Download as pdf or txt
Download as pdf or txt
You are on page 1of 44

Weighing Accuracy

Estimating Measurement Bias & Uncertainty


Of A Weighing

© METTLER TOLEDO
Arthur Reichmuth V2.1 Sept. 2001
Abstract
In many applications, the weighing result needs to be
qualified, i.e., the measurement bias and uncertainty
accompanying the weighing process are required. These
values are usually not readily available, not least because
they are dependent on the application at hand. At other
times, the operator needs to know the minimal amount of
mass (a.k.a. ”minimal weight”) that can be weighed
conforming to a required relative uncertainty and confidence
level.
This paper discusses influences on the balance and the
weighing object, and explains how the weighing uncertainty
and minimal sample mass can be estimated. The assumptions
and restrictions, under which this deduction is valid, and
under which conditions influences may be neglected, are
explained. Two examples with actual data from analytical
balances are given as illustrations.
The theory and examples provided here enables the user to
asses the uncertainty and estimate figures of uncertainty of,
or minimal sample mass for, a weighing.

Contents
Introduction ........................................................... 1
Influences Originating From The Balance ................. 1
Influences Affecting The Weighing Object ................ 2
Influences by the Environment ................................. 3
Modelling The Weighing ........................................ 4
Characterization Of The Individual Influences .......... 6
Combined Weighing Bias ....................................... 17
Combined Weighing Standard Uncertainty.............. 18
Expanded Weighing Uncertainty ............................ 19
Example 1............................................................. 21
Uncertainty Charts ................................................. 25
Minimal Sample Mass ............................................ 31
Example 2............................................................. 33
Chart of Minimal Sample Mass ............................... 33
Conclusions ........................................................... 35
Appendices ........................................................... 37
References............................................................. 42

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—1—

Introduction
The primary aim of a weighing is to determine the mass of
the weighing object.
Measurement processes are generally subject to influences of
all kinds of origins. The weighing is no exception, as it may
get distorted by many influences which introduce bias and
uncertainty in its result. Such influences may affect the
weighing object, the balance, or both. The distortions may be
caused by properties of the weighing object, the balance, the
environment and the procedure, by which the weighing is
performed.
Influences originating from balance are manifold. They
include effects introduced by the balance’s quantized digital
display, the balance’s limited capability to repeat, its non-
linear characteristic, its sensitivity to eccentric loading, its
deviation of sensitivity from the correct value, as well as its
temperature dependence.
Among the influences that afflict the weighing object, the
most prominent is buoyancy, which is caused by the fact that
usually weighings are carried out in air (instead of an empty
space, the vacuum). Other influences are magnetic and
electrostatic forces.
Ambient conditions like air temperature, humidity, pressure
and air velocity will influence the weighing object and the
balance. Influences caused by vibration, inclination, or other
conditions may be present.

Influences Originating From The Balance


Sources of weighing deviations and uncertainties with
laboratory balances are:
Readability
Rounding of the weighing value to the last digit inherently
introduces quantization noise.
Repeatability
is introduced by…
•noise of the electronic circuits (especially by the A/D
converter’s reference)

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—2—

•thermal agitation caused by power dissipated by electronic


circuits
•wind draft at the site of the balance (especially with resolu-
tions of 1mg and below)
•vibrations of the support on which the balance stands
•temperature differences
•pressure fluctuations of the ambient air
Non-Linearity
is caused by…
•load dependent deformations of the weighing cell
•the electrodynamic transducer’s inherent non-linearity
between current and force
•a non-linear characteristic of the A/D conversion
Eccentric Load
is caused by…
•finite tolerances in manufacturing, assembly and adjustment
•deformations of the weighing cell due to the position of the
weighing object relative to the weighing pan
•load dependent deformations of the weighing cell.
Sensitivity Accuracy & Drift
is caused by…
•adjustment tolerance, or determination accuracy, of the
calibration weight
•buoyancy of the calibration weight
•spontaneous and temperature induced drift, mainly of the
lever's mechanical advantage, the electrodynamic
transducer’s magnetic flux, and the A/D converter's reference
(if the balance is not re-calibrated and adjusted).
•mechanical shocks
Zero Point Drift
is caused by…
•spontaneous or temperature induced drift of the weighing
cell, mainly of the weighing cell’s mechanical and electronic
components.
•mechanical shocks
•air buoyancy

Influences Affecting The Weighing Object


The most prominent influence that acts on the weighing object
is air buoyancy. If a weighing is carried out in a

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—3—

”natural”environment, i.e., in the atmosphere that surrounds


the earth, a buoyancy force acts on the weighing object. This
force is equal to the weight of the air displaced by the
weighing object, and is opposed to its weight force; that way
always diminishing it. The factor by which the weight force is
reduced is dependent on the densities of the weighing object
and the air.
Sometimes, the weighing object’s mass is not well defined in
the first place 1). Mass gain or loss may take place, for
example caused by hygroscopic exchange of water, or by
evaporation. Thus, we have to be aware that the measurand
of a weighing, the mass of an object, may not be the
unalterable physical quantity as it is often believed to be.
Varying or contradicting weighing values need not
necessarily be caused by the measuring instrument, the
balance, but are often resulting from the fact that the mass (or
its weight force) is not well defined and varying. We do
assume here that such spurious influences are recognized
before the weighing is carried out, or that the measurement
result is corrected for if the amount of influence is known.
The handling of the weighing object, i.e., its preparation, the
way how it is loaded onto the balance, may have an
influence on the weighing result. There may also be
influences brought about by the operator 2).
Last not least, it has to be defined whether the true mass, the
conventional (i.e., apparent) mass, or just the weighing value
shall be reported [R33]. All three values are generally
different for one and the same weighing object.

Influences by the Environment


Environmental conditions are critical for every weighing. The
air surrounding the weighing object and the balance has a
major influence on the weighing result. Air temperature,
humidity, pressure and draft (caused by ventilation of fans)
may deteriorate the weighing process. Heat radiation is also
critical. Direct sunlight, the rays of incandescent lights, or
other radiation sources should be prevented from reaching
the balance or the weighing objects. With high-resolution

1) see [Niels]
2) see [Clark]

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—4—

balances, even the radiation of the operator’s body may,


after 1/2 hour to one hour, disturb the weighing.

Modelling The Weighing


The measurand of a weighing process is the mass of the
weighing object, the result is the indication of the balance.
Ideally, the reading R should be identical to the mass m of
the weighing object. For this case, we can write
R=m .
This relationship largely holds, indeed. At a closer look,
however, the influences discussed do affect, although to a
small extent, the weighing.
Some of the influences—such as readability, repeatability,
non-linearity, zero drift—are additive influences, i.e., their
contribution simply adds to the reading. We can model them
mathematically as follows
R = m+yRD+yRP+yNL+yZ ,
Where yRD , yRP , yNL and yZ are the contributions of
readability, repeatability, non-linearity and zero drift.
Others—such as eccentric load, sensitivity offset, sensitivity
drift with time, sensitivity temperature drift and air
buoyancy—are multiplicative in their effect as they are
proportional to the sample mass and therefore can be
considered to influence the sensitivity of the transfer
characteristic. Sensitivity is the proportionality factor between
the mass of the weighing object and the displayed value.
A sensitivity drift of the balance can be caused by long-term
drift, or by temperature change. Internally, the balance is
equipped with a calibration factor c which relates the
weighing force FW acting upon the transducer to the
displayed value, according to
R = c FW .
On the surface of the earth, the weight force of an object
originates from the attraction force caused by gravity g ,
according to
FW = gm , where g ≈ 9.81 N/kg .
It is well known that the earth is covered by a layer of air, at
the bottom of which an object is usually weighed. Its weight

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—5—

force is therefore reduced by buoyancy, and hence the object


appears to have less mass 3). Modelling this effect, we have
a
R = cg 1– ρ m ,
where a and ρ are the densities of air, and the object
weighed, respectively, where
a ≈ 1 .. 1.2 kg/m3 .
Combining these influences, we are able to model the
weighing process with the following mathematical expression
a
R = cg 1– ρ m+yRD+yRP+yNL+yZ .
We did not consider eccentric load in this model, for two
reasons: first, the mathematical expression to estimate the
influence of eccentric load is a rather complicated one, and
second, the effect is easily suppressed by placing the
weighing object (i.e., its center of gravity) in the center of the
weighing pan.
Having thus determined the transfer function of a balance
from load to reading, we now direct our interest to the
process of weighing. A regular weighing consists of two
loadings (load instances) and readings, namely
1) one with the empty pan (or with a tare), which constitutes
the weighing of the tare mass, and
2) one with the object (or with the tare and the object), which
constitutes the weighing of the gross mass.
The mass of the object is obtained by subtracting the tare
reading from the gross reading, resulting in the (net) mass of
the object 4), namely
R = R2–R1 .

3) see ”Buoyancy” in the appendix


4) Even though one might argue that it it possible to weigh by simply
loading the object and reading once, this does but suppress the fact
that the state of no load (or tare mass) represents a load instant of the
balance, which serves as a reference value for the subsequent
reading(s) with the object placed on the pan.
The subtraction of the two readings can also be achieved, at the
operators discretion, by re-zeroing the balance as soon as the tare
mass is on the pan.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—6—

We obtain the tare reading R1 from the balance while either


no load, or the tare load, is on the on the weighing pan,
constituting the tare load mt 5)
a
R1 = cg1 1– ρ1 mt+yRD +yRP +yNL +yZ ,
t 1 1 1 1
with the parameters of the influences acting on the weighing
at the time of this tare loading and reading. Analogously,
placing the sample m in addition on the weighing pan, we
get for the gross reading
a a
R2 = c g2 1– ρ2 mt+g2 1– ρ2 m +yRD +yRP +yNL +yZ ,
t 2 2 2 2
again with the parameter values at the time of this gross
loading and reading. The sample mass is obtained by
subtracting the tare reading from the gross reading.
Before we execute the subtraction of these readings, we will
take a closer look at these expressions. We do assume that
the loadings are executed on the same balance in sequence,
within a short time interval between the two readings. From
this assumption we conclude that the same air density and
gravity apply to both loadings, namely
a1 = a2 = a and g1 = g2 = g .
Unfortunately, this equality does not apply for the readability,
repeatability and non-linearity contributions of the respective
readings; the contributions of non-linearity will be dealt with
separately later. We thus obtain for the net value of the
sample mass
a
R = cg 1– ρ m+yRD –yRD +yRP –yRP +yNL –yNL +yZ –yZ .
2 1 2 1 2 1 2 1
This expression allows to determine the effects on the
weighing by the influences considered. The next step is to
characterize the behavior of the individual contributing
influences.

Characterization Of The Individual Influences


What we are interested in is the question whether, and how
the individual influences affect accuracy and precision of the
weighing result. We thus have to ascertain whether they

5) Even with no load on the weighing pan, there is always weight acting
on the balance, constitued by the mass of the weighing pan, and other
dead loads, such as mechanical parts of the weighing cell. This
defines the ”no load” condition.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—7—

provide a bias, i.e., have an average value different from


zero, and what their contribution to uncertainty is.
Readability
concerns the term yRD –yRD .
2 1
To indicate a value on a display with a finite number of digits
it must be rounded. If the step with of the display is d , and if
the value is rounded halfway between two steps 6), this
operation adds quantization noise yRD to the value. The
average value of this noise is zero
δ RD = yRD = 0 ,
while its variance can be approximated by
2 = 1 d2 7).
sRD 12
The average contribution of the readability term yRD2–yRD1 ,
comprising two readings, is therefore zero. As the two
roundings may be considered uncorrelated, their uncertainty
variance evaluates to
2 = 1 d2 .
u2RD = 2 sRD 6
The quantization noise is a random contribution.
Repeatability
concerns the term yRP –yRP .
2 1
Repeatability is the capability of a balance to produce the
same reading, provided the same load is placed several
times in sequence on the platform.
The bias of this measurement series is the average of all net
readings R = R2–R1 , namely
n n
R= n 1 Σ R = 1 Σ R –R
i=1 i
ni = 1 2 1 i
and represents the best estimate for the sample mass that has
been repeatedly weighed. As a consequence, the contribution
of the repeatability term yRP –yRP is considered free of bias
2 1
δ RP = yRP = 0 .
Or stated in different words: its average ”deviation” is the
measurement value. The variance of this series, however, is

6) a.k.a. ”4/5-rounding”
7) This is valid for a ”4/5-rounding”, with a busy measurement value,
i.e., a weighing value that ”bounces” at least some steps around its
average value.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—8—

generally nonzero, and is obtained from the weighing series


as
n
2 = 1
sRP Σ
n–1 i = 1
Ri–R
2
.
Hence, repeatability does not introduce a bias, but it
contributes to uncertainty, namely
u2RP = sRP
2 .
Repeatability is a random contribution.
Sensitivity 8)
concerns the proportionality term between sample mass and
displayed value.
Sensitivity is the ratio of change in reading versus change of
load. The slope of a straight line through the readings at the
no-load and full load points of a balance may therefore be
interpreted as its (global 9) sensitivity.
S= m R .
To evaluate sensitivity deviations, the following decomposition
will prove useful
S = S0 1+∆S .
The sensitivity of a balance is accurate, if its reading
coincides with the value of the mass loaded. Therefore, the
following relationship must hold
R=m .
Provided the sensitivity deviation ∆S is zero, it follows that a
balance’s accurate value of sensitivity S0 is
S0 = 1 .
We thus have
S = 1+∆S .
The transfer function of a balance whose sensitivity has been
adjusted is
a
g 1– ρ
R= 1 m 10 ).
a
gcal 1– cal 1+∆µ cal
ρ cal
The parameters with the index ”cal”denote values at the time
of calibration.

8) Definition of term: see [VIM], item 5.10


9) Sensitivities measured over small(er) intervals are an issue of linearity.
10) see chapter ”Sensitivity Adjustment” in the appendix

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
—9—

Relating this expression with the definition of sensitivity, we


obtain for the sensitivity
a
g 1– ρ 1
S= a = 1+∆S ,
gcal 1– cal 1+∆µ cal
ρ cal
from which the sensitivity deviation can be expressed as
a
g 1– ρ 1
∆S = a –1 ≈
gcal 1– cal 1+∆ µ cal
ρ cal
g a a
≈ 1– ρ 1+ ρcal 1–∆µ cal –1 11)
gcal cal

From this expression we conclude that gravity and


buoyancy—both at the time of adjustment, as well as at the
time of weighing—and calibration weight accuracy, influence
sensitivity accuracy.
Gravity
The gravity g at the location of weighing does influence the
sensitivity of the balance. We do assume here, however, that
the sensitivity of the balance has been adjusted at the site of
use 12). Thus, we have
g = gcal ,
thereby eliminating the influence of local gravity.
Calibration Weight
The sensitivity proper, namely the sensitivity of the balance,
depends on the calibration weight deviation, as
∆S ∆µ ≈ 1–∆µ cal –1= –∆µ cal 13).
Its average deviation vanishes
δ∆µ = ∆µ cal = 0 ,

11) This approximation is based on the assumptions that g ≈ g ,


cal
acal << ρcal and ∆µ cal << 1 .
12) see chapter ”Sensitivity Adjustment” in the appendix
13) We do assume for this and the following derivations that g ≈ gcal ,
a << ρ , acal << ρcal , and ∆µ cal << 1 apply, meaning that the
nominal values of the terms not investigated in the formula are close to
1.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 10 —

while its variance can be estimated from the calibration


weight tolerance
∆µ lim ≤ max ∆µ cal
as
∆m
s∆µ 2 = 13 ∆µ 2lim with ∆µ lim= mlim .
rel 0
This influence is systematic in nature, although it will vary
from balance to balance.
Air Buoyancy
Buoyancy does not influence the sensitivity proper of a
balance 14), but it does influence the overall sensitivity of a
weighing. Both the calibration standard, as well as the
sample is affected by buoyancy.
A first order approximation of the influence of buoyancy on
the overall sensitivity is
a a a a
∆S buoy ≈ 1– ρ 1+ ρcal –1 ≈ ρcal – ρ .
cal cal
Sensitivity deviation thus depends on the densities of the
calibration standard ρ cal and the sample ρ , as well as the
air densities prevailing at the time of sensitivity adjustment
acal and at the time of weighing a .
The density of a calibration weight (internal or external) is
kept around the conventional density of 8000kg/m 3 15). If a
sample being weighed happens to be of the same density,
and the air density at the time of weighing is the same as it
was at the time of sensitivity adjustment, the buoyancy
influence cancels as can be concluded from the expression of
sensitivity deviation 16). However, this assumption does hold
only in few special applications, the dissemination of mass
standards being the most important of them. For other
applications, where the weighing object has a density
different from the conventional one, buoyancy does influence
the weighing result.

14) With the exception, when weighing the calibration standard for the
purpose of adjusting the balance’s sensitivity.
15) see [R33]
16) This is not to say that buoyancy has ceased to occur. It means, that the
effect on the weighing result is greatly reduced and can be neglectec
but in a few cases.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 11 —

If the conventional value of mass of the sample is to be


expressed, then the influence of buoyancy is to be
forgone 17). However, if the true, physical mass of the sample
is to be determined, buoyancy must be corrected for.
According to this expression, the average value of sensitivity
deviation due to buoyancy is
a a
δbuoy = ∆S buoy = ρcal – ρ .
rel cal
The variance resulting from buoyancy at the time of weighing
the reference standard can be derived by taking the partial
derivatives of the sensitivity expression. This yields
acal 2 2 a 2 1 2
sbuoy = 2 sρ + ρ1 sa + 2 sρ + ρ
2
sa 18 ).
rel ρ cal cal cal cal ρ
While the density variance of the calibration standard may
be obtained from the balance manufacturer, the user must
provide the density information of the object weighed.
A completely different issue is the determination of the density
variance of air. Air density depends largely on air pressure
and temperature
p 19
a= m ).
nk T
If no specific air density data is available, one could estimate
it from pressure and temperature. As a first order
approximation, the air density depends on air pressure
∂a a
∂p = p
and on air temperature
∂a
=– a .
∂T T
From this, the variance of air density can be determined to be

17) The conventional value of mass is understood to be the value obtained


from a weighing in air of density 1.2kg/m3 on a balance, whose
sensitivity was calibrated with a mass standard of 8000kg/m3. Hence,
the exact conventional value is only obtained if the air density at the
time of weighing the sample is 1.2kg/m3.
Details are given in [R33].
18) see cahpter ”Variance Of Relative Buoyancy” in the appendix
19) To a lesser extent, air density is also dependent on air humidity, its
compressibility and its composition. See ”Dependence Of Air Density
On Temparature & Pressure” in the appendix.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 12 —

a 2 a 2 sp 2 s 2
sa2 = p s p + – sT = a2 p + T 20 ).
T T
The standard uncertainty of air density is then
sp 2 s 2
sa = a T .
p + T
Buoyancy introduces both a bias and an uncertainty, mainly
stemming from the limited knowledge of the density of the
sample and the air density at the time of weighing. Its
influence is systematic as far as the average buoyancy goes,
and is random to the extent, as the true sample and air
densities are unknown 21).
Sensitivity Temperature Drift
Temperature may also influence sensitivity, as the weighing
transducer of the balance may exhibit a temperature
dependent characteristic. This influence is expressed by the
(sensitivity) temperature coefficient (TC) of the balance and is
defined as the ratio of relative change of sensitivity ∆S/S0
and the temperature difference ∆T causing it, namely
∆S
S
TCS = 0 .
∆T
The TC of sensitivity only states how sensitive the slope of the
balance’s characteristic curve reacts to temperature. To
provoke a slope change, a change of ambient temperature,
with respect to the temperature at the last sensitivity
adjustment, must occur. The sensitivity change can be
expressed as
∆S TC = TCS∆T S0 .
The influence of ambient temperature on sensitivity is of
systematic nature. However, because the actual TC of a
balance are usually unknown to the operator, this influence is
treated as a random contribution 22).

20) see chapter ”Variance Of Relative Buoyancy” in the appendix


21) Unless special means are provided, such as a controlled atmoshperes
in an enclosed volume
22) uncertainty of systematic influence, type B uncertainty, according to
[GUM], chapter 4.3

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 13 —

If at least some information about the TC and the behavior of


ambient temperature is available, this can be used to estimate
their influence. Typically, the operator may get some
tolerance band of the TC from the data sheet,
TCS ≤ max TCS 23 ) ,
lim
and from experience or observation, a maximum excursion
of the ambient temperature may be known
∆T lim ≤ max ∆T .
As a tolerance specification, the average value of the
sensitivity TC of a set of balances (not of a single one) is zero,
i.e.,
TCS = 0 .
The average value of sensitivity drift due to temperature of
this set of balances is therefore zero as well
δS = ∆S TC = 0 .
TC rel
Its variance, however, is not zero. As an approximation it
can be obtained as the product of the variances of TC and
temperature excursion
sS 2 = sTCS
2 sT2 .
TC rel
To obtain the variance of the TC, we assume—for lack of
other information—that the TCs of a set of balances be
uniformly distributed within the interval of the specified limits.
We thus get for the variance
2
sTCS = 13 TCS 2 24).
lim
The variance of the ambient temperature can be obtained
likewise from its maximum excursion, with the same
reasoning. This yields
sT2 = 13 ∆T 2lim .
Sensitivity Long-Term Drift
Except for the influences modeled mathematically, there is the
balance’s long-term drift of sensitivity that may occur. This
means that the sensitivity over time may change
spontaneously, if only slightly. This property is sometimes
specified in the data sheet as a tolerance value
∆S LTD ≤ max ∆S LTD .
lim

23) the subscript ”lim” refers to the specified limit value


24) see [GUM], chapter 4.3.7

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 14 —

As with other influences before, the average value of long-


term drift can be assumed to vanish
δS = ∆S LTD = 0 ,
LTD rel
while its variance may be estimated by
2
sS 2 = 13 ∆S LTD ,
LTD rel lim
assuming a uniform probability distribution.
Sensitivity Drift: Summary
When the sensitivity of a balance is adjusted with an internal
or external calibration standard, the long-term drift and
temperature drift of sensitivity are eliminated. The influence of
long-term drift starts anew, and any temperature change with
respect to the temperature at the time of adjustment produces
a temperature drift again.
We assume here that the balance is adjusted occasionally.
Some balance models provide sensitivity adjustment means,
by which the balance automatically adjusts sensitivity at
proper time intervals, or when the ambient conditions
change.
We therefore drop the influence of long-term drift here; only
if a balance does not get adjusted anymore, it should be
considered.
On the other hand, we will still consider temperature drift, for
two reasons: First, temperature changes may occur within
adjustment intervals, for example daily fluctuations, and
second, even with automatic adjustment features, there has to
occur a certain temperature change to trigger an automatic
adjustment.
The total sensitivity bias thus amounts to
a a
δS = δbuoy = ρcal – ρ ,
rel rel cal
while the total sensitivity variance is equal to
uS 2 = s∆µ 2 + sS 2 + sbuoy 2 .
rel rel TC rel rel
Sensitivity is the proportionality factor between load and
reading. Therefore, sensitivity deviation must be multiplied by
the sample mass to obtain absolute bias
δS = m δS .
rel

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 15 —

Because the variance is a quadratic measure, it must be


multiplied with the square of the sample mass to obtain
absolute variance
sS2 = m2 sS 2 .
rel
Non-Linearity
concerns the term yNL –yNL .
2 1
The relationship between the input (the load) and the output
(the reading) of a balance is called its characteristic curve. In
an ideal case, this curve is a straight line, as defined in the
chapter about sensitivity. Non-linearity is the deviation of the
characteristic curve from this straight line going through the
no-load and full-load (maximum capacity) point. This
deviation is systematic in nature for a given balance unit,
although generally different from balance to balance.
However, linearity deviation as a function of load is generally
unknown to the user; in the best case a maximum deviation
may be known from the data sheet.
NLlim ≤ max NL .
The actual linearity deviations of the tare and the gross
readings for large sample loads may be considered random
values that are weakly correlated, if at all. Without additional
information, non-linearity deviations are considered to be
evenly distributed on both sides of the ideal straight line.
Thus, the bias introduced by linearity deviation yNL vanishes
δ NL = yNL ≈ 0 .
Without further information about the actual shape of the
linearity deviation, we assume a uniform distribution
(generally an overly pessimistic assumption 25) within the
tolerance interval specified in the data sheet. From this we
can estimate the non-linearity variance to be
2 = 1 NL2 .
sNL 3 lim
valid for one reading. For the net weighing value, obtained
by taking the difference of two readings, we therefore have to
take twice this variance
2 = 2 NL2 .
u2NL = 2sNL 3 lim

25) A triangularly shaped distibution might be appropriate in some cases.


See [Rth1] und [Rth2].

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 16 —

For small sample loads, the correlation between the two


linearity deviations increases. It can be shown that the
difference of the two non-linearity deviations, picked up with
the two readings required for a sample weighing, decreases
with diminishing sample mass, at least in a statistical sense.
We will make use of this behavior when discussing the issue
of minimal sample mass.
Zero Drift & Zero Temperature Drift
concerns the term yZ –yZ .
2 1
The majority of weighings is probably carried out according
to one of the following procedures:
i) the weighing takes place within a short interval of time,
without re-zeroing the balance between the two readings. If
this is the case, virtually no zero drift will occur, signifying
that the two contributions of zero are about equal, namely
yNL ≈ yZ ;
1 2
ii) the two readings take place at different times, yet the
balance is re-zeroed before the net reading, as well as
before the gross reading. In this case, there might have
occurred a zero drift; however, because the balance was re-
zeroed, the two contributions of zero vanish, namely
yNL ≈ yZ ≈ 0 .
1 2
In either case, the difference
yZ –yZ = 0
2 1
of the two contributions will be zero.
Die Varianz der Differenz der Nullpunktdrift verschwindet
ebenfalls, da die Differenzen indentisch Null sind
sZ 2= 0 .
It is a different issue, if for example a container is weighed,
and after a long time—hours or days—say, this container
with the sample added is weighed again, and from the two
readings the net value is determined and assigned to the
sample mass. Here, there might have occurred a zero drift,
and the two zero contributions may not longer be equal
yNL ≠ yZ ,
1 2
thereby introducing a difference into the net value that stems
from zero drift
δ Z = yNL –yZ ≠ 0 .2
1

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 17 —

The same is true when the measurement value needs to be


monitored continuously, as in sedimentation, evaporation or
other applications, where the sample mass versus time is of
interest.
These considerations apply independently of the source of the
zero drift.

Combined Weighing Bias


The combined bias picked up in a sample weighing is
obtained by arithmetically adding up all individual biases
discussed in the previous chapters. We thus have for the
combined bias
δ = δ RD+δ RP+δ NL+δ Z+δS ,
where we have used the indices already introduced.
Contributions from readability, repeatability and non-linearity
and zero point to bias are all zero. Whereas bias, resulting
from the sensitivity proper of the balance, is also zero (both
initial and temperature induced), apparent sensitivity bias is
caused by buoyancy, since
a a
δS = δbuoy = m ρcal – ρ .
cal
We thus obtain for the combined bias of a sample weighing
a a
δ = δS = m ρcal – ρ ,
cal
stemming exclusively from air buoyancy.
The reason the balance does not provide bias is owed to the
fact that systematic deviations of the balance’s transfer
characteristic—provided, they are of systematic origin and
invariable—are eliminated either through adjustment after
assembly, or measured and stored in the balance for on-line
compensation. These include:
• non-linearity correction
• correction of temperature influence (with on-line measure-
ment of the temperature)
• correction of the calibration weight’s adjustment deviation
• on-site adjustments for sensitivity and sometimes non-line-
arity
The remaining deviations after adjustment or compensation…
• are too small by definition to be compensated (had they
been large enough, they would have been compensated);

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 18 —

• are time dependent in an unknown manner (unknown


systematic deviations);
• are caused by unknown ambient conditions (such as tem-
perature or humidity);
• or are of entirely unknown origin—neither their source or
amount, nor their course over time are known—and
therefore are by definition not identifiable as systematic
deviations.
Hence, these influences, although of systematic origin, must
be regarded as uncertainties and are included in the
uncertainty estimate.
Only air buoyancy introduces a significant bias into the
weighing result.

Combined Weighing Standard Uncertainty


To obtain the combined variance of a sample weighing, all
individual variances earlier discussed are added. This
provides a reasonable estimate of the combined variance if
the individual contributions are mutually independent, or at
least statistically uncorrelated 26). This seems to be a
reasonable assumption, since the individual causes for the
balance’s deviation from its ideal performance may be
considered independent from each other, indeed. We thus
have for the combined variance
u2 = u2RD+u2RP+u2NL+uZ2 +uS2 .
The variance caused by rounding (readability) is
2 = 1 d2 ,
u2RD = 2 sRD 6
the variance caused by repeatability
u2RP = sRP
2 ,
and the variance caused by non-linearity
2 = 2 NL2 .
u2NL = sNL 3 lim
The variance due to the sensitivity proper of the balance is
s∆µ 2 = 13 ∆µ 2lim ,
rel
and the variance due to sensitivity drift caused by
temperature is
2
sS,TC 2
= sTCS 2 m2 ,
s∆t
with

26) see [GUM], chapter 5.1

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 19 —

2
sTCS = 13 TCS 2 and sT2 = 13 ∆T 2lim ,
lim
while the variance due to apparent sensitivity, caused by
buoyancy, is
acal 2 2 a 2 1 2
sbuoy = 2 sρ + ρ1 sa + 2 sρ + ρ
2
sa .
rel ρ cal cal cal cal ρ
Eventually the combined variance equals
u2 = u2RD+u2RP+u2NL+m2 s∆µ 2 +sS,TC
2 + sbuoy 2 .
rel rel
Taking the square root of the uncertainty variance yields the
standard uncertainty
u = u2RD+u2RP+u2NL+m2 s∆µ 2 +sS,TC 2 + sbuoy 2 .
rel rel

Expanded Weighing Uncertainty


It is common practice to express the result of a measurement
as
Y = y±U ,
where y is the best estimate of the measurand, and U the
expanded uncertainty. The expanded uncertainty is obtained
by multiplying the combined standard uncertainty by a
coverage (or expansion) factor k , such that
U = k⋅u .
The expanded uncertainty ”±U is an interval that may be
considered to encompass a large fraction of the distribution
of values that could reasonably be attributed to Y ” 27).
The coverage probability or level of confidence P is the
measure of how large that fraction is. The coverage factor
and the coverage probability stay in relation to each other.
This relation is mediated by the probability distribution of the
measured value. Unless there is a clear indication for a
different distribution, a normal distribution is adequate.
This may be assumed for the weighing result. After all, some
of the uncertainty contributions are themselves normally distr-
ibuted, e.g. the one of repeatability. Moreover, as there are
multiple sources of uncertainty contributions, their combined
distribution tends to a normal distribution.
The normal probability distribution is therefore used to define
the relationship
27) [GUM], chapter 6.2.1

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 20 —

k=kP
between coverage factor k and coverage probability) P 28).
Sometimes, the missing probability is used, according to
Q = 1–P .
This number states the fraction of those values which are
expected to lie outside the uncertainty interval. Frequently
used coverage factors are 2 and 3, which corresponds to a
coverage probability of about 95% and 99.7%, respectively.
Thus, between a sample of mass m and the result R (single
observation) of its weighing, the following relationship holds
R –U ≤ m ≤ R +U ,
where the expanded uncertainty
U = k⋅u 29),
is obtained by multiplying the standard uncertainty u with
the coverage factor k , the latter being an appropriate choice
(2 or 3, for example), from which the coverage probability
k —> P = f k
follows, or vice versa, the coverage probability is
appropriately chosen, and from it the coverage factor is
determined
P —> k = f P ,
according to the table given.
Eventually, the weighing result can be expressed as
28) Coverage Factor Level of Confidence Missing Probability
(Single Sided) (Coverage Probability)
k P Q = 1–P
——————— ———————–– ———————
1 68.27% 31.73%
1.645 90% 10%
1.960 95% 5%
2 95.45% 4.55%
2.576 99% 1%
3 99.73% 0.27%
4 99.994% 0.006%
5 99.99994% 0.00006%
(A normal probability distribution is assumed for this table.)
29) To keep things simple, we have consequently refrained from deter-
mining, or correcting for, the degree of freedom. Of course, nothing
stands against the notion of correcting for the degree of freedom, if it
is known of all individual contributions. An instruction for how to de-
termine the correction factor can be found in [GUM].

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 21 —

m = R –δ ±U .
We now have all pieces in place to discuss the weighing
result of an example.

Example 1: Determination Of Weighing Bias And


Uncertainty
The mass of approximately 1g of aluminum is weighed in a
190g container on a 200g semi-micro balance 30). What are
the resulting bias and uncertainty of the weighing result
(considering a coverage factor of 2)? The temperature in the
laboratory is controlled to ±2°C.
Additional information:
Density of aluminum (standard dev.) ≈ 2700kg/m3 (≈1%)
Density of surrounding air ≈ 1.2kg/m3
Laboratory Temperature (22±2)°C
Bias
First, we determine the bias of the weighing result. We have
a a
δ = m ρcal – ρ .
cal
The density of the sample and the calibration standard are
given with
ρ = 2700 kg/m3 and
30) AT201 semi-micro balance:

Excerpt from data sheet specifications:


Property Specification
Readability 0.01mg
Repeatability up to 50g 0.015mg
50-200g 0.04mg
Non-Linearity within 10g 0.03mg
within 200g 0.12mg
Temperature Coefficient 1.5ppm/K
Sensitivity Long Term Drift 1.5ppm

Additional information about this balance


Property Specification
Calibration Weight Tolerance *) 1.5ppm
Calibration Weight Density 8006kg/m3
Std. Dev. of Cal. Weight Density 10kg/m3
_____________________
*) adjusted to conventional value

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 22 —

ρ cal = 8006 kg/m3 ,


respectively, while we apply a density of air at normal
conditions and assume it to be the same at the time of
calibration, as well as at the time of weighing
acal = a ≈ 1.2 kg/m3 .
The sample weighs
m = 1g .
Hence, we get for the bias of the weighing result
1.2 kg/m3 1.2 kg/m3
δ = 1g – =
8006 kg/m3 2700 kg/m3
= 1g – 295ppm ≈ – 0.30mg
While the result R obtained from the balance represents
approximately the conventional mass mcon of the sample 31),
its bias corrected result R–δ renders the physical mass m .
In this example, this means that sample actually holds 30mg
more mass than what is reported from the balance. This loss
is resulting from buoyancy.
Uncertainty
Second, we determine the uncertainty of the weighing result.
We have for the standard uncertainty
u = u2RD+u2RP+u2NL+m2 s∆µ 2 +sS,TC 2 + sbuoy 2 .
rel rel
The individual contributions are:
Readability
d = 0.01mg
u2RD = 16 0.01mg 2 = 17×10–12g 2
Repeatability
As the repeatability specification at 190g is unavailable, we
use the 200g specification instead:
sRP = 0.04mg
200g
2 = 0.04mg 2 = 1.6×10–9g 2
u2RP = sRP
Non-Linearity
The sample mass is smaller than the interval of the more
favorable non-linearity specification (10g). We therefore may
use it.
31) exactly, if the air density is 1.2kg/m3

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 23 —

NLlim = 0.03mg
10g
2 = 2s2 = 2 NL2 = 2 0.03mg 2 = 600×10–12g 2
uNL NL 3 lim 3
Calibration weight tolerance
From the specification, we have
∆µ cal = 1.5 ppm
s∆µ 2 = 13 ∆µ 2lim = 13 1.5 ppm 2= 0.75ppm2
rel
Temperature drift
From the balance specifications we know the TC
TCS = 1.5 ppm/K
lim
sTCS = 13
2 TCS 2 = 13 1.5 ppm/K 2 = 0.75 ppm/K 2
lim
and from the environmental conditions the temperature
excursion we have
∆T lim = 2K sT2 = 13 ∆T 2lim = 13 2K 2 = 1.33K 2 .
Hence, the variance of sensitivity temperature drift amounts to
2
sS,TC 2
= sTCS sT2 = 0.75 ppm/K 2⋅1.33K 2= 1ppm2 .
Buoyancy
From the environmental conditions, we have the laboratory
temperature
22°C → T = 295K ,
and we know already the laboratory temperature variance
sT2 = 1.33K 2 .
Considering air pressure, no information is available. We
consider a standard pressure of
p = 1bar .
Assuming that the laboratory is essentially in free pressure
exchange with the outside atmosphere, we could try and find
some data about the latter. From weather data we can
estimate a pressure standard deviation of atmospheric
pressure fluctuations to be in the order of
s p = 10mbar .
We thus get for the air density standard deviation
sp 2 s 2
10mbar 2 1.33K 2
sa = a p + T
T = 1.2 kg/m3 + ≈
1bar 295K 2
≈ 1.2 kg/m3 10–4+1.5×10–5 ≈ 12.9g/m3 .

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 24 —

(In this example the variance introduced by pressure


fluctuation dominates the variance introduced by temperature
change.)
The density of the sample is given above, and its standard
deviation is
sρ = 1%⋅2700 kg/m3 = 27 kg/m3 .
The density of the calibration standard is also given above,
and its standard deviation is specified as
sρ = 10 kg/m3 .
cal

From these givens we calculate the relative variance of


buoyancy
acal 2 2 a 2 1 2
sbuoy = 2 sρ + ρ1 sa + 2 sρ + ρ
2
sa =
rel ρ cal cal cal cal ρ
2
1.2 kg/m3 12.9g/m3 2 1.2 kg/m3 2
12.9g/m3 2
= 10 kg/m 3
+ + 27 kg/m 3
+ ≈
8006 kg/m3 2cal 8006 kg/m3 2700 kg/m3 2 2700 kg/m3
≈ 35×10–15+2.6×10–12+20×10–12+23×10–12 ≈ 46×10–12= 46ppm2 .
The corresponding standard deviation amounts to
sbuoy ≈ 6.8ppm .
rel
The sample amounts to
m ≈ 1g .
Adding all variances yields the combined variance
u2 = u2RD+u2RP+u2NL+m2 s∆2µ +sS,TC2 + sbuoy 2 =
rel rel
= 17×10–12g 2+1.6×10–9g 2+600×10–12g 2+ 1g 2 0.75ppm2+1ppm2+46×10–12 ≈
–9 2 2 –9 2 –12 2
≈ 2.2×10 g +1g 46ppm ≈ 2.2×10 g +24×10 g ≈ .
2

≈ 2.2×10–9g 2 ,
from which we obtain the standard uncertainty
u ≈ 47µg .
With the required coverage factor
k=2 ,
corresponding to a coverage probability of
P ≈ 95% ,
we get for the expanded uncertainty
U ≈ 0.094mg .
Conclusion Of Example 1
The major contributors to weighing uncertainty in this
example are repeatability and non-linearity, in that order.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 25 —

Negligible effects, in their order of increasing irrelevance, are


readability, and all the influences which are proportional to
the sample mass, namely air buoyancy, temperature drift and
calibration weight tolerance. Air buoyancy per se is
dominated by the uncertainties of sample density and air
density. The expanded weighing uncertainty of the 1g sample
(weighed in a 190g container) thus amounts to 0.094mg.
Besides, the mass obtained from the weighing is too small by
0.3mg because of buoyancy; the actual mass of the sample is
larger by that amount.
We thus we obtain for the true mass of the sample
m = R–δ ±U = R+0.30mg ±0.094mg .
If we were to forego buoyancy correction, the weighing result
reflects the apparent mass of the sample
mapp = R±U = R±0.094mg ,
which approximately corresponds to the conventional mass.
In this example, measurement bias exceeds measurement
uncertainty by about a factor of three 32).

Uncertainty Charts
The example discussed just pinpoints one weighing situation,
given by circumstances of the particular weighing object and
container, as delineated in example 1. Instead, we are
interested in the uncertainties for all possible combinations of
sample masses and gross weights applicable to a given
balance.
Uncertainty Chart Of A Semi-Micro Balance
To chart the weighing uncertainty obtained with a balance,
we use the formula of standard uncertainty, derived in the
previous chapter, and multiply it with the coverage factor.
This yields for the expanded uncertainty
U = k u2RD+u2RP+u2NL+m2 s∆µ 2 +sS,TC 2 .
rel
To prevent the cluttering of the chart with too many
parameters, in the above expression the contribution of
buoyancy was dropped. However, the influence of buoyancy
is individually shown in the chart for several sample densities.
Ubuoy = kubuoy = km ubuoy =
rel
32) siehe [Kehl]

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 26 —

acal 2
1 s 2+ a s 2+ 1 s 2
= km s
2 ρ cal + ρ cal acal ρ a
ρ cal ρ2 ρ
As instrument we will use here the AT201 semi-micro balance
introduced in the previous example, hence we will continue to
employ the same specifications.
We have to be aware that sometimes the data sheet may state several
values for a single property, as these may depend on the
circumstances of the weighing. This is the case for the AT201, since its
repeatability and linearity depend on the gross weight (tare plus
sample) and the sample weight, respectively:
Repeatability up to 50g 0.015mg
50-200g 0.04mg
Non-Linearity within 10g 0.03mg
within 200g 0.12mg
This constitutes four operational cases:
I) gross below 50g & sample less than 10g
II) gross 50…200g & sample less than 10g
III) gross below 50 & sample 10…50g
IV) gross 50…200g & sample 10…200g
In cases I) and II), the more favorable non-linearity specification
applies, while in cases I) and III) the more favorable repeatability
specification applies. As it turns out, the numerical values for case III)
are almost identical with those of case IV); case III) is therefore omitted
in the chart.
All other parameters and assumptions in example 1, such as
specifications of the balance and environmental conditions
and assumptions, are used unaltered for this chart (except
sample mass, which is the independent variable, and
container mass, which is a parameter, of the chart).
Relative Uncertainty Chart Of A Semi-Micro Balance
Often it is more convenient, or it may even be required, to
use the relative uncertainty, i.e., the uncertainty normalized to
the sample mass
u
urel = m .
This yields
1
urel = m u2RD+u2RP+u2NL+m2 s∆µ 2 +sS,TC 2 + sbuoy 2 =
rel rel
= 1 u2 +u2 +u2 + s 2 +s2 + s 2
m2 RD RP NL ∆ µ rel S,TC buoy rel

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 27 —

Diagram 1: Absolute weighing uncertainty on an AT201 semi-micro balance


as a function of sample mass.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 28 —

and
Urel = kurel = k 1 u2 +u2 +u2 + s 2 +s2 + s 2
m2 RD RP NL ∆ µ rel S,TC buoy rel
Again, in this chart we dropped the influence of buoyancy.
Urel = k 1 u2 +u2 +u2 + s 2 +s2
2m RD RP NL ∆µ S,TC
rel
to represent buoyancy separately in the chart
Ubuoy = k ubuoy = k sbuoy .
rel rel rel
These uncertainty charts of the AT201 illustrate that there are
three characteristic behaviors of the uncertainty functions that
are related to three ranges with respect to the sample mass.
First, for small sample masses, below 10g, uncertainty stays
constant. This is attributable to the contributions which
dominate uncertainty and are independent of sample mass.
These are mainly repeatability, non-linearity and readability,
usually in that order of importance. This means that the
weighing uncertainty stays essentially constant and is not a
function of the sample weight. Therefore, as the sample mass
decreases, the relative uncertainty increases in this range.
Second, for large sample masses, larger than half the
maximal weighing capacity, the contributions that are
proportional to sample mass dominate uncertainty. These are
temperature drift 33), initial accuracy (calibration weight
tolerance), long-term drift and buoyancy. Therefore, the
relative uncertainty stays essentially constant in this range. If
the sample density is low, buoyancy may become by far the
largest contributor to uncertainty in this region.
Third, between the two ranges is a transition range where the
constant contributions as well as the ones proportional to
sample mass are of the same order of magnitude.
Although there may be specific differences between balance
models, this behavior is typical for most laboratory balances.
For ”small”samples—i.e., sample masses below 1/10 to
1/20 of the weighing capacity—the (absolute) uncertainty
stays constant, while the relative decreases with increasing
sample mass). For ”large”samples—i.e., sample masses of
1/4 to 1/2 of the weighing capacity and larger—the

33) Unless the balance provides an means to automatically re-adjust its


sensitivity, or there is no temperature change.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 29 —

Diagram 2: Relative weighing uncertainty on an AT201 semi-micro balance


as a function of sample mass.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 30 —

(absolute) uncertainty increases proportionally with sample


mass, while the relative uncertainty remains constant.
Having said this, it becomes clear that for small samples it is
usually unnecessary to evaluate the contributions that are
proportional to the sample mass. They may be neglected in
favor of the contributions that are constant. In the original
expression of the standard uncertainty
u = u2RD+u2RP+u2NL+m2 s∆µ 2 +sS,TC 2 + sbuoy 2
rel rel
we may therefore drop the term that gets multiplied with the
sample mass, and we obtain a much simpler expression for
the standard uncertainty, namely
NL (valid for
u = uRD 2 +u2 +u2
RP
m << mmax ),
which is valid for small samples. From this we get for the
relative standard uncertainty
u = 1 u2 +u2 +u2
NL (valid for
urel = m m RD RP
m << mmax ).
On the other hand, with large samples, the constant terms
may be dropped in favor of the one that gets multiplied with
the sample mass. We then have
u = m2 s∆µ 2 +sS,TC
2 + sbuoy 2 =
rel rel
2
=m s∆µ 2
+sS,TC+ sbuoy 2 (valid for
rel rel
m ≈ mmax ).
In this case, the relative standard uncertainty becomes
urel = mu = s∆µ 2 +sS,TC
2 + sbuoy 2 (valid for m ≈ mmax ).
rel rel
Moreover, the diagrams illustrate that the uncertainty which is
caused by air buoyancy—especially with large sample
masses—can easily exceed all other uncertainty
contributions. A perfectly adjusted balance with a high
resolution is no guarantee for a small weighing uncertainty,
especially if the sample has a low mass density. Measuring
the air density (usually inferred from air temperature,
humidity and pressure) and using it for the compensation of
air buoyancy, or more sophisticated, a pressure and
temperature controlled atmosphere, reduces this uncertainty.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 31 —

Minimal Sample Mass


Often, a user is interested in the amount of mass that can be
weighed on a balance under the condition that the relative
uncertainty meets a required level. For example, the US
Pharmacopeia (USP) requires that weighings have to be
performed with a relative uncertainty of smaller than 0.1%,
observing an coverage factor of three 34). The smallest
possible amount of mass that meets such a requirement is
known as ”minimal sample mass”, ”minimal sample weight”,
or simply ”minimal weight”.
The minimal sample mass can be determined from the
formula of relative uncertainty. Instead of determining the
relative uncertainty, we start with the required expanded
relative uncertainty Urel and coverage factor k 35). From
these givens we determine the standard uncertainty to be met
U
urel < rel .
k
We insert this expression into the formula of the relative
standard uncertainty—and we use the approximation
formula, as the minimal sample mass clearly is a small mass.
1 u2 +u2 +u2 .
urel = m RD RP NL
This yields
Urel 1
> m u2RD+u2RP+u2NL .
k
Finally, we solve this equation for the sample mass. We get
for the minimal sample mass (minimal weight)
mmin > k u2RD+u2RP+u2NL .
Urel

34) Excerpt from the US Pharmacopeia,


USP24-NF19, <41> Weights And Balances:
«Unless otherwise specified, when substances are to be ”accurately
weighed” for Assay the weighing is to be performed with a weighing
device whose measurement uncertainty (random plus systematic error)
does not exceed 0.1% of the reading. Measurement uncertainty is
satisfactory if three times the standard deviation of not less than ten
replicate weighings divided by the amount weighed, does not exceed
0.001.»
35) If instead of the coverage factor the coverage probability is given, the
coverage factor can be obtained by aid of the table given.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 32 —

Considering non-linearity, we have here the situation, that the


sample mass m is, by definition, rather small. In a previous
chapter we have pointed out that for small sample masses the
linearity deviations picked up by the tare and the gross
reading are highly correlated. It is fair to expect that, with
decreasing sample mass, the difference of these two
deviations decreases, too, at least in a statistical sense. This
means that for small sample weight, the difference tends
towards zero
yNL –yNL → 0 ,
2 1
or that it may at least be neglected with respect to other
contributions. In fact, linearity measurements with laboratory
balances support this assumption 36). It can be concluded,
then, that the uncertainty contribution of non-linearity is
smaller than the remaining contributions, especially the one
of repeatability
RP .
2 << s2
sNL
We thus get as the following simplified expression for the
minimal mass
mmin > k u2RD+u2RP (valid for m << mmax ).
Urel
If we have the situation, as is mostly the case with high-
resolving laboratory balances, that the repeatability is at least
equal to, but usually larger than, the readability
sRP ≥ d ,
then the variance of repeatability is at least 6 times larger
than the variance of readability
2 ≥ d2 = 6s2
sRP .
RD
One might thus even neglect the contribution of
readability 37). This leads to an even simpler expression,

36) see [Rth1] and [Rth2]


37) It might be argued why the contribution of readability was considered
in the first place.
If repeatability is evaluated from the readings of a measurement
series, then clearly, because the indicated values were displayed by
the balance, they were rounded, and hence, repeatability already
includes the effect of readability. From this point of view, the variance
of readability should not be included.
On the other hand, it sometimes happens with balances that have not
that high a resolution, such a precision balance, for example, when
evaluation repeatability, all values of a measurement series have the

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 33 —

whereby the minimal sample mass depends only on one


property of the balance, namely its repeatability
mmin > k 2 = k s
sRP (valid for sRP ≥ d ).
Urel Urel RP

Example 2: Determination Of The Minimal Sample Mass


A substance is weighed in a container. What is the required
minimal amount of this substance, if a relative weighing
uncertainty of 0.1% with an expansion factor of 2 must be
observed? The weighing is performed on a 200g semi-micro
balance 38). The container weighs 190g.
To solve this task, we use the simplified expression for the
minimal sample mass. The coverage factor k = 2 and the
required relative uncertainty Urel are given. Beyond that we
need the readability of the balance, which is d = 0.01mg .
As the container weighs 190g, we have to use the
repeatability specification valid for 200g, namely
sRP = 0.04mg .
200g
Repeatability is clearly larger than readability; we therefore
resort to the simpler expression of minimal sample mass.
From the data we calculate the minimal amount of substance
in this example to be
k
mmin > sRP = 2 0.04mg ≈ 80mg .
Urel 0.1%

Chart of Minimal Sample Mass


As example 1 depicts just one solution in the uncertainty
domain, so does example 2 give just one solution in the
same value. This would lead to a repeatability equal to zero, a value
which is certainly not realistic. In fact, it might even have been
partially provoked by the rounding of the indicated result. In any case,
including the readability variance in this case makes sens, as it reflects
at least the contribution of rounding.
It might therefore be sensible to include the variance of readability in
any case: if repeatability is larger than readability, then the variance
of the former negligible compared to the latter; if repeatability
happens to be very small or to vanish, then the variance of readability
is a valid representation of the situation.
38) The specification of the AT201 semi-micro balance is given in a
previous footnote.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 34 —

domain of minimal sample mass. Here as there, it is desirable


to have a chart from which the actual requirement may be
derived. Therefore, we might want to create a chart for the
minimal sample mass, as well as get an overview.
There is no need for a new chart of sample mass, as the chart
of relative uncertainty already contains the required
information, indeed. We only have to interpret is the other
way round. Choosing the required relative uncertainty (given
the chart reflects the correct coverage factor) on the y-axis,
the corresponding minimal sample weight can be read from
the x-axis. We thus may choose the 0.1% level of uncertainty
in the chart of relative uncertainty shown above, and find the
crossing of this gridline with the bold curve (which represents
the 200g repeatability). We read a value of approximately
100mg at this crossing, a value slightly higher than what we
have just calculated. The reason for this is that the uncertainty
charts as given above include the full influence of non-
linearity even for small weights, something we have neglected
in example 2. These uncertainty curves are therefore
somewhat conservative for small sample masses; however,
they still will deliver fair estimates of uncertainty for small
sample mass.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 35 —

Conclusions
Weighings, i.e., the determination of the mass of a weighing
object, are subject to influences from various sources.
Generally, they cause a measurement bias as well as a
measurement uncertainty.
The influences can originate from the balance and from the
weighing object. Thereby, the environment plays a major
role. The weighing result is influenced by non-ideal
characteristics of the balance, such as readability,
repeatability, corner load error, non-linearity and long-term
drift, or by disturbances caused by environmental conditions
such as temperature drift. The weighing sample gets affected
mainly by ambient conditions, such as air temperature, air
humidity, air draft, and substantially so, by air buoyancy.
Often, the mass of the object weighed can be the culprit, as it
may vary. One should be aware that the mass of an object
may not be the unchanging physical quantity, for which it is
often taken.
Whether these influences are of relevance for the weighing
result depends on several conditions, which should be
checked for every application. As a rule of thumb one can
say that the higher the required measurement resolution of
the weighing is, the more likely the influences will manifest
themselves.
Contrary to a maintained balance which as a rule does not
introduce a measurement bias that is to the detriment of the
measurement accuracy, air buoyancy can cause a substantial
measurement bias. There is an international regulation about
conventional mass as how to deal with this influence.
However, it was primarily issued to establish rules for the
dissemination of mass standards. For weighing objects with
distinctly different densities than the conventional density
(8000kg/m3 ) defined is this regulation it is not applicable.
Dependent on the application, the measurement bias caused
by air buoyancy can exceed measurement uncertainty by far.
The measurement uncertainty is determined by the
performance of the balance, the properties of the weighing
object and the state of the environment. The properties of the
balance are partly of random nature (such as readability,
repeatability) and partly of systematic nature (such as non-

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 36 —

linearity, sensitivity offset, temperature drift or long-term


drift). The individual values of the systematic deviations are
unknown. However, within a set of balances large enough,
their expectation is zero, as these deviations get adjusted for
a zero target value. Therefore the systematic deviations can
be treated like random deviations and may be included into
the uncertainty determination. It turns out that for small
sample weights (compared to the weighing capacity of the
balance) the (absolute) uncertainty is virtually constant , while
for large sample masses uncertainty increases with sample
load. Therefore, the relative uncertainty, i.e., the uncertainty
normalized to sample mass, remains constant for large
sample masses, while it increases for decreasing sample
masses. With small sample masses, uncertainty is dominated
mainly by repeatability, and to some extent by non-linearity,
while with large sample masses the influences of sensitivity
offset, sensitivity drift with temperature and air buoyancy
prevail. Especially with small sample densities, variations of
the ambient air density, caused by atmospheric conditions,
can influence buoyancy heavily, such that this influence
dominates uncertainty with large sample masses.
Relative uncertainty increases with decreasing sample mass.
On the other hand, with large sample masses, uncertainty is
dominated by sensitivity accuracy and, depending on the
sample density and the conditions of the ambient air, air
buoyancy. Relative uncertainty rolls off and stays essentially
constant with increasing sample mass.
Sometimes, when a weighing is performed, a given weighing
uncertainty should be reached, while at the same time only as
little substance as required should be weighed. The minimal
mass of a sample (also called ”minimal sample mass” or
simply ”minimal weight”) is the amount of mass that just
meets this condition. This amount can be derived from the
estimation of uncertainty and is—being a small mass—
essentially a function of repeatability, determined by the
performance of the balance, the properties of the weighing
object and the environmental conditions.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 37 —

Appendices
Buoyancy
If a weighing takes place in a vacuum on the surface of
earth, a body with mass m produces a weight force of
G = mg .
If the same body is weighed in the atmosphere, it will be
subject to an air buoyancy force which is opposed to the
weight force and diminishes the latter. According to
Archimedes’ principle 39), a body receives a buoyant force
equal to the weight force of the fluid displaced. This buoyant
force can be determined from the density of the air a and the
volume V displaced, according to
GF = aVg .
Since the volume of displacement is equal to the volume of
the body weighed, we can determine it from the density ρ
and the mass m of the body
m
V= ρ
which yields for the buoyant force
m
GF = a ρ g .
For the sum of all forces acting on the buoyant body
(apparent weight) we have
a
GA = G – GF = mg–a m ρ g = g 1– ρ m .
Sometimes, within the above expression the term
a
m A = 1– ρ m
is referred to as ”apparent mass”.
Sensitivity Adjustment (Determination of the calibration factor)
The sensitivity of a balance is established with the aid of a
calibration mass (a.k.a. calibration standard).
Upon loading a mass m onto a balance, it will display a
value
a
R = cg 1– ρ m .

39) Archimedes, 287?-212 B.C. Greek mathematician, engineer, and


physicist.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 38 —

Loading the same balance with a known calibration mass


mcal will produce the following result
a
Rcal = cgcal 1– ρcal mcal ,
cal
where the parameters with index ”cal” denote values at the
time of this calibration weighing. From this measurement the
calibration factor c can now be determined
Rcal
c= acal .
gcal 1– ρ mcal
cal
The calibration mass may deviate from its nominal value. The
actual mass of the calibration standard may be expressed as
mcal = mcal 1+∆µ cal ,
0
where mcal is the nominal value of the standard, with
0
∆µ cal being its relative deviation from the nominal value,
according to
∆mcal
∆µ cal=
mcal
0 ,
where ∆mcal is the (absolute) mass deviation.
Inserting this calibration mass into the expression for the
calibration factor provides
Rcal
c= acal .
gcal 1– ρ mcal 1+∆µ cal
cal 0
It seems so obvious that we almost forget to mention that we
expect the displayed value to be equal to the nominal mass of
the calibration standard loaded on the balance, namely
Rcal = mcal .
0
We thus obtain for the calibration factor
1
c= acal .
gcal 1– ρ 1+∆µ cal
cal
Putting this calibration factor to work will adjust the sensitivity
of the balance to its correct value, and we get for the transfer
characteristic the adjusted balance

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 39 —

a
g 1– ρ 1
R= a m .
gcal 1– cal 1+∆µ cal
ρ cal
Dependence Of Air Density On Temperature & Pressure
The density of air depends mainly on air temperature and
pressure; to a lesser extent also on its relative humidity and its
composition. For the sake of uncertainty estimation, we give
here a simplified theory for the dependence of air density on
temperature and pressure.
If we consider air to be an ideal gas—not a severe
simplification for our purposes here—the parameters of an
amount m of air, such as its volume V , absolute
temperature T and pressure p , obey the law of an ideal
gas, according to
pV = nkT ,
where n is the number of molecules in that amount, and k is
Boltzman’s constant. By definition, the density a of this air
equals
m
a= ,
V
and substituting the volume of air from the gas law, we get
p
a= m .
nk T
From this expression we take the partial derivative with
respect to pressure
∂a m 1
∂ p = nk T
and with respect to temperature
∂a –p
= m 2 .
∂T nk T
Substituting the constant term m/ nk in this expressions with
the aid of the original formula for air density, we get
∂a aT 1 a
∂p = p T = p
and
∂a = aT – p = – a 40 ).
∂T p T2 T

40) As a rule of thumb, at normal conditions the relative density change


amounts to about 1/1000 per mbar and about 1/300 per °C.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 40 —

Thus, at approximately normal atmospheric conditions with


p0 ≈ 1 bar and T0 ≈ 290K = 17°C ,
resulting in an air density of
a0 ≈ 1.20 kg/m3 (at 50%rH) 41),
we get for its sensitivity with respect to pressure
∂a a0 3
∂p 0
= p ≈ 1.20 kg/m /bar ,
0
and with respect to temperature
∂a a
= – 0 ≈ – 4.1×10–3 kg/m3 /K .
∂T 0 T0
For a comprehensive theory on air density see [Giac] and
[Davis].
Variance Of Relative Buoyancy
The relative change of the weight force under the influence of
air buoyancy on a sample with density ρ , being weighed on
a balance, including the balance’s sensitivity adjustment by
means of a calibrations standard with density ρ cal is an
(apparent) sensitivity bias given by
a a a a
b ≈ 1– ρ 1+ ρcal –1 ≈ ρcal – ρ ,
cal cal
where a and acal are the air densities at the time of
weighing and at the time of sensitivity adjustment,
respectively, and b is the relative change in sensitivity (or
apparent mass).
If the corresponding standard deviations sρ , sa , sρ and
cal cal
sa of the densities are known, the variance of the buoyancy
influence can be determined. From the partial derivatives
∂b = – acal
∂ρ cal 2
ρ cal
∂b 1
∂acal = ρ cal
∂b a
∂ρ = ρ 2
∂b 1
∂a = – ρ

41) Air density calculated from [Davis].

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 41 —

we can now determine the combined variance by adding up


all individual variances, multiplied by the squares of their
respective derivatives, assuming that the contributing
influences are statistically uncorrelated. This yields
a 2
1 s 2+ a s 2+ 1 s 2 .
sb 2 = cal s
2 ρ cal + ρ cal acal ρ a
ρ cal ρ2 ρ

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001
— 42 —

References
[Niels] Henrik S. Nielsen: How long is a piece of string?
NCSL Workshop and Conference, Toronto, Canada, July 17-
20, 2000.
[Clark] John P. Clark: Evaluation Of Methods For Estimating The
Uncertainty Of Electronic Balance Measurements.
Weighing 2000, South Yorkshire Trading Standards Unit,
Sheffield (UK), June 13-14, 2000.
[R33] Conventional value of the result of weighing in air.
International Organization of Legal Metrology (OIML), Paris,
France. International Rrecommendation No 33, 1979.
[VIM] International Vocabulary Of Basic And General Terms In
Metrology
International Organization for Standardization, Geneva,
1993
ISBN 92-67-01075-1
[GUM] Guide To The Expression Of Uncertainty In Measurement
First edition 1995
International Organization for Standardization, Geneva
ISBN 92-67-10188-9
[Rth1] Arthur Reichmuth: Non-Linearity Of High Resolution
Laboratory Balances.
Weighing 2000, South Yorkshire Trading Standards Unit,
Sheffield (UK), June 13-14, 2000.
[Rth2] Arthur Reichmuth: Non-Linearity Of Laboratory Balances And
Its Impact On Uncertainty.
NCSL Workshop & Symposium, Toronto (CAN), July 17-20,
2000.
[Kehl] Karl Kehl, Klaus Weirauch, Samuel Wunderli, Veronika
Meyer: The influence of variations in atmospheric pressure on
the uncertainty budget of weighing results.
The Analyst, 2000, Vol. 125, P. 959-962.
[Giac] P. Giacomo: Equation for the Determination of the Density of
Moist Air (1981). Metrologia [1982], 18, pp. 33-40.
[Davis] R.S. Davis: Equation for the Determination of the Density of
Moist Air (1981/91). Metrologia [1992], 29, pp. 67-70.

© Mettler Toledo, A. Reichmuth Weighing Accuracy 2.1 Prtd.: September 25, 2001

You might also like