GWP® - The Weighing Standard
Why We Should Challenge the Established Way We
Calibrate and Test Weighing Instruments
NCSLI Nashville
July 2013
Dr. Klaus Fritsch
Manager Compliance
Weighing – Essential for all Processes
Regulations affecting Scales and Balances - ISO
How do we put this
into practice?
“Measuring equipment shall be calibrated and/or verified at
specified intervals … against measurement standards
traceable to international or national measurement
standards.”
ISO 9001:2008, 7.6 Control of Monitoring and Measuring Devices
3
Regulations Affecting Scales & Balances - cGMP
How do we put this
into practice?
“Automatic, mechanical or electronic equipment […]
shall be routinely calibrated, inspected or checked
according to a written program designed to assure
proper performance.”
21 CFR part 211.68 (a), US GMP for Pharma
Regulations affecting Balances - USP
What does this
mean for my
balances?
"Weighing shall be performed using a balance that is calibrated […] and
meets the requirements defined for repeatability and accuracy."
"Repeatability is satisfactory if two times the standard deviation of the
weighed value, divided by the nominal value of the weight used, does not
exceed 0.10%."
USP 36–NF 31 Second Supplement, General Chapter 41 “Balances“
5
New USP General Chapters 41 & 1251
Publication in the Second Supplement to USP 36–NF 31 June 3rd 2013
 General Chapter 41 "Balances"
 General Chapter 1251 "Weighing on an Analytical Balance"
After a six month transition period the new chapters will be official
from December 1st 2013 onwards.
6
GWP® - The Weighing Standard
The science-based global
standard for efficient life cycle
management of weighing
instruments.
It applies a risk-based
approach that allows improving
control of the whole measuring
process, which in turn helps to
avoid costly OoS.
GWP® gives clear answers to
questions such as how to specify
or calibrate balances correctly.
It covers every relevant step of
the instrument's life cycle from
evaluation to routine operation.
The Weighing Instrument is being Calibrated…
Uncertainty of Weighing Instruments
 Nominal Property:
Readability
RD
 Measurement Properties:
Sensitivity
SE
Nonlinearity
NL
Eccentricity
EC
Repeatability
RP
All these properties contribute to the overall measurement
uncertainty of the respective weighing instrument.
The Perfect Weighing Instrument…
…Impaired by Readability, …
(RD)
…Sensitivity Offset, …
(SE)
…Nonlinearity, …
(NL)
…Eccentricity, …
(EC)
…and Repeatability
(RP)
Behavior of Weighing Uncertainty
U [g] = U0 + Constant x Weight
Relative Measurement Uncertainty [%]
Absolute Measurement
Uncertainty [mg]
(= Absolute measurement uncertainty / weight)
Uncertainty U [mg or %]
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Relative measurement uncertainty
increases as sample mass decreases
0
1
2
3
4
5
Load [g]
6
7
8
9
10
Max
For small sample weights, the relative uncertainty can become
so high that the weighing results cannot be trusted anymore!
Uncertainty U [mg]
Example of a Calibration Certificate
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
1
2
3
4
5
6
7
Load [g]
8
9
10
Max
Semi-micro
balance
Uncertainty U [mg or %]
Example of a Calibration Certificate
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
1
2
3
4
5
6
7
Load [g]
XP205
Max
8
9
The examples show the increase of the relative measurement
uncertainty upon decrease of load.
10
The Accuracy Limit = Minimum Weight
Relative Measurement Uncertainty [%]
Uncertainty U [mg or %]
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Required weighing accuracy [%]
0
1
2
3
Accuracy
limit:
Minimum sample weight
4
5
Load [g]
6
7
8
9
10
Max
When weighing below the minimum weight, the measurement uncertainty is larger than the accuracy required:  Inaccurate results
The Accuracy Limit = Minimum Weight
Relative Measurement Uncertainty [%]
Uncertainty U [gm or %]
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Required weighing accuracy [%]
0
1
2
3
Accuracy
limit:
Minimum sample weight
4
5
Load [g]
6
7
8
9
10
Max
Max
Weight [g]
When weighing above the minimum
weight, the measurement
uncertainty is smaller than the accuracy required:  Accurate results
Minimum Weight is a Function of the Accuracy
Relative Measurement Uncertainty [%]
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
1%
0.8
0.7
0.6
0.5
0.4
0.1%
0.3
0.2
0.1
0
0
1
2 for
Minimum
weight
an accuracy of 1%
Required weighing accuracy [%]
Required weighing accuracy [%]
3Minimum
4
5
6for
weight
7
8
9 [g] 10
Load
an accuracy of 0.1%
More stringent accuracy = Increased minimum sample weight
Minimum Weight varies over Time
Relative Measurement Uncertainty [%]
Uncertainty U [mg or %]
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Weighing accuracy [%]
0
1
2
3
4
5
Load [g]
6
7
8
9
10
Max
Safety
Max
Weight [g]
The minimum weight depends
on its environment,
the
operator, and varies over time => Apply a safety factor
22
Minimum Weight from Calibration Certificate
Variability of Minimum Weight – Safety Factor
Minimum weight
Smallest net weight
adjustment
(by service)
SAFETY
calibration period
calibration period
time
Calibration
at Installation
As-found calibration As-left calibration
USP <1251>: Variability of Minimum Weight
 "Factors that can influence repeatability while the balance is in use
include:
- The performance of the balance and thus the minimum weight can vary over
-
time because of changing environmental conditions
Different operators may weigh differently on the balance—i.e., the minimum
weight determined by different operators may be different
The standard deviation of a finite number of replicate weighings is only an
estimation of the true standard deviation, which is unknown
The determination of the minimum weight with a test weight may not be
completely representative for the weighing application
The tare vessel may also influence minimum weight […]."
 "For these reasons, when possible, weighings should be made at larger
values than the minimum weight."
USP General Chapter 1251 suggests the application of a "safety
factor".
Accuracy = Weighing above Minimum Weight
Most important measure for accurate weighing:
Know the minimum weight of your instruments and
always weigh above it, considering a safety factor!
Minimum weight
unknown
20 mg
Minimum weight
unknown
200 g
A Risk-based Approach of Testing
 USP General Chapter 1251: "The balance check is performed
at appropriate intervals based on applicable standard operating
procedures. The frequency of the balance check depends on
the risk of the application and the required weighing
tolerance."
FDA – Questions and Answers on cGMPs
http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/
Guidances/ucm124777.htm
Supplier's Question
“Many leading analytical balance manufacturers provide built-in "auto
calibration" features in their balances. Are such auto-calibration
procedures acceptable instead of external performance checks? If not,
then what should the schedule for calibration be?”
FDA Answer
 External performance checks still have to be carried out, but less
frequently.
 Risk analysis (criticality and tolerance of the process) and frequency
of use determine the frequency of performance checks.
 The calibration of an “auto-calibrator” should be periodically verified –
usual frequency is once per year - using […] traceable standards.
Uncertainty U [mg or %]
Log scale is better suited to depict uncertainties
2
1.9
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
U_rel
U_abs
29
1
2
3
4
Load6 [g]
5
7
8
9
10
Max
Balance Contributions to Uncertainty
Relative Expanded Weighing Uncertainty [%]
Balance XP204: Individual Uncertainty Contributions (@k=2)
1
Analytical
balance
0.1
U_tot
U_RP
U_EC
U_NL
U_SE
0.01
0.001
0.0001
0.00001
0.01
0.1
1
10
Sample Mass [g]
Repeatability dominates uncertainty
100
1000
Sensitivity & Eccentricity
dominate uncertainty
Test Weights for User Tests
Weight 1: Largest OIML or ASTM nomination which
equals or is smaller than the capacity of the
balance or scale
Weight 2: Largest OIML or ASTM nomination which
equals or is smaller than 5% of the capacity of the
balance or scale
Semi-micro balance Ind. bench scale
Capacity 60 kg
Capacity 220 g
200 g and 10 g
50 kg and 2 kg
Assessing Systematic Deviations with Weight 1
 Test sensitivity with a test weight close to nominal capacity (Weight 1).
 Using a very small test weight results in the sensitivity deviation to be
completely buried within repeatability.
Characteristic curve with
sensitivity deviation
Indication
Correct sensitivity
No accuracy test
at working point.
Not allowed
Stipulated
Load
< 5%
5%
Capacity
USP General Chapter 41: "A test weight is suitable if it has a mass
between 5% and 100% of the balance's capacity."
Assessing Repeatability with Weight 2
Indication
At the lower end of the
measurement range, the
repeatability is almost
constant.
Smaller weight:
Feasible, but
difficult to handle
(especially for micro
and analytical
balances)
Test weight up to a few percent of
the balance's capacity.
Load
« 5%
5%
USP <41>: "To facilitate handling, the test weight that is used for the repeatability test
does not need to be at the minimum weight value but can be larger because the
standard deviation of repeatability is only a weak function of the test weight value."
Repeatability Test Performed with a Robot
Working point
Test point
Maintain Accuracy with Performance Verification
To maintain continuous accuracy, it is important
to carry out regular performance verification.
However, perform only meaningful tests and
avoid unnecessary testing!
Doing the wrong tests?
 Quality Risk!
Doing too much testing?
 Waste of Money!
GWP® - The Weighing Standard
The science-based global
standard for efficient life cycle
management of weighing
instruments.
It applies a risk-based
approach that allows improving
control of the whole measuring
process, which in turn helps to
avoid costly OoS.
GWP® gives clear answers to
questions such as how to specify
or calibrate balances correctly.
It covers every relevant step of
the instrument's life cycle from
evaluation to routine operation.
Publications in Journals
 Pharmaceutical Engineering, November/December 2009
 Pharmaceutical Engineering, January/February 2012
 Pharmaceutical Formulation & Quality, February/March 2012
Publication in September 2013 Edition of MEASURE.
GWP® - The Weighing Standard
Why We Should Challenge the Established Way We
Calibrate and Test Weighing Instruments
NCSLI Nashville
July 2013
Dr. Klaus Fritsch
Manager Compliance