KAN Pd-02.

05 Revisi: 0

Table of Content

1. INTRODUCTION ............................................................................................................ 1
2. MEASUREMENT ASSURANCE CONCEPT ................................................................... 1
3. THE USE OF CHECK STANDARD................................................................................. 4
4. DEFINING AND CREATING CONTROL CHART ............................................................ 5
5. USING CONTROL CHART ............................................................................................. 9
6. SOURCE OF MEASUREMENT DATA FOR MEASUREMENT ASSURANCE .............. 19
7. REFERENCES ............................................................................................................. 20

Date of issue : 2 January 2019 ii

 KAN Pd-02.05 Revisi: 0

KAN GUIDE ON MEASUREMENT ASSURANCE

1. INTRODUCTON
Chapter 7.7 of ISO/IEC 17025:2017 “General requirements for the competence of
testing and calibration laboratories” on “Ensuring the validity of results” requires:

7.7.1 The laboratory shall have a procedure for monitoring the validity of results. The
resulting data shall be recorded in such a way that trends are detectable and, where
practicable, statistical techniques shall be applied to review the results. This
monitoring shall be planned and reviewed and shall include, where appropriate, but
not be limited to:
a) use of reference materials or quality control materials;
b) use of alternative instrumentation that has been calibrated to provide
traceable results;
c) functional check(s) of measuring and testing equipment;
d) use of check or working standards with control charts, where applicable;
e) intermediate checks on measuring equipment;
f) replicate tests or calibrations using the same or different methods;
g) retesting or recalibration of retained items;
h) correlation of results for different characteristics of an item;
i) review of reported results;
j) intralaboratory comparisons;
k) testing of blind sample(s).

7.7.2 The laboratory shall monitor its performance by comparison with results of other
laboratories, where available and appropriate. This monitoring shall be planned and
reviewed and shall include, but not be limited to, either or both of the following:
a) participation in proficiency testing;

NOTE ISO/IEC 17043 contains additional information on proficiency tests and

proficiency testing providers. Proficiency testing providers that meet the requirements
of ISO/IEC 17043 are considered to be competent.

b) participation in interlaboratory comparisons other than proficiency testing.

This document gives the recommended method to implement the above requirement
that is applicable for calibration and testing laboratories that willing to be accredited
by National Accreditation Body of Indonesia or Komite Akreditasi Nasional (KAN)
based on ISO/IEC 17025:2017

2. MEASUREMENT ASSURANCE CONCEPT

Measurement assurance is a process to ensure adequate measurement results that
may include, but is not limited to:
 the use of good experimental design principle so that the entire measurement
process, its components and relevant influence factors can be well characterized,
monitored and controlled;
 complete experimental characterization of the measurement process uncertainty
including statistical variations, contribution from all known, or suspected influence
factors, imported uncertainties and the propagation of uncertainties throughout the
measurement process;
Date of issue : 2 January 2019 1
 KAN Pd-02.05 Revisi: 0

 continuously monitoring the performance and state of statistical control of

measurement process with proven statistical process control techniques, including
the measurement of well characterized check standard along with the normal
workload and use of appropriate control chart

The purpose of statistical control is to guarantee the ‘goodness’ of measurement

results within predictable limits and to validate the statement of uncertainty of the
measurement result. Statistical control methods can be used to test the measurement
process for change with respect to bias and variability from historical levels. However,
if the measurement process is improperly specified or calibrated, then the control
procedures can only guarantee comparability among measurements.

The assumptions that relate to measurement process apply statistical control; namely
that the errors of measurement are uncorrelated over time and come from a
population with a single distribution. The tests for control depend on the assumption
that the underlying distribution is normal (Gaussian), but the test procedures are
robust to slight departures from normality. Practically speaking, all that is required is
that the distribution of measurements be bell-shaped and symmetric.
Measurements on a check standard provide the mechanism for controlling the
measurement process. Measurement on a check standard should produce identical
results except for the effect of random errors, and tests for control are basically tests
of whether or not the random errors from the process continue to be drawn from the
same statistical distributions as the historical data on the check standard. Changes
that can be monitored and tested with the check standard database are:
 changes in bias and long-term variability; and
 changes in instrument precision or short-term variability

The concept of statistical process control is based on comparing how measurement

process is performing today (often a single measurement) to how it has performed in
the best (a database of measurement results). From our knowledge of statistics, we
have an idea what to expect from a stable process.

In a stable measurement process, we expect the majority of measurement to fall within

control limits that were established based on statistical evaluation of historical data.
Measurements that are fall outside the control limits are assumed to be “out of control”.
An investigation is then needed to find the cause(s) and suitable corrective action must
be taken.

Bias and long-term variability are controlled by monitoring measurements on a check

standard over time. A change in the measurement on the check standard that persists
at a constant level over several measurement sequences indicates possible:
 change or damage to the reference standards;
 change or damage to the check standard artifact;
 procedural change that vitiates the assumption of the measurement process

A change in the variability of the measurement on the check standard can be due to
one of many causes such as:
 loss of environmental controls;
 change in handling techniques
 severe degradation in instrumentation

Date of issue : 2 January 2019 2

 KAN Pd-02.05 Revisi: 0

Short-term variability or instrument precision is controlled by monitoring standard

deviations from repeated measurements on the instrument(s) of interest. The database
can come from measurements on a single artifact or representative set of artifacts.

The artifacts must be of the same type and geometry as items that are measured in the
workload, such as:
 items from the workload
 a single check standard chosen for this purpose
 a collection of artifacts set aside for this specific purpose

Calibration and testing can be thought as a production process in which the

measurement result and resulting report is the final product, then PLAN – DO –
CHECK – ACT (PDCA) process flow, which was originally developed for production
process and is often used to monitor change and to measure improvement of process
may be applied in assuring calibration and testing results.

The implementation of PDCA cycle in a measurement assurance system may be

described as follows:
 PLAN: identify the measurement process and procedures; use modeling tools to
help diagram the process; select and calibrate appropriate check standard
 DO: collect initial data to characterize the process; make sure that data are good,
stable, accurately reflect reference values and under control, use checklist and
data collection form to ensure consistency in gathering data; plot initial data on a
control chart; and establish limits
 CHECK: periodically gather additional data and immediately plot them on control
chart; monitor data on ongoing basis
 ACT: evaluate and analyze the data; use statistical tools for consistency;
implement corrective action or improvement as needed or desired; at this point, the
cycle continues.

EXAMPLE:

PLAN:

a. Define the measurement process: We are establishing a measurement

assurance program for the calibration of 200 g OIML F1 class. We use a double
substitution procedure using a 1 kg capacity AT 1005 comparator with 0.01 mg
readability and will use a 1 kg OIML E2 class reference standard for the
calibration
b. Establish the objectives: remember that whenever possible it is a good idea to
monitor both variability of the measurement process and the value of the
reference standard
c. Diagram (model) the process: the model of calibration process can be described
as follows:
S : reference standard
X : unit under test
X : unit under test
S : reference standard

The duplication of the above calibration process for a check standard can be
described as:

Date of issue : 2 January 2019 3

 KAN Pd-02.05 Revisi: 0

S : reference standard
Sc : check standard
Sc : check standard
S : reference standard

We can use one check standard to monitor this process because we can’t
incorporate check standard into the measurement and achieve the goals of the
procedure, we will duplicate the process by substituting a check standard into the
place of unit under test and repeating the process for a check standard.

3. THE USE OF CHECK STANDARD

Check standard is a standard that is used as part of measurement assurance

program to provide check on the standard and process to ensure that standard,
measurement result, and measurement process are within acceptable statistical
limits.

The check standard should be thought of in terms of a database of measurements. It

can be defined as an artefact or as a characteristic of the measurement process
whose value can be replicated from measurements taken over the life of the process.
Examples ofm check standards are:
 measurements on a stable artefact;
 difference between values of two reference standards as estimated form
calibration experiment;
 values of a process characteristic such as bias term, which is estimated from
measurement on reference standards and/or test items.

To control the measurement process and determine a valid uncertainty, we should

select a check standard that:
 duplicate the item being tested/calibrated
 evaluates the maximum random variation and bias of the process over an
extended period of time;
 has an established mean or calibrated reference value, with an uncertainty
statement; and
 its calibrated values is independent from the reference standards to ensure
that bias due to the process or calibration standards may be defected.

EXAMPLE:

PLAN:

d. Select and calibrate check standard: For the measurement described before in
the example, we will choose a standard that will monitor bias, and determining
the variability of the process. We have an older set of stainless steel standards
that have demonstrated stability over time, and have an independent calibration
value from the other accredited calibration laboratory. In this case we use a 200
g stainless steel mass standard of F1 class from this set to be a check standard
for the calibration described in this example. The calibration value of this
standard is used as the reference value for this check standard.

Date of issue : 2 January 2019 4

 KAN Pd-02.05 Revisi: 0

4. DEFINING AND CREATING CONTROL CHART

Control chart is a graphical tool used to visualize data when monitoring, evaluating
and improving a measurement process:
 monitoring and evaluating may include the need for controlling process
through corrective action and process improvement
 the measurement process include the evaluation of many contributing factors
including standards, measurement, variability, uncertainties, environmental
condition and staff performance

The type of chart, which is commonly used for monitoring, evaluating and improving
the measurement process, using control shart allow us to:
 ensure agreement with reference value, or to detect bias, offset, or change in
the set point of references;
 ensure ongoing stability of process variability and reported uncertainty or
detec change in the variability of the process that effects the uncertainty
analysis;
 ensure and document ongoing stability of the standards and process for
accuracy and traceability and to predict future measurement values;
 make decision regarding corrective action or process improvement

The question is often asked in defining and creating control chart, “how many initial
measurements do we need?“, considering our knowledge about statistic and the
measurement process, we can answer the question as follows:
 2 to 3 measurements, we need at least 2 measurement to calculate mean and
standard deviation, however, only this amount of data provides limited values
for our process
 7 to 12 measurements, we need at least 7 to 12 data to obtain initial estimates
on sample and population to begin developing control chart. The first set of 7
to 12 measurements will be used to establish initial reference limits. 10 is good
number of data points to use as the baseline for our initial control
measurements. Be sure to gather data over a period of time, with not more
than 1 measurement from the same day
 25 to 30 measurements, we need at least 25 to 30 data points on which to
have statistical decision about the data. We will need this many points to
establish whether the data are stable, randomly distributed and symmetric
about the mean. We should also have at least this much data for estimating
and calculating measurement uncertainty
 50 to 100 measurements, once we have maintained data for a while, it is good
idea to keep at least 50 to 100 data points or at least 3 years data to allow
monitoring trends. We will need to determine and document how long data will
need in our laboratory and the basis for archiving our old data

There is specific information that should always be on a control chart and some
information should be always obtained with the data and maintained in a database.
The data collection forms and a spreadsheet template or specialty software program
Date of issue : 2 January 2019 5
 KAN Pd-02.05 Revisi: 0

should be used to gather complete information consistently. There are some essential
information and good information to have in a control chart:
 Essential information:
1. Titles: identify laboratory, standard operating procedure, standard(s) and/or
check standard(s), nominal value, and time of measurement
2. Data: measured or calculated values, number of data point(s), mean,
standard deviation
3. X-axis: identify observations by date or time
4. Y-axis: observations or calculated values, with measurement unit identified
5. Central line: mean, and reference value whenever available, though it may
not be at the center
6. Limit: identification of upper and lower warning and control (action) limits.

 Good information:
1. Legend: when more tha
2. n one series present
3. Tolerances: when applicable
4. Uncertainties: for reference value, check standard and process output
5. Equipment information: device readability, configuration setting
6. Standard information: calibration date and interval information
7. Responsible staff: need on chart or in database
8. State of Control
9. Information about previous limits and history of the chart/data: if available

It is critical to understand the concept associated with each following step in creating
control chart no matter what tools are used to create the chart. Various tools may be
used, but the most common are spreadsheet or a specialty software product created
for this application. The following is the explanation to create a control chart, with an
example on the measurement assurance of the mass standard calibration using
Microsoft Excel Spreadsheet:

EXAMPLE:

DO:

a. STEP 1: Make measurement

We will collect observations and perform calculations and will start to create control
chart when we have 10 corrections (mass values) to enter. Remember that we
need 7 to 12 data points to construct the initial chart. We do not plot balance
observations, but the correction (mass value) of the check standard because we
want to evaluate the entire measurement process and not just the balance
variation over time.

We should normally enter data in columns when using a Microsoft Excel

Spreadsheet as follows:

Date of issue : 2 January 2019 6

 KAN Pd-02.05 Revisi: 0

description : 10 points - control chart initial data

SOP : direct comparison
balance used : AT 1005 / 0,01 mg
load : 200 g
unit : mg
check standard : stainless steel F1 OIML class
measurement process : calibration of 200 g F1 OIML class

date correction reference

01/02/02 0,27 0,276
02/02/02 0,30 0,276
03/02/02 0,25 0,276
04/02/02 0,31 0,276
05/02/02 0,28 0,276
06/02/02 0,29 0,276
07/02/02 0,30 0,276
08/02/02 0,26 0,276
09/02/02 0,30 0,276
10/02/02 0,28 0,276

b. STEP 2: Plot values

Plotting values on the y – axis versus dates of measurements on the x – axis is the
standard approach for plotting chart values.

Plotting values in a spreadsheet will require (at a minimum) selecting column

measurement data for series 1 and the chart type and then select the dates of
measurement for the x – axis.

This is the simplest approach to creating a chart. We will need to enter labels and
titles next. Be sure to label your values with the appropriate measurement units. To
create control chart in Microsoft Excel, you should use “LINE chart”.

c. STEP 3: Calculate statistics

The initial statistics in creating a control chart are central tendency and dispersion.

The mean and standard deviation are used to represent the central tendency and
dispersion. Spreadsheet allows us to calculate these statistics more easily than
Date of issue : 2 January 2019 7
 KAN Pd-02.05 Revisi: 0

using calculator. However, it will help to be able to validate our spreadsheet

equation with a calculator.

description : 10 points - control chart initial data

SOP : direct comparison
balance used : AT 1005 / 0,01 mg
load : 200 g
unit : mg
check standard : stainless steel F1 OIML class
measurement process : calibration of 200 g F1 OIML class

date correction reference mean std dev

01/02/02 0,27 0,276 0,28 0,02
02/02/02 0,30 0,276 0,28 0,02
03/02/02 0,25 0,276 0,28 0,02
04/02/02 0,31 0,276 0,28 0,02
05/02/02 0,28 0,276 0,28 0,02
06/02/02 0,29 0,276 0,28 0,02
07/02/02 0,30 0,276 0,28 0,02
08/02/02 0,26 0,276 0,28 0,02
09/02/02 0,30 0,276 0,28 0,02
10/02/02 0,28 0,276 0,28 0,02

d. STEP 4: Plot “mean” and “reference value”

The centre line for a control chart is often used as, or represents, an accepted
value. As a minimum, we plot the mean value of the data to show the central
tendency. But, also having an independently calibrated value allows us to evaluate
the reference standard and process for the presence of bias.

The accepted value may be the mean values as determined from the data, may be
calibration value determined by another laboratory, or may be a calibration value
determined in-house using a higher level procedure than the one being monitored.
If the reference value is the mean of corrections, we can only monitor variability in
the process and change of the standard, no monitoring of bias is possible.

If the reference value comes from a higher-level independent laboratory, bias from
reference value can be monitored as well as process variability. Having a
reference value with the fewest measurement possible in traceability chain (e.g.
from an NMI) will provide a measure of bias with the highest level of confidence. If
the reference value comes from a higher level measurement inhouse, bias within
the laboratory can be monitored also as well as process variability, but with lower
level of confidence that is possible when using an independence reference value
from an NMI or higher level accredited calibration laboratory.

Date of issue : 2 January 2019 8

 KAN Pd-02.05 Revisi: 0

5. USING CONTROL CHART

The overall goal in using a control chart is to efficiently ensure the quality of the
measurement results by minimizing the loss of data, integrating the system into
normal laboratory workload, and monitoring the standard and process for accuracy
and uncertainty.

Measurement assurance should be a real time monitoring process and not one in
which data are saved for entry into control charts just prior to laboratory audit or when
customer calls with measurement discrepancy. Data should be plotted and evaluated
immediately.

After plotting data, look for stability and randomness, be sure the latest point entered
is within the established limits, or we will need to take appropriate action. Plotting real
time data on the control charts can prevent the release of questionable data and
possible recall of certificate for items you have already tested or calibrated.

Normal distribution is probably the most important and most frequently used
distribution, both in the theory and the application of statistics. We expect our
measurement data to come from a normal distribution. We looked at this distribution
and considered confidence intervals and probabilities. In this chapter, think about the
probability that a point will come from this distribution and be within limits on the
control chart unless there is something wrong. We will consider what might be wrong
and what action steps may be needed to be taken to correct the problem.

When all data incorporated on the same control chart, we can see approximately
where the change occurred and predict future value, barring additional damage. The
unique pattern shown on the chart is the result of a change in the value of the check
standard. To conduct further analysis of these data, we could separate the before and
after data and conduct t-test and F-test, or we could compare the histograms of the
before and after data. The differences in the two distributions are quite obvious when
they are viewed graphically.

Date of issue : 2 January 2019 9

 KAN Pd-02.05 Revisi: 0

The calculation of F-value and interpretation of F-test result can be described as

follows:

sold is the standard deviation used to establish existing limit in the control chart, and
snew is the standard deviation of the most recent data. Our measurement system is
considered to be “out of control” if we meet the following conditions

The calculation of t-value and interpretation of t-test result can be described as

follows:

xold is the mean of the old data existing in the control chart; xnew is the mean of most
recent data; sold is the standard deviation used to establish existing limit in the control
chart; snew is the standard deviation of the most recent data; nold the number of data
points existing in the control chart; nnew is the number of data points of most recent
data. Our measurement system is considered to be “out of control” if we meet the
following conditions:

tcritical is the critical tvalue that depends on degrees of freedom calculated using the
following formula:

Our control charts helps to identifying stability or change in the measurement process
and in the standards we need to monitor the process and the standards to ensure the
quality of our measurements. In the following we identify the number of examples of
“things to look for” in the control charts that show instability or changes:

EXAMPLE:

CHECK and ACT

PROCESS DEGRADATION

Date of issue : 2 January 2019 10

 KAN Pd-02.05 Revisi: 0

The most recent data have wider dispersion than the earlier data on which the limits
were based as shown in the following control chart.

Possible assignable causes:

• Our balance need service
• We have new calibration officer in the laboratory who needs training
• Our environmental controls stopped working properly

Possible action steps:

• Obtain service or repair for balance
• Get new staff trained as soon as possible
• Fix environmental control as soon as possible

PROCESS IMPROVEMENT

The most recent data have a narrower dispersion than the earlier data on which the
limits were based as shown in the following control chart.

Date of issue : 2 January 2019 11

 KAN Pd-02.05 Revisi: 0

Possible assignable causes:

• We obtained service for your balance
• We have a new calibration officer who has many years of previous experience
• Our environmental controls were repaired or improved
• We have been practicing in balance operation (more care in centering the
weight on the pan, gently releasing the beam, allowing more stability before
taking a reading)

Possible action steps:

• Recalculate your standard deviation for the measurement process and update
your chart
• Recalculate tour standard deviation for the measurement process and
updated your uncertainty values

PROCESS COMPARED TO SPECIFICATION / TOLERANCES

When evaluating measurement process against specifications and tolerance, we

should consider whether or not the process is acceptable. If the process is
acceptable, there is no need to look for assignable causes. If it is not acceptable, then
we need to look for assignable causes and take action.

Date of issue : 2 January 2019 12

 KAN Pd-02.05 Revisi: 0

In the above cases, the control chart shows that the variability of the most recent data
is out of the action limits and the result of F-test also shows “out of control” condition,
However, the variability is still acceptable compared the tolerance limits, then there is
no need to look for possible assignable causes if we used the tolerance limits as a
action limits for our measurement processes.
If our control chart shows that the variability of most recent data exceeds tolerance
limits as shown in the following control chart:

Date of issue : 2 January 2019 13

 KAN Pd-02.05 Revisi: 0

Possible assignable causes:

 Our process is not achieving the requirements needed to meet the tolerances
because our balances has degraded and can not reach its normal
performance
 The process may be acceptable for the clients needs even though it is larger
than the tolerances we have assigned to the process

Possible action steps:

 Select or purchase a better balance that will meet the needs of our clients
and/or our stated tolerance
 Talk with our client to determine real, rather than stated measurement
requirements

Date of issue : 2 January 2019 14

 KAN Pd-02.05 Revisi: 0

SHIFT IN VALUES OF STANDARD

All of the most recent data are above the mean value of the earlier data as shown in
the following control chart:

Possible assignable causes:

 Our standard is damage
 Our check standard is damage
 We are using calculation software that has not been validated
 We have errors in entering data and not to double check data
 Our equation for the correction due to environmental condition were drifted
and correction changed
 Our standard was recalibrated and a new value was assigned without
starting a new control chart

Possible action steps:

 Correct for each of the above causes
 For damage to standard and check standard, a recalibration of the
standards is not enough, you must determine how and why the standard
was damage and make sure that it does not happen again

Date of issue : 2 January 2019 15

 KAN Pd-02.05 Revisi: 0

DRIFT OF VALUES OF STANDARD

The recent values seem to be steadily increasing above the previous mean values as
shown in the following control chart:

Remember that the stability is less important than the predictability. In fact, some
laboratories apply correction to their check standard based on a consistent history of
drift, they have a level of confidence for predicting drift and incorporate an uncertainty
associated with the predicted correction.
Possible assignable causes:
 Our working standard worn from use
 Our check standard getting dirty
 Our standards are not stable enough for this process
Possible action steps:
 Evaluate drift rate of standard, which may be used to evaluate calibration
intervals
 Incorporate the level of offset at some future points in time
 Include the uncertainty of the offset from reference values now and in the
future

Date of issue : 2 January 2019 16

 KAN Pd-02.05 Revisi: 0

BIAS OF STANDARD COMPARED TO REFERENCE

In our example, there is difference between the mean value and the reference value
obtained from the independent calibration of check standard as shown in the following
control chart:

Possible assignable causes:

 Repeated observation simply shows a measurement bias that is within the
uncertainty of calibrated values
 An error in the density of standard that affect buoyancy correction
 An error of environment correction equipment causing bias
 Standard changed at some point in the past and the shift is not reflect the
current chart
 It may be using standard operating procedure that does not compensate
for drift

Possible action steps:

 Evaluate bias with respect to the uncertainty and if it is small enough,
incorporate as an uncorrected systematic error
 Obtain a calibration for standard, check standard and environmental
equipment
 Get density measurement for our standard or check standard
 Use standard operating procedure that compensate for drift

Date of issue : 2 January 2019 17

 KAN Pd-02.05 Revisi: 0

OUTLIERS

There are a few of data from a recent measurement fall outside the action limits as
shown in the following control chart:

If we can identify the cause of outliers and take appropriate action, we can make a
note of the action in our control database and flag the points so they are not included
in the statistical calculations. Do not delete these points from control database, if we
find problems that cause outliers every 10th data point, it means the corrective action
steps are not effective. If we arbitrarily delete observations, we will not be able to
track erratic readings.

Possible assignable causes:

 Balance not exercised normally / properly
 We do not allow adequate thermal equilibrium to the standard prior to
begin calibration
 We have data entry error or miscalculation
 Our environmental system is out of control
 Our standard was recently damaged and this is the first indication of
problem (we would only see the data “out” on one side of the chart, if were
damaged, which is not the case here)

Possible action steps:

 Immediately make another measurement to check the validity of the
current measurement result and then make a determination about what to
do
 Another means for calculating the data may be used if an error was found
Date of issue : 2 January 2019 18
 KAN Pd-02.05 Revisi: 0

6. SOURCES OF MEASUREMENT DATA FOR MEASUREMENT ASSURANCE

Every laboratory has a number of measurement processes that should be identified

during the planning stage of PLAN – DO – CHECK – ACT cycle. Each measurement
process produces unique data that we can use to assure the quality of our
measurements. Each of these sources of data may be plotted, tracked and/or controlled
using control chart technique
It is important to integrate the analysis of our data sources and consider whether data
from one process is related to another, and whether one set of data supports our
assessment and tentative conclusions regarding another set of data. All of our data
sources are unique but they fit together like pieces of a puzzle. The sources of
measurement data we can used to monitor measurement process may include, but not
limited to:
a. Calibration of reference standards
 We can check the values for our reference standards over time
 The stability of reference standards is the main thing we should look for because
the values may be determined using different processes and different
laboratories over a long period of time. This is unlike tracking the values for our
check standards, because tracking the checks standard values represents a
process that we control within our laboratory. We don not always control the
process for the calibration of our standards
 By analyzing drift rates (due to possible wear) and variability (stability) of our
standards, we can determine suitable calibration intervals for our standards. We
can also determine if a change occurs
b. Proficiency testing data
 Data obtain for proficiency testing by inter-laboratory comparison often reflect
characteristics common to data from control charts for other standards and/or
measuring equipments maintained in the laboratory. Bias that is observed in a
proficiency testing results may also be evident in a control chart that monitor the
standards and/or measuring equipments that were used for particular proficiency
testing program
 For example: we have participated in inter-laboratory comparison with a 200 g
mass standard using substitution procedure, and our results showed a bias of
0.68 mg compared with the accepted reference value. We take a look at the chart
we have for the calibration history of our working standard and note that we have
a noticeable drift in mass values and that the calibration interval was nearly
exceeded. Then, we estimate the drift rate and find that the mass value for our
working standard is likely off by 0.60 mg to 0.70 mg in the same direction as the
bias observed in the inter-laboratory comparison. The values from two sources
agree and we need to calibrate our working standard
c. Intermediate check
 Intermediate check is the checking of our standards and/or measuring
equipments between official calibrations scheduled in the calibration program, to
provide additional assurance that all values continue to be stable. It provides an
intermediate verification of the values for our standards and/or measuring

Date of issue : 2 January 2019 19

 KAN Pd-02.05 Revisi: 0

equipments Intermediate check may be conducted that are internal to our

laboratory but which are used with limited frequency, perhaps only for this
purpose intermediate check may also be conducted with purpose of intermediate
check of our standards and/or measuring equipments

d. Calibration of working standards or measuring equipments

 We can track the values of working standards and/or measuring equipments over
time. We will look
 for the values and we may also be able to analyze different process by tracking
the procedures that
 have been used in the calibration of working standards and/or measuring
equipments over a long
 time
 By analyzing drift rates (due to possible wear) and variability (stability) of our
working standards
 and/or measuring equipments, we can determine appropriate calibration intervals
for working
 standards and/or measuring equipments (just as we can do with our reference
standards)
 We can also determine if a change in the values of our working standards and/or
measuring
 equipments occurs if measurement assurance program for our routine
calibrations and/or testing
 are modeled to monitor our working standards and/or measuring equipments,
then our charts
 provide warning signs for recalibration

e. Calibration or testing of customer item

 Every process used to calibrate and/or test customer items should have a
measurement assurance system applied. A check standard is needed to reflect
the standard deviation of the process and the validity of the standards and/or
measuring equipments on the date of the test
 However, as you have seen, calibration and/or testing of customer items is not
the only measurement process in our laboratory where measurement assurance
methodologies are useful

7. REFERENCES
1. Harris G. L, “NIST Special Publication 1001: Basic Mass Metrology CD ROM”,
National Institutes of Standards and Technologies, Technology Administration, US
Department of Commerce, September 2003
2. NIST/SEMATECH, “Engineering Statistics Handbook”, electronic file web based
handbook at http://www.itl.nist.gov/div898/handbook/, National Institutes of Standards
and Technology, Technology Administration, US Department of Commerce,
November 2003

Date of issue : 2 January 2019 20