JFA (J URNAL F ISIKA DAN A PLIKASINYA ) VOLUME 17, N UMBER 1, JANUARY 2021

Evaluation and Analysis of Uncertainty Measurement of

The Sound Level Meter Calibration by Coupler Method
Bondan Dwisetyo and Denny Hermawanto
Badan Standardisasi Nasional, Kompleks PUSPIPTEK Gedung 420, Serpong, Tangerang Selatan 15314

Abstract: The calibration of sound level meter (SLM) using a coupler method has been introduced. This
method utilizes a multifunction acoustic calibrator as a portable laboratory instrument. By the reason of its
convenient application, it is appropriate to be realized by the secondary calibration laboratory. To complete the
requirement of calibration result provided by ISO 17025, an analysis of the uncertainty measurement related to
this calibration is strongly needed. Therefore, this work aims to provide an evaluation and analysis of some ele-
ments that influence to the coupler method of SLM calibration using the statistical approximation to determine
the major parameters such as standard uncertainty, combine uncertainty, degree of freedom, and expanded un-
certainty in accordance with The Guide to the expressions of Uncertainty in Measurement (GUM). Moreover, a
case study related to this instrument calibration also discussed along this paper. From the result, the uncertainty
budgets that contribute to the SLM calibration has been obtained, and generally, it is found that the standard
instrument has the highest contribution to the determination of the expanded uncertainty. For the whole results,
the obtained uncertainty values are still acceptable according to the acceptance limit values that required by IEC
61672-1.
Keywords: calibration of sound level meter; coupler method; GUM; uncertainty measurement.

*Corresponding author: [email protected]

Article history: Received 2 May 2020, Accepted 25 August 2020, Published January 2021.
http://dx.doi.org/10.12962/j24604682.v17i1.6988
2460-4682 Departemen
c Fisika, FSAD-ITS

I. INTRODUCTION the calibration result is necessary to be discussed comprehen-

sively. According to the requirement of calibration and test-
The calibration of the acoustical instruments such as a mi- ing provided by ISO 17025, the uncertainty measurement as a
crophone, a sound calibrator, and a sound level meter, has be- main part of the calibration result needs strongly to be evalu-
come the trending topic in a Metrology area in recent years ated and analyzed [5].
[1]. The sound level meter, known as SLM, is considered Therefore, the objective of this work is to provide an eval-
as the most applicated device by the users among these in- uation of the uncertainty measurement of SLM calibration by
struments, and the customer needs related to this instrument the coupler method using the statistical approximation to de-
calibration tend to increase [2]. To accommodate their neces- termine the major parameters such as standard uncertainty,
sity, the calibration of SLM has been provided by the National combine uncertainty, degree of freedom, and expanded un-
Standardization Agency of Indonesia (BSN) as the represen- certainty in accordance with The Guide to the expressions of
tative of the National Metrology Institute (NMI) of Indonesia Uncertainty in Measurement (GUM). Moreover, a case study
by using the absolute method. To realize this method, it is related to this instrument calibration is also discussed along
necessary to utilize a full anechoic chamber as a calibration with this paper, where the result will be compared with the
medium, and the laboratory standard of a microphone as a acceptance limit value required by IEC 61672-1. In addition,
reference device [3]. Even though this method is capable to the calibration and measurement capability (CMC) of sound
assure the calibration result so accurately, this method is diffi- level meter calibration using this method also is planned to be
cult to be applied because of its complexity to be implemented proposed to the Committee of National Accreditation (KAN)
and have a very high financial budget, especially for the sec- as the representative organization for local accreditation.
ondary calibration laboratory.
Furthermore, the other method has been introduced, where
it is known as the coupler method that uses a multifunction II. MATHEMATICAL MODEL
acoustic calibrator as the portable laboratory instrument, and
utilize a small insulation box as a calibration medium [4]. This The calibration of sound level meter by coupler method is
method is quite simpler than the previous method, and it is ap- conducted using the multifunction acoustic calibrator as the
propriate to be applied by the private calibration laboratories laboratory precision instrument, and diagram of the calibra-
to comply with the industrial necessity for the SLM calibra- tion system is shown in Fig. 1.
tion that tends to increase by a year. However, this method is Additionally, to optimize the calibration process, it is nec-
considered as the new method in Indonesia relatively, where essary to make use of an insulation box as the calibration
Bondan Dwisetyo et al. / J.Fis. dan Apl., vol. 17, no. 1, pp. 14-19, 2021 15

measured by the standard instrument (repeatability),

and its readability. The former is categorized as A-type
of uncertainty budget. The number of the measurement
data of L for the required frequencies is 5 times. After
that, the standard deviation of the corresponding data is
calculated later. Meanwhile for the later, it is classified
as B-type of uncertainty budget. Further, it depends on
the type of sound level meter used, where the class 1 of
this handheld device has the maximum resolution up to
0.01 dB, while for the lower class, its readability is 0.1
FIG. 1: Diagram of calibration system of sound level meter dB.
by coupler method
2. Meanwhile for the standard instrument as the next pa-
rameter that represented by multifunction acoustic cal-
medium. In the process, sound level meter, as a unit under test
ibrator, the component of the uncertainty budget com-
(UUT), is putted into this box, and therefore, its sound pres-
prises the calibration of a nominal sound pressure level,
sure indication is compared directly to sound pressure level
drift due to its use, the sound pressure response, the
that generated by the calibrator.
level of accuracy, and the total harmonic distortion with
Subsequently, the mathematical model of this calibration
noise. The first mentioned component of this parame-
procedure can be defined, where it is stated as the deviation
ter is subsumed as B-type of uncertainty budget. It is
between the sound pressure level (L) that measured by sound
obtained from the result of calibration using the labo-
level meter that later is mentioned as Ldut , and the generated
ratory standard of microphone as primary reference of
L by the multifunction acoustic calibrator (Lstd ). In addition,
acoustic measurement by insert voltage method. Next,
these devices have a correction values as the function of fre-
the second components also is characterized as B-type
quency, where it can be obtained from the calibration result of
of uncertainty budget, where it is taken from the mea-
the calibrator, meanwhile for the UUT, it is provided by the
surement result of an intermediate check that conducted
manufacturer generally. Later, these corrections are denoted
annually using the same standard and method as the
as δstd and δdut respectively. Hence, the mathematical model
previous component. Afterwards, the last three men-
for the sound level meter calibration is expressed as follow
tioned of this parameter is provided by the manufacturer
[4]:
through the initial certificate and manual book, and fur-
∆L = (LDU T + δDU T ) − (LST D + δST D ) (1) ther, these components are grouped as B-type of uncer-
tainty budget.
For some cases, the information of correction values of
sound level meter is not assigned by the manufacturer. Thus, it
3. Subsequently, an additional budget also be added in this
can be ignored from the Eq.(1). Subsequently, the stipulation
evaluation, and it is considered as a capability of the
of uncertainty budgets can be applied based on this mathemat-
used insulation box to insulate the background noise
ical model.
from outside. It can be obtained by conducting mea-
surement of L inside and outside this box separately,
III. UNCERTAINTY BUDGETS and later calculating deviation among these conditions.
In addition, the insulation capability of this box also can
be determined by performing the sound insulation m.
The budget of uncertainty is divided into two categories,
A-type and B-type. The former is the parameter that can
be obtained by applying some measurement series and gen-
erally solving with a statistical procedure, while the latter is IV. EVALUATION STANDARD UNCERTAINTY
the source that can be defined through a scientific judgment or
other information that is considered to contribute to the result According to GUM, the evaluation method to determine a
of measurement or calibration [6]. For the convenient, there is standard uncertainty of the corresponding budgets comprises
an additional means to simplify the determination of the un- A-type and B-type [6]. The former can be applied by using
certainty budgets beside the use of mathematical model, and the statistical analysis of series of measurements. Meanwhile
therefore, it can be carried by applying the ishikawa diagram for the later, it is obtained by evaluation by means other than
[7] that is shown in Fig.2 for the case of sound level meter the statistical analysis of series of observations. Therefore, de-
calibration by coupler method. termination of the standard uncertainty of the aforementioned
From Fig.2, it is shown that the components of the uncer- components is as follow:
tainty budgets of this calibration method can prescribe as fol-
low:
1. The repeated measurement of SLM that measured by
1. According to the sound level meter as DUT, the com- the standard instrument (repeatability). it is obtained by
ponent consists of the repeated measurement of L that calculating the standard deviation after taking the data
16 Bondan Dwisetyo et al. / J.Fis. dan Apl., vol. 17, no. 1, pp. 14-19, 2021

FIG. 2: Determination of uncertainty budgets by Ishikawa diagram

serially, where it can be expressed as follow: there is no statement and information what the used dis-
r tribution and its confidence level. Hence, the appropri-
1 ate distribution that should be applied of this budget is
Stdev = Σ(Li − L̄) (2)
N −1 the square. Therefore, this uncertainty budget (u4 ) can
be calculated as follow:
where, Li is the individual of sound pressure level mea-
Udrif t
surement, while is the mean value of L measurements, u4 = √ (6)
and N is number of measurement series at the same con- 3
ditions. After that, the standard uncertainty of this bud-
get (u1 ) can be calculated using the Eq.(3). 5. Sound pressure response measurement of the standard.
It is found from the manual book of the corresponding
Stdev instrument. By the reason of there are no information
u1 = √ (3) that explained the confidence level, so the square distri-
N
bution is appropriate to be applied for this component.
2. Readability of UUT. It is determined based on its res- Therefore, the standard uncertainty (u5 ) can be calcu-
olution, and therefore, the standard uncertainty of this lated as follow:
component (u2 ) can be assigned as follow:
a Uresponese
u2 = √ (4) u5 = √ (7)
3 3
√
where, a is a half of UUT resolution, and 3 is a divi- 6. Accuracy level of the standard. The same source is
sor of the square distribution according to GUM [6]. In obtained as the previous budget. After that, the stan-
addition, an uncertainty budget is stated to be used this dard uncertainty according to this component (u6) can
distribution if there is no information further related to be calculated respectively as follow::
the corresponding budget [6].
Uaccuracy
u6 = √ (8)
3. The calibration of nominal SPL of the reference instru- 3
ment. It is taken from the last certificate of calibration
as mentioned above, where it is stated the confidence 7. Total harmonic distortion and noise (THD + N). This
level has a value is 95%, so it should use a normal distri- is a part of calibration of nominal SPL of the reference
bution that has divisor of 2, and the standard uncertainty instrument, and it can be calculated respectively as fol-
(u3 ) calculated as follow: low:
Ucertif icate UT HD
u3 = (5) u7 = (9)
2 2

4. Drift of the reference instrument. As mentioned above, 8. Capability of the insulation box to insulate the back-
it is obtained from the annual check, and moreover, ground noise from outside. The same step also applied,
Bondan Dwisetyo et al. / J.Fis. dan Apl., vol. 17, no. 1, pp. 14-19, 2021 17

and it can be calculated as follow: Welch-Satterthwaite formula as follow [6]:

Uinsulation u4c (L)
u8 = √ (10) vef f = (16)
3 Σi = 1N
u4i (L)
vi

In addition, the other of principal parameters also is deter- Afterwards, k can be calculated using the programmable
mined, and it consists of sensitivity coefficient for the budgets software for the convenient, where in this work, a spreadsheet
of uncertainty (ci ), and a degree of freedom (vi ). The first pa- excel is used that has capability to calculate k using the func-
rameter is described how the obtained measurand varies with tion of TINV (probability; vef f ), where the probability is con-
changes in the values of the other parameters. In particular, sidered as a level of hesitancy that has value of 5%, and it is as-
the change in the main measurand produced by a small change signed from the normal distribution with the confidence level
of another parameter, and is given by calculating the partial of 95% . Finally, the expanded uncertainty (U) is determined
derivative of the Eq.(1) to the input parameter [6]. Therefore, by using the formula as follow:
it can be expressed mathematically as follow:
U (L) = kuc (L) (17)
∂L
ci = (11)
∂xi Later, this quantity will be compared with the acceptance limit
value required by IEC 61672-1.
Meanwhile for the second parameter, it depends on the type
of used uncertainty method. It can be calculated by subtract-
ing total amount of the measurement data (N) with 1 for A-
VII. CASE STUDY
type, meanwhile for the other, it should be infinite accord-
ing to JCGM and an estimation result of the published paper.
Therefore, these parameters can be written serially as follow: In this work, the case study of uncertainty evaluation of
SLM calibration using the coupler method was provided for
vi = N − 1 for type-A (12) the A-frequency weighting parameter at the sound pressure
vi = ∝ for type-B (13) level of 114 dB, where the calibration process and its calcula-
tion were applied separately. The calibration was conducted
in a laboratory of acoustics and vibration BSN using the sys-
V. DETERMINATION OF COMBINED STANDARD tem apparatus consists of a class-2 SLM as a unit under test
UNCERTAINTY (UUT), a reference instrument (Multifunction Acoustic Cali-
brator B&K 4226), and an insulation box.
During the calibration, there was no alteration of environ-
Furthermore, the combined standard uncertainty can be cal-
mental conditions, where it was recorded as 24.2◦ C, 63%RH,
culated using the equation as follow [8]:
and 100.0 kPa for ambient conditions of temperature, relative
u2c (L) = ΣN 2 2 humidity, and air pressure respectively inside the insulation
i=1 ci ui (14)
box. The data was taken five times for the corresponding fre-
From this equation, coefficient of sensitivity of the afore- quencies, where the details of the result shown in Table I.
said budgets is calculated using the equation (10), where in The determination of uncertainty measurement for SLM
this work, it is found that this value is 1. Hence, the combined calibration by the coupler method is written in Table I. For
standard uncertainty is written as follow [8]: the first parameter represented by the UUT, the uncertainty
component consists of the repeated measurement and its read-
ability. The former has the standard uncertainty values vary
q
uc (L) = u21 + u22 + u23 + u24 + u25 + u26 + u27 + u28 (15) for the measurement frequencies, where the maximum value
is found at the frequency of 63 Hz, 125 Hz, 1 kHz, 2 kHz, and
8 kHz that its value is about 0.02 dB. Meanwhile for the later,
VI. DETERMINATION OF EXPEANDED UNCERTAINTY the standard uncertainty value is seen to be equal for the cor-
responding frequencies that are obtained and calculated using
The expanded uncertainty can be determined by multiply B-type.
the combined uncertainty that has the confidence level is 67%, Subsequently, the second parameter represented by the
with a coverage factor (k). Furthermore, there are some guide Multifunction Acoustic Calibrator provides the most compo-
to obtain k value, and it can be conducted by using t-student nents that contribute to the uncertainty calculation, and it can
table or calculating the other parameter that is considered as be seen that the standard uncertainty tends to be steady for the
effective of degree of freedom that expressed as vef f . For the component that is obtained from SPL calibration and THD+N
first guide, the table has mentioned that the coverage factor report. Subsequently, the remain components that comprise
(k) vary in the corresponding of confidence level, where it is the drift value of this standard instrument therewith sound
found that this parameter has the value 1.96 for the confidence pressure response and level of accuracy are shown that the
level of 95%. Subsequently, the other guide also can be im- standard uncertainty values lean to go up and down beyond
plemented by calculating effective degree of freedom using these frequencies. Afterward, the last parameter is given by
18 Bondan Dwisetyo et al. / J.Fis. dan Apl., vol. 17, no. 1, pp. 14-19, 2021

FIG. 3: Expanded uncertainties comparison (Blue: IEC 61672-1:2013, Red: Proposed coupler method)

TABLE I: Calculation of uncertainty measurement of sound level meter calibration for SPL 114 dB.
No Uncertainty budged S t a n d a r d U n c e r t a i n t y (dB)
Parameter Component 63 Hz 125 Hz 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 8 kHz

1 Sound Level Repeatability (Tipe A) 0.02 0.02 0.00 0.00 0.02 0.02 0.00 0.02
Meter UUT Readability 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03
2 Multifunction SPL of 114 dB 0.06 0.06 0.05 0.05 0.05 0.06 0.06 0.05
Calibrator Drift 0.04 0.04 0.04 0.01 0.04 0.03 0.02 0.02
B&K 24226 Sound pressure response 0.07 0.06 0.06 0.06 0.06 0.08 0.11 0.17
Level of accuracy 0.09 0.09 0.06 0.06 0.09 0.09 0.09 0.14
THD + N 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
3 Insulation box Insulation capacity 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06
uc (dB) 0.15 0.14 0.13 0.12 0.14 0.15 0.17 0.24
vef f 207.30 202.41 255.97 218.66 203.73 200.68 167.85 124.69
k 95% 1.97 1.97 1.97 1.97 1.97 1.97 1.97 1.98
u (dB) 0.29 0.28 0.25 0.23 0.28 0.29 0.33 0.48

the insulation box that assists the standard uncertainties is coupled to this standard. It can be seen in the datasheet of
about 0.06 dB for all frequencies. multifunction acoustic calibrator that shown that the level of
After that, the combined uncertainty (uc ) and the effective accuracy and SPL response has the biggest correction value at
of degree of freedom (vef f ) is calculated, and therefore, the the high frequencies.
multifunction acoustic calibrator withal the sound pressure re- In addition, comparison of the determined expanded uncer-
sponse and level of accuracy has the highest contribution to tainties and its acceptance values provided by IEC 61672-1:
calculate the combined uncertainty at the frequency of 8000 2013 [9] is shown in Fig.3. The obtained expanded uncer-
Hz that its value is 0.24 dB. After that, by determining k value tainty values in this work that represented by the red triangles
for the confidence level of 95%, the expanded uncertainty (Uc ) are below to the maximum of acceptance values that expressed
can be obtained, and its value reaches up to 0.48 dB of this fre- by the blue dots for all frequencies as seen in Fig.3. The two
quency. Meanwhile for the other frequencies, it has a value at curves have a difference of its trend slightly, where it is shown
the range of 0.24-0.33 dB. that the obtained Uc start to increase about 0.1 dB at frequency
In contrast to using the absolute method that has the ex- of 4000 Hz and go up to 8000 Hz with the same increasing
panded uncertainty tend to be flat in this frequency range, the value. Meanwhile for the curve that given by IEC 61672-1:
coupler method of SLM calibration tends to increase at the 2013, it is equal from the initial frequency to 4000 Hz, af-
high frequencies. It can be explained that at this frequency, ter that, the similar condition is found as well as the previous
the sound that generates by the standard instrument is easier curve. Hence, the results of this work are still acceptable and
to be flanking transmission when the microphone of SLM is reasonable to be applicated for the calibration of sound level
Bondan Dwisetyo et al. / J.Fis. dan Apl., vol. 17, no. 1, pp. 14-19, 2021 19

meter further. In addition, the expanded uncertainty is esti- tained expanded uncertainty values in this work are accept-
mated to has a lower value when using the sound level meter able, and therefore, it is reasonable to be proposed to the Com-
test that has better resolution, and it has alteration about 0.1 mittee of National Accreditation (KAN) as the representative
dB for whole frequencies. organization for local accreditation. In addition, this result is
strongly recommended to be compared with another method
to evaluate uncertainty measurement such as the Monte-Carlo
VIII. CONCLUSION method.

The evaluation of uncertainty measurement of sound level

meter calibration by coupler method has been conducted by
Research Group for Acoustics and Vibration at laboratory Acknowledgments
acoustics and vibration-BSN. Besides, the determination of an
uncertainty budget, and other important attributes also have The authors acknowledge the financial support from Min-
been provided. The multifunction acoustic calibrator along istry of Research and Technology/National Agency for Re-
with the sound pressure response and level of accuracy has the search and Innovation of the Republic of Indonesia through
highest contribution to determine the expanded uncertainty at the scheme of Incentive Research Program for the National
the high frequency. This result also has shown that the ob- Innovation System in 2020/2021.

[1] C. Series, ”Measurement traceability of acoustics and vibration [6] JCGM, ”Evaluation of measurement data Guide to the expres-
instruments in Indonesia Measurement traceability of acoustics sion of uncertainty in measurement”, no. September, 2008.
and vibration instruments in Indonesia”, 2018. [7] D. A. Giardino and J. P. Seiler, ”Uncertainties associated with
[2] D. Hermawanto et al., ”Development of Decibel Stepped Atten- noise dosimeters in mining”, vol. 100, no. May 1996, pp.
uator for Automated Sound Level Meter Calibration”, Mapan - 15711576, 2014.
J. Metrol. Soc. India, vol. 35, no. 1, pp. 8186, 2020. [8] I.MA.2.06.U-E.2-R.0, ”Evaluasi Ketidakpastian Kalibrasi Sound
[3] N. Garg, et al., ”Measurement Uncertainty in Microphone Free- Level Meter Menggunakan Multifunction Acoustic Calibrator”,
Field Comparison Calibrations”, MAPAN, 2019. 2019.
[4] I.MA.2.06, ”Calibration Procedure for Sound Level Meter by [9 ] IEC 61672-1, ”Electroacoustics-Sound Level Meters-Part 1:
Using Multi Frequency Acoustic Calibrator”, 2015. Spesification”, International Electroacoustics Commision, 2013.
[5] SNI ISO/IEC 17025:2017 ”Persyaratan umum kompetensi labo-
ratorium pengujian dan kalibrasi”, 2019.