant

uncertainty 𝑋𝑖 𝑥𝑖 /volt /volt
/volt
Repeatability δ𝑉rep 0 0.0051
Calibration ∆𝑉cal 0.001 0.0035
Resolution δ𝑉res 0 0.00005
Calibration drift δ𝑉drift 0 0.0001
Thermal effects δ𝑉T 0 0.00012
D.2.4 For a model that has several measured input quantities the repeatability can be treated as a
separate input for each individual quantity. Such a treatment is most likely to be useful when the
inputs are obtained separately.
For example, the outer surface area of a smooth regular cylinder can be modelled as
𝐴 = 𝜋 × 𝑑ave × ℎave
which, with extension to include metrological terms might be written as
𝐴 = 𝜋(𝑑ave + δ𝑑rep + δ𝑑cal + ⋯ )(ℎave + δℎrep + δℎcal + ⋯ )
where terms for inputs such as drift and resolution have been omitted for the sake of clarity.
If the budget is presented in a table, it might appear as shown in the example below.

Source of Quantity Estimate Uncertainty PDF Divisor 𝑢(𝑥𝑖 ) 𝑐i 𝑢𝑖 (𝐴)

uncertainty 𝑋𝑖 𝑥𝑖 /m2
Repeatability of
δ𝑑rep 0m 0.0081 m
diameter
Calibration of
δ𝑑cal 0m 0.00061 m
diameter gauge
⋮ ⋮ ⋮ ⋮
Repeatability of
δℎrep 0m 0.0012 m
height
Calibration of
δℎcal 0m 0.00044 m
height gauge
⋮ ⋮ ⋮ ⋮
Note that 𝑑ave and ℎave do not need to appear in the table if all uncertainty associated with these
quantities is accounted for by other terms.
D.2.5 Alternatively, repeatability is often evaluated and modelled in terms of the output quantity. This is
more likely to be the case in situations where reliable individual repeatability estimates are
difficult to obtain (as is very often the case), and repeatability is instead estimated from multiple
realisations of the measurand (either during the measurement or during a previous repeatability
assessment).
For example, the outer surface area of the smooth regular cylinder can be estimated from the
average of several repeat values
𝑛 𝑛
1 1
𝐴= ∑ 𝐴𝑖 = ∑ 𝜋 × 𝑑𝑖 × ℎ𝑖
𝑛 𝑛
𝑖=1 𝑖=1
which, with extension to include metrological terms might be written as
𝑛
1
𝐴= ∑ 𝜋(𝑑𝑖 + δ𝑑cal + ⋯ )(ℎ𝑖 + δℎcal + ⋯ ) + δ𝐴
𝑛
𝑖=1
again, terms such as drift and resolution have been omitted for the sake of clarity.
If the budget is presented in a table, it might appear as shown in the example below

uncertainty 𝑋𝑖 𝑥𝑖 /m2
Calibration of
δ𝑑cal 0m 0.00061 m
diameter gauge
Calibration of
δℎcal 0m 0.00044 m
height gauge
Repeatability δ𝐴 0 m2 0.0012 m2
⋮ ⋮ ⋮ ⋮

D.3 Correlated input quantities
D.3.1 The expressions for the standard uncertainty of the output given in equations (1) and (8) only apply
when the input quantities are independent of each other, in other words, when there is no
correlation between any of the input estimates.
It may however be the case that some input quantities are affected by a common factor which
introduces correlation into the equation. For example, common temperature effects, or common
measurement errors from an instrument that is used to measure several of the inputs to a process
introduces correlation.
D.3.2 Sometimes it is possible to construct the equation in such a way as to avoid correlated inputs.
For example, weights taken from the same set are usually assumed to have correlated values.
Suppose that 3 weights 𝑚1 , 𝑚2 , and 𝑚3 from such a set are combined to form a load that is
applied to a piston. The applied pressure can be written as:
(𝑚1 + 𝑚2 + 𝑚3 )𝑔
𝑃=
𝐴
where A is the cross sectional area over which the weight acts and g is the acceleration due to
gravity.
This equation includes 3 correlated terms, 𝑚1 , 𝑚2 , and 𝑚3 , so it would be incorrect to combine

the uncertainties of all items using equation (1)
2 (𝑝) + 𝑢 2 (𝑝) + 𝑢 2 (𝑝) + 𝑢 2 (𝑝) + 𝑢 2 (𝑝) + 𝑢 2 (𝑝)

𝑢𝑐 (𝑝) = √𝑢𝑚 (incorrect)
1 𝑚1 𝑚2 𝑚3 𝑔 𝐴
A common approach to this problem would be to evaluate separately the uncertainty for the total
load, 𝑊 and incorporate this into an evaluation involving only independent terms
2 (𝑝)
𝑢𝑐 (𝑝) = √𝑢𝑊 + 𝑢𝑔2 (𝑝) + 𝑢𝐴2 (𝑝)
where 𝑊 = 𝑚1 + 𝑚2 + 𝑚3
When there is full correlation between several input quantities, 𝑥𝑗 , their combined standard
uncertainty is found by summation rather than combining them using equation (1)
∑ 𝑐𝑗 𝑢(𝑥𝑗 ) (9)
𝑗
It is common practice to assume that there is full correlation between the values of weights in a
set; therefore the uncertainty 𝑢(𝑊) is found from
𝑢(𝑊) = 𝑢(𝑚1 ) + 𝑢(𝑚2 ) + 𝑢(𝑚3 )
D.3.3 In cases where measurement errors combine so as to increase uncertainty, as in the above
example, this is referred to as positive correlation. In other cases, the effects of correlated input
quantities may serve to reduce the combined uncertainty, such as when an instrument is used as a
comparator between a standard and an unknown - this is referred to as negative correlation.

D.3.4 A more detailed discussion of correlation is beyond the scope of M3003. The GUM (Annex F.1.2)
should be consulted for a more detailed description of approaches for dealing with correlation.
NOTE: a further example of the treatment of correlated contributions can be found in paragraph K6.4.

Appendix E Some Sources of Error and Uncertainty in Electrical Calibrations
The following is a description of the more common sources of systematic error and uncertainty
(after correction) in electrical calibration work, with brief comments about their nature. Further,
more detailed, advice is given in specialised technical publications and manufacturers’ application
notes, as well as other sources.
E.1 Imported uncertainty
E.1.1 The uncertainties assigned to the values on a calibration certificate for the calibration of an
instrument, whether measuring equipment or a reference standard, are all contributors to the
uncertainty budget.
E.2 Secular stability
E.2.1 The performance of all instruments, and the values of reference standards, must be expected to
change to some extent with the passage of time. Passive devices such as standard resistors or
high-grade RF and microwave attenuators may be expected to drift slowly with time. An estimate
of such a drift has to be assessed on the basis of values obtained from previous calibrations. It
cannot be assumed that a drift will be linear. Data can be assimilated readily if displayed in a
graphical form. A curve fitting procedure that gives a progressively greater weight to each of the
more recent calibrations can be used to allow the most probable value at the time of use to be
assessed. The degree of complexity in curve fitting is a matter of judgement; in some cases,
drawing a smooth curve through the chosen data points by hand can be quite satisfactory.
Whenever a new calibration is obtained the drift characteristic will need re-assessment. The
corrections that are applied for drift are subject to uncertainty based on the scatter of data points
about the drift characteristic. The magnitude of the drift and the random instability of an
instrument, and the accuracy required will determine the calibration interval.
E.2.2 With complex electronic equipment it is not always possible to follow this procedure as changes in
performance can be expected to be more random in nature over relatively long periods. Checks
against passive standards can establish whether conformity to specification is being maintained or
whether a calibration with subsequent equipment adjustment is needed. The manufacturer’s
specification can be a good starting point for assigning the uncertainty due to instrument drift but
should be confirmed by analysis of quality control and calibration data.
E.3 Environmental conditions
E.3.1 The laboratory measurement environment can be one of the most important considerations when
performing electrical calibrations. Ambient temperature is often the most important influence and
information on the temperature coefficient of, for example, resistance standards has to be sought
or determined. Variations in relative humidity can also affect the values of unsealed components.
The influence of barometric pressure on certain electrical measurement standards can also be
significant. At RF and microwave frequencies, ambient temperature can affect the performance of,
for example, attenuators, impedance standards that depend on mechanical dimensions for their
values and other precision components. Devices that incorporate thermal sensing, such as power
sensors, can be affected by rapid temperature changes that can be introduced by handling or
exposure to sunlight or other sources of heat.
E.3.2 It is also necessary to be aware of the possible effects of electrical operating conditions, such as
power dissipation, harmonic distortion, or level of applied voltage being different when a device is
in use from when it was calibrated. Resistance standards, resistive voltage dividers and
attenuators at any frequency are examples of devices being affected by self-heating and/or applied
voltage. It should also be ensured that all equipment is operating within the manufacturer's stated
range of supply voltages.

E.3.3 The effects of harmonics and noise on ac calibration signals may have an influence on the
apparent value of these signals. Similarly, the effects of any common-mode signals present in a
measurement system may have to be accounted for.
E.4 Interpolation of calibration data
E.4.1 When an instrument with a broad range of measurement capabilities is calibrated, there are
practical and economic factors that limit the number of calibration points. Consequently, the value
of the quantity to be measured and/or its frequency may be different from any of the calibration
points. When the value of the quantity lies between two calibration values, consideration needs to
be given to systematic errors that arise from, for example, scale non-linearity.
E.4.2 If the measurement frequency falls between two calibration frequencies, it will also be necessary to
assess the additional uncertainty due to interpolation that this can introduce. One can only proceed
with confidence if:
(a) a theory of instrument operation is known from which one can predict a frequency characteristic, or
there is additional frequency calibration data from other models of the same instrument,
and wherever reasonable,
(b) the performance of the actual instrument being used has been explored with a swept frequency
measurement system to verify the absence of resonance effects or aberrations due to
manufacturing or other performance limitations.
E.5 Resolution
E.5.1 The limit to the ability of an instrument to indicate small changes in the quantity being measured,
referred to as resolution or “digital rounding error”, is treated as a systematic component of
uncertainty.
E.5.2 Many instruments with a digital display use an analogue-to digital converter (ADC) to convert the
analogue signal under investigation into a form that can be displayed in terms of numeric digits.
The last displayed digit will be a rounded representation of the underlying analogue signal. The
error introduced by this process will be from -0.5 digit (else the last digit would be one lower) to
+0.5 digit (else the last digit would be one higher). A quantisation error of ±0.5 digit is therefore
present. As there is no way of knowing where within this range the underlying value is, the
resulting error is assumed to be zero with limits of ±0.5 digit.
E.5.3 This “digital rounding error” of ±0.5 digit may not apply in all instances and an understanding of
instrument operation is needed if the assigned uncertainty is to be realistic. For example, a
direct-gating frequency counter has a digital rounding error of ±1 digit, due to the random
relationship between the signal being measured and the internal clock. Some instruments may
also display hysteresis that, although not necessarily a property of the display itself, may result
in further uncertainties amounting to several digits.
E.5.4 In an analogue instrument the effect of resolution is determined by the practical ability to read the
position of a pointer on a scale. In either case, the last digit actually recorded will always be subject
to an uncertainty of at least ±0.5 digit. The presence of electrical noise causing fluctuations in
instrument readings will commonly determine the usable resolution, however it is possible to make
a good estimate of the mean position of a fluctuating pointer by eye.
E.6 Apparatus layout

E.6.1 The physical layout of one item of equipment with respect to another and the relationship of these
items to the earth plane can be important in some measurements. Thus a different arrangement
between calibration and subsequent use of an instrument may be the source of systematic errors.
The main effects are leakage currents to earth, interference loop currents, and electromagnetic
leakage fields. In inductance measurements it is necessary to define connecting lead configuration
and be conscious of the possible effects of an earth plane or adjacent ferromagnetic material. The
effect of mutual heating between apparatus may also need to be considered.
E.7 Thermoelectric voltages
E.7.1 If an electrical conductor passes through a temperature gradient, then a potential difference will be
generated across that gradient. This is known as the Seebeck effect and these unwanted, parasitic
voltages can cause errors in some measurement systems – in particular, where small dc voltages
are being measured.
E.7.2 They can be minimised by design of connections that are thermally symmetrical, so that the
Seebeck voltage in one lead is cancelled by an identical and opposite voltage in the other. In
some situations, e.g., ac/dc transfer measurements, the polarity of the dc supply is reversed, and
an arithmetic mean is taken of two sets of dc measurements.
E.7.3 Generally, an allowance has to be made as a Type B component of uncertainty for the presence of
thermal EMFs.
E.8 Loading and cable impedance
E.8.1 The finite input impedance of voltmeters, oscilloscopes and other voltage sensing instruments may
so load the circuit to which they are connected as to cause significant systematic errors.
Corrections may be possible if impedances are known. In particular, it should be noted that some
multi-function calibrators can exhibit a slightly inductive output impedance. This means that when a
capacitive load is applied, the resulting resonance may cause the output voltage to increase with
respect to its open-circuit value.
E.8.2 The impedance and finite electrical length of connecting leads or cables may also result in
systematic errors in voltage measurements at any frequency. The use of four-terminal connections
minimises such errors in some dc and ac measurements.
E.8.3 For capacitance measurements, the inductive properties of the connecting leads may be
important, particularly at higher values of capacitance and/or frequency. Similarly, for inductance
measurements the capacitance between connecting leads may be important.
E9 RF mismatch errors and uncertainty
E.9.1 At RF and microwave frequencies the mismatch of components to the characteristic impedance of
the measurement system transmission line can be one of the most important sources of error and
of the systematic component of uncertainty in power and attenuation measurements. This is
because the phases of voltage reflection coefficients are not usually known and hence corrections
cannot be applied.
E.9.2 In a power measurement system, the power, 𝑃0 , that would be absorbed in a load equal to the
characteristic impedance of the transmission line has been shown (by Harris and Warner [17]) to
be related to the actual power, 𝑃𝐿 , absorbed in a wattmeter terminating the line by the equation
𝑃𝐿
𝑃0 = (1 − 2|Γ𝐺 ||Γ𝐿 | cos 𝜙 + |Γ𝐺 |2 |Γ𝐿 |2 ) E1
1 − |Γ𝐿 |2

where 𝜙 is the relative phase of the generator and load voltage reflection coefficients Γ𝐺 and Γ𝐿 .
When Γ𝐺 and Γ𝐿 are small, this becomes
𝑃𝐿
𝑃0 = (1 − 2|Γ𝐺 ||Γ𝐿 | cos 𝜙) E2
1 − |Γ𝐿 |2
E.9.3 This expression for absorbed power can have limits:
𝑃𝐿
𝑃0 (limits) = (1 ± 2|Γ𝐺 ||Γ𝐿 |) E3
1 − |Γ𝐿 |2
E.9.4 The calculable mismatch error is (1 − |Γ𝐿 |2 ) and is accounted for in the calibration factor, while
the limits of mismatch uncertainty are (±2|Γ𝐺 ||Γ𝐿 |). Because a cosine function characterises the
probability distribution for the uncertainty, Harris and Warner show that the distribution is U-
shaped with a standard deviation given by
2|Γ𝐺 ||Γ𝐿 |
𝑢(mismatch) = = 1.414 |Γ𝐺 ||Γ𝐿 | E4
√2
E.9.5 When a measurement is made of the attenuation of a two-port component inserted between a
generator and load that are not perfectly matched to the transmission line, Harris and Warner have
shown that the standard deviation of mismatch, 𝑀, expressed in dB is approximated by
8.686
𝑀= [𝑀𝐺 + 𝑀𝐿 + 𝑀𝐺𝐿 ]0.5 E5
√2
where
𝑀𝐺 = |Γ𝐺 |2 (|𝑠11𝑎 |2 + |𝑠11𝑏 |2 )
𝑀𝐿 = |Γ𝐿 |2 (|𝑠22𝑎 |2 + |𝑠22𝑏 |2 )
𝑀𝐺𝐿 = |Γ𝐺 |2 |Γ𝐿 |2 (|𝑠21𝑎 |4 + |𝑠21𝑏 |4 )
and Γ𝐺 and Γ𝐿 are the source and load voltage reflection coefficients respectively and 𝑠11 , 𝑠22 , 𝑠21
are the scattering coefficients of the two-port component with the suffix 𝑎 referring to the starting
value of the attenuator and 𝑏 referring to the finishing value of the attenuator. Harris and Warner
concluded that the distribution for 𝑀 would approximate to that of a normal distribution due to the
combination of its component distributions.
E.9.6 The values of Γ𝐺 and Γ𝐿 used in equations E(4) and E(5) and the scattering coefficients used in
equation E(5) will themselves be subject to uncertainty because they are derived from
measurements. This uncertainty has to be considered when calculating the mismatch uncertainty
and it is recommended that this is done by adding it in quadrature with the measured or derived
value of the reflection coefficient; for example, if the measured value of Γ𝐿 is 0.03 ± 0.02 then the
value of Γ𝐿 that should be used to calculate the mismatch uncertainty is √0.032 + 0.022 i.e. 0.036.

E.10 Directivity
E.10.1 When making voltage reflection coefficient (VRC) measurements at rf and microwave frequencies,
the finite directivity of the bridge or reflectometer gives rise to an uncertainty in the measured value
of the VRC, if only the magnitude and not the phase of the directivity component is known. The
uncertainty will be equal to the directivity, expressed in linear terms, e.g., a directivity of 30 dB is
equivalent to an uncertainty of 0.0316 VRC.
E.10.2 As with E9.6 above it is recommended that the uncertainty in the measurement of directivity is
taken into account by adding the measured value in quadrature with the uncertainty, in linear
quantities; for example, if the measured directivity of a bridge is 36 dB (0.016) and has an
uncertainty of +8 dB -4 dB (±0.01) then the directivity to be used is
√0.0162 + 0.012 = 0.019, (i.e. 34.4 dB).
E.11 Test port match
E.11.1 The test port match of a bridge or reflectometer used for reflection coefficient measurements will
give rise to an error in the measured VRC due to re-reflection. The uncertainty, 𝑢(𝑇𝑃), is
calculated from
𝑢(𝑇𝑃) = 𝑇𝑃. |Γ𝑋 |2 , where 𝑇𝑃 is the test port match, expressed as a VRC, and Γ𝑋 is the measured
reflection coefficient. When a directional coupler is used to monitor incident power in the calibration
of a power meter it is the effective source match of the coupler that defines the value of Γ𝐺 referred
to in E9. As with E9.6 and E10, the measured value of test port match will have an uncertainty that
should be taken into account by using quadrature summation.
E.12 RF connector repeatability
The lack of repeatability of coaxial pair insertion loss and, to a lesser extent, voltage reflection
coefficient is a problem when calibrating devices in a coaxial line measurement system and
subsequently using them in some other system. Repeatedly connecting and disconnecting the
device can evaluate the repeatability of particular connector pairs in use.

Appendix F Some Sources of Error and Uncertainty in Mass Calibrations
This Appendix describes the more common sources of errors and uncertainties in mass
calibration with brief comments about their nature. They may not all be significant at all levels of
measurement, but their effect should at least be considered when estimating the overall
uncertainty of a measurement. Further information about mass calibration and the calibration of
weighing machines can be found in UKAS LAB 14 [13] and EURAMET CG-18 [12].
F.1 Reference weight calibration
F.1.1 The uncertainties assigned to the values on a calibration certificate for the calibration of the
reference weights are all contributors to the uncertainty budget.
F.2 Secular stability of reference weights
F.2.1 It is necessary to take into account the likely change in mass of the reference weights since their
last calibration. This change can be estimated from the results of successive calibrations of the
reference weights. If such a history is not available, then it is usual to assume that they may
change in mass by an amount equal to their uncertainty of calibration between calibrations. The
stability of weights can be affected by the material and quality of manufacture (e.g., ill-fitting screw
knobs), surface finish, unstable adjustment material, physical wear and damage and atmospheric
contamination. The figure adopted for stability will need to be reconsidered if the usage or
environment of the weights changes. The calibration interval for reference weights will depend on
the stability of the weights.
F.3 Weighing machine/weighing process
F.3.1 The performance of the weighing machine used for the calibration should be assessed to estimate
the contribution it makes to the overall uncertainty of the weighing process. The performance
assessment should cover those attributes of the weighing machine that are significant to the
weighing process. For example, the length of arm error (assuming it is constant) of an equal arm
balance need not be assessed if the weighing process only uses substitution techniques (Borda's
method). The assessment may include some or all of the following:
(a) repeatability of measurement;
(b) linearity within the range used;
(c) digit size/weight value per division, i.e., resolution;
(d) eccentricity (off centre load), especially if groups of weights are placed on the weighing pan
simultaneously;
(e) magnetic effects (e.g., magnetic weights, or the effect of force balance motors on cast iron
weights);
(f) temperature effects, e.g., differences between the temperature of the weights and the weighing
machine;
(f) length of arm error.

F.4 Air buoyancy effects
F.4.1 The accuracy with which air buoyancy corrections can be made depends on how well the density
of the weights is known, and how well the air density can be determined. Some laboratories can
determine the density of weights, but for most mass work assumed figures are used. The air
density is usually calculated from an equation (e.g., see [13]) after measuring the air temperature,
pressure and humidity. For the highest levels of accuracy, it may also be necessary to measure the
carbon dioxide content of the air. The figures that follow are based upon an air density range of
1.079 kg m-3 to 1.291 kg m-3 which can be produced by ranges of relative humidity from 30 % to
70 %, air temperature from 10 C to 30 C and barometric pressure from 950 mbar to 1050 mbar.
F.4.2 For mass comparisons a figure of ±1 part in 106 of the applied mass is typical for common weight
materials such as stainless steel, plated brass, German silver and gunmetal. For cast iron the
figure may be up to ±3 parts in 106 and for aluminium up to ±30 parts in 106. The uncertainty can
be reduced if the mass comparisons are made within suitably restricted ranges of air temperature,
pressure and humidity. If corrections are made for the buoyancy effects the uncertainty can be
virtually eliminated, leaving just the uncertainty of the correction.
F.4.3 Certain weighing machines display mass units directly from the force they experience when
weights are applied. It is common practice to reduce the effects of buoyancy on such devices by
the use of an auxiliary weight, known as a spanning weight, which is used to normalise the
readings to the prevailing conditions, as well as compensating for changes in the machine itself.
This spanning weight can be external or internal to the machine. If such machines are not spanned
at the time of use the calibration may be subject to an increased uncertainty due to the buoyancy
effects on the loading weights. For weighing machines that make use of stainless steel, plated
brass, German silver or gunmetal weights this effect may be up to ±16 parts in 106. For cast iron
weights the figure may be up to ±18 parts in 106 and for aluminium weights up to ±45 parts in 106.
F.4.4 For the ambient conditions stated above the uncertainty limits due to buoyancy effects may be
±110 parts in 106 and ±140 parts in 106 respectively for comparing water and organic solvents with
stainless steel mass standards, and ±125 parts in 106 and ±155 parts in 106 respectively for direct
weighing.
F.4.5 Apart from air buoyancy effects, the environment in which the calibration takes place can introduce
other uncertainties. Temperature gradients can give rise to convection currents in the balance
case, which will affect the reading, as will draughts from air conditioning units. Rapid changes of
temperature in the laboratory can affect the weighing process. Changes in the level of humidity in
the laboratory can make short-term changes to the mass of weights, while low levels of humidity
can introduce static electricity effects on some comparators. Dust contamination also introduces
errors in calibrations. The movement of weights during the calibration causes disturbances to the
local environment.

Appendix G Some Sources of Error and Uncertainty in Temperature Calibrations
The more common sources of systematic error and uncertainty in the measurement of temperature
are described in this section. Each source may have several uncertainty components.
G.1 Reference thermometer calibration
G.1.1 The uncertainties assigned to the values on a calibration certificate for the calibration of the
reference thermometer are all contributors to the uncertainty budget.
G.2 Measuring instruments
G.2.1 The uncertainty assigned to the calibration of any electrical or other instruments used in the
measurements, e.g., standard resistors, measuring bridges and digital multimeters.
G.3 Further influences
G.3.1 Additional uncertainties in the measurement of the temperature using the reference thermometers:
(a) Drift since the last calibration of the reference thermometers and any associated measuring
instruments;
(b) Resolution of reading; this may be very significant in the case of a liquid-in-glass thermometer or
digital thermometers;
(c) Instability and temperature gradients in the thermal environment, e.g., the calibration bath or
furnace, including any contribution due to difference in immersion of the reference standard from
that stated on its certificate of calibration;
(d) When platinum resistance thermometers are used as reference standards any contribution to the
uncertainty due to self-heating effects should be considered. This will mainly apply if the measuring
current is different from that used in the original calibration and/or the conditions of measurement
e.g., `in air' or in stirred liquid.
G.4 Contributions associated with the thermometer to be calibrated
G.4.1 These may include factors associated with electrical indicators as well as some of the further
influences already mentioned. When partial immersion liquid-in-glass thermometers are to be
calibrated an additional uncertainty contribution to account for effects arising from differences in
depth of immersion should be included even when the emergent column temperature is measured.
G.4.2 When thermocouples are being calibrated any uncertainty introduced by compensating leads and
reference junctions should be taken into account. Similarly, any thermal EMFs introduced by
switches or scanner units should be investigated. Unknown errors arising from inhomogeneity of
the thermocouple being calibrated can give rise to significant uncertainties. Ideally this should be
evaluated at the time of calibration, possibly by varying the immersion depth of the thermocouple
in an isothermal enclosure. For many calibrations, however, this will not be practical. In such
cases, a figure of 20 % of the maximum permissible error for the particular thermocouple type is
considered reasonable.

G.5 Mathematical interpretation
G.5.1 Uncertainty arising from mathematical interpretation, e.g. in applying scale corrections or
deviations from a reference table, or in curve-fitting to allow for scale non-linearity, should be
assessed.

Appendix H Some Sources of Error and Uncertainty in Dimensional Calibrations
The more common sources of systematic error and uncertainty in dimensional measurements are
described in this section.
H.1 Reference standards and Instrumentation
H.1.1 The uncertainties assigned to the reference standards and those for the measuring instruments
used to make the measurements.
H.2 Secular stability of reference standards and instrumentation
H.2.1 The changes that occur over time must be taken into account, usually by reference to the
calibration history of the equipment. This is particularly important when the equipment may be
exposed to physical wear as part of normal operation.
H.3 Temperature effects
H.3.1 The uncertainties associated with differences in temperature between the gauge being calibrated
and the reference standards and measuring instruments used should be accounted for. These will
be most significant over the longer lengths and in cases involving dissimilar materials. Whilst it may
be possible to make corrections for temperature effects there will be residual uncertainties resulting
from uncertainty in the values used for the coefficients of expansion and the calibration of the
measuring thermometer.
H.4 Elastic compression
H.4.1 These are uncertainties associated with differences in elastic compression between the materials
from which the gauge being calibrated and the reference standards were manufactured. They are
likely to be most significant in the more precise calibrations and in cases involving dissimilar
materials. They will relate to the measuring force used and the nature of stylus contact with the
gauge and reference standard. Whilst mathematical corrections can be made there will be residual
uncertainties resulting from the uncertainty of the measuring force and in the properties of the
materials involved.
H.5 Cosine errors
H.5.1 Any misalignment of the gauge being calibrated, or reference standards used, with respect to the
axis of measurement, will introduce errors into the measurements. Such errors are often referred to
as cosine errors and can be minimised by adjusting the attitude of the gauge with respect to the
axis of measurement to find the relevant turning points that give the appropriate maximum or
minimum result. Small residual errors can still result where, for instance, incorrect assumptions are
made concerning any features used for alignment of the datums.
H.6 Geometric errors
H.6.1 Errors in the geometry of the gauge being calibrated, any reference standards used, or critical
features of the measuring instruments used to make the measurements can introduce additional
uncertainties. Typically, these will include small errors in the flatness or sphericity of stylus tips, the
straightness, flatness, parallelism or squareness of surfaces used as datum features, and the
roundness or taper in cylindrical gauges and reference standards. Such errors are often most
significant in cases where perfect geometry has been wrongly assumed and where the
measurement methods chosen do not capture, suppress or otherwise accommodate the geometric
errors that prevail in a particular case.

Appendix J Some Sources of Error and Uncertainty in Pressure Calibrations using

Dead Weight Testers
The more common sources of systematic error and uncertainty in the generation of known
pressures, using dead weight testers (DWT), are described in this section.
J.1 Reference DWT
J.1.1 The uncertainties assigned to the values on a calibration certificate for the reference dead weight
tester are all contributors to the uncertainty budget. These include the following:
(a) Area uncertainty including any uncertainty in the distortion. This uncertainty will often vary with
pressure;
(b) Piston and weight carrier mass.
J.2 Secular stability of the reference dead weight tester
J.2.1 It is necessary to account for likely changes in the area and mass of the reference DWT since
the last calibration. This change can be estimated from successive calibrations of the reference
DWT. The secular stability uncertainty for the area will depend on the calibration interval and can
be larger than the calibration uncertainty. It may also vary with pressure and should be evaluated
over the range of use of the DWT. The variation between calibrations in the area of a DWT will
depend on its usage, design, and material composition and is therefore a best estimate from
actual data. Where this is not available it is recommended that a pessimistic estimate is made,
and a short calibration interval set.
J.2.2 The drift of the piston mass will be larger in oil DWTs as this will reflect the difficulties in repeat
weighting of pistons that have been immersed in oil. These difficulties arise from incomplete
cleaning processes and possible instability due to the evaporation of solvents.
J.3 Reference DWT mass set uncertainty
J.3.1 The uncertainties assigned to the values on a calibration certificate for the weights in the
reference dead weight tester mass set are all contributors to the uncertainty budget. The
uncertainty of the mass stack used to generate pressure should be evaluated over the range of
the DWT. The relative uncertainty is often higher at lower pressures.
J.4 Secular stability of the reference DWT mass set
J.4.1 It is necessary to account for likely changes the mass set of the reference DWT since the last
calibration. Paragraph F.2.1 addresses the subject of secular stability of reference weights.
J.5 Uncertainty of local gravity determination
J.5.1 The pressure generated by a DWT is directly affected by the local acceleration due to gravity, 𝑔.
With care, this can be measured with an uncertainty of less than 1 ppm. It is possible for an
estimate of the 𝑔 value to be obtained from a reputable geological survey organisation based on
a grid reference; this would attract an uncertainty of around 3 ppm. It can also be calculated from
knowledge of latitude and altitude; however, the uncertainty will be much larger - around 50 ppm
in the UK. Some knowledge of the Bouguer anomalies is required to achieve these levels of
uncertainty from such calculations.

J.6 Air buoyancy effect
J.6.1 Air buoyancy affects the mass set of a DWT in the same way as described in paragraph F.4
J.7 Temperature effect on DWT area
J.7.1 The area on a DWT changes with temperature; its temperature coefficient of expansion being
related to the particular materials that the piston and cylinder are made from. Consideration has
to be given to any variation in temperature from the reference temperature when the DWT was
calibrated, variation in temperature during a calibration and uncertainty in the determination of
the piston temperature.
J.8 Uncertainty due to head correction
J.8.1 Any difference between the height of the reference DWT datum level and that of the item being
calibrated will affect the pressure generated at that item. For pneumatic calibrations this effect is
proportional to pressure and normally equates to about 116 ppm/m. For hydraulic calibrations
the effect is a fixed pressure effect that will depend on the density of the fluid used, local
acceleration due to gravity and the height difference (fluid head pressure = 𝜌𝑔ℎ). For most DWT
oils the effect is between 8 Pa/mm and 9 Pa/mm.
J.8.2 The float height position of the piston will also contribute to the head correction uncertainty. This
effect will be related to the fall rate of the piston and the particular measurement procedure in
use.
J.9 Effects of fluid properties
J.9.1 For hydraulic calibrations the effect of the fluid properties on fluid head corrections, buoyancy
volume corrections and surface tension corrections will also need to be considered. These
figures are usually reported on calibration certificates for DWTs. However, care must be taken to
convert any quoted correction to the actual oil used if different from that used during the
calibration of the reference DWT. In most circumstances the uncertainty of these influence
quantities can be treated as negligible.
J.10 Non-verticality of the DWT piston
An uncertainty arises due to the fact that the piston may not be perfectly vertical. If it were, then all
of the force would act on the area. Any departure from vertical will reduce the force and therefore
the generated pressure. The effect in terms of generated pressure is proportional to the cosine of
the angle from true vertical.
J.11 Uncertainties arising from the calibration process
J.11.1 Any uncertainty arising from the calibration process will need to be evaluated. These could
include the resolution and repeatability of the unit being calibrated and the effects of the
environment on it. Uncertainties due to calculation or data fitting of the calibration results may
also have to be considered.

Appendix K Examples of Application
Introduction:
(a) This Appendix presents a number of example uncertainty budgets in various fields of
measurement. The examples are not intended as preferred or mandatory requirements. They
are presented to illustrate the principles involved in uncertainty evaluation and to show how the
common sources of uncertainty in the various fields can be analysed in practice. They are,
however, believed to be realistic for the particular measurements described.
(b) An uncertainty budget is a statement of measurement uncertainty evaluation, of the components

of that uncertainty, and of their calculation and combination. It should include the measurement
model, estimates and uncertainties associated with the quantities in the measurement model,
covariances, type of assumed probability density functions, degrees of freedom, type of
evaluation, coverage interval, coverage probability and coverage factors. An uncertainty budget
is not simply a summary table; it has to include all these factors. The example uncertainty
budgets presented in this Appendix comply with this definition.
(c) These examples may also be used for the purpose of software validation. If an uncertainty
budget has been prepared using a spreadsheet, the configuration of the spreadsheet can be
verified by entering the same values and comparing the output of the spreadsheet with the
results shown in the examples.
K.1 Measurement of a 10 kΩ resistor by voltage intercomparison
K.1.1 A high-resolution digital voltmeter is used to measure the voltages developed across a standard
resistor and an unknown resistor of the same nominal value as the standard when the series-
connected resistors are supplied from a constant current dc source. Both resistors are immersed in
a temperature-controlled oil bath maintained at 20.0 °C. The value of the unknown resistor, 𝑅𝑋 , is
given by
𝑉𝑋
𝑅𝑋 = (𝑅𝑆 + δ𝑅𝐷 + 𝑅𝑇𝐶 ∆𝑡)
𝑉𝑆
where
𝑅𝑆 = calibration value for the standard resistor,

δ𝑅𝐷 = drift in RS since the previous calibration,
𝑅𝑇𝐶 = temperature coefficient of resistance for 𝑅𝑆 ,
∆𝑡 = maximum variation in oil bath from nominal temperature,
𝑉𝑋 = voltage across 𝑅𝑋 ,
𝑉𝑆 = voltage across 𝑅𝑆 .
K.1.2 The calibration certificate for the standard resistor reported an uncertainty of ±0.5 ppm at a
coverage probability of approximately 95% (𝑘 = 2).
K.1.3 No correction was made for drift in the value of 𝑅𝑆 i.e., the drift is assumed to be δ𝑅𝐷 = 0.
Records indicate that the relative drift in 𝑅𝑆 is unlikely to exceed ±0.5 ppm.
K.1.4 The temperature coefficient of resistance for the standard resistor was obtained from a graph of
temperature versus resistance. Such curves are normally parabolic in nature, however using a
linear approximation over the small range of temperature variation encountered in the bath, a value

of ±2.5 ppm per ºC was assigned. This value was included in the uncertainty budget as a
sensitivity coefficient.
K.1.5 Records of evaluation of the oil bath characteristics showed that the maximum temperature
deviation from the set point did not exceed ±0.1 ºC at any point within the bath.
K.1.6 The same voltmeter is used to measure 𝑉𝑋 and 𝑉𝑆 and although the uncertainty contributions will
be correlated the effect is to reduce the uncertainty and it is only necessary to consider the relative
difference in the voltmeter readings due to linearity and resolution, which was estimated to have
limits of ±0.2 ppm for each reading. Each of these is assigned a rectangular distribution.
K.1.7 Type A evaluation: The repeatability for 𝑅𝑋 can be estimated from the repeatability of the measured
voltage ratio, 𝑉𝑋 ⁄𝑉𝑆 . Five measurements were made to record the departure from unity in the ratio
The measured departures were:
+10.4 ppm, +10.7 ppm, +10.6 ppm, +10.3 ppm, +10.5 ppm
From equation (3), the mean departure from unity = +10.50 ppm
The repeatability standard deviation 𝑠(𝑅𝑋 ) is estimated from the five measured departure values.
Applying equation (5), gives 𝑠(𝑅𝑋 ) = 0.158 ppm.
The reported measurement result is to be calculated using the mean of the 𝑛 = 5

measurements. So, from equation (4), the repeatability uncertainty is
𝑠 ( 𝑅𝑋 ) 0.158
𝑢rep (𝑅𝑋 ) = = = 0.0707 ppm
√𝑛 √5
𝑉𝑋
K.1.8 Summary table for 𝑅𝑋 = (𝑅𝑆 + δ𝑅𝐷 + 𝑅𝑇𝐶 ∆𝑡)
𝑉𝑆
Probability 𝑢𝑖 (𝑅𝑋 ) 𝜐𝑖 or
Symbol Source of uncertainty Uncertainty Divisor 𝑐𝑖
distribution ppm 𝜐eff
𝑅𝑆 Calibration of standard resistor 0.5 ppm Normal 2 1 0.25 ∞

δ𝑅𝐷 Uncorrected drift since last calibration 0.5 ppm Rectangular √3 1 0.289 ∞
∆𝑡 Temperature effects 0.1 ºC Rectangular √3 2.5 ppm/ºC 0.144 ∞
𝑉𝑆 Voltmeter across RS 0.2 ppm Rectangular √3 1 0.115 ∞
𝑉𝑋 Voltmeter across RX 0.2 ppm Rectangular √3 1 0.115 ∞
𝑅𝑋 Repeatability of indication 0.071 ppm Normal 1 1 0.071 4
𝑢c (𝑅𝑋 ) Combined standard uncertainty Normal 0.445 >500
Normal
𝑈95% (𝑅𝑋 ) Expanded uncertainty 0.891 >500
( 𝑘 = 2)
K.1.9 Reported result
The measured value of the 10 kΩ resistor at 20 ºC ± 0.1 ºC was 10 000.1050 Ω ± 0.0089 Ω
The reported expanded uncertainty is based on a standard uncertainty multiplied by a coverage factor 𝑘 = 2,
providing a coverage probability of approximately 95 %. The uncertainty evaluation has been carried out in
accordance with UKAS requirements.
NOTE: The temperature coefficient of the resistor being calibrated is not included here, as it is an “unknown” quantity. The
relevant temperature conditions are included in the reporting of the result. Best practice would be to estimate a value
together with a suitable uncertainty, and to include these details with the reported results.

K.2 Calibration of a coaxial power sensor at a frequency of 18 GHz
K.2.1 The measurement involves the calibration of an unknown power sensor against a standard
power sensor by substitution on a stable, monitored source of defined source impedance. The
measurement is made in terms of Calibration Factor, defined as,
Incident power at reference frequency

𝐾𝑋 =
Incident power at calibration frequency
for the same power sensor response. It is determined from the following:
Calibration Factor,
𝐾𝑋 = (𝐾𝑆 + δ𝐷𝑆 ) × 𝑓DC × 𝑓M × 𝑓REF
where
𝐾𝑆 = Calibration Factor of the standard sensor,

δ𝐷𝑆 = error due to drift in standard sensor since the previous calibration,
𝑓DC = DC voltage output non linearity factor,
𝑓M = ratio of Mismatch Losses,
𝑓REF = ratio of reference power source (short-term stability of 50 MHz reference).
K.2.2 Four separate measurements were made which involved disconnection and reconnection of both
the unknown sensor and the standard sensor on a power transfer system. All measurements were
made in terms of voltage ratios that are proportional to calibration factor.
K.2.3 None of the uncertainty contributions are considered to be correlated to any significant extent.
K.2.4 There will be mismatch uncertainties associated with the source/standard sensor combination
and with the source/unknown sensor combination. These will be 200. Γ𝐺 . Γ𝑆 % and 200. Γ𝐺 . Γ𝑋 %
respectively, where
Γ𝐺 = 0.02 at 50 MHz and 0.07 at 18 GHz,

Γ𝑆 = 0.02 at 50 MHz and 0.10 at 18 GHz,
Γ𝑋 = 0.02 at 50 MHz and 0.12 at 18 GHz.
These values include the uncertainty in the measurement of Γ as described in paragraph E.9.6.
K.2.5 The standard power sensor was calibrated by an accredited laboratory 6 months before use; the
expanded uncertainty of 1.1 % was quoted for a coverage factor 𝑘 = 2.
K.2.6 The long-term stability of the standard sensor was estimated from the results of 5 annual
calibrations. No predictable trend could be detected so drift corrections could not be made. The
error due to secular stability was therefore assumed to be zero with limits, in this case, not greater
than ±0.4 % per year. A value of ±0.2 % was used as the previous calibration was within 6 months.
K.2.7 The instrumentation linearity uncertainty was estimated from measurements against a reference
attenuation standard. The expanded uncertainty for 𝑘 = 2 of 0.1 % applies to ratios up to 2:1.
K.2.8 Type A evaluation: The four measurements resulted in the following values of Calibration Factor:
93.45 %, 92.20 %, 93.95 %, 93.02 %

From equation (3), the mean value ̅̅

𝐾̅̅
𝑋 = 93.16 %
The repeatability standard deviation 𝑠(𝐾𝑋 ) is estimated from the four measured values. Applying
equation (5), gives 𝑠(𝐾𝑋 ) = 0.7415 %.
The reported measurement result is to be calculated using the mean of 𝑛 = 4 measurements.

So, from equation (4), the repeatability uncertainty is
𝑠(𝐾𝑋 ) 0.7415
𝑢rep (𝐾𝑋 ) = = = 0.3707 %
√𝑛 √4
K.2.9 Summary table for 𝐾𝑋 = (𝐾𝑆 + δ𝐷𝑆 ) × 𝑓DC × 𝑓M × 𝑓REF
Uncertainty Probability 𝑢𝑖 (𝐾𝑋 ) 𝜐𝑖 or

Symbol Source of uncertainty Divisor 𝑐𝑖
% distribution % 𝜐eff
𝐾𝑆 Calibration factor of standard 1.1 Normal 2.0 1.0 0.55 ∞

δ𝐷𝑆 Drift since last calibration 0.2 Rectangular √3 1.0 0.116 ∞
𝑓DC Instrumentation linearity 0.1 Normal 2.0 1.0 0.05 ∞
𝑓REF Stability of 50 MHz reference 0.2 Rectangular √3 1.0 0.116 ∞
Mismatch:
𝑀1 Standard sensor at 50 MHz 0.08 U-shaped √2 1.0 0.06 ∞
𝑀2 Unknown sensor at 50 MHz 0.08 U-shaped √2 1.0 0.06 ∞
𝑀3 Standard sensor at 18 GHz 1.40 U-shaped √2 1.0 0.99 ∞
Unknown sensor at 18 GHz 1.68 U-shaped √2 1.0 1.19 ∞
𝑀4
𝐾𝑋 Repeatability of indication 0.37 Normal 1.0 1.0 0.37 3
𝑢c (𝐾𝑋 ) Combined standard uncertainty Normal 1.69 >500

Normal
𝑈95% (𝐾𝑋 ) Expanded uncertainty 3.39 >500
( 𝑘 = 2)
K2.10 Reported result
The measured calibration factor at 18 GHz was 93.2 % ± 3.4 %.
The reported expanded uncertainty is based on a standard uncertainty multiplied by a coverage factor
𝑘 = 2, providing a coverage probability of approximately 95%. The uncertainty evaluation has been carried out
in accordance with UKAS requirements.
NOTES:
1 For the measurement of calibration factor, the uncertainty in the absolute value of the 50 MHz reference source need
not be included if the standard and unknown sensors are calibrated using the same source, within the timescale
allowed for its short-term stability.
2 This example illustrates the significance of mismatch uncertainty in measurements at relatively high frequencies.
3 In a subsequent use of a sensor further uncertainty contributions may arise due to the use of different connector pairs.

K.3 Measurement of a 30 dB coaxial attenuator
K.3.1 The measurement involves the calibration of a coaxial step attenuator at a frequency of 10 GHz
using a dual channel 30 MHz IF substitution measurement system. The measurement is made in
terms of the attenuation in dB between a matched source and load from:
𝐴𝑋 = 𝐴𝑏 − 𝐴𝑎 + δ𝐴IF + δ𝐷IF + δ𝐿M + δ𝑅D + δ𝑀 + δ𝐴L
where
𝐴𝑋 = value for the attenuator,

𝐴𝑏 = indicated attenuation with unknown attenuator set to zero,
𝐴𝑎 = indicated attenuation with unknown attenuator set to 30 dB,
δ𝐴𝐼𝐹 = error due to calibration of reference IF attenuator,
δ𝐷𝐼𝐹 = drift in reference IF attenuator since last calibration,
δ𝐿𝑀 = departure from linearity of mixer,
δ𝑅𝐷 = error due to resolution of detection system when measuring 𝐴𝑏 or 𝐴𝑎 ,
δ𝑀 = mismatch error,
δ𝐴𝐿 = effect of signal leakage.
K.3.2 The result is corrected for the calibrated value of the IF attenuator using the results from a
calibration certificate, which gave an uncertainty of ±0.005 dB at a coverage probability of 95 %
(𝑘 = 2).
K.3.3 No correction is made for the drift of the IF attenuator. The limits of ±0.002 dB were estimated from
the results of previous calibrations.
K.3.4 No correction is made for mixer non-linearity. The uncertainty was estimated from a series of
linearity measurements over the dynamic range of the system to be ±0.002 dB / 10 dB. An
uncertainty of ±0.006 dB was therefore assigned at 30 dB. The probability distribution is assumed
to be rectangular.
K.3.5 The resolution of the detection system was estimated to cause possible rounding errors of one-half
of one least significant recorded digit i.e., ±0.005 dB. This occurs twice - once for the 0 dB
reference setting and again for the 30 dB measurement. Two identical rectangular distributions with
semi-range limits of a combine to give a triangular distribution with semi-range limits of 2𝑎. The
uncertainty due to resolution is therefore 0.01 dB with a triangular distribution.
K.3.6 No correction is made for mismatch error. The mismatch uncertainty is calculated from the
scattering coefficients using the equation given at E.9.5. The values used were as follows:
Γ𝐿 = 0.03 Γ𝐺 = 0.03
𝑠11𝑎 = 0.05 𝑠11𝑏 = 0.05 𝑠22𝑎 = 0.05 𝑠22𝑏 = 0.01 𝑠21𝑎 = 1 𝑠21𝑏 = 0.31
K.3.7 Special experiments were performed to determine whether signal leakage had any significant
effect on the measurement system. No effect greater than ±0.001 dB could be observed for
attenuation values up to 70 dB. The probability distribution is assumed to be rectangular.
K.3.8 Type A evaluation: Four measurements were made which involved setting the reference level with
the step attenuator set to zero and then measuring the attenuation for the 30 dB setting. The
results were as follows:
30.04 dB, 30.07 dB, 30.03 dB, 30.06 dB

From equation (3), the mean value ̅̅̅̅

𝐴𝑋 = 30.050 dB
The repeatability standard deviation 𝑠(𝐴𝑋 ) is estimated from the four measured values. Applying
equation (5), gives 𝑠(𝐴𝑋 ) = 0.018 dB.
The reported measurement result is to be calculated using the mean of 𝑛 = 4 measurements.

𝑠(𝐴 𝑋 ) 0.018
𝑢rep (𝐴𝑋 ) = = = 0.009 dB
√𝑛 √4
K.3.9 Summary table for 𝐴𝑋 = 𝐴𝑏 − 𝐴𝑎 + δ𝐴IF + δ𝐷IF + δ𝐿M + δ𝑅D + δ𝑀 + δ𝐴L
Uncertainty Probability 𝑢𝑖 (𝐾𝑋 ) 𝜐𝑖 or

dB distribution dB 𝜐eff
δ𝐴IF Calibration of reference attenuator 0.005 Normal 2.0 1.0 0.0025 ∞

δ𝐷IF Drift since last calibration 0.002 Rectangular √3 1.0 0.0012 ∞
δ𝐿M Mixer non-linearity 0.006 Rectangular √3 1.0 0.0035 ∞
δ𝑅D Resolution of indication 0.010 Triangular √6 1.0 0.0041 ∞
δ𝑀 Mismatch 0.022 Normal 1 1.0 0.022 ∞
δ𝐴L Signal leakage effects 0.001 Rectangular √3 1.0 0.0006 ∞
𝐴𝑋 Repeatability of indication 0.009 Normal 1.0 1.0 0.009 3
𝑢c (𝐴𝑋 ) Combined standard uncertainty Normal 0.0245 >150

Normal
𝑈95% (𝐴𝑋 ) Expanded uncertainty 0.0491 >150
( 𝑘 = 2)
Note that 𝐴𝑏 and 𝐴𝑎 do not need to appear in the table as all uncertainty associated with these quantities is accounted for by
other terms.
The measured value of the 30 dB attenuator at 10 GHz was 30.050 dB ± 0.049 dB.
𝑘 = 2, providing a coverage probability of approximately 95 %. The uncertainty evaluation has been carried
out in accordance with UKAS requirements.
NOTES:
1 Combination of relatively small uncertainties expressed in dB is permissible since loge(1+x) ≈ x when x is small and
2.303log10(1+x) ≈ x. For example: 0.1 dB corresponds to a power ratio of 1.023 and 2.303log10(1+0.023) = 0.0227.
Thus, relatively small uncertainties expressed in dB may be combined in the same way as those expressed as
linear relative values, e.g., percentage.
2 For attenuation measurements, the probability distribution for RF mismatch uncertainty is dependent on the
combination of at least three mismatch uncertainties and can be treated as having a normal distribution. For further
details see paragraph E.9.5.
3 In a subsequent use of an attenuator further uncertainty contributions may arise due to the use of different connector
pairs.

K.4 Calibration of a weight of nominal value 10 kg of OIML Class M1
K.4.1 The calibration is carried out using a mass comparator whose performance characteristics have
previously been determined, and a weight of OIML Class F2. The unknown weight is obtained
from:
𝑊𝑋 = 𝑊𝑆 + ∆𝑊 + δ𝐷𝑆 + δ𝐼𝑑 + ∆𝐶 + ∆𝐴𝑏 + δ𝑊𝑟
where
𝑊𝑋 = value of unknown weight,

𝑊𝑆 = calibration of the standard weight,
∆𝑊 = measured weight difference weight,
𝑊𝑆 = calibration of the standard weight,
δ𝐷𝑆 = drift of standard weight since last calibration,
δ𝐼𝑑 = the rounding error of the value of the least significant digit of the two indications,
∆𝐶 = comparator non-linearity,
∆𝐴𝑏 = correction for air buoyancy,
δ𝑊𝑟 = repeatability error.
K.4.2 The calibration certificate for the standard mass gives an uncertainty of 30 mg at a coverage
probability of approximately 95 % (𝑘 = 2).
K.4.3 The allowed monitored drift limits for the standard mass have been set equal to the expanded
uncertainty of its calibration, and are ±30 mg. A rectangular probability distribution has been
assumed.
K.4.4 The least significant digit 𝐼𝑑 for the mass comparator represents 10 mg. Digital rounding has
limits of ±0.5 𝐼𝑑 for the indication of the values of both the standard and the unknown weights.
Combining these two rectangular distributions gives a triangular distribution, with uncertainty
limits of ±𝐼𝑑 , that is ±10 mg.
K.4.5 The linearity error of the comparator over the 2.5 g range permitted by the laboratory's
procedures for the comparison was estimated from previous measurements to have limits of
±3 mg. A rectangular probability distribution has been assumed.
K.4.6 No correction is made for air buoyancy, for which limits were estimated to be ±1 ppm of nominal
value i.e., ±10 mg. A rectangular probability distribution has been assumed.
K.4.7 A previous Type A evaluation of the repeatability of the measurement process, comprising
𝑚 = 10 comparisons between the standard and unknown weight, gave a repeatability standard
deviation, 𝑠(δ𝑊𝑟 ), of 8.7 mg, with 𝜐 = (𝑚 − 1) = 10 degrees of freedom.
This evaluation replicates the normal variation in positioning single weights on the comparator,
and therefore includes effects due to eccentricity errors.

K.4.8 Three results were obtained for the unknown weight using the conventional technique of
bracketing the reading with two readings for the standard. The results were as follows:
No. Weight on pan Comparator Standard mean unknown - standard

reading
standard + 0.01 g
1 unknown + 0.03 g + 0.015 g + 0.015 g
standard + 0.02 g
2 unknown + 0.04 g + 0.015 g + 0.025 g
standard + 0.01 g
3 unknown + 0.03 g + 0.010 g + 0.020 g
standard + 0.01 g
Mean difference: ∆𝑊 = + 0.020 g
From the calibration certificate, the mass of the standard is 10 000.005 g. The calibrated value of
the unknown is therefore 𝑊𝑋 = 10 000.005 g + 0.020 g = 10 000.025 g.
K.4.9 The reported measurement result is to be calculated using the mean of 𝑛 = 3 measurements
(since three comparisons between standard and unknown are made). So, from equation (4), the
repeatability uncertainty is
𝑠(δ𝑊𝑟 ) 8.7
𝑢rep (δ𝑊𝑟 ) = = = 5.0 mg
√𝑛 √3
K.4.10 Summary table for 𝑊𝑋 = 𝑊𝑆 + ∆𝑊 + δ𝐷𝑆 + δ𝐼𝑑 + ∆𝐶 + ∆𝐴𝑏 + δ𝑊𝑟
Uncertainty Probability 𝑢𝑖 (𝑊𝑋 ) 𝜐𝑖 or

mg distribution mg 𝜐eff
𝑊𝑆 Calibration of standard weight 30.0 Normal 2.0 1.0 15.0 ∞

𝐷𝑆 Drift since last calibration 30.0 Rectangular √3 1.0 17.32 ∞
δ𝐼𝑑 Digital rounding error, comparison 10.0 Triangular √6 1.0 4.08 ∞
Δ𝐶 Comparator non-linearity 3.0 Rectangular √3 1.0 1.73 ∞
Δ𝐴𝑏 Air buoyancy (1 ppm of nominal value) 10.0 Rectangular √3 1.0 5.77 ∞
δ𝑊𝑟 Repeatability of indication 5.0 Normal 1.0 1.0 5.02 9
𝑢c (𝑊𝑋 ) Combined standard uncertainty Normal 24.56 >500

Normal
𝑈 Expanded uncertainty 49.12 >500
( 𝑘 = 2)
Note that all uncertainty associated with ∆𝑊 is accounted for by δ𝐼𝑑 and δ𝑊𝑟 .
The measured value of the 10 kg weight was 10 000.025 g ± 0.049 g.
𝑘 = 2, providing a coverage probability of approximately 95 %. The uncertainty evaluation has been
carried out in accordance with UKAS requirements.
NOTE: The degrees of freedom shown in the uncertainty budget are derived from a previous evaluation of repeatability,
for which 10 readings were used (see paragraph B4).

K.5 Calibration of a weighing machine of 205 g capacity by 0.1 mg digit
K.5.1 The calibration is carried out using weights of OIML Class E2. Checks will normally be carried out
for linearity of response across the nominal capacity of the weighing machine, eccentricity effects
of the positioning of weights on the load receptor, and repeatability of the machine for repeated
weighing near full load. The span of the weighing machine has been adjusted using its internal
weight before calibration. The following uncertainty evaluation is carried out for a near full loading
of 200 g.
The machine indication errors are obtained from
∆𝐼𝑋 = 𝐼𝑋 − 𝑊𝑆 + δ𝐷𝑆 + δ𝐼𝑑0 + δ𝐼𝑑 + ∆𝐴𝑏 + δ𝐼𝑟
where
∆𝐼𝑋 = measurement error (‘error of indication’) for indication 𝐼𝑋 ,

𝐼𝑋 = indication (for a weight of ‘unknown’ value),
𝑊𝑆 = weight of the standard,
δ𝐷𝑆 = drift of standard since last calibration,
δ𝐼𝑑0 = the rounding error of the value of one digit at the zero reading,
δ𝐼𝑑 = the rounding error of the value of one digit of the indicated value,
∆𝐴𝑏 = correction for air buoyancy,
δ𝐼𝑟 = repeatability error of the indication.
K.5.2 The calibration certificate for the stainless steel 200 g standard mass gives an uncertainty of
0.1 mg at a coverage probability of approximately 95 % (𝑘 = 2).
K.5.3 No correction is made for drift, but the calibration interval is set so as to limit the drift to ±0.1 mg.
The probability distribution is assumed to be rectangular.
K.5.4 No correction can be made for the rounding due to the resolution of the digital display of the
machine when zeroing. The least significant digit on the range being calibrated corresponds to 0.1
mg and there is therefore a possible rounding error of ±0.05 mg. The probability distribution is
assumed to be rectangular.
NOTE: It is often the case that when a weighing machine is zeroed, or tared, it may do so to a greater resolution than that
provided by the digital readout. The above contribution may be reduced where justified, for example by determining the error
at zero By OIML R76-1, A.4.2.3.2.
K.5.5 No correction can be made for the rounding due to the resolution of the digital display of the
machine when loaded. The least significant digit on the range being calibrated corresponds to 0.1
mg and there is therefore a possible rounding error of ±0.05 mg. The probability distribution is
assumed to be rectangular.
K.5.6 No correction is made for air buoyancy. As the span of the machine was adjusted with its internal
weight before calibration, the uncertainty limits were estimated to be ±1 ppm of the nominal value,
i.e., ±0.2 mg.
K.5.7 The repeatability of the machine was established from a series of 𝑚 = 10 readings (Type A
evaluation), which gave a repeatability standard deviation, 𝑠(𝐼𝑟 ), of 0.05 mg, with 𝜐 = (𝑚 − 1) = 9
degrees of freedom.
K.5.8 Only one reading was taken to establish the weighing machine indication for each linearity and
eccentricity point. For this particular calibration point the indication, 𝐼𝑋 , was 199.9999 g when the
200 g standard mass was applied.

The reported measurement result is to be this single, 𝑛 = 1 measurement. So, from equation (4),
the repeatability uncertainty is:
𝑠(δ𝐼𝑟 ) 0.05
𝑢rep (δ𝐼𝑟 ) = = = 0.05 mg.
√𝑛 √1
K.5.9 Summary table for ∆𝐼𝑋 = 𝐼𝑋 − 𝑊𝑆 + δ𝐷𝑆 + δ𝐼𝑑0 + δ𝐼𝑑 + ∆𝐴𝑏 + δ𝐼𝑟
Uncertainty Probability 𝑢𝑖 (𝐼𝑋 ) 𝜐𝑖 or

mg distribution mg 𝜐eff
𝑊𝑆 Calibration of standard weight 0.1 Normal 2.0 1.0 0.05 ∞

δ𝐷𝑆 Drift since last calibration 0.1 Rectangular √3 1.0 0.058 ∞
δ𝐼𝑑0 Digital rounding error (at zero) 0.05 Rectangular √3 1.0 0.029 ∞
δ𝐼𝑑 Digital rounding error (for indicated value) 0.05 Rectangular √3 1.0 0.029 ∞
∆𝐴𝑏 Air buoyancy (1 ppm of nominal value) 0.2 Rectangular √3 1.0 0.115 ∞
δ𝐼𝑟 Repeatability of indication 0.05 Normal 1.0 1.0 0.05 9
𝑢c (𝐼𝑋 ) Combined standard uncertainty Normal 0.150 >500

Normal
( 𝑘 = 2)
K5.10 Reported result
For an applied weight of 200 g:

the error of indication of the weighing machine was 0.10 mg ± 0.30 mg
The reported expanded uncertainty is based on a standard uncertainty multiplied by a coverage

factor 𝑘 = 2, providing a coverage probability of approximately 95 %. The uncertainty evaluation
has been carried out in accordance with UKAS requirements.
K.6 Calibration of a Grade 2 gauge block of nominal length 10 mm
K.6.1 The calibration was carried out using a comparator with reference to a grade K standard gauge
block of similar material. The length of the unknown gauge block, 𝐿𝑋 , was determined from
𝐿𝑋 = 𝐿𝑆 + δ𝐿𝐷 + ∆𝐿 − [𝐿(𝛼 ∆𝑡 + ∆𝛼 ∆𝑇 )] + ∆𝐷𝐶 + δ𝐶 + 𝐿𝑉(𝑋) + δ𝐿𝑟
where
𝐿𝑆 = certified length of the standard gauge block at 20ºC,

δ𝐿𝐷 = drift with time of certified length of standard gauge block,
∆𝐿 = measured difference in length,
𝛼 = mean thermal expansion coefficient of the standard and unknown gauge blocks,
∆𝑡 = difference in temperature between the standard and unknown gauge blocks,
∆𝛼 = difference in thermal expansion coefficients of the standard and unknown gauge blocks,
∆𝑇 = difference in mean temperature of gauge blocks and reference temperature of 20 ºC
when ∆𝐿 is determined,
∆𝐷𝐶 = discrimination and linearity of the comparator,

δ𝐶 = difference in coefficient of compression of standard and unknown gauge blocks,

𝐿𝑉(𝑋) = variation in length with respect to the measuring faces of the unknown gauge block,
δ𝐿𝑟 = repeatability error of measurement.
K.6.2 The value of 𝐿𝑆 was obtained from the calibration certificate for the standard gauge block. The
associated uncertainty was 0.03 μm (𝑘 = 2).
K.6.3 The change in value 𝐿𝐷 of the standard gauge block with time (drift) was estimated from previous
calibrations to be zero with an uncertainty of ±15 nm. From experimental evidence and prior
experience, the value of zero was considered the most likely, with diminishing probability that the
value approached the limits. A triangular distribution was therefore assigned to this uncertainty
contribution.
K.6.4 The coefficient of thermal expansion applicable to each gauge block was assumed to have a
value, 𝛼, of 11.5 μm m-1 ºC-1 with limits of ±1 μm m-1 ºC-1. Combining these two rectangular
distributions, the difference in thermal expansion coefficient between the two blocks, is
±2 μm m-1 ºC-1 with a triangular distribution. For 𝐿 = 10 mm this corresponds to ±20 nm ºC-1. This
difference will have two influences:
(a) The difference in temperature, δ𝑡, between the standard and unknown gauge blocks was
estimated to be zero with limits of ±0.08 ºC, giving rise to a length uncertainty of ±1.6 nm.
(b) The difference, δ𝑇, between the mean temperature of the two gauge blocks and the reference
temperature of 20 ºC was measured to be zero and was assigned limits of ±0.2 ºC, giving rise to a
length uncertainty of ±4 nm.
As the influence of the expansion coefficient appears directly in both of these uncertainty
contributions they are considered to be correlated and, in accordance with paragraph D.3.2, the
corresponding uncertainties have been added before being combined with the remaining
contributions. This is included in the uncertainty budget as δ𝑇𝑆,𝑋 .
K.6.5 The error due to the discrimination and non-linearity of the comparator, ∆𝐷𝐶 , was taken as zero
with limits of ±0.05 μm, assessed from previous measurements. Similarly, the difference in elastic
compression δ𝐶 between the standard and unknown gauge blocks was estimated to be zero with
limits of ±0.005 μm.
K.6.6 The variation in length of the unknown gauge block, 𝐿𝑉(𝑋) , was considered to comprise two
components:
(a) Effect due to incorrect central alignment of the probe; assuming this misalignment was within a
circle of radius 0.5 mm, calculations based on the specifications for grade C gauge blocks indicted
uncertainty limits of ±17 nm.
(b) Effects due to surface irregularities such as scratches or indentations; such effects have a
detection limit of approximately ±25 nm when examined by experienced staff.
Little is known about the PDF for either effect, but it is considered very unlikely that the extremes of
both effects will be encountered. Their combined effect is therefore modelled by a triangular PDF
with limits of ±42 nm.
K.6.7 The repeatability of the calibration process was established from previous measurements using
gauge blocks of similar type and nominal length. This Type A evaluation, based upon 𝑚 = 11
measurements and using equation (5), yielded a repeatability standard deviation 𝑠(δ𝐿𝑟 ) of 16 nm
with 𝜐 = (𝑚 − 1) = 10 degrees of freedom.

K.6.8 The calibration of the unknown gauge block was established from a single measurement; however,
as the conditions were the same as for the previous evaluation of repeatability, the standard
uncertainty due to repeatability can be obtained from equation (4) using the previous estimate of
repeatability standard deviation 𝑠(δ𝐿𝑟 ), and 𝑛 = 1 (because only one reading is made for the
actual calibration).
𝑠(δ𝐿𝑟 ) 16
𝑢rep (𝐿𝑟 ) = = = 16 nm
√𝑛 √1
K.6.9 The measured length of the unknown gauge block was 𝐿𝑋 = 𝐿𝑆 + ∆𝐿 = 9.999 940 mm.
K.6.10 Summary table for 𝐿𝑋 = 𝐿𝑆 + δ𝐿𝐷 + ∆𝐿 − [𝐿(𝛼 ∆𝑡 + ∆𝛼 ∆𝑇 )] + ∆𝐷𝐶 + δ𝐶 + 𝐿𝑉(𝑋) + δ𝐿𝑟
Uncertainty Probability 𝑢𝑖 (𝐿𝑋 ) 𝜐𝑖 or

nm distribution nm 𝜐eff
𝐿𝑆 Calibration of the standard gauge block 30 Normal 2.0 1.0 15.0 ∞

δ𝐿𝐷 Drift since last calibration 15 Triangular √6 1.0 6.1 ∞
∆𝐷𝐶 Comparator 50 Rectangular √3 1.0 28.9 ∞
δ𝐶 Difference in elastic compression 5.0 Rectangular √3 1.0 2.9 ∞
δ𝑇𝑆,𝑋 Temperature effects 5.6 Triangular √6 1.0 2.3 ∞
𝐿𝑉(𝑋) Length variation of unknown gauge block 42 Triangular √6 1.0 17.2 ∞
δ𝐿𝑟 Repeatability 16 Normal 1.0 1.0 16.0 10
𝑢c (𝐿𝑋 ) Combined standard uncertainty Normal 40.7 >350
Normal
( 𝑘 = 2)
Note that all uncertainty associated with ∆𝐿 is accounted for by ∆𝐷𝐶 and δ𝐿𝑟 .
The measured length of the gauge block was 9.999 940 mm ± 0.081 μm.


K.7 Calibration of a Type N thermocouple at 1000 ºC
K.7.1 A Type N thermocouple is calibrated against two reference standard Type R thermocouples in a
horizontal furnace at a temperature of 1000 ºC. The EMFs generated by the thermocouples are
measured using a digital microvoltmeter via a selector/reversing switch. All the thermocouples
have their reference junctions at 0 ºC. The unknown thermocouple is connected to the reference
point using compensating cables.
K.7.2 The temperature 𝑡𝑋 of the hot junction of the unknown thermocouple is given by
δ𝑡0𝑆
𝑡𝑋 = 𝑡𝑆 (𝑉𝑖𝑆 + δ𝑉𝑖𝑆1 + δ𝑉𝑖𝑆2 + δ𝑉𝑅 − ) + δ𝑡𝐷 + δ𝑡𝐹 + δ𝑡𝑟
𝐶𝑆0
𝐶𝑆
≈ 𝑡𝑆 𝑉𝑖𝑆 + 𝐶𝑆 δ𝑉𝑖𝑆1 + 𝐶𝑆 δ𝑉𝑖𝑆2 + 𝐶𝑆 δ𝑉𝑅 − δ𝑡 + δ𝑡𝐷 + δ𝑡𝐹 + δ𝑡𝑟
𝐶𝑆0 0𝑆
The voltage 𝑉𝑋 (𝑡) across the thermocouple wires with the reference junction at 0 ºC during the
calibration is
∆𝑡 δ𝑡0𝑋 δ𝑡 δ𝑡0𝑋
𝑉𝑋 (𝑡) ≈ 𝑉𝑋 (𝑡𝑋 ) + − + δ𝑉𝑟 = 𝑉𝑖𝑋 + δ𝑉𝑖𝑋1 + δ𝑉𝑖𝑋2 + δ𝑉𝐿𝑋 + δ𝑉𝑇𝐻 + δ𝑉𝑅 + − + δ𝑉𝑟
𝐶𝑋 𝐶𝑋0 𝐶𝑋 𝐶𝑋0
where
𝑡𝑆 (𝑉) = temperature of the reference thermometer in terms of voltage with the cold
junction at 0 ºC. The function is given in the calibration certificate,
𝑉𝑖𝑆 , 𝑉𝑖𝑋 = indication of the microvoltmeter,
δ𝑉𝑖𝑆1 , δ𝑉𝑖𝑋1 = voltage corrections due to the calibration of the microvoltmeter,
δ𝑉𝑖𝑆2 , δ𝑉𝑖𝑋2 = rounding errors due to the resolution of the microvoltmeter,
δ𝑉𝑅 = voltage error due to contact effects of the reversing switch,
δ𝑡0𝑆 , δ𝑡0𝑋 = temperature corrections associated with the reference junctions,
𝐶𝑆 , 𝐶𝑋 = sensitivity coefficients of the thermocouples for voltage at the measurement
temperature of 1000 ºC,
𝐶𝑆0 , 𝐶𝑋0 = sensitivity coefficients of the thermocouples for voltage at the reference
temperature of 0 ºC,
δ𝑡𝐷 = drift of the reference thermometers since the last calibration,
δ𝑡𝐹 = temperature correction due to non-uniformity of the furnace,
𝑡 = temperature at which the unknown thermocouple is to be calibrated,
∆𝑡 = 𝑡 − 𝑡𝑋 = deviation of the temperature of the calibration point from the temperature of
the furnace,
δ𝑉𝐿𝑋 = voltage error due to the compensation leads,
δ𝑉𝑇𝐻 = error due to inhomogeneity of the unknown thermocouple,
δ𝑡𝑟 = repeatability error for established reference temperatures,
δ𝑉𝑟 = repeatability error of voltage measurements.
K.7.3 The reported result is the output EMF of the test thermocouple at the temperature of the hot
junction. The measurement process consists of two parts - determination of the temperature of
the furnace and determination of the EMF of the test thermocouple. The evaluation of
uncertainty has therefore been split into two parts to reflect this situation.

K.7.4 The Type R reference thermocouples are supplied with calibration certificates that relate the
temperature of their hot junctions with their cold junctions at 0 ºC to the voltage across their
wires. The expanded uncertainty is 0.3 ºC with a coverage factor 𝑘 = 2.
K.7.5 No correction is made for drift of the reference thermocouples since the last calibration but an
uncertainty of ±0.3 ºC has been estimated from previous calibrations. A rectangular probability
distribution has been assumed.
K.7.6 The voltage sensitivity coefficients of the reference and unknown thermocouples have been
obtained from reference tables as follows:
Sensitivity coefficient at temperatures of

Thermocouple
0 ºC 1000 ºC
Reference (type R) 𝑐𝑆0 = 0.189 ºC/μV 𝑐𝑆 = 0.077 ºC/μV
Unknown (type N) 𝑐𝑋0 = 0.038 ºC/μV 𝑐𝑋 = 0.026 ºC/μV
K.7.7 The least significant digit of the microvoltmeter corresponds to a value of 1 μV. This results in
possible rounding errors, δ𝑉𝑖𝑆2 and δ𝑉𝑖𝑋2 , of ±0.5 μV for each indication.
K.7.8 Corrections were made to the microvoltmeter readings by using data from its calibration
certificate. Drift and other influences were all considered negligible, therefore only the calibration
uncertainty of 2.0 μV (𝑘 = 2) is to be included in the uncertainty budget.
K.7.9 Residual parasitic offset voltages due to the switch contacts were estimated to be zero within
±2.0 μV.
K.7.10 The temperature of the reference junction of each thermocouple is known to be 0 ºC within
±0.1 ºC. For the 1000 ºC measurements, the sensitivity coefficient associated with the
uncertainty in the reference junction temperature is the ratio of those at 0 ºC and 1000 ºC, i.e.,
𝑐𝑠
= −0.407.
𝑐𝑠𝑜
K.7.11 The temperature gradients inside the furnace had been measured. At 1000 ºC deviations from
non-uniformity of temperature in the region of measurement are within ±1 ºC.
K.7.12 The compensation leads had been tested in the range 0 ºC to 40 ºC. Voltage differences
between the leads and the thermocouple wires were estimated to be less than ±5 μV.
K.7.13 The error due to inhomogeneity of the unknown thermocouple was determined during the
calibration by varying the immersion depth. Corrections are not practical for this effect therefore
the error was assumed to be zero within ±0.3 ºC.
K.7.14 The sequence of measurements is as

follows:
1 First standard thermocouple
2 Unknown thermocouple
3 Second standard thermocouple
4 Unknown thermocouple
5 First standard thermocouple
The polarity is then reversed, and the sequence is repeated. Four readings are thus obtained for
all the thermocouples. This sequence reduces the effects of drift in the thermal source and
parasitic thermocouple voltages. The results were as follows:

Thermocouple: First standard Unknown Second standard
Corrected voltages: + 10500 μV + 36245 μV + 10503 μV

+ 10503 μV + 36248 μV + 10503 μV
– 10503 μV – 36248 μV – 10505 μV
– 10504 μV – 36251 μV – 10505 μV
Absolute mean values: 10502.5 μV 36248 μV 10504.0 μV
Temperature of hot junctions: 1000.4 ºC 1000.6 ºC
Mean temperature of furnace: 1000.5 ºC
K.7.15 The thermocouple output EMF is corrected for the difference between the nominal temperature of
1000 ºC and the measured temperature of 1000.5 ºC. The reported thermocouple output is
1000
𝑉𝑋 = 36248 × μV = 36230 μV.
1000.5
K.7.16 In this example it is assumed that the procedure requires that the difference between the two
standards must not exceed 0.3 ºC. If this is the case, then the measurement must be repeated
and/or the reason for the difference investigated.
K.7.17 From the four readings on each thermocouple, one observation of the mean voltage of each
thermocouple was deduced. The mean voltages of the reference thermocouples are converted to
temperature observations by means of temperature/voltage relationships given in their calibration
certificates. These temperature values are highly correlated. By taking the mean they are combined
into one observation of the temperature of the furnace at the location of the test thermocouple. In a
similar way one observation of the voltage of the test thermocouple is extracted.
K.7.18 In order to determine the repeatability standard deviation associated with these measurements
Independent, Type A evaluations had been carried out on a previous occasion. In each case a
series of 𝑚 = 10 measurements had been undertaken at the same temperature of operation.
Using equation (5), this gave estimates of the repeatability standard deviations for the temperature
of the furnace, 𝑠(δ𝑡𝑟 ) of 0.10 ºC and for the voltage of the thermocouple to be calibrated, 𝑠(δ𝑉𝑟 ), of
1.6 μV.
In each case there are 𝜐 = (𝑚 − 1) = 9 degrees of freedom
Only one calibration measurement is made, so from equation (4), the repeatability uncertainties are
𝑠(δ𝑡𝑟 ) 0.10
𝑢rep (δ𝑡𝑟 ) = = = 0.10 ℃
√𝑛 √1
and
𝑠(δ𝑉𝑟 ) 1.6
𝑢rep (δ𝑉𝑟 ) = = = 1.6 μV
√𝑛 √1
(The value of 𝑛 = 1 is used to calculate the standard uncertainty because in the normal procedure
only one sequence of measurements is made at each temperature.)

K.7.19 Summary table - temperature of the furnace

𝑐𝑆
𝑡𝑋 = 𝑡𝑆 𝑉𝑖𝑆 + 𝑐𝑆 δ𝑉𝑖𝑆1 + 𝑐𝑆 δ𝑉𝑖𝑆2 + 𝑐𝑆 δ𝑉𝑅 − δ𝑡 + δ𝑡𝐷 + δ𝑡𝐹 + δ𝑡𝑟
𝑐𝑆0 0𝑆
Probability 𝑢𝑖 (𝑡𝑋 ) 𝜐𝑖 or
distribution ºC 𝜐eff
𝑡𝑆 Calibration of standard thermocouples 0.3 ºC Normal 2.0 1.0 0.150 ∞

δ𝑡𝐷 Drift in standard thermocouples 0.3 ºC Rectangular √3 1.0 0.173 ∞
δ𝑉𝑖𝑆1 Voltmeter calibration and drift 2.0 μV Normal 2.0 0.077 0.077 ∞
δ𝑉𝑅 Switch contacts 2.0 μV Rectangular √3 0.077 0.089 ∞
δ𝑡0𝑆 Determination of reference point 0.1 ºC Rectangular √3 -0.407 -0.024 ∞
δ𝑡𝑟 Repeatability 0.1 ºC Normal 1.0 1.0 0.100 9
δ𝑉𝑖𝑆2 Voltmeter resolution 0.5 μV Rectangular √3 0.077 0.022 ∞
δ 𝑡𝐹 Furnace non-uniformity 1.0 ºC Rectangular √3 1.0 0.577 ∞
𝑢c (𝑡𝑋 ) Combined standard uncertainty Normal 0.641 >500
K.7.20 Summary table - EMF of unknown thermocouple

δ𝑡 δ𝑡0𝑋
𝑉𝑋 (𝑡) = 𝑉𝑖𝑋 + δ𝑉𝑖𝑋1 + δ𝑉𝑖𝑋2 + δ𝑉𝐿𝑋 + δ𝑉𝑇𝐻 + δ𝑉𝑅 + − + δ𝑉𝑟
𝑐𝑋 𝑐𝑋0
Probability 𝑢𝑖 (𝑉) 𝜐𝑖 or
distribution μV 𝜐eff
𝑡𝑋 Furnace temperature 0.641 ºC Normal 1.0 38.5 24.65 >500

δ𝑉𝐿𝑋 Compensating leads 5.0 μV Rectangular √3 1.0 2.89 ∞
δ𝑉𝑖𝑋1 Voltmeter calibration and drift 2.0 μV Normal 2.0 1.0 1.00 ∞
δ𝑉𝑅 Switch contacts 2.0 μV Rectangular √3 1.0 1.15 ∞
δ𝑡0𝑋 Determination of reference point 0.1 ºC Rectangular √3 26.3 1.52 ∞
δ𝑉𝑟 Repeatability 1.6 μV Normal 1.0 1.0 1.60 9
δ𝑉𝑖𝑋2 Voltmeter resolution 0.5 μV Rectangular √3 1.0 0.29 ∞
δ𝑉𝑇𝐻 Inhomogeneity of thermocouple 0.3 ºC Rectangular √3 38.5 6.66 ∞
𝑢c (𝑉) Combined standard uncertainty Normal 25.9 >500
Normal
( 𝑘 = 2)
The type N thermocouple generates an EMF of 36 230 μV ± 52 μV at a temperature of 1000 C

with its cold junction at a temperature of 0 ºC.


K.8 Calibration of a Digital Pressure Indicator (DPI) at a nominal pressure of 2 MPa using a
reference hydraulic dead weight tester
K.8.1 The calibration pressure was generated using a dead weight tester (DWT) the performance
characteristics of which had previously been determined. The indication was approached with
increasing pressure to account for the existence of possible hysteresis in the DPI.
The measurement error for the unknown DPI is obtained from:
∆𝐼𝑋 = 𝐼𝑋 − 𝑃𝑋 + δ𝐼𝑑 + δ∆𝑟
𝜌
[∑ 𝑚𝐿 𝑔 (1 − 𝑎 )] + (−𝐵𝑉 (𝜌𝑓 − 𝜌𝑎 )𝑔) + 𝜙 × 𝐶
𝜌𝑚
∆𝐼𝑋 = 𝐼𝑋 − { + (𝜌𝑓 − 𝜌𝑎 )𝑔ℎ + δ𝑒𝑉 } + δ𝐼𝑑 + δ∆𝑟
𝐴0 (1 + 𝑎𝑝 )(1 + 𝜆(𝑡 − 20))
where
∆𝐼𝑋 = Error in the indication of the unknown DPI,

𝐼𝑋 = Indication of the DPI,
𝑃𝑋 = Generated pressure at the reference level of the DPI,
𝑚𝐿 = Mass of the component parts of the load, including the piston,
𝜌𝑎 = Density of ambient air,
𝜌𝑚 = Density of the mass, 𝑚, and can be significantly different for each load component,
𝑔 = The value of the local acceleration due to gravity,
ℎ = Height different between reference levels for standard and for generated pressure,
𝐵𝑉 = Buoyancy volume of the reference piston – from calibration certificate,
𝜌𝑓 = Density of hydraulic fluid,
𝜙 = Surface tension coefficient of hydraulic fluid,
𝐶 = Circumference of reference piston,
𝐴0 = Effective area at zero pressure of reference piston,
𝑎𝑝 = Distortion coefficient of reference piston (pressure dependant term),
𝜆 = Temperature coefficient of piston and cylinder,
𝛿𝑒𝑉 = Error due to the piston not being perfectly vertical,
𝛿 𝐼𝑑 = The rounding error corresponding to the least significant digit of the indication,
δ∆𝑟 = Repeatability error when measuring ∆𝐼𝑋 .
K.8.2 The calibration certificate for the reference DWT gives the piston area and its expanded
uncertainty as:
𝐴0 = 80.6516 mm2 ± 0.0026 mm2.
This results in a relative expanded uncertainty (𝑘 = 2) in 𝐴0 of 32.2 ppm.
K.8.3 The calibration certificate for the reference DWT gives the distortion coefficient of the reference
piston as 𝑎𝑝 = 6.0 x 10-6/MPa ± 0.5 x 10-6/MPa, where the coverage interval is defined in terms
of (±) the expanded uncertainty (𝑘 = 2).
K.8.4 The allowed drift limit in the effective area of the DWT, based on results from previous
calibrations, has been set to ±30 ppm, to be treated as the limits of a rectangular distribution.
K.8.5 Mass uncertainties

K.8.5.1 The mass of the piston is shown on the calibration certificate as 0.567 227 kg with an expanded
uncertainty (𝑘 = 2) of 0.000 010 kg.
The values of the mass set, as shown on the calibration certificate for the three weights “A”, “B”
and “C” used to generate 2 MPa are:
A = 0.255 242 kg ± 0.000 010 kg

B = 7.402 137 kg ± 0.000 050 kg
C = 8.224 784 kg ± 0.000 050 kg
where the coverage intervals are defined in terms of (±) the expanded uncertainty (𝑘 = 2)
As all calibrated mass values are likely to be strongly correlated the uncertainty for the combined
load is found by summing the individual uncertainties. Expressed as a relative value for the 2
MPa load this is:
𝑈piston + 𝑈A + 𝑈B + 𝑈C
Relative expanded uncertainty in mass =
Total mass of piston + A + B + C
(10 + 10 + 50 + 50) mg
Relative expanded uncertainty in mass =
(0.567 227 + 0.255 242 + 7.402 137 + 8.224 784) kg
Relative expanded uncertainty in mass = 7.3 ppm
to be treated as the semi-range of a rectangular distribution
K.8.5.2 The allowed drift limit of the piston mass, based on previous calibrations, has been set to
±0.000 015 kg.
The allowed limits to the drift of the mass set have been set to be equal to the expanded
uncertainty of its calibration, i.e., 10 mg, 50 mg and 50 mg respectively.
Again, as all calibrated mass values are likely to be strongly correlated, the uncertainty for the
combined load is found by summing the individual uncertainties. Expressed in relative terms the
value is:
(15 + 10 + 50 + 50)mg
Relative mass drift limit = ± = 7.6 ppm
(0.567 227 + 0.255 242 + 7.402 137 + 8.224 784)kg
to be treated as the semi-range of a rectangular distribution.
K.8.6 The uncertainty in the temperature of the piston is estimated to be no more than 0.5 ºC. This will
affect the pressure generated in proportion to the temperature coefficient of the piston and
cylinder combination. In this case a steel piston and cylinder has a temperature coefficient of
23 ppm/ºC. This sensitivity coefficient was obtained from the calibration certificate for the DWT.
K.8.7 A correction has already been made for the value of the local acceleration due to gravity. The
correction was estimated from knowledge of the measurement location and known (Bouguer)
anomalies. The expanded uncertainty associated with this estimate, is 3 ppm (𝑘 = 2).
K.8.8 A standard air buoyancy correction has been made assuming an air density of 1.2 kg.m-3 and a
mass density of 8000 kg.m-3. As all masses have an assumed density of 7800 kg.m-3, and
assuming normal laboratory ambient conditions, the additional uncertainty for this approximation
is estimated to be 13 ppm, to be treated as the semi-range of a rectangular distribution.

K.8.9 The uncertainty relating to fluid head effects arises from the height different between the
reference level of the reference DWT and the datum level for generated pressure, estimated as
+2 mm with an uncertainty of 1 mm. No correction is made; therefore, a limit value of 3 mm has
been assigned for the uncertainty associated with fluid head effects. Assuming that the density
of the oil used is 917 kg/m3 and the local value of g is 9.81 m.s-2, then the uncertainty associated
with the fluid head effect is 917 x 9.81 x 0.003 = 27.0 Pa. In relative terms this corresponds to an
uncertainty of 13.5 ppm at 2 MPa, to be treated as the semi-range of a rectangular distribution.
K.8.10 The uncertainty contribution arising from the combined effects of the uncertainty in the
knowledge of buoyancy volume, surface tension and fluid density has been estimated as 2 ppm
based on a relative uncertainty in each of ±10 %, to be treated as the semi-range of a
rectangular distribution.
K.8.11 An uncertainty arises due to the fact that the piston may not be perfectly vertical. If it were, then all
of the force would act on the area. Any departure from vertical will reduce the force, and therefore
the pressure, by the cosine of the angle. In this example, it is assumed that, after levelling, the
piston is vertical to within 0.15°. The effect in terms of generated pressure is proportional to the
cosine of the angle from true vertical. The cosine of 0.15° is 0.999 996 6. The maximum error is
therefore -3.4 ppm of the generated pressure. This is to be represented by a rectangular
distribution with limits of ±3.4 ppm.
NOTE: This effect always acts in one direction, i.e., the generated pressure will always be smaller than that obtained if
the piston were truly vertical. As this uncertainty is small compared with others in this particular calibration, it is
convenient to treat it as bilateral.
K.8.12 No correction can be made for the rounding error, δ𝐼𝑑 , due to the resolution of the digital display of
the DPI. The least significant digit on the range being calibrated of this particular DPI changes in
steps of 200 Pa and there is therefore a possible rounding error of ±100 Pa or, in relative terms,
±50 ppm. The probability distribution is assumed to be rectangular
K.8.13 The repeatability standard deviation of the calibration process had been established from previous
measurements using DPIs of similar type and range. This Type A evaluation, based upon 𝑚 = 10
measurements, yields a relative standard deviation of 16 ppm, with 𝜐 = (𝑚 − 1) = 9 degrees of
freedom.
This calibration of the unknown DPI was established from a single measurement. As the conditions
were the same as previously, the repeatability uncertainty can be obtained using the previously
obtained repeatability standard deviation, and 𝑛 = 1 (as only one measurement is made for the
calibration).
𝑠(δ ∆ 𝑟 ) 16
𝑢rep (δ∆𝑟 ) = = = 16 ppm
√𝑛 √1
K.8.14 All uncertainties associated with the indicated value 𝐼𝑋 are accounted for by the error terms δ𝐼𝑑
and δ∆𝑟 , so no further contributions need be considered.
K.8.15 The generated pressure was calculated from the mass of the piston, that of the mass set, and
the following quantities:
Temperature of Piston: 21 ºC
Buoyancy volume of piston: 3.0 x 10-7 m3
Air density: 1.2 kg.m-3
Local gravity: 9.811812 m.s-2
Oil surface tension coefficient: 0.0315 N/m

This results in an applied pressure of P𝑋 = 2.000 806 MPa
K.8.16 Summary table for

𝜌
[∑ 𝑚𝐿 𝑔 (1 − 𝑎 )] + (−𝐵𝑉 (𝜌𝑓 − 𝜌𝑎 )𝑔) + 𝜙 × 𝐶
𝜌𝑚
∆𝐼𝑋 = 𝐼𝑋 − { + (𝜌𝑓 − 𝜌𝑎 )𝑔ℎ + δ𝑒𝑉 } + δ𝐼𝑑 + δ∆𝑟
𝐴0 (1 + 𝑎𝑝 )(1 + 𝜆(𝑡 − 20))
Probability 𝑢𝑖 (∆𝐼𝑋 ) 𝜐𝑖 or
distribution ppm 𝜐eff
𝐴0 Calibration of DWT (area) 32.2 ppm Normal 2.0 1.0 16.25 ∞
𝑎𝑝 Calibration of DWT (distortion) 0.5 ppm/MPa Normal 2.0 2.0 MPa 0.500 ∞
𝐴0 Drift in area 30 ppm Rectangular √3 1.0 17.32 ∞
𝑚𝐿 Calibration of total load 7.3 ppm Normal 2 1.0 3.65 ∞
𝑚𝐿 Drift of total load 7.6 ppm Rectangular √3 1.0 4.38 ∞
𝑡 Temperature of the piston 0.5 ºC Rectangular √3 23 ppm/ºC 6.64 ∞
𝑔 Local gravity determination 3.0 ppm Normal 2 1.0 1.50 ∞
BV Air buoyancy 13 ppm Rectangular √3 1.0 7.51 ∞
ℎ Fluid head effects 13.5 ppm Rectangular √3 1.0 7.79 ∞
𝜙 Other fluid effects 2.0 ppm Rectangular √3 1.0 1.15 ∞
δ𝑒𝑉 Levelling effects 3.4 ppm Rectangular √3 1.0 1.96 ∞
δ𝐼𝑑 Digital rounding error 50 ppm Rectangular √3 1.0 28.9 ∞
δ∆𝑟 Repeatability 16 ppm Normal 1 1.0 16.0 9
𝑢c (∆𝐼𝑋 ) Combined standard uncertainty Normal 43.0 >350
Normal
( 𝑘 = 2)
The pressure was applied in an increasing direction until it reached a final value of
2.000 806 MPa. The indication of the digital pressure indicator was 2.000 84 MPa.
The corresponding measurement error ∆𝐼𝑋 is 17 ppm ± 86 ppm (34 Pa  170 Pa)

NOTES:
1 In this example, the uncertainty due to resolution, δ𝐼𝑑 , is larger than any other contribution and is assigned a
rectangular distribution. Nevertheless, the combined standard uncertainty is still normally distributed, due to the
presence of the other uncertainties, even though they are of smaller magnitude. This has been verified by MCS.
2 The resolution uncertainty is based on the least significant digit of the DPI. For this particular instrument it changes
in steps of 2 digits, therefore the uncertainty is 1 digit.
3 This uncertainty budget has been constructed in relative terms (ppm), as most of the errors that arise are
proportional to the generated pressure and it is the convention in this particular field of measurement to express
uncertainties in this manner. If it is required that the uncertainty is reported in absolute units, it can be calculated
from the reported value and the relative uncertainty. In this case, the expanded uncertainty in absolute terms is
0.000 17 MPa.

Appendix L Expression of Uncertainty for a Range of Values
L.1 Introduction
L.1.1 On occasions it is convenient to provide a statement of uncertainty that describes a range of

values rather than a single result.
L.1.2 The GUM [1] deals with expression of uncertainties for the reporting of a single value of a
measurand, or more than one parameter derived from the same set of data. In practice many
measuring instruments are calibrated at several points on a range and the use of an expression
describing the uncertainty at any of these points can be desirable.
L.1.3 This Appendix therefore describes the situations when this can occur, explains how it can be
dealt with using the principles of this guide and provides an illustration of the process using a
worked example.
L.2 Principles
L.2.1 When measurements are made over a range of values and the corresponding sources of
uncertainty are examined it may be found that some are absolute in nature (i.e., they arise in a
manner that is independent of the value of the measurand) and some are relative in nature (i.e.
they arise in a manner that makes them proportional to the value of the measurand).
L.2.2 It is possible, of course, to follow this guide and to calculate a unique value for the
measurement uncertainty at each tabulated point over the range. This is often the least
ambiguous approach and arguably provides the clearest form of reporting.
L.2.3 However, for many users there are circumstances where it is useful to be able to separately
identify the absolute and relative constituents of the measurement uncertainty. For example, so
that these can be imported into a subsequent uncertainty evaluation as separate inputs.
The user can achieve this by performing a mathematical fit to tabulated values, if these have
been provided.
L.2.4 The process for evaluating the measurement uncertainty describing a range of values is
identical to that for single values except that the absolute and relative input quantities are
separated, and in effect, an uncertainty evaluation is carried out for each type in the manner
already described in this guide.
L.2.5 Traditionally the measurement uncertainty has sometimes (incorrectly) then been expressed as
a ‘linear’ combination of the relative and absolute components
𝑈 = 𝑈rel + 𝑈abs
but this linear addition of quantities is not in accordance with the principles embodied in the
GUM (unless there happens to be a very high degree of correlation between the absolute and
relative terms - which is not usually the case).
Instead, the measurement uncertainty should be expressed as a quadrature sum (root of the
sum of the squares of the components). This can either be in a mathematical form or as an
equivalent statement in words.

L.3 Example of uncertainty evaluation for a range of values
L.3.1 In this example a 6½ - digit electronic multimeter (DMM) is calibrated on its 1 V dc range using a
multi-function calibrator.
L.3.2 The calibrations were carried out in both polarities at 0.1 V increments from zero to 1 V. Only
one measurement was carried out at each point and therefore reliance was placed on a previous
evaluation of repeatability using similar multimeters.
L.3.3 No corrections were made for known errors of the calibrator as these were identified as being
small relative to other sources of uncertainty. The uncorrected errors are assumed to be zero
with an uncertainty obtained by analysis of information obtained from the calibration certificate
for the calibrator.
L.3.4 The measurement error ∆𝐼DMM associated with indication 𝐼DMM of the multimeter under test, can
be described as follows:
∆𝐼DMM = 𝐼DMM − 𝑉REF + δ𝐼RES + δ∆𝑟
with
𝑉REF = 𝑉CAL + δ𝑉D + δ𝑉UE + δ𝑉TC + δ𝑉LIN + δ𝑉T + δ𝑉CM
i.e.,
∆𝐼DMM = 𝐼DMM − {𝑉CAL + δ𝑉D + δ𝑉UE + δ𝑉TC + δ𝑉LIN + δ𝑉T + δ𝑉CM } + δ𝐼RES + δ∆𝑟
where
∆𝐼𝐷𝑀𝑀 = error in the indication of the DMM,

𝐼𝐷𝑀𝑀 = indication of the DMM,
𝛿𝐼𝑅𝐸𝑆 = rounding error due to the resolution of DMM being calibrated,
𝛿∆𝑟 = repeatability error of indication when DMM is measuring 𝑉REF ,
𝑉𝑅𝐸𝐹 = reference voltage presented to the DMM,
𝑉𝐶𝐴𝐿 = calibrated voltage setting of multifunction calibrator,
𝛿𝑉𝐷 = error due to drift in voltage of multifunction calibrator since last calibration,
𝛿𝑉𝑈𝐸 = error due to uncorrected errors of multifunction calibrator,
𝛿𝑉𝑇𝐶 = error due to temperature coefficient of multifunction calibrator,
𝛿𝑉𝐿𝐼𝑁 = error due to linearity and zero offset of multifunction calibrator,
𝛿𝑉𝑇 = error due to thermoelectric voltages generated at junctions of connecting leads
calibrator and multimeter,
𝛿𝑉𝐶𝑀 = effects due to imperfect common-mode rejection characteristics of the measurement
system.
L.3.5 The calibration uncertainty was obtained from the certificate for the multi-function calibrator. This
had a value of 2.8 ppm as a relative uncertainty and a further 0.5 μV in absolute units (𝑘 = 2).
L.3.6 The manufacturer's 1-year performance specification for the calibrator included information
relating to the following effects: 𝑉𝐷 , 𝑉UE , and 𝑉TC - these contributions are seen to be relative in
nature; 𝑉𝐿𝐼𝑁 - this contribution is seen to be absolute in nature.

The specification for the calibrator on the 1 V dc range was ±8 ppm of reading ±1 ppm of full-
scale. On this particular multifunction calibrator, the full-scale value is twice the range value;
therefore, the absolute term is ±2 x 10-6 V = ±2 μV. The performance of the calibrator had been
verified by examining its calibration data and history, using internal quality control checks and
ensuring that it was used within the temperature range and other conditions as specified by the
manufacturer. A rectangular distribution was assumed.
L.3.7 The effects of thermoelectric voltages for the particular connecting leads used had been
evaluated on a previous occasion. Thermoelectric voltages are independent of the voltage
setting are therefore an absolute uncertainty contribution. An uncertainty of 1 μV was assigned,
based on previous experiments with the leads. The probability distribution was assumed to be
rectangular.
L.3.8 Effects due to common-mode signals had also been the subject of a previous evaluation and an
uncertainty of 1 μV, with a rectangular distribution, was assigned. This contribution is absolute in
nature, as the common-mode voltage is unrelated to the measured voltage.
L.3.9 No correction can be made for the rounding due to the resolution of the digital display of the
multimeter. The least significant digit on the range being calibrated corresponds to 1 μV and
there is therefore a possible rounding error δ𝐼RES of ±0.5 μV. The probability distribution is
assumed to be rectangular, and this term is absolute in nature.
L.3.10 The repeatability standard deviation of the calibration process had been established from previous
measurements using a similar multimeter. This Type A evaluation, based upon 𝑚 = 10
measurements at each value, was performed at 0 V and at 1.0 V (as well as other nearby values).
Repeatability at the zero point was found to be insignificant compared with other absolute
contributions, whereas for measurements at 1.0 V, the repeatability standard deviation 𝑠(δ∆𝑟 )
was found to be 2.5 ppm, (consistent with previous evaluations at other voltage levels) with
𝜐 = (𝑚 − 1) = 9 degrees of freedom.
This calibration of the unknown DMM was established from a single measurement. However,
assuming that the previous evaluation of repeatability using a similar multimeter is representative
of this later measurement, the repeatability uncertainty can be obtained using the previously
obtained repeatability standard deviation, and 𝑛 = 1 (as only one reading is made for the actual
calibration).
𝑠(δ ∆ 𝑟 ) 2.5
𝑢rep (δ∆𝑟 ) = = = 2.5 ppm
√𝑛 √1
L.3.11 All uncertainties associated with the indicated value 𝐼DMM are accounted for by other terms, so
no further contributions need be considered.

L.3.12 Summary table for

∆𝐼DMM = 𝐼DMM − {𝑉CAL + δ𝑉D + δ𝑉UE + δ𝑉TC + δ𝑉LIN + δ𝑉T + δ𝑉CM } + δ𝐼RES + δ∆𝑟
value value
Probability 𝑢𝑖 (∆𝐼DMM ) 𝑢𝑖 (∆𝐼DMM ) 𝜐𝑖 or
Symbol Source of uncertainty (rel) (abs) Divisor 𝐶𝑖
ppm μV
distribution ppm μV 𝜐eff
𝑉CAL Calibration uncertainty 2.8 0.5 Normal 2.0 1 1.40 0.25 ∞
Specification of
𝑉SPEC 8.0 2.0 Rectangular √3 1 4.62 1.15 ∞
calibrator
Thermoelectric
𝑉T voltages
1.0 Rectangular √3 1 0.58 ∞
𝑉CM Common mode effects 1.0 Rectangular √3 1 0.58 ∞
δ𝐼RES Voltmeter resolution 0.5 Rectangular √3 1 0.29 ∞
δ∆𝑟 Repeatability 2.5 Normal 1.0 1 2.5 9
𝑢𝑐 (∆𝐼DMM ) 5.42 1.46

Standard uncertainty Normal >100
Normal
𝑈 Expanded uncertainty 10.8 2.92 >100
(𝑘 = 2)
L.3.13 It is assumed that the results of this calibration will be presented in tabular form (as is common
practice).
When the uncertainty is not reported in the same table, one of the following equivalent
statements regarding uncertainty can instead be given:
Mathematical representations:
The expanded uncertainty for the above measurements is
𝑈(∆𝐼DMM ) = √(𝑣 × 11 ppm)2 + (2.9 μV)2
where 𝑣 is the corrected indication of the DMM
So, for example,

if 𝑣 = 500 mV we find that
𝑈(∆𝐼DMM ) = √(500 mV × 11 ppm)2 + (2.9 μV)2 = √(5.5 μV)2 + (2.9 μV)2 = 6.2 μV
Alternatively, this expression can be written in terms of a defined function, e.g.,
𝑈(∆𝐼DMM ) = 𝑄[𝑣 × 11 ppm, 2.9 μV]
where
𝑄[𝑎, 𝑏] = [𝑎2 + 𝑎𝑏 ]1⁄2
Note that in practice the parameter (𝑣 in this case) is often omitted, and expressions such as
𝑈 = √(11 ppm)2 + (2.9 μV)2
or
𝑈 = 𝑄[11 ppm, 2.9 μV]
might be encountered.

Although this simplification doesn’t usually create any confusion, it should be recognised that
these expressions are not mathematically complete or dimensionally consistent (in this example
the relative term is ‘dimensionless’ whereas the absolute term has dimensions corresponding to
voltage).
The simplification should be avoided when there is any possibility of ambiguity in its use.
Non-mathematical representation:
The expanded uncertainty for the above measurements is found by taking the square root of the
sum of the squares of two values corresponding to:
a) Relative uncertainty of 11 ppm of the corrected indication of the DMM, and
b) Absolute uncertainty of 2.9 μV
L.3.14 In each case the result should be accompanied by the usual statement:

L.3.15 Uncertainties reported in this form can be imported into subsequent evaluations as single
enumerated values i.e., as a single input value computed from the expression.
Alternatively, they can be imported as two separate (independent) inputs corresponding to the
relative and the absolute quantity.

Appendix M Assessment of Conformity with Specification
M.1 In test reports and to some extent in calibration certificates, there will be occasions where it
becomes necessary to make a statement about whether or not the measured result indicates
conformity of the measurand with some form of specification.
M.2 Guidance on performing and reporting assessment of conformity with specifications is given in
UKAS document LAB 48: ‘Decision Rules and Statements of Conformity’ [14] and the
references therein.

Appendix N Uncertainties for Test Results
N.1 Introduction
N.1.1 ISO/IEC 17025 [5] requires that testing laboratories shall have and apply procedures for
estimating measurement uncertainty.
N.1.2 Testing laboratories should therefore have a defined policy covering the evaluation and reporting
of the uncertainties associated with the tests performed. The laboratory should use documented
procedures for the evaluation, treatment and reporting of the uncertainty.
N.1.3 The methodology for estimation of uncertainty in testing is no different from that in calibration
and therefore the procedures described in this document apply equally to testing results.
N.1.4 It is however recognised that in certain areas of testing it may be known that a significant
contribution to uncertainty exists but that the nature of the test precludes a rigorous evaluation of
this contribution. In such cases, ISO/IEC 17025 requires that a reasonable estimation be made
and that the form of the reporting does not give an incorrect impression of the uncertainty.
NOTES:
1. In some fields of testing, it may be the case that the contribution of measuring instruments to the overall uncertainty
can be demonstrated to be insignificant when compared with the repeatability of the process. Nevertheless, such
instruments have to be shown to comply with the relevant specifications, normally by calibration.
2. It will sometimes be the case that the procedure requires standard reference materials to be subject to the same
process, the result being the difference between the readings for the analyte and the reference material. In such
cases, much of the process can be considered to be negatively correlated and the measurement uncertainty can be
evaluated from the resolution and repeatability of the process; matrix effects may also have to be considered.
N.1.5 Many tests involve some form of examination (or inspection) where the outcome of the test is a
nominal property (e.g., colour, shape, species, sequence of markers…). Others might involve
establishing a position on an ordinal scale (e.g., Rockwell C, Richter, Beaufort, octane…). In the
case of these ‘qualitative’ tests the concept of measurement uncertainty does not readily apply.
But that is not to say that measurement uncertainty doesn't play a role in such tests… in fact, in
most cases, such tests are performed under defined conditions that are themselves subject to
measurement.
For example:
A test requires an inspector to examine the colour of a fluid sample after preparation according
to a defined procedure and processing in an oven at (40  1) C for between 60 min and 65 min.
For this test, the examination of the colour involves a subjective judgement from a trained and
competent examiner whose reliability can be established by proficiency testing. However, the
oven temperature and elapsed time are both measurable quantities for which a value and a
measurement uncertainty can be established.

N.2 Common sources of uncertainty in testing
N.2.1 The following general issues will apply to many areas of testing:
(a) Incomplete definition of the test - the requirement may not be clearly described, e.g., the
temperature of a test may be given as 'room temperature'.
(b) Imperfect realisation of the test procedure; even when the test conditions are clearly defined it
may not be possible to produce the theoretical conditions in practice due to unavoidable
imperfections in the materials or systems used.
(c) Sampling - the sample may not be fully representative. In some disciplines, such as
microbiological testing, it can be very difficult to obtain a representative sample.
(d) Inadequate knowledge of the effects of environmental conditions on the measurement process,
or imperfect measurement of environmental conditions.
(e) Personal bias and human factors; for example:
- Reading of scales on analogue indicating instruments.
- Judgement of colour.
- Reaction time, e.g. when using a stopwatch.
- Instrument resolution or discrimination threshold, or errors in graduation of a scale.
(f) Values assigned to measuring equipment and reference materials.
(g) Changes in the characteristics or performance of measuring equipment or reference materials

since the last calibration.
(h) Values of constants and other parameters used in data evaluation.
(i) Approximations and assumptions incorporated in the measurement method and procedure.
(j) Variations in repeated observations made under similar but not identical conditions - such
random effects may be caused by, for example, electrical noise in measuring instruments,
short-term fluctuations in the local environment, e.g. temperature, humidity and air pressure,
variability in the performance of the person carrying out the test and variability in the
homogeneity of the sample itself.
These sources are not necessarily independent.
In addition, unrecognised systematic effects may exist that are not considered, but contribute to
error. This is one reason that participation in inter-laboratory comparisons, proficiency testing
schemes and internal cross-checking of results by different means are encouraged.
N.2.2 Information on some of the sources of these errors can be obtained from:
(a) Data in calibration certificates - this enables corrections to be made and uncertainties to be
assigned.
(b) Previous measurement data - for example, history graphs can be constructed and can yield
useful information about changes with time.

(c) Experience with or general knowledge about the behaviour and properties of similar materials
and equipment.
(d) Accepted values of constants associated with materials and quantities.
(e) Manufacturers' specifications.
(f) All other relevant information.
These are all referred to as Type B evaluations because the values were not obtained by
statistical means. However, the influence of random effects is often evaluated by the use of
statistics; if this is the case then the evaluation is designated Type A.
N.3 The ‘top-down’ approach
N.3.1 The organisation of the uncertainty budget around a measurement equation as described in this
guidance document is commonly described as a ‘bottom-up’ approach. However, in some
situations, it is not feasible to establish a detailed measurement equation and a different ‘top-
down’ approach is taken.
N.3.2 Guidance documents such as, EURACHEM/CITAC Guide CG 4 “Quantifying Uncertainty in

Analytical Measurement” [10] and ISO 21748 “Guidance for the use of repeatability,
reproducibility and trueness estimates in measurement uncertainty evaluation” [8] describe a
process for evaluating measurement uncertainty using data obtained from collaborative studies
conducted in accordance with ISO 5725-2 “Accuracy (trueness and precision) of measurement
methods and results — Part 2: Basic method for the determination of repeatability and
reproducibility of a standard measurement method” [7]
N.3.3 Measurement methods that have been evaluated by this ‘collaborative study’ approach are
characterised in terms of ‘precision’ i.e., by a ‘repeatability standard deviation’ and a
‘reproducibility standard deviation’. In some cases, a ‘method bias’ and its associated
uncertainty will also be established.
N.3.4 In its simplest form, a laboratory that demonstrates control of precision and bias, and introduces
no additional factors (from operations not conducted during the collaborative study) may use the
reproducibility standard deviation for estimating the standard uncertainty.

Appendix P Comparing Independent Results using En ratio
P.1 The equivalence of two independent measurement results is often compared by computing the
Normalised Error ratio (𝐸𝑁 ratio), where
𝐿1 − 𝐿2
𝐸𝑁 =
√(𝑈1 )2 + (𝑈2 )2
𝐿1 = the value of measurement result 1

𝐿2 = the value of measurement result 2
𝑈1 = the expanded uncertainty associated with measurement 1
𝑈2 = the expanded uncertainty associated with measurement 2
It is common to perform the test with uncertainties expanded for 95 % coverage probability. In
this case, if the results are equivalent, the 𝐸𝑁 ratio should be within the range ±1, on
approximately 95 % of occasions.
If the analysis reveals that 𝐸𝑁 lies outside this range more frequently than expected, some
investigation and corrective action will be required.
NOTES:
1 The commonly used formation of the 𝐸𝑁 ratio described above, involving the direct combination of expanded
uncertainties, is not consistent with the GUM [1]
Instead, it should be evaluated from
𝐿1 − 𝐿2
𝐸𝑁 =
𝑘𝑝 √(𝑢1 )2 + (𝑢2 )2
where 𝑢1 and 𝑢2 are the standard uncertainties associated with 𝐿1 and 𝐿2 , and 𝑘𝑝 is the coverage factor established
for the PDF of the quantity (𝐿1 − 𝐿2 ).
In situations where the PDFs for both 𝐿1 and 𝐿2 have normal distributions this gives identical values to the
commonly used equation.
2 It is not physically realistic for comparisons to produce consistently small 𝐸𝑁 values. This usually indicates that 𝑈1
and/or 𝑈2 are significant overestimates. This can arise from failure to identify correlation between 𝐿1 and 𝐿2 ,or
simply from the (poor) practice of basing uncertainty evaluations on ‘safe’ or ‘conservative’ estimates rather than on
realistic estimates for input quantities (as required by the GUM).
3 The 𝐸𝑁 ratio (as defined above) is not usually appropriate for testing the equivalence of operators or processes in
the same laboratory, as the measurement results are not likely to be independent. To use it in such cases an
approximation can be obtained by omitting all common inputs (such as those relating to temperature effects and
calibration traceability) from each uncertainty evaluation, so that only independent quantities, such as errors of
resolution and repeatability remain.

Appendix Q Symbols
The symbols used are taken mainly from the GUM [1]. The meanings have also been described further in the
text, usually where they first occur, but are summarised here for convenience of reference.
𝑓 Functional relationship between the measurand 𝑌 and the input quantities 𝑋𝑖 on which 𝑌
depends 𝑌 = 𝑓(𝑋𝑖 ), and between output estimate 𝑦 and input estimates 𝑥𝑖 , thus, 𝑦 = 𝑓(𝑥𝑖 ).
𝜕𝑓 Sensitivity coefficient 𝑖 th input quantity.

𝑐𝑖 =
𝜕𝑥𝑖 Partial derivative of the functional relationship 𝑓, taken with respect to input quantity 𝑋𝑖 ,
evaluated at 𝑋𝑖 = 𝑥𝑖 .
𝑎𝑖 Estimated semi-range of a probability distribution for an input quantity 𝑥𝑖 with width 2𝑎𝑖 .
𝑎R = 𝑐R 𝑎 Semi-width of a rectangular input distribution, corresponding to the semi-width 𝑎 of the input

distribution scaled by the corresponding sensitivity factor 𝑐R .
𝑘 Coverage factor (general).
𝑘𝑝 Coverage factor used to calculate the expanded uncertainty 𝑈𝑝 for a defined coverage
probability 𝑝., e.g., 𝑘95% .
𝑚 Number of readings or observations that are used for the evaluation of 𝑠(𝑞𝑗 ).
𝑛 Number of readings or observations that contribute to a mean value.
𝑁 Number of input quantities 𝑥𝑖 on which the value of the measurand depends, 𝑖 = 1 to 𝑁.
𝑞𝑗 𝑗th repeated observation of randomly varying quantity 𝑞.

See Note.
𝑞̅ Arithmetic mean or average of 𝑛 repeated observations of randomly varying quantity 𝑞.
𝑝 Coverage probability or ‘level of confidence’ (0 ≤ 𝑝 ≤ 1).

Often expressed in percentage terms (0 % ≤ 𝑝 ≤ 100 %).
𝜎 The standard deviation that characterises the variation in a measurement process

(The limit value of 𝑠 that would be found were it possible to repeat a measurement a very large
number of times).
𝑠 Best available estimate of 𝜎.

Referred to here as the repeatability standard deviation.
𝑠 Repeatability uncertainty.
𝑢rep =
√𝑛
𝑠(𝑞𝑗 ) Value of 𝑠 obtained from a sample of 𝑚 observations {𝑞1 , … , 𝑞𝑚 }.

See Note.
𝑡𝑝 (𝜈eff ) Student t-factor for 𝜈eff degrees of freedom corresponding to a given coverage probability 𝑝.
𝑢(𝑥𝑖 ) Standard uncertainty of input estimate 𝑥𝑖 .
𝑢𝑖 (𝑦) Standard uncertainty of input estimate 𝑥𝑖 .expressed in terms of the measurand,

𝑢𝑖 (𝑦) = |𝑐𝑖 |𝑢(𝑥𝑖 )

𝑢c (𝑦) Combined standard uncertainty of output estimate 𝑦.
𝑈𝑝 , 𝑈 Expanded uncertainty of output estimate 𝑦 that defines an interval 𝑦 ± 𝑈, with a defined

coverage probability, 𝑝 of containing the measurand 𝑌.
𝜈𝑖 Degrees of freedom of standard uncertainty 𝑢(𝑥𝑖 ) of input estimate 𝑥𝑖 .
𝜈eff Effective degrees of freedom of 𝑢𝑐 (𝑦) used to obtain 𝑡𝑝 (𝜈eff ).
δ𝑥 The prefix δ is used to demote a quantity whose most likely value is zero, although there is an
associated (non-zero) uncertainty.
∆𝑥 The symbol ∆ is used to represent a difference, a measurement error or a correction for which
a non-zero value is usually known.
N, R, T, U Labels, respectively used to designate, normal, rectangular, triangular and U-shaped

probability distributions.
N, R, T, U Subscripts, respectively used to designate, normal, rectangular, triangular and U-shaped

probability distributions.
NOTE The GUM uses the symbols 𝑞𝑘 and 𝑠(𝑞𝑘 ) whereas 𝑞𝑗 and 𝑠(𝑞𝑗 ) are used here. M3003 uses the subscript 𝑗 instead
of 𝑘 in order to avoid any possible confusion with the coverage factor 𝑘.

Appendix R References
[1] JCGM 100:2008,

Evaluation of measurement data - Guide to the Expression of Uncertainty in Measurement.
Known as ‘The GUM’.
[2] JCGM 101:2008,

Evaluation of measurement data - Supplement 1 to the “Guide to the expression of uncertainty in
measurement” - Propagation of distributions using a Monte Carlo method.
[3] JCGM GUM-6:2020,

Guide to the expression of uncertainty in measurement — Part 6: Developing and using
measurement models.
[4] JCGM 200:2008,

International vocabulary of metrology – Basic and general concepts and associated terms.
Known as ‘The VIM’.
[5] BS EN ISO/IEC 17025,

General requirements for the competence of testing and calibration laboratories.
[6] BS EN ISO 15189,

Medical laboratories - Requirements for quality and competence.
[7] ISO 5725-2:2019,

Accuracy (trueness and precision) of measurement methods and results – Part 2: Basic method
for the determination of repeatability and reproducibility of a standard measurement method.
[8] ISO 21748:2017,

Guidance for the use of repeatability, reproducibility and trueness estimates in measurement
uncertainty evaluation.
[9] ISO/TS 20914:2019,

Medical laboratories - Practical guidance for the estimation of measurement uncertainty.
[10] EURACHEM/CITAC Guide CG 4

Quantifying Uncertainty in Analytical Measurement,
QUAM:2012. P1, 3rd Ed, 2012. ISBN 978-0-948926-30-3.
[11] ILAC-P14:2020,
ILAC Policy for Uncertainty in Calibration.
[12] EURAMET Calibration Guide No. 18,

Guidelines on the Calibration of Non-Automatic Weighing Instruments
[13] UKAS LAB 14,

Calibration of Weighing Machines.
[14] UKAS LAB 48,

Decision Rules and Statements of Conformity.
[15] UKAS LAB 45,

Schedules of Accreditation for Calibration Laboratories.

[16] DIETRICH, C.F.

Uncertainty, Calibration and Probability: The Statistics of Scientific and Industrial Measurement
Uncertainty,
Institute of Physics, January 1991,
ISBN 0-7503-0060-4.
[17] HARRIS, I.A and WARNER, F.L.

Re-examination of mismatch uncertainty when measuring microwave power and attenuation.
IEE Proceedings, Vol 128 Pt H No 1, February 1981.


M3003 The Expression of Uncertainty and Confidence in Measurement

Uploaded by

Copyright:

Available Formats

M3003 The Expression of Uncertainty and Confidence in Measurement

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

M3003 The Expression of Uncertainty and Confidence in Measurement

Uploaded by

Copyright:

Available Formats

F

Edition 5 September 2022

The Expression of Uncertainty and Confidence in

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

Date Details of amendment

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

“Is the thermometer accurate?”

“How well can I read it?”

“Is the reading changing?”

“I am holding the thermometer in my hand. Am I warming it up?”

“Does it matter whereabouts in the room I take the measurement?”

2.7 Is the thermometer accurate?

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

usually achieved by an unbroken chain of comparisons originating at a national metrology

2.8 How well can I read it?

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

2.9 Is the reading changing?

So, what can be done about this?

NOTE: Standard deviation of the mean is also known as standard error.

In Microsoft Excel the STDEV.S cell function can be used.

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

2.9.11 We obtain an estimated standard deviation of 0.2915 °C.

2.10 I am holding the thermometer in my hand. Am I warming it up?

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

2.12 Does it matter where in the room I make the measurement?

2.13 “What exactly is it that I am trying to measure?”

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

3.17 In the resolution example, 𝑎 = 0.5 (for the resolution of 1.0).

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

Table 1 gives the expressions for various situations.

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

If a coverage probability is not stated, then a rectangular

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

The height ℎ of a flagpole is determined

Both ℎ and 𝑑 are in units of length but

If the measured distance is 7.0 m and 𝛷

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

3.37 The combined standard uncertainty is usually calculated from:

𝑢c (𝑦) = √∑ 𝑐𝑖2 𝑢2 (𝑥𝑖 ) = √∑ 𝑢𝑖2 (𝑦) (1)

𝑢𝑖 (𝑦) = |𝑐𝑖 |𝑢(𝑥𝑖 )

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

3.45 For example:

Coverage probability Coverage factor

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

4 Type A Evaluation of Standard Uncertainty

4.1 If an uncertainty is evaluated by statistical analysis of a series of observations, it is known as a

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

The estimated standard deviation,

The estimated repeatability is therefore,

where in this case 𝑛 = 𝑚.

The estimated repeatability is therefore,

w: www.ukas.com | t: +44(0)1784 429000 | e: [email protected]

5 Type B Evaluation of Standard Uncertainty

5.1 If an uncertainty is evaluated by non-statistical analysis, it is known as a Type B evaluation.

(e) The operational procedure.