Diaphragm Area and Mass Nonlinearities of Cone Loudspeakers 4082 (I-2)

Erling Sandermann Olsen

Technical University of Denmark
Lyngby, Denmark

Knud Thorborg
Peerless Fabrikkeme A/S
Karlslunde, Denmark

Presented at ^uD,
the 99th Convention
1995 October 6-9
New York
Thispreprinthas been reproducedfrom the author'sadvance
manuscript,withoutediting,correctionsor considerationby the
Review Board. The AES takes no responsibilityfor the
contents.

Additionalpreprintsmay be obtainedby sendingrequestand

remittanceto theAudioEngineeringSociety,60 East42nd St.,
New York,New York 10165-2520, USA.

All rightsreserved.Reproductionof thispreprint,or any portion

thereof,is not permittedwithoutdirectpermission from the
Journalof the Audio EngineeringSociety.

AN AUDIO ENGINEERING SOCIETY PREPRINT

 Diaphragm area and mass nonlinearities of cone loudspeakers

ERLING SANDERMANN OLSEN, AES member

The Acoustics Laboratory, Tecnieal University of Denmark

and

KNUD THORBORG, AES member

Peerless Fabrikkerne A/S, Karlslunde, Denmark

In the investigation of mechanical parameter nonlinearities of low frequency

loudspeakers with 'single-roll' suspension a systematic variation of the mass with
diaphragm position has been revealed. The variation turns out to be associated
with a variation in the effective area of the diaphragm. The variations are
described and documented and some consequences of the variations are discussed.
The work presented is a part of the Danish LoDist project.

0 INTRODUCTION

In the investigation of nonlinearities of electrodynamic loudspeakers it is normally assumed that

the diaphragm area and the mass are constants that do not vary with the diaphragm position.

In the Lodist project [1] different methods for determining the position dependence of the
mechanical parameters have been investigated. At an early point it became clear that the dynamic
mass of the diaphragm of the loudspeaker units investigated varied significantly with the
diaphragm position, and that consistent results could not be obtained if the same diaphragm area
was used in the calculations for all diaphragm positions. These nonlinearities have been
discussed during the project.

The most common geometrical configuration of the outer suspension of modem woofers is the
"single roll", that is, the cross section of the suspension is, or is close to, a semicircle. The
loudspeaker units that have been used in the investigations of measurement methods were units
with this type of suspension. The observed nonlinearities lead to the geometrical considerations
of the behaviour of the the single roll suspension that are presented here.

The considerations presented in this paper are all restricted to frequencies at which the
suspension vibrates in phase with the diaphragm, that is, below the resonance frequency of the
suspension. The term 'suspension' is used for the outer suspension throughout the paper,
although the outer suspension often only is a part of the diaphragm suspension.
E. S. Olsen and K. Thorborg: Diaphragm area and mass nonlinearities 2

1 THE SINGLE-ROLL SUSPENSION

The geometry of a typical single roil suspension is shown in figure 1. Typical materials used in
these suspensions are rubber, foam and fabric. These materials can bend, but normally they can
only be elongated a very little. Thus, the diaphragm can move between the two positions where
the suspension is totally stretched. In practical loudspeakers the movement is limited to a smaller
range between these positions because of the thickness and bending stiffness of the material.

In figure 2 the geometry of the suspension is shown schematically at the rest position and at the
two maximum excursion points. At the the rest position the suspension has the shape of the half
of a toms. The volume of air under this shape is (see e.g. [2])

Vsus
p = _2Rr2 (1)

When the diaphragm moves from the rest position to the outer extremum point the volume of
air moved is the difference between the volume of the frustum of a cone that is formed by the
suspension and the volume under the suspension at the rest position

Vout = _2rI___l(3R2+r2 ) _ _2Rr2 (2)

Similarily, when the diaphragm moves from the rest position to the inner extremum point the
volume of air moved, Vi,, is the (negative) sum of these volumes. For a typical 6_h'' woofer
with a nominal diaphragm diameter of 150 mm and a supension width of 10 mm (R = 75 mm
and r = 5 mm) Vo_,is 196 em3 while Vi, is 233 cm3, a difference of 37 em 3 or 17 %.

The analysis above shows that the volume of air moved when the diaphragm is moved to an
extreme position is rather different in the two directions. This indicates that the effective
diaphragm area changes with the diaphragm position, but not how the area changes with the
position. When a diaphragm with a single roll suspension is moved slowly it appears that the
suspension does not move as a whole but rather 'rolls' around a bend so that the suspension is
divided into a part that moves with the diphragm and a part that does not move. In this way the
diaphragm area changes from the area of the diaphragm alone to the area of the diaphragm and
the suspension together when the diaphragm-is moved between its extreme positions. If the
change in the diaphragm area is assumed to be proportional to the excursion, the diaphragm area
is

Sd(x) = _R 2 - _----f---R
x

(3)

If this assumption is true the volume of air that is moved can be calculated by integrating the
expression from x = 0 to the actual position. The volumes of the 6_/z" woofer from the example
above calculated in this way are 200 em3 and 228 em 3. The difference is a little smaller than the
E. S. Olsen and K. Thorborg: Diaphragm area and mass nonlinearities 3

difference calculated from the geometrical shape, that is, the area may be assumed to change
with a higher slope near the mst position and with a lower slope near the maximum excursion
points. Note that the difference between an assumption of linear diameter variation and the
assumption of linear area variation is very small.

Associated with the area variation will be a variation in the radiation mass of the diaphragm. The
radiation mass per side for a piston with radius a in an infinite baffel is [3]

mr
= p-_a 3 (4)

In the case of an unbaffied piston the radiation mass is half that value [3]

The radiation mass of the diaphragm from the example above varies from 1,051 g to 0,657 g
in the case of an unbaffied unit or from 2,102 g to 1,304 g when the unit is mounted in a baffle.

A much larger mass variation is associated with the way the suspension moves. As mentioned
above the suspension seems to roll around a bend that divides the diaphragm into a part that
moves with the diaphragm and a part that does not move. Thus, only the mass of the moving
part of the suspension will contribute to the dynamic mass of the diaphragm. The mass of the
suspension is normally a significant part of the total dynamic mass of the diaphragm. For a
typical 6'A" woofer with a rubber single roll suspension, the (small signal) dynamic mass is 15 g
and the mass of the roll of the suspension is 4 g. Thus, the mass should be expected to vary
from 13 g to 17 g between the two maximum excursion points. A first approximation is a linear
variation with position of the mass

nh(x) = mo+px (5')

2 EXPERIMENTAL VERIFICATION

In the LoDist project the mechanical nonlinearities of a number of woofers have been
investigated with a direct mechanical measurement method [4]. In this method the mechanical
parameters can be determined with high accuracy at different diaphragm positions, and it is not
necessary to assume linearity of any of the parameters. In figure 3 the dynamic mass of a 6_//''
woofer with single roll suspension is shown as a function of diaphragm position. The mass
variation with diaphragm position is clearly close to be a linear function. The mass of the roll
of the suspension of this loudspeaker unit is approximately 6 g and the radius of the roll is 5.5
mm. The (theoretical) maximum excursion of this suspension is 13.3 mm. The extrapolation of
the line to +13.3 mm actually results in a variation that is very close to 6 g.

In order to verify the diaphragm area varition a supplemental experiment has been carried out.
After the measurement of the mechanical parameters of the unbaffled unit, a small box was
mounted on the unit and the parameters were determined again. Knowing the compliance of the
loudspeaker unit and the volume of the box the diaphragm area of the loudspeaker unit can be
calculated from the measured compliance. In figure 4 the area measured in this way is shown
E. S. Olsen and K. Thorborg: Diaphragm area and mass nonlinearities 4

as a function of the diaphragm position. Also shown in figure 4 is the diaphragm area calculated
from (3). From approximately x = -2 mm to x = 6 mm the calculated line seems to be a good
approximation to the measured data. At values of x below -2 mm the area does not change as
rapid as expected from the geometrical considerations. At extreme positive values ofx the area
changes a with a little higher slope than the slope of the tine. The difference between the
measuremem results and the results of the geometrical considerations are reasonable. The robber
of the suspension has a finite thickness and the diaphragm is not infinitely rigid. At extreme
negative values of x the diameter can not increase further and the movement is rather due to
elastic changes in the suspension. In the opinion of the authors, at present a linear relation is a
good approximation to the actual changes in the surface area, the slope of line should be
determined by linear regression to measured data.

The results are similar for all the single-roll suspension loudspeakers investigated in the LoDist
project, although the magnitude of the effects vary for the individual units depending on the
weight of the suspension and the details of its geometrical configuration.

3 THE INFLUENCE OF THE NONLINEARITIES

The nonlinearities of the dynamic mass and the diaphragm area of electrodynamic loudspeakers
have some consequences for the equations of the lumped parameter model of the loudspeaker
unit.

When the mass is not a constant but a function of the diaphragm position the change in the
momentum must be considered when the force across the mass is calculated. The force across
the mass is then

F= _t (mv) = --_v:+m dv
-dr _ v2+ma (6)

The effects of the area nonlinearities are more complex. Since the diaphragm area is the
radiating surface, the area variation also results in a diaphragm position dependence of the sound
pressure generated by the loudspeaker. The sound pressure at distance l when the loudspeaker
is radiating into solid angle fl is

p R d (VSd) = -- (7)

When the unit is mounted in an enclosure the compliance of the enclosure will also be affected
by the area nonlinearities. If it is assumed that the box compliance is much smaller than the
suspension compliance the sound pressure generated at frequencies below the resonance
frequency is
E. S. Olsen and K. Thorborg: Diaphragm area and mass nonlinearities 5

where F is the force applied by the B1 motor. Here it is assumed that the change in the box
volume because of the diaphragm movement can be ignored. If the box is very small this is not
the case. If, on the other hand, the box is large it can no longer be assumed that the box
compliance is much smaller than the suspension compliance.

From the equations above it can be seen that the diaphragm area and mass nonlinearities leads
to distortion. The v2 term in (6) and (7) and the Fv term in (8) indicates that even order
distortion and a shift in the mean position (a 'DC-shift') of the diaphragm is generated when a
signal is applied. Since the area and the mass are factors in the expressions, the nonlinearities
will also lead directly to intermodulation distortion. In order to calculate the distortion, the
complete set of equations for the loudspeaker unit and the box must be set up and solved.
However, this is beyond the scope of this paper. The calculations are under preparation in the
LoDist project.
t

4 DISCUSSION

As shown in the previous sections, the single roll suspension type often used in electrodynamic
loudspeakers have some inherent nonlinearities additional to the well known nonlinearities of the
compliance and the losses. The diaphragm area and the dynamic mass of loudspeakers with a
single roll suspension depends on the diaphragm position. The knowledge of these nonlinearities
is important for the understanding of the behaviour of loudspeakers with this type of suspension.
The mass and area nonlinearities contribute to the distortion from the loudspeakers, and should
be taken into consideration when the distortion is calculated. Of course, the importance of these
nonlinearities depends on the magnitude of the generated distortion and on the interaction
between the mass and area nonlinearities and other nonlinearities in the loudspeaker.

At frequencies above the resonance frequency of the system the mass and area nonlinearities will
to some extend be self-compensating. At these frequencies the sound pressure is proportional to
the ratio between the diaphragm area and the mass [3]. If the relative changes in the mass and
in the area are the same the sound pressure will not be affected, that is, a large amplitude low
frequency will not modulate a low amplitude high frequency. Similarily, since the signs of dm/dx
and dS/dx are the same, the v2 term of the force across the mass will tend to ctmpensate for the
v2 term of the sound pressure. An interesting aspect of this is that the diaphragm area
nonlinearity will not be compensated for with a motional feedback system. On the contrary,
since the motional feedback system keeps the acceleration proportional to the input voltage at
frequencies above the cutoff frequency, the compensation will be cancelled out by the system.

At frequencies at and below the resonance frequency the effect of the nonlinearities is more
complicated. The total effect will depend on the ratio between the box compliance and the
suspension compliance, and it is not simple to evaluate.

It may be possible to change the magnitude of the nonlinearities described in this paper by
changing the geometry of the roll. Also, other basic geometries with two or more bendings may
reduce the nonlinearities. However, many other aspects must be taken into consideration in the
design of the suspension for a loudspeaker unit. With a well designed single roll suspension the
resonance of the suspension can be kept under control, and the necessary braking of the motion
E. S. Olsen and K. Thorborg: Diaphragm area and mass nonlinearities 6

before the voice coil is stopped by the bottom of the motor system can be achieved in a way
where only moderate distortion is generated. This is difficult to achieve with other suspension
configurations. Thus, the single roll suspension may still be the best compromise in suspension
design. The purpose of this paper is not to say that the single roll is a bad suspension design,
but to present the presence of the nonlinearities so that they can be taken into consideration in
future discussions on loudspeaker design.

An important question related to the nonlinearities described in this paper is of course whether
the distortion due to these nonlinarities is audible. The audibility of the effects described is one
of the subjects that will be investigated in the listening experiments of the LoDist project.

5 CONCLUSIONS

In this paper two types of nonlinearity inherent to the single roll suspension often used in
electrodynamic loudspeakers have been discussed.

The effective area and the dynamic mass of a loudspeaker diaphragm with single roll suspension
varies with the diaphragm position.

The nonlinearities discussed cause distortion, but it might to some extent be possible to balance
out this distortion.

The nonlinearities should be included in the equations of the loudspeaker model.

The single roll is a good geometrical configuration, it should not necessarily be avoided because
of the nonlinearities described in this paper. It is possible that a modification of the geometry
could lead to a better compromise between the desired properties and the associated
nonlinearities.

6 ACKNOWLEDGEMENTS

The authors wish to thank the other partners in the LoDist project, Bang & Olufsen, AudioNord,
Dynaudio and Vifa-Speak who, by joining the project, have made the work possible and who
have contributed with many interesting details in the discussions of the subject of this paper. The
authors also wish to thank Loc A Han and Lars E. Nyegaard for letting us use the remits from
measurements they have made during M.Sc. courses on the Acoustics Laboratory and at
Peerless.

8 REFERENCES

[1] Press release about the LoDist project. Published in J. Audio. Eng. Soc. vol.42, p. 286
(1994 april).
.E.S. Olsen and K. Thorborg: Diaphragm area and mass nonlinearities 7

[2] M.R. Spiegel, "Mathematical Handbook of Formulas and Tables", McGraw-Hill Book
Company, pp. 9-10 (1968).

[3] L.L. Beranek, "Acoustics", American Institute of Physics, Chapters 5 and 7 (1986).

[4] E.S. Olsen, "Measurement of the mechanical parameter nonlinearities of electrodynamic

loudspeakers", presented at the 98. convention of the Audio Engineering Society, J.
Audio Eng. Soc. (Abstracts), vol. 43, p. 400 (1995 May), preprint 4000.

Figure 1. Cross section of a typical single roll suspension, a) chassis, b) suspension.

c) diaphragm.
 I
Z
I

Figure 2. The geometry of a diaphragm with single roll suspension at the rest position and
at the two maximum excursion points.
 25

o_
20 __
E
._o
E
m 10
c-
O

0 i i i i i i i
-10 -8 -6 -4 -2 0 2 4 6 8 10
Diaphragm position, mm

Figure 3. Measured dynamic mass of 6_h" woofer.

2OO
180

160
<
E 140

120-
100
E
80
,,¢:
_- 60
.m_
c3 40
20

0 i i _ i i i i
-10 -8 -6 -4 -2 0 2 4 6 8 10
Diaphragm position, mm

Figure 4. Measured diaphragm area of 6_h" woofer. Also shown (bold) is the straight line
of expression (3).