Professional Documents
Culture Documents
Vkronique Adam
LEMA (EPFL), Lausanne, Switzerland
Presented at AUDIO
the 108th Convention
2000 February 19-22
Paris, France -
This preprint has been reproduced from the author’s advance
manuscript, without editing, corrections or consideration by the
Review Board. The AES takes no responsibility for the
contents.
Veronique Adam
Laboratory of Electromagnetism and Acoustics (LEMA)
Swiss Federal Institute of Technology (EPFL), Ecublens, Lausanne, Switzerland
veronique.adam@epf.ch
Abstract
In an outdoor active noise control project, the need arose to design a loudspeaker system
based not only on its amplitude response but also on its phase linearity and therefore on its
group delay. This study presents different analyses and synthesesof both closed and vented
loudspeaker systems, adapted to this rather particular requirement rarely considered during
the design of conventional loudspeaker systems,but which could have implications for their
future design.
1. Introduction
W ithin the framework of an outdoor active noise control project, the specifications of the
anti-noise electroacoustic sources - closed-box or vented-box loudspeakers - contained an
essential requirement relating to the quality of the phase response of the systems. The
methods of analysis and synthesishave therefore to be adapted.
The first stage of this study was to calculate, from a given excitation, the pressure responses
(amplitude, phase and group delay) of vented and closed-box loudspeaker systems.
The second stage was to write under Matlab, different standard analyses of systems for a
given driver. These calculations were validated through the design, construction and
characterisation in an anechoic chamber of closed-box and vented loudspeaker systems.
Finally, the last stage involved the synthesisof a system that meets all the specifications.
The design of loudspeaker systems is based on its equivalent acoustical circuit, valid in the
low-frequency range, and determined from Thiele and Small’s small signal parameters
(Figures l-2). This representation readily enables the volume velocities of the system to be
calculated as a function of frequency, e.g:
The calculation of the far field sound pressure was carried out without using the radiated
acoustic power, with the distinctive advantage of obtaining the phase response. This way of
proceeding can be generalised to other more complex source models.
In the case of a vented-box, the computation of the phase excludes the use of the hypothesis
of coincidence of the radiating openings usually adopted. The pressure is therefore calculated
via the principle of superposition, by adding pressuresrelating to the different source volume
velocities. W ithin the framework of this study, the effect of the mutual coupling between
sourcesis not taken into account.
The pressure at the point of observation (farfield - 4nsteradian) is therefore given as a
function of the volume velocities of the system:
Zkc e(-jkd)
l For a closed-box: PC =4d _ j4&
l For a vented-box:
Z,k e(-,k’d+0.5/cos(~)-,,,, )) Z,k (-/~(d-oslcos(Q)-o,,jj
+9 e
PP =ydP 4d(ka, -j) --P 4d(ka, -j)
An indication of the quality of the phase responseof the system is given by calculation of the
group delay, which corresponds to the reaction time of the system. This delay is calculated as
the w phase pressure derivative, without bringing into play the propagation phase rotation,
e.g.
z = -dq/dm [2] with cp the pressure argument.
This delay, which is typically in the order of a few ms at the amplitude maximum around the
closed-box resonance (or anti-resonance of a vented system), can not be neglected in the
integration of the system in an open loop active control chain. At comparable performance,
this delay is in general higher with vented systemsas opposed to closed-box systems.
3. Analyses
The study began with the implementation - under Matlab - using the usual analyses [l] of
vented and closed-box loudspeaker systemsfor a given driver. This theoretical part led to the
designing of boxes which were built with the aim of verifying experimentally the accuracy of
the calculations.
The parameters of the chosen driver were measured on an IEC baffle and in test closed-box:
S,= 0.0142 m* V,, = 0.01518 m3 f, = 70 Hz Re = 3.8 L2 Q, = 0.63 Q “,,,= 2.1 Q,,,= 0.49.
The acoustic power is fixed at 0.001 W and an estimation has to be made of certain
parameters: R, = 0.5 R p = 1 Q, = 15. It was possible to simplify the construction phase of
the boxes by choosing an identical volume of 12.11 dm3 for both systems (vented and closed).
The design of the vent was, for its part, carefully determined according to the available space,
whilst avoiding any undesirable coupling or vent noise: up= 2.5 cm et $= 14 cm.
The amplitude, phase and group delay measurementswere carried out m an anechoic chamber
at a distance of 1 metre from the source with a voltage between terminals of 1 V. The
measured and calculated results (Figure 3) are in relatively close agreement for frequencies
above 50 Hz (resolution limit of the instrumentation in the anechoic chamber).
4. Syntheses
Once the fundamental calculations were validated experimentally, the study focused on the
different synthesis procedures for vented and closed-box systems, according to the values
imposed on the system and chosen amongst the supernumerary data in the specifications:
These syntheses lead to design solutions that still have to fulfil the technical feasibility
requirements of the some of the parameters linked to the loudspeaker, amplifier as well as to
the open loop active control:
l group delay: z c 10 ms
0 mass of the moving system: mS > 60 g
0 electrodynamic coupling coefficient: BZ c 25 Tm
l peak electrical power: PC< 500 W
The value of ms is fixed with the aim of keeping the most rigid cone possible, whilst
maintaining the values of BZ and the electrical power in order to use an existing magnetic
circuit and amplifier.
In view of the unusually high number of imposed values, the design of the system lead to
numerous iterations seeking to come as close as possible to the requirements. Other types of
syntheses can be carried out by modifying the imposed values and those that need to be
verified, according to the importance given to the various values. Thus, in this application
where the peak electrical power can not be greater than 500 W, this specification was added
as a given value of the synthesis in place of the cut-off frequency which in turn became a
parameter to be verified.
The solution found corresponds to a vented-box that has to verify the following parameters:
As was be expected, the solution retained does not respect the maximum value of the cut-off
frequency. However, we have decided thoughtfully to admit this value which bears no
consequencesfor this particular application. The curves exhibited in Figure 4 represent the
pressure responses(amplitude and phase) and group delay.
5. Conclusion
The syntheses have to be rethought through each time new requirements are made, such as
phase linearity. No synthesis can be conceived from an overdetermined system. Due to this
fact and with the aim of obtaining solutions that can be implemented, the requirements have
to be sorted in such a way as to keep the most important ones as formal given values of the
synthesis, whilst still leaving a small range of variation on the values of the other
specifications. According to this method, our industrial partner developed a loudspeaker
system fulfilling the requirement of the group delay among the others (see paper entitled “
Current controlled vented box loudspeaker system with motional feedback” by P. Robineau
and M. Rossi).
6. Glossary of terms
1 compliance factor
kg/m3 air density
S group delay
rad/s angular frequency (2@ )
rad angle to the loudspeaker
rad pressureresponseargument
m radius of the projected surface
radius of the port surface
iIS speed of sound
distance to the loudspeaker
Ez frequency
HZ lower cut-off frequency (-3 dB)
Hz resonancefrequency (vented-box, closed box, baffle)
rad/m wave number
m distance between centre of the membrane and port
length of the port
pg/m” acoustic mass corresponding to the closed-box
kg/m4 acoustic mass corresponding to the port
kg/m” radiation acoustic mass (closed-box)
kg/m4 radiation acoustic mass of the driver (vented-box)
kg/m4 radiation acoustic mass of the port (vented-box)
kg/m4 equivalent acoustic mass of the moving system
kg/m4 m~s+ mm+ mob
kg/m4 mm+ mm,+ mab
kg/m4 map+ mw2
7. References
[II M. Rossi, “Acoustics and electroacoustics”, Artech House, Boston London, 1988
PI J.E. Benson, “Theory and design of loudspeaker enclosures”, Synergetic Audio
Concepts, 1993
8. Equations
U*Bl
R, = ’
‘, = Sd(R, i-R,) Qmac,
Bl* 1 1 1
47, = R,, =- Qb mlac= - mVnoz - mVop= ~
Sd2V$ +&.I Ob ‘ab %*cBc ~s2cns wb2c,b
teq. 11
P
-8
4 =
-dp teq. 21
1 h-m’,
R,, +R, + johao+-+
in,
l+(jw)2(m’, Cab)+ jco>
0,
-4dp
4
-P
= teq. 31
(jw)2(m’OpC,b)+ jcoy+l
0,
9. Figures
CLOSED-BOX VENTED-BOX
/
dB
I
,,
Hz HZ
50 80100 200 300 500 50 80100 200 300 500
Phase
HZ
sO”:r;j
180y-- ,
a) I I b)
110
109
108
106
dB
c) 0.015 /
S
Figure 4 Calculations a) pressure level at 1 meter of the port, the driver and the sum b) pressure phase
without the propagation of the port, the driver and the sum
c) group delay without the propagation