See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/278771171

An additional configuration to standard ASTM E2611-09 for measuring the

normal incidence sound transmission loss in a modified impedance tube

Article · April 2012

CITATIONS READS

4 1,148

3 authors:

Olivier Doutres R. Panneton

École de Technologie Supérieure Mecanum Inc, and Université de Sherbrooke
103 PUBLICATIONS   895 CITATIONS    171 PUBLICATIONS   3,028 CITATIONS   

SEE PROFILE SEE PROFILE

Yacoubou Salissou
Université de Sherbrooke
18 PUBLICATIONS   406 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Porous Material View project

Study of the sound attenuation of earmuffs, numerical modelling and experimental validation View project

All content following this page was uploaded by R. Panneton on 15 October 2015.

The user has requested enhancement of the downloaded file.

 An additional configuration to standard ASTM
E2611-09 for measuring the normal incidence sound
transmission loss in a modified impedance tube
Olivier DOUTRES, Raymond Panneton, Yacoubou Salissou

To cite this version:

Olivier DOUTRES, Raymond Panneton, Yacoubou Salissou. An additional configuration to
standard ASTM E2611-09 for measuring the normal incidence sound transmission loss in a
modified impedance tube. Société Française d’Acoustique. Acoustics 2012, Apr 2012, Nantes,
France. <hal-00810671>

HAL Id: hal-00810671

https://hal.archives-ouvertes.fr/hal-00810671
Submitted on 23 Apr 2012

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Proceedings of the Acoustics 2012 Nantes Conference 23-27 April 2012, Nantes, France

An additional configuration to standard ASTM

E2611-09 for measuring the normal incidence sound
transmission loss in a modified impedance tube
O. Doutres, R. Panneton and Y. Salissou
Groupe d’Acoustique de l’Université de Sherbrooke, 2500 Boulevard de l’Université, Faculté
de Génie, Sherbrooke, AB, Canada J1K2R1 QC
[email protected]

2029
23-27 April 2012, Nantes, France Proceedings of the Acoustics 2012 Nantes Conference

This paper presents a three-microphone two-load (3M2L) method for measuring the normal incidence sound
transmission loss of a noise control system or material in a modified impedance tube. In the standard ASTM
E2611-09, the downstream section of the impedance tube includes two microphones flush with the interior
surface of the tube and the proposed four-microphone two-load (4M2L) method to assess the normal incidence
sound transmission loss requires the use of two arbitrary acoustic terminations. The method presented here is
conceptually identical but (i) the downstream section is reduced to a simple movable rigid backing with one
microphone flush mounted on it and (ii) the two arbitrary acoustic terminations are replaced by two air cavities.
It thus requires one microphone less and fewer transfer functions. The standard switching technique used to
correct the variations between the three microphones is validated on a symmetrical air layer. The proposed
3M2L method is then applied to a non-symmetrical specimen and validated compared to the standard 4M2L
method.

1 Introduction
The normal incidence sound transmission loss (nSTL) is
an important indicator used to assess the sound insulation
property of noise control systems (e.g., sound barrier
samples, mufflers, expansion chambers, and resonators).
This measurement of the nSTL has been standardized in
ASTM E2611-09[1]. It uses a plane wave tube
instrumented with four microphones and a termination of
Figure 1: Measurement configuration.
adaptable acoustic load. For any specimen, the four-
microphone two-load (4M2L) standard method can be
applied for measuring its transfer matrix and nSTL. The Now, suppose a unit amplitude incident plane wave
two loads are typically a minimally reflecting termination with time dependence of the form exp(j ωt), where j = −1 ,
(e.g., anechoic termination), and a termination reflecting a ω is the angular frequency, and t is the time. The acoustic
portion of incident wave (e.g., open termination). A pressure pi(x) and particle velocity ui(x) upstream (x≤0) and
minimum of 6 transfer functions need to be measured. downstream (x≥d) the test sample are respectively given by
An additional configuration to those described in
standard ASTM E2611-09 (see Table 2 in [1]) has been
recently published [2] and is described here. This pi ( x ) = e − jkx + Ri e jkx , ui ( x ) =
Zs
(e + Ri e jkx ),
1 − jkx (1)
configuration uses three microphones in a slightly modified
classical impedance tube, the third microphone being flush
mounted on the movable hard termination. Furthermore, the and
proposed method requires fewer transfer function pi ( x ) = 2 Ai e − jkLi cos(k ( x − Li )),
measurements (4 in the general case) than the standard Ai − jkLi (2)
method. The two acoustic loads are in this case two ui ( x ) = − j 2e sin(k ( x − Li ))
Zs
different air cavities. The purpose of this paper is to present
this simple additional configuration and how it can be used where subscript i=a,b and refers to a value obtained with an
in conjunction with (or in complement to) standard ASTM air cavity of thickness Di, Li=d+Di, d is the thickness of the
E2611-09. sample, Zs and k are the complex specific acoustic
impedance and complex wave number of the air in the tube,
Ri is the complex sound reflection coefficient at the surface
2 Theory of the sample (i.e. at x = 0), and 2Ai is the maximum
pressure amplitude of the standing wave downstream the
A schematic view of the modified impedance tube is sample. The two air cavities Da and Db are the two loads of
shown in Figure 1. The apparatus consists of a finite-length the proposed 3M2L method. The geometrical variables are
hard walled impedance tube with uniform inner cross- also defined in Figure 1. Note that Zs and k account for
section. The tube features a loudspeaker (source) at one viscous and thermal dissipation effects at the tube walls.
end and a movable hard termination at the other end. The For the setup shown in Figure 1, the reflection coefficient is
loudspeaker is used to generate a plane wave field in the given by
impedance tube. There are two microphones flush mounted H ( D )e jks − 1 2 jkl
upstream the test sample and one microphone flush Ri = 12 i e , (3)
mounted on the hard termination. Downstream the sample, 1 − H 12 ( Di )e − jks
an air cavity is added. The thickness of the cavity is where H12(Di) is the measured transfer function between
adjusted with the movable hard termination. microphones 2 and 1 (pi(µ2)/pi(µ1)) with an air cavity of
thickness Di, s is the spacing between microphones 1 and 2,
and l is the distance between microphone 2 and the front
face of the sample. In a similar manner, one can deduce
coefficient Ai by introducing transfer function H13(Di)
between microphones 3 and 1 (pi(µ3)/pi(µ1)). This yields

2 Ai e − jkLi = H 13 ( Di )(e jk ( l + s ) + Ri e − jk ( l + s ) ). (4)

2030
Proceedings of the Acoustics 2012 Nantes Conference 23-27 April 2012, Nantes, France

According to Eqs. (1) and (2), calculating Ri and Ai from Figure 1 without sample and placing the movable piston at
Eqs. (3) and (4) allows a complete description of the sound position x=0. The cavity thickness downstream the sample
field in the air inside the impedance tube presented in is fixed by moving the hard termination conveniently. With
Figure 1. the used setup, the working frequency range for this study
For any specimen, Eq. (22) of standard ASTM E2611- is 150–4100 Hz. To implement the 4M2L standard method,
09 gives the transfer matrix T of the specimen in terms of the same setup is used; however the downstream section is
the pressure and particle velocity on its both faces (at x=0 replaced by a two-microphone instrumented 360-mm long
and at x=d). For the sake of clarity, this equation is repeated tube. The two downstream microphones are separated by a
here with the used notations: distance s=25.6 mm and flush mounted on the tube
extension. The distance between the back surface of the
1 sample and microphone 3 is l2. The two termination loads
T= × are selected to have relatively different reflection
p a ( d )u b ( d ) − p b ( d )u a ( d ) (5)
coefficients to yield good results. The first load is a
⎡ p a (0)u b (d ) − p b (0)u a (d ) p b (0) p a (d ) − p a (0) pb ( d )⎤ partially anechoic termination. It is constructed using a 1.5-
⎢ u (0)u (d ) − u (0)u (d ) p b (d )u b (0) − p b (d )u a (0) ⎥⎦ m long cylindrical tube filled with low density wool. The
⎣ a b b a
wool is arranged in a way that its density increases
where pressures and velocities are derived from Eqs. (1)-(4) gradually as the acoustic wave propagates in the tube. The
for two different loads. For cavity load i=a,b, they are second load is a 25.4-mm thick melamine foam backed by a
given by rigid cap.
The same four-channel analysis system and ¼-inch
H 12 ( Di ) sin(k (l + s)) − sin(kl ) MPA416 BSWA microphones are used to conduct all
pi (0) = −2 je jkl , measurements. The analysis system uses a USB Fireface
H 12 ( Di )e − jks − 1 UC sound card driven by a MATLAB script which
2 je jkl H 12 ( Di ) cos(k (l + s)) − cos(kl ) generates, acquires, and processes the signals. The input
ui (0) = , signals from the microphones are stored on 24 bits by the
Zs H 12 ( Di )e − jks − 1 (6)
sound card with a sampling frequency of 44.1 kHz. The
H 13 ( Di ) sin(ks) cos(kDi ) MATLAB script processes the signals to obtain the
pi (d ) = −2 je jkl ,
H 12 ( Di )e − jks − 1 required transfer functions. The output signal is a white
noise and the source is a 4-in. loudspeaker. Each
2 je jkl H 13 ( Di ) sin(ks) sin(kDi )
ui (d ) = . characterization test is the average of 50 repeated
Zs H 12 ( Di )e − jks − 1 measurements, and uses a linear weighting. For the 3M2L
. method, three microphones (µ1, µ2, µ3) and four channels
Consequently, the application of the 3M2L method to (ch1, ch2, ch3, ch4) are used for measuring the required four
obtain the transfer matrix T of any specimen will require transfer functions (H12(Da), H13(Da), H12(Db), H13(Db)).
only 4 transfer function measurements, which is two Each microphone µn (n=1, 2, 3) is connected to channel chn
measurements less than the 4M2L standard method. to form measurement line µnchn, and ch4 is the output
Before testing experimentally the three-microphone source signal. For correcting the measured transfer
configuration, a singularity specific to the 3M2L method is functions for amplitude and phase mismatches between the
underlined. Combining Eqs. (5) and (6), one can show that three measurement lines, the sensor-switching technique as
T is not determined when described in ASTM E2611-09 is used. Here line µ1ch1 is
cos(kDa)sin(kDb)=cos(kDb)sin(kDa). To avoid this, a the reference line. Consequently the calibration is
condition on the difference between the depths of the successively made between µ1ch1 and µ2ch2 and between
cavities must be fulfilled. This condition is |Da- µ1ch1 and µ3ch3 using microphone positions 1 and 2. For
Db|<π/k≈172/fu, where fu is the upper frequency limit of the the 4M2L method, the one-microphone two-channel
tube in hertz to ensure plane wave propagation. configuration given in Table 2 of ASTM E2611-09 is used.
In this case, the output source signal on ch4 is the transfer
function reference. Here, µ1 is connected to ch1 and moves
3 Experimental setup successively to microphone locations 1 to 4 for measuring
the required eight transfer functions (i.e., four for each
To implement the 3M2L method, a 44.45-mm diameter
load). Since the same microphone is used, no correction on
tube is used. The tube has three sections: upstream (235-
the transfer functions is required.
mm long), sample holder, and downstream. The
downstream section is terminated with a 30-mm thick
sliding piston – the piston acts as a hard wall and is made of 4 Results
steel. In the upstream section, two microphones are flush
mounted as shown in Figure 1. A third microphone is flush
mounted directly on the hard termination, as depicted in 4.1 nSTL of an air layer and calibration
Figure 1. The distance between microphone 1 and 2 is of the setup
s=25.2 mm and the distance between microphone 2 and the
front surface of the test sample is l=45.5 mm. Precise
measurements of these distances have been carried out
With a view to validate both the transfer function
using the method proposed by Katz [3]. This latter method
correction technique discussed in the previous section and
is based on the frequency determination of the nulls of the
the 3M2L test procedure, an air layer (d=80 mm) is first
transfer functions between microphones 1, 2 and
tested. The air layer is placed at the sample position shown
microphone 3, respectively H31 and H32. These transfer
in Figure 1. The sound transmission loss of an air layer is 0
functions are easily measurable using the proposed setup of

2031
23-27 April 2012, Nantes, France Proceedings of the Acoustics 2012 Nantes Conference

dB (at least on a small distance when neglecting losses) and

its transfer matrix is theoretically known. Its coefficients
are: T11=T22=cos(kd), T12=jZssin(kd), and T21= jsin(kd)/Zs.
The thicknesses of the two air cavities (i.e., the two loads)
are Da=25.4 mm and Db=50.8 mm. These cavities respect
the singularity condition of the 3M2L method (i.e. | Da - Db
|<172/fu Æ 25.4 mm<42 mm). Figures 2 and 3 present the
3M2L measurement results between 150 and 4100 Hz in
terms of the real and imaginary parts of the transfer matrix
coefficients respectively. Figure 4 presents the 3M2L
measurement of the nSTL.

With the used microphones, one can note that if the

transfer functions are not corrected, poor results are Figure 4: nSTL of the 80mm air layer.
obtained compared to the theoretical results. On the
contrary, using the standard sensor-switching correcting
technique, a good correlation is obtained between the 3M2L
and the theory. Moreover, one can note that the symmetry These first results show that using the standard sensor-
of the material is preserved with the 3M2L measurements, switching technique as described earlier is appropriate to
i.e. T11=T22. Nevertheless, a slight divergence can be correct transfer functions H12 and H13 of the 3M2L method,
observed at high frequencies on the nSTL (less than 0.3 dB) even if microphone 3 is used at 90° compared to
due to viscous losses around the movable termination, and a microphones 1 and 2.
loss in accuracy when approaching the lower frequency
limit of the system.
4.2 nSTL of an non-symmetric bi-layer
sample
A two-layered poroelastic sample made up from a 51-
mm thick melamine foam (porosity: 0.980±0.005 and
airflow resistivity: 10800±100 N.s.m-4) and a 37-mm thick
mineral wool (porosity: 0.970±0.005 and airflow resistivity:
18000±500 N.s.m-4) is used to validate the proposed 3M2L
method on a non-symmetrical sample.

New 3M2L
4M2L

Figure 2: Real part of the 80mm air layer transfer matrix.

Figure 3: nSTL of the two-layered sample. Comparison

between the results obtained with the proposed three-
microphone method (3M2L) and the general four-
microphone two-load method (General 4M2L).

Figure 3 shows that the two methods are in good

agreement. The deviation observed between the two
transmission loss curves around 700 Hz is due to the
change of the boundary conditions that affect the elastic
Figure 3: Imaginary part of the 80mm air layer transfer frame resonance of the sample. In fact, the sample was
matrix. reinstalled between the 3M2L and 4M2L tests.

2032
stats noitacilbup weiV

Proceedings of the Acoustics 2012 Nantes Conference 23-27 April 2012, Nantes, France

5 Conclusion
This paper presented a simple three-microphone
configuration which complies with standard ASTM E2611-
09 for measuring the sound transmission loss and transfer
matrix of acoustical materials. This configuration uses three
microphones in a slightly modified classical impedance
tube, the third microphone being flush mounted on the
movable hard termination. The standard sensor-switching
technique is used without any modification, even if
microphone 3 is used at 90° compared to the other 2
microphones. ASTM E2611-09 can be complemented with
this additional three-microphone configuration, if its Eq.
(22) is used with Eq. (6) of this paper.

Acknowledgments

This work was supported in part by grants-in-aid from the

National Sciences and Research Council of Canada
(N.S.E.R.C.).

References
[1] ASTM E2611-09, “Standard test method for
measurement of normal incidence sound transmission
of acoustical materials based on the transfer matrix
method,” American Society for Testing and Materials
(2009)
[2] Y. Salissou, R. Panneton, O. Doutres, “Complement to
standard method for measuring normal incidence sound
transmission loss with three microphones”, J. Acoust.
Soc. Am. 131(3), EL 216 (2012)
[3] B.F.G. Katz, “Method to resolve microphone and
sample location error in the two-microphone duct
measurement method,” J. Acoust. Soc. Am. 108, 2231-
2237 (2000)

2033