Loudspeaker Manifolds for High-Level Concert Sound Reinforcement

LOUDSPEAKER
MANIFORLDS FOR HIGH-LEVEL
CONCERT
SOUND
REINFORCEMENT
David
David
2387
(D-ii)
Carlson
Gunness
Electro-Voice,
Inc.
Buchunan,
Michigan
Presented at
the 81st Convention
A U D IO
1986November 12-16
Los Angeles, California
Th/spreprinthas beenreproducedfrom theauthor'sadvance
manuscript,withoutediting,correctionsor consideration
by
the ReviewBoard.TheAEStakesno responsibility
for the
contents.
Additionalpreprintsmaybe obtainedby sendingrequest
and remittanceto the AudioEngineeringSociety,60 East
42ndStreet,New York,New York10165USA.
All rightsreserved.Reproductionof thispreprint,or any
portionthereof,is notpermittedwithoutdirectpermission
fromtheJournalof theAudioEngineering
Society.
AN AUDIO ENGINEERING SOCIETY PREPRINT
LOUDSPEAKER
FOR HIGH-LEVEL CONCERT
MANIFOLDS
SOUND REINFORCEMENT
By:
David Carlson and David Gunness
ABSTRACT
The acoustical
manifold is presented as a device for improving
the performance of high-level concert sound reinforcement speaker
systems.
Recently developed bass, midbass,
and high-frequency
devices are discussed,
each of which effectively sums the output
of four loudspeakers,
producing from the four a single coherent
source. The advantages of manifolds over multiple
sources
are
improved audience coverage, reduced polar response lobtng, smallfrontal area arrays,
small transportable
enclosures,
and in
certain cases reduced distortion and increased efficiency.
INTRODUCTION
High-level
concert
sound reinforcement
places the greatest
demands on current loudspeaker technology in the area of maximum
acoustic
power output. The sound pressure levels and component
reliability
required have been attained in the past by creating
large arrays
of sources.
This practice reduces coverage pattern
control
and produces polar-response lobtng.
In addition,
the
practice
results
in large-frontal-area
arrays and bulky
loudspeaker systems to transport.
By combining the outputs of multiple drivers within a carefully
defined space,
a manifold,
these effects can be reduced. The
result is extremely high-level sound with minimized
distortion,
small-frontal-area
arrays,
small transportable
enclosures,
greatly
reduced
directive response anomalies,
and in certain
cases increased efficiency.
BACKGROUND
Manifolding is not a new idea. In the lg40's there was an RCA
Y-throat
for coupling two one-inch-throat drivers to a special
horn. This and various other Y-throats have been used with some
success in siren and voice warning systems, where the redundancy
of drivers is a decided advantage. Their use in quality musical
sound reinforcement
has been limited primarily
to midrange
applications
hecause of cancellations
at high frequencies.
Figure ! is the summation response _ of a p.opular large Y-throat
(drivers with 4gmm exits into a special horn), which reveals why
many who have experimented with Y-throats describe them as "dullsounding". We submit that this performance
is characteristic
o_
specific manifold designs and is not indicative of the potential
of manifolding
in general.
The general concept of manifolding
is hased on the prediction
that two like drivers mounted on a manifold will be indistinguishable from a single driver with twice the diaphragm area and
an equivalent motor. The two-driver version has the advantages of
smaller diaphragms (which are capahle of pistonic motion through
the passband), greater power-handling (thermal-limit) per square
centimeter of diaphragm, and greater reliability.
Furthermore, Henricksen [1] has shown that given currently known
materials,
a very abrupt limit is reached in the ultimate
performance
of compression
drivers.
As a result,
a given
system
requirement
may call
for
a compression
driver
that
is
not
physically
realizable,
necessitating
the use of multiple
drivers.
Compared to multiple
horns,
a smaller
number of horns utilizing
manifolds
will
produce
less
polar
response
lobing
and better
audience
coverage.
They will
also
form smaller
arrays
and cost
less tO implement.
But what happens
when "like"
drivers
vary from one another,
as
all
real
drivers
will?
At low frequencies,
that is frequencies
at
which there
is mutual
coupling
between driver_,
there is usually
very little
unit-to-unit
variation
between drivers.
Furthermore,
the type of variations
that
occur
generally
sum mathematically
(and acoustically)
to an output
that
is between
that
of the two
contributors.
At high frequencies
on the other
hand, compression
drivers
may show significant
unit-to-unit
variations,
not
1.
Summation
response,
as used
herein
ts defined
as the
difference
between
the 1/3-octave
smoothed
response
of a
horn-driver
combination
and the 1/3-octave
smoothed response
of the same horn with
a manifold
and similar
drivers.
The
reason
for
the smoothing
is to avoid
attributing
local
response
changes to the manifold
which are really
due to the
change in horn length
that
occurs
when a length
of throat
section
is replaced
with a manifold.
necessarily
in
average
level
but in the exact
frequencies
at
which
particular
peaks
and
dips
occur.
Simple
mathematical
summation
of
randomly
selected
drivers
may yteld
a frequencyresponse
that
is
worse
than
the
individual
contributors.
Fortunately
though,
as unit-to-unit
variations
become significant
at
higher
frequencies,
directive
behavtor
(ray
propogation)
within
the manifold
occurs.
It has been found that
a manifold
which effectively
transfers
sound "rays",
as well
as sound waves,
performs
well even with unmatched drivers
and up to much higher
frequencies
than one which only addresses
wave behavfor.
The
discussion
so far has been concerned
with
manifolding
of
high-frequency
compression
drivers,
since
historically
this
has
been the most prevalent
application
of manifolding.
However,
the
concept
has
been
found
to
be equally
valuable
at
lower
frequencies,
where the benefits
are not as Intuitively
apparent.
After
a compression-driver
mantfold
design
bas been described,
a
mid-bass
and then a low-frequency
mantfold
will
be discussed
to
Illustrate
the
unique benefits
of this
design
approach
over
a
wide range of frequencies.
DESIGN OF A MANIFOLD FOR FOUR HIGH-FREQUENCY
COMPRESSION
DRIVERS
A manifold
has been developed utilizing the concepts mentioned
above
as well as several others to achieve a summation
response
suitable for critical applications.
The drivers are arranged
as
close
together
as possible
to maintain
small-cross-section
acoustic
paths. The wavefronts entering the manifold
encounter
reflective surfaces,
rather than curved tubes. Additionally, two
different entry-angles are used to obtain a smoother summation.
As a starting
point, an array of drivers was created which
produced the minimum possible spacing of the driver throats. This
minimum
spacing occurs
when the four drivers are arranged
radially with magnets nearly touching and exits facing a central
point
(like spokes of a wheel).
Two facts become
immediately
obvious.
First,
a relatively severe bend must occur within the
manifold
(a 90-degree bend for the closest exit spacing).
And
second, convex-drive
compression drivers are inherently
better
suited to this type of design than equivalent
concave-drive,
through-the-pole-piece
drivers,
since the convex-drive
magnets
are behind the diaphragms,
permitting closer packing of driver
exits.
Using the concept
of ray tracing,
a right-angle
reflective
bend
was modelled
and found to work very well,
as shown in Figure
2.
This
curve
shows
the difference
tn
frequency
response
of
a
compression
driver
and horn measured with and without a rightangle reflective bend in a round:cross-sectton
connecting
tube
between the horn and driver.
This data shows that by using an
appropriate
reflector the bend can be traversed with negligible
loss of response. However,
when several of these reflectors are
brought together
in a manifold,
the path of each wavefront
leaving tts reflector is not constrained hy a tube, as it ts in
the pure case. The resulting summation response is not as smooth.
The solution to this unevenness of response is based on the
observation that the frequencies of cancellation may be predicted
with a loosely related formula. The off-axis response of an ideal
piston,
as presented
in Kinsler and Frey [2], shows a good
correlation with these cancellation frequencies if a piston twice
the size of the input tube is assumed.
The ideal-piston off-axis
response is described by:
p =
A
·
2.Jl(k.a.sinO)
k.a.sln_
(1)
in which;
a = the radius of the piston (use twice the
radius of the tube),
t - the angle of the bend,
k = the wave number (w/c),
A = the unperturbed response (same source without bend),
Ol(x) = the first order Bessel function of the first kind.
The important aspect of this observation is that it holds for
changing angle,
i. Furthermore,
if the function ts plotted for
various angles,
it may he seen that certain
pairs of these
functions appear to he complementary - dips in one correspond to
peaks tn the other. Consequently, the manifold design was refined
one step further by tilting two of the drivers 45 degrees, which
still maintains the minimum exit spacing. Figure 3 shows a fourdriver manifold with this configuration - notice
the different
angles of entry.
Equation I ts plotted in Figure 4 for a 23.6mm
tube (47.2mm piston) at angles of gO and 45 degrees.
It shows
that the first two nulls in the gO-degree response correspond
to
lobes in the 45-degree response.
Likewise, the first null in the
4S-degree
response
corresponds
to a lobe in the gO-degree
response.
If the contributions
from both sets of drivers were highly
correlated and well represented by equation (1), their summation
would
be similar
to either contribution
singly
(the sum of
equation
[1] for two different
angles
is another
function
representable
by equation
[1]).
Fortunately,
though,
the
combining waves
have some directionality
and the individual
contributions
have much shallower dips and very little highfrequency roll-off. As a result, the two characteristics do have
a general smoothing
effect, which
is further augmented
by
adjusting the shape of the reflective detail to distribute
highfrequency energy evenly over the area of the manifold exit. The
final result is the summation response in Figure 5. A photograph
of a prototype manifold showing
the driver
arrangement
and
reflective interior detail appears in Figure 6.
DIRECTIVE PROBLEMS ASSOCIATED
WITH LARGE-THROAT CONSTANT DIRECTIVITY
HORNS
As a result of manifolding multiple drivers, the entrance to a
horndesigned
to accept the manifold must be larger than it would
have been for a single driver.
In the case described
above,
changing from a 23.6mm entrance to a 4gmm entrance caused
some
directive anomalies
which required a more complex
horn throat
design.
Figure
7 is a beamwtdth plot that is typical of constant
directivity
horns with 49mm entrances;
The curious behavior
at
the upper end of its range is an on-axis
dropout causing
an
apparent widening of coverage in the octave around 10.5 kHz. The
collapsing
beamwidth
ahove 15 kHz is due _:o the excessive
diameter of the throat (if the throat is too large, the initial
dispersion
is narrower than the coverage angle of the horn, so
the horn walls have very little effect on the beamwidthJ.
Figure
8 shows the on-axis dropout.
Once again,
ray tracing provides the explanation.
A circularcross-section
wavefront
encountering a change to a rectangular
cross-section
will "attempt" to redistribute its energy so that
it is distributed rectangularly. Since the area of a rectangle is
distributed
more toward the corners (as opposed to the center),
the wavefront acquires an intensity component that is directed
across the desired propagation path, toward the sides of the
horn. When this component reflects off the straight walls of a
constant-directivity
horn. an interference
pattern
develops,
resulting in on-axis dropout. Figure g shows, graphically,
the
paths
of
the
rays.
This
postulate
has
been
supported
experimentally
as it predicts,
accurately,
the dropout
frequency
for varying
wall
angles and horn entrance
diameters.
For a 4gmmthroat
horn,
this
frequency
varies
from 7 kHz in a 60-degree
waveguide
to 16 kHz in a 20-degree
waveguide.
Once the
mechanism
causing
the
problem
was understood
the
solution
was straightforward.
A pair of vanes positioned
in
the
throat
of
the
horn
cures both of the
directive
problems
by
creating
a sectoral
horn for a few inches - just
long enough
to
allow
the wavefronts
to take on the shape of the horn, hut not so
long
as to
produce
the
response
ripples
characteristic
of
sectoral
horns.
The position
and size of these vanes is critical
to the solution
of the problem
and to the maintenance
of
smooth,
lossless
response.
DESIGN OF A FOUR-DRIVER MANIFOLD
FOR THE MID-BASS FREQUENCY RANGE
Some of the concepts used in the design of
manifold
were also applied to the design of
Close driver packing was utilized to minimize
cross-sectional
areas of the manifold,
and
were used to achieve good summation,
the high-frequency
a mid-bass manifold.
the path length and
reflecting
surfaces
Before
the problems of manifolding could be tackled,
several
other problems had to be addressed. For a maximum-power system, a
driver was required that would provide flat frequency
response
from
150 to 2000 Hz on a constant-directivity
horn.
A
conventional high-power, 25cm cone-type loudspeaker was chosen as
the basic drive unit, and a special phase plug was designed.
Details of the development of the phase plug are beyond the scope
of this paper, but to summarize; the phase plug makes use of the
fact that at higher frequencies the amplitude of vibration
is
greater
near the voice coil than it is near the cone edges.
By
loading
the cone asymmetrically,
high-frequency
output is
maximized.
Additionally,
the phase plug has a rectangular exit,
making manifold
construction easier and permitting
a simpler
interface to a wood mtdbass horn with a rectangular-throat
opening.
As a demonstration
of mid-bass
manifolding,
a stngle
driver
mounted
on a mid-bass
horn (a 60-degree
by 40-degree
horn with a
mouth
area
of .47 square meters - see Figure
10)
was compared
with
two drivers
manifolded
on a similar
horn (also a 60-degree
by 40-degree
horn with a .47-square-meter
mouth
- see Figure
11).
In
the manifolded
system the drivers
face each
other
and
address
right-angle
reflective
bends.
The frequency
response
of
the single-driver
system is presented
in Figure
12 along with the
response
of the manifolded
system in Figure
13. The increased
low
frequency
level
of the manifold
system is due to mutual coupling,
while
the higher-frequency
differences
are due to a slight
change
in
directtvity
as a result
of the increased
throat
dimension
of
the manifold
horn.
It is apparent
that no significant
power-loss
occurs
in the manifold.
The
final
form of the system employs two of the above
manifolds
arrayed
vertically
and loaded by a sltghtly
larger
60-degree
by
40-degree
horn. The frequency
response
of this
system is shown in
Figure
14.
Note especially
the sensitivity.
This system, at 1200
watts continuous power handling capability, can deliver 138dB SPL
(144dB peaks) at one meter over its full frequency range.
DESIGN OF A LOW-FREQUENCY
SYSTEM UTILIZING A MANIFOLD CHAMBER
The advantages
of manifolding
woofers
are not
intuitively
obvious,
since phase cancellations and destructive interference
typically
are not problems at low frequencies (because of the
long wavelengths involved); however, several benefits are derived
that will become apparent later.
Over the years,
two basic loudspeaker designs have proliferated
in
the concert
sound industry
for the reinforcement
and
reproduction of low frequencies;
horn-loaded systems and directradiating
vented-box systems.
A horn-loaded
design usually
exhibits more efficiency
than a direct-radiating
vented-box
design, but at the expense of a larger enclosure.
When the
smaller direct
radiators
are assembled
into large
arrays,
however, their efficiency
increases due to mutual couplin9 and
begins to approach that of horn-loaded systems. Keele [3] has
shown that direct-radiating woofers hold an advantage over hornloaded woofers in acoustic power output per bulk volume occupied.
Consequently, a direct radiating vented-box design was chosen for
the low-frequency manifold.
To demonstrate the effects of manifolding woofers,
twff identical
vented-box
enclosures
containing a pair of 46 cm woofers were
built.
Each had
a net internal volume of 174 liters and was
tuned to 44 Hz. The two enclosures were initially positioned
side-by-side,
facing
forward as shown in Figure 15. Next the
enclosures
were turned so that the loudspeakers faced each other
as shown
in Figure 16. The space between
the enclosures was
sealed on three sides (top, bottom, and rear) so that the loudspeakers
radiated into a common chamber (the manifold chamber)
with a single exit, and the ports were relocated to the new front
of the cabinets
(outside the manifold
chamber).
Acoustic
measurements were made on both configurations.
The on-axis frequency response of the direct-radiating system is
shown in Figure 17 and the response of the manifolded system is
shown in Figure 18. Comparing the two curves,
there are several
differences that are readily apparent; below 70 Hz the manifolded
system has substantially more output than the direct-radiating
system; above 100 Hz the manifold has slightly less output; and
above 200 Hz the manifold rolls off abruptly.
The explanation
for the manifold's 3-dB advantage below
70 Hz
lies in the analysis of the mutual radiation impedance.
In the
classical analysis of the mutual radiation impedance between two
adjacent identical pistons in an infinite baffle, Pritchard
[4]
points
out that the resistive component is absolutely convergent
for all values of a (the radius of the pistons)
and d (the
center-to-center
distance
between
the two
pistons),
and
approaches that of a single radiating piston when the wavelength
is much larger than d. This effective doubling of radiation
resistance
gives us the well known 3-dB increase
in efficiency
when
the number of closely-spaced loudspeakers is doubled.
This
effect
holds equally
true for both the dire,ct-radiating and
manifolded systems, with the difference between the two designs
lying
in the reactive
component of the mutual
radiation
impedances.
Pritchard points
out that the reactive
component
increases
rapidly with decreasing frequency when the wavelength
exceeds the spacing d, and additionally, approaches infinity as d
approaches a. In the case of the manifolded system both of these
cases
are encountered.
As a consequence,
the additional mass
loading
provided by the manifold is significant enough to raise
the Q over the direct radiating case and increase the sensitivity
below 70 Hz. The design of the manifold is critical to limit the
additional mass loading to the low-frequency rolloff
region.
If
the loading extends
into the passband a decrease
in midhand
efficiency will result.
From 70 Hz to 100 Hz the manifolded
response
is very stmilar
to
the
direct-radiating
case.
As frequency
increases
the effect
of
the
additional
mass loading
due
to
manifolding
decreases,
producing
little
effect
ahove 70 Hz.
Ahove 100 Hz the
on-axis
sound-pressure
level
is slightly
higher
for the
direct-radiator
system
because
of
increased
directlvtty
due to
the
larger
radiating
area.
At yet higher
frequencies,
the manifold
chamher
acts as a low-pass
filter.
This high-frequency
rolloff
is of
no
consequence
for low-frequency
operation
(less
than 200 Hz),
but
offers
the
unexpected
benefit
of
significantly
reducing
distortion
in
the passhand by acoustically
attenuating
higher
harmonics.
A properly-designed
frequency
efficiency
internal
volume
or,
output
from a smaller
mantfold
chamber can offer
increased
lowover
direct
radiators
with
equivalent
conversely,
offer
the same low-frequency
box.
DESIGN OF A FULL-RANGE,
ALL-MANIFOLDED SYSTEM
An example of this
new manifold
technology
is the MT-4,
a fourway-active,
two-box
system shown in prototype
form in Figure
lg.
There
are
four
drivers
in each frequency
band for a total
of
sixteen
drivers.
The enclosures
have identical
dimensions
of .91m
X .91m X .76m (36in.
X 36in.
X 30in.),
providing
integer
divisors
of
a standard
truck-bed.
The prototype
shown
is
of
slightly
different
proportions,
The lower box,
the MTL-4,
is a vented-box
design that
contains
four
46 cm woofers,
each facing
into
a manifold
chamber at
the
center
of the enclosure.
Additionally,
the woofers
are mounted
"magnets-out"
for more efficient
heat transfer.
In the frequency
range
below
70 Hz,
the
manifold
design
is
nearly
twice
as
efficient
as typical
born-loaded
systems of equivalent
size.
[ts
sensitivity
of
101dB(IW/lm)
and
1600
watt
continuous
power
handling
capabi)ity
enable it to produce
133dB SPL (13gdB
peaks)
at one meter in full-space.
The sensitivity
can,
of course,
be
increased
by placing
the enclosure
on the floor
and/or
using
multiple
enclosures.
The upper box,
the MTH-4, is a three-way
mid-bass/midrange/highfrequency
system,
The mid-bass
section
is the four-driver
system
described earlier. The mid-range section consists of four
51mm
titanium-diaphragm
compression
drivers manifolded onto a 60-
degree by 40-degree constant-directivity horn. The high-frequency
section consists of four 32mm titanium-diaphragm
compression
drivers
on a horn identical to that used for midrange.
Using
identical horns
and manifolds for hoth the mtdrange
and high
frequencies results in equal path lengths for the two sections.
The frequency response for the complete system with recommended
crossover, equalization and delay compensation is shown in Figure
20. Crossovers are fourth-order Linkwttz-Riley filters at 160 Hz,
1600 Hz, and 8 kHz. Beamwidth plots for each section are shown in
Figure 21.
CONCLUSIONS
The use of an acoustic manifold to combine the outputs of several
loudspeaker
drivers
to provide a single acoustic
source with
high-quality performance has been demonstrated,
and designs were
presented
for manifolding
four drivers
in various
frequency
ranges. The merits of manifolding include improved
directivity
control
and
audience
coverage
(mid-bass
through
high
frequencies),
increased
efficiency
(low
frequencies),
and
extremely
high power-to-enclosure-volume
ratios
(all ranges).
Applications
such as concert sound reinforcement,
which require
extremely high acoustic power output, should benefit greatly from
these new forms of manifold technology.
The devices mentioned in this paper are the subjects of a
of patents currently pending or approved.
number
ACKNOWLEDGEMENTS
This project involved many people at Electro-Voice over the past
several years,
most notably Ray Newman,
Chief Engineer
- Loudspeakers, whom the authors
thank for continued
guidance
and
contributions.
Thanks also to Cliff Henrtcksen for inspiration
and advice,
to Ketth Walker for prototyptng
and setting up
numerous
"1/2-ton listening tests",
and to Sheldon
Crapo, who
participated in some of the earlier designs.
10
REFERENCES
[1]
Cltfford
A.
Henrtcksen,
anical
Specifications
Vol.
11, No. 2 (1983).
of
"Sound System Designs
Using
MechDrivers",
S_naudcon Tech
Topics,
[2]
Lawrence
Acoustics,
E.
Ktnsler
and Austin
Wiley,
1966.
R.
Frey,
Fundamentals
of
[3]
D.B.
Keele,
"An Efficiency
Constant
Comparison
Between LowFrequency Horns and Otrect
Radiators"
presented
at the
S4th
Convention
of The Audio En 9.
Soc.,
Preprint
no.
1127, -1_
1976.
[4]
R. L. Prttchard,
In an Inftnite
32, No. 6, pp.
"Mutual
Acousttc
Impedance Between Radiators
Baffle",
J.
Acoustical
Soc. of Amertca,
Vol.
730-737 (June 1950).
11
+1o:
:
I
I
Reference
I
:
:
I
]
Ii
il
:, : :::::,,
',
]_llil
Il]Ill
....
:
:
I I
I
I
ii
I
T:
! ]
_1o
: : Itlil,
]
Illlli
[ '
l't'
I ! I,lllE
J I
,
I
I 20I 30; I 6oIIitll
lOO
I ; I;[I,1
I 200
I , I soo
II,I
I
tii ii. .
I
_,
:_
I ' I
&_-q_[_--]
* I
' I _..Z..,.,.,_.,'_'%
_ .......
-
]
' l--T _,/,',/.__ZL_
II
10
I
i I
,
i
I
!
lOOO
?
,
_:
r 2O00
I i :50O0
j,li.
I
',
1000O 20OOO 4OO00
Figure1. SummationResponseof Large Y-Throat
._o_.....
+10
iFiii..........
i 'i....
i
[ _;'l[
:::::
I
i
:;:::
![![B
i
-Io
i
:
[
--
,IlL
'
20
30
50
100
-
2oo
500
1000
I
i
I I
i
I
I
'I
I fI
_
I
'I --
I
I
-I-q-, !
i'
_
......
10
? rtl 'i'r'_I
_,__ i _:_L !
20oo
5000
100o0
200oo
40000
Figure2. Right-AngleReflectiveBend; DifferenceBetweenHorn with
Straight Tube, Horn with Reflective Bend.
ENTRANCE_
ENTRANCE
NTRANC[
ENTRANCE<_
EXII
Figure 3.
Manifold for Four Compression Drivers.
+20
"_ ? ' i' T Yt_i':-_ ....
*lo
:
i i i ,'.I
!
II
Iii!
-_o--,
_--_"'_'_%'l'd?_f:Li'L--i':_
_ __ _' --:'L-re'--"
'-=
,:
I,,
I
i_Tt--.-'--r-l-\_.Ti-__--I_
,
_
, , -'I(-'142---___1 ;
I '' _
I ! ,1_
;
-20
I i_;,,t
10
20
30
50
100
,
I ,
'
200
500
t ' I-_-,_/_-----
,
i
iq'
i , i :_',
I
i
1000
,
I
I i-i-!:t--I
_
_
! :- :
--
5O00
10000
:._l_t_
:_-I-I---
--_-*-_:r--:-_
-_
_--- _r _
2000
20000
a)
b)
Figure 4.
Off-Axis Response, Piston Source;
b) 45 Degrees.
a) 90 Degrees.
40000
+20
,
I
I
-o
I
_
(
]
[ I
! I [[
1. : :
--:0'
= .
I
I
I'1
_
[
.
.
I
I
,o
I
)
I
I
[
i
llll
IIIlil
!
'k i
i I
! I l_
_
]
I
:
)
I
I
I
_
II:*[l
I
i
[
I i : i :Ii]ii
-291
i
i
//l-[-'[-H-_Z
]
i
j_---I
,_il
i ....
,'
I
[
l',ll!l
i , , lil
i
I
[
,'_l: i
-_:_._-:_
i
:
,
l
[_
I
[iii'ii
Il [
i l_---,--_i-?:-'
_---
]::i[ i
--
_:_
::IZ! :l:
I !.$i,i',:I
I
I
I .
....
.o
-_o
I
IF ,
[ [ i[illl
....
I
_, .
l l_lL,I I I ! i ti':]i!___r
I
I
_ .
IIII!ll
j I IIIll
[
=:=::=:=
i IHdlil
I
c)
Figure 5. Four-DriverManifoldSummation;a) Single Driver on Horn,
e_ualized, b) Manifold on Same Horn with Same Equalization.
c) SummationResponse.
Figure
6.
Photograph
---
Manifold
with
Four
Compression
Drivers.
{Fl [TiFTi ri rllril rT rllrTT.'rT'rl
360
300 ,'
20O
of Prototype
.
.
.
I
Degrees
I
,._,., ,oo{.
i} {iiiiii i{J
60 I
I
401
{ I
I [
II
I
I
[ I I I II
I II1_11
I I
I I
!t 111111II
"1 il !111[
10
·
;iQ 30
60
.
}/l/l/l} l/l/
-'l'--I
] I I I I!I
I IIIIII
I /
/
I lllltl
t/
Illllil
100
200
_00
1000
II I/
! I I
LIA,_LI_
A. '2_IIZ_T_"_'T_
}
TT'Bk_'I{I S.,I I I
! - :
2000
'
50,DO 10000 20000 40000
FREQUENCY
(Hz)
Figure
7.
Beamwidth
Plot
Of Typical
49mm
Throat
Feeding
40-degree
Horn
O"
60
270 240___.
IllI
90
180
Figure 8.
Polar Response, Typical 49mm Throat Feeding 40-degree Horn
Figure 9.
·
Path of rays in Horn Throat
Figure 10. Single Mid-Bass Driver on a 60 o X 40o Horn.
40
o
Figure 11.
THo Mid-Bass Drivers Manifolded on a 60o X 40© Horn.
i--_:F:-t_:_--_F, IH/
_
_ ___
+1o
hre erence
[---
,-
-I-
--=
I I
,
I I I4-H-H---+-+---F--H-H-t-H_
.........
tUi-_l
_ .
I LL;J-""i'
I !-- I IIIII
'
ELT-_LFUIlI-'
-
/F-_T_"_
I
i
r
,'-±-p-_'
I I I I I
I ! I I h_nh-m
....
II
_--
'
_--
--
,m-m_-r--ri-%i
i _i4-N41J
,o......,,o.oo
-.oo,o.o
.o
.......o o0o
FREQUENCY
Figure
12.
+20
E)+i0
E]xlO
Frequency Response of Single Mid-Bass
(1 w @ I m; 100-dB Reference).
1::::_2Y-'
-_
'L
_
_1
_llg_lg_L
I , ,
I
I
4
!i '!iiiiiii
Jl_L__-i
IIIIIIIJ
½t-t
Driver
i _i iilll i i__,'_r*--_
'
+
IIIll----
Referen=e
_"---'_-_----_--_-H_/'l
,--_"I I i.iii!!'
-lol--_-t--t--H_t--i
ii
-
-
I
'
t_l
I I 1 I II II
,I
I tl---F---F-t-_-I_j--_
,,',, :' _i
ii i_
', :'
I', II
, I
,, ,,,,,, ,,
_i iiiHi tl
Ii i iiiii
I _ I :::ii
:' ',', ',I',I',II
Illll
II I', IIl,It
, I i
I---_--I
20
[ I I-IIIII
30
50
100
i i i-Id Il Ill
I I
ii-d-,l-I I
,,,,,,__LJJ_iJ' :',_t_-t=_
1 I IIIlll
200
FREQUENCY
13.
II II
i i i i iii,,,,:
ii ii iliiii
_:
i_J_.Li_=LLLUi
'
! ! ! ! iijjii
i ii ii_i
_=,
10
J_
i ii',',:"_--h-_l_
illliii:
,: ::
,_-_=¢\-1/_=1,==i x_-ml
-ool_:_-L____lm-,
Figure
(Hz)
500
1000
(Hz)
[3 + 10
I I I I I-_m_-?,
,_-,1
I I
I I-tillll
2000
[3x
0000
I I---t--I
10000
20000
10
Frequency Response of Double-Manif61ded Mid-Bass
Drivers (1 w @ I m; lO0-dB Reference).
40000
k_-I
--L_-_[_I;
*lo__l
I -I
I i III[II
I
i I I I [tllll
', I i Ir!l
I ! !!!!!,
', I
I I
I I I I I_q-__l
I I : IIIF[
I
T_.[_I_LJ_LLI___/
I
I
I I Jill[
I ![[I
L__L_I____L_LLLLL__
! ! ! !!!!!!
g-!_ _ ; ii_i[i
I---H-t--_--t-t-HH_i
I
I-T-I
I
II
I I-IILLLi__i i
..I_I-_--H-4-F-H_III
-'_--H-_-t-t-H-H_l--_r-i-l-i-_
_l_H
i ,,ltlll
I
B
[
[
I
I
f
/I
I
I I Ill]l
I I Illll
I
I
] I
! I
I I
I I
I I ! !1 f_-- ,-q--I_
t I I III/I,I
B
I I
'. I
_
I f
! I
il
I I
I
I
I
I I IiHtl
! I
I]
II I+__tl--_-I I Illll
I I
',
i
i
I
I
I
i
:
I
I
i
:
; I
r
I
I ]
t11111
I i ...........
__
_t=:t=l_lll
I
I i
I
10
....
i
I
20 30
i iiJ
ii
[ I11[11
50
100
I
i
:
I
i
I
200
',
I
i
:
',
i
i
:
'.',;',I',
IIIIII
iiiirJ
111',II
500
1000
FREQUENCY(Hz)El+lO
Figure
14.
'.
i
i
I
2000
III/'_1
tltllk___l
Jlrl',l
I ....
I
', _
_ i
5000 10000 200oo 40000
E]xlO
Frequency
Response
of Quad-Manifolded
Mid-Bass
Drivers
(1 w @ I m; lDO-dB Reference),
FRONTVIEW
_O
Figure 15.
O _
SECTION
X-X
O-]
C)x
Four 46 cm Woofers Direct Radiating.
FRONTVIEW
Y
Figure 16.
SECTION
Y-Y
¥
Four 46 cm Woofers Manifolded.
.,o
_t
_,_,-_t-_-__
I--_-t--_'-x_J-t_'__,
-- __
-,_
_,_k_,'_-'_
_t
_-_
'--t_-_4-.,_---_
- -_ ---_-._-_
_
_.___
,.,......1-_- I--*- ,_,4-f-y_--k-t_--_ldd-_--i--_
-lO
;--_-
10
20
-
30
'
50
100
200
FREQUENCY
Figure 17.
.
500
1000
{Hz)
[] + 10
'I_}m:F-_--F ::m
' _
2000
-
5000
_.__-7___
10000
00000
40000
[3 x 10
Frequency Response of Four D_ect-Rad_ating
(! w @l m; 100-dB Reference),
Woofers.
-ol_K-_i ii iili
10
2O
30
50
100
200
FREQUENCY(Hz)
FJgure ]8.
500
1000
2000
r1+10
[]xlO
5000
10000
20000
Frequency Response of Four Man_folded _oofers
(1 w @ 1 m; [00-dB Reference).
40000
Figure 19.
Photograph of Prototype MT-4 Four-Way Manifolded
Loudspeaker System
10
20
Figure 20.
360
30
50
100
600
1000
FREQUENCY (HzJ
200
[] + 10
2000
5000
100o0
20000
40000
[] x 10
Frequency Response of MT-4 Loudspeaker System with
Recommended Crossover, Equalization and Time Delay
(10 w 0 3 m into Low-Frequency Section; lO0-dB Reference).
.
LOW _REODENCY
MID-BASS
MID-RANGE
[NIGH FREQUENCY
80°[i-[Th_,-l-FiT[,_,,.
':fTTrlrlr';,'_'
Beamwldlh
Peg
(-6 ....
dB)
vE
! AL:
......
_
100
80
60
.....
20
"%-'"
_'
"
I I Iii ',', ,. ,
:----k.._[ I] ._?-'_'
30
60
100
200
600
11300
2000
6000
FREQUENCY (Hz)
Figure 21.
, ,
' '
,l_ _"_'
,_T_-':_"-'"
__'"''_i'"
'
_o
10
'
Beamwidth of MT-4 Loudspeaker System.
10000
2o000
4000o