CONFIDENTIAL

RESEARCH DEPARTMENT

The design of
studio monitoring loudspeakers
types LS5/5 and LS5/6

TECHNOLOGICAL REPORT No. PH-13

UDC 621. 395. 623. 842 1967157

THE BRITISH BROADCASTING CORPORATION

ENGINEERING DIVISION
 CONFIDENTIAL

RESEARCH DEPARTMENT

THE DESIGN OF STUDIO MONITORING LOUDSPEAKERS TYPES LS5/5

AND LS5/6

Technological Report No. PH-13

1"", UDC 621.395.623.842 1967/57

H.D. Harwood, B.Sc. J?lMaA'''~

S.A. Hughes for Head of Research Department
This Report is the property of the British
Broadcasting Corporation and may not be
reproduced or disclosed to a third party in
any form without the written permission of
the Corporation.
- - -.. _... - - - - _ .__._.. _ - - - - - - - - - - - - '

This Report uses SI units in accordance

with B.S. document PD 5686.
_ _ _ _ _.--1
 Technological Report NO. PH·13

THE DESIGN OF STUDIO MONITORING LOUDSPEAKERS TYPE LSSIS AND LSS/6

Section Title Page

SUMMARY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1. INTRODUCTION................................. 1

2. DESIGN CONSIDERATIONS ........................... .

3. BASS EQUALI ZATION. . 2

3.1. General . . . . . . 2
3.2. Experimental Detai Is 2

4. DESIGN DETAILS 4

4.1. Cabinet . . . 4
4.2. Use of SI it . . 4
4.3. Details of Units 6
4.3.1. Bass Unit 6
4.3.2. Middle-Frequency Units 7
4.3.3. High-Frequency Units . 8

5. DESIGN OF COMPLETE LOUDSPEAKER 8

5.1. General . . . . . . 8
5.2. Type A Loudspeaker 9
5.3. Type B Loudspeaker 9
5.4. Type C Loudspeaker 10

6. LISTENING TESTS . . . . . . . . . . . . . . . . . 10

7. REPEATABILlTY OF THE TYPE LS5/5 LOUDSPEAKER 11

8. DIRECTIVITY.................................. 13
Section Title Page

9. IMPEDANCE AND DISTORTION CHARACTERISTICS 13

10. POWER AMPLIFIER 14

11. DIMENSIONS. . . . 14

12. CONCLUSIONS.................................. 15

13. REFERENCES.................................. 15

14. APPENDIX ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 16

 CONFIDENTIAL

November 1967 Technological Report No. PH-13

UDC 621.395.623.842 1967/57

THE DESIGN OF THE STUDIO MONITORING LOUDSPEAKERS TYPES LS5/5

AND LS5/6

SUMMARY

Details are given of the various factors which have led to the design of
two new monitoring loudspeakers suitable for use in studios and outside
broadcasts. The loudspeakers are much smaller than those of the present
type; a floor-standing model is designated type LSS/S, and one intended for
hanging is called type LSS/6. In the course of the design, the questions of
bass pre-emphasis and of directivity have been examined in some detail.

The quality of reproduction and the directional properties are appreciab-

ly in advance of those obtained for the LSS/1A and the maximum sound level
is also higher. The spread in frequency characteristics between development
specimens is extremely small, and the level of non-linear distortion is low.

1. INTRODUCTION should serve both for studios and outside broad-

casts.
The present studio monitoring loudspeaker type
LS5/1A was developed in 1959 and employs a
special Goodmans 380 mm low-frequency unit and
two Rola Celestion 58 mm high-frequency units. 2. DESIGN CONSIDERATIONS
Although Some 250 have been built, considerable
difficulty has been experienced in securing adequate In a modern monitoring loudspeaker the choice
supplies of low-frequency units which meet the lies in practice between two- and three-unit designs.
tolerances applied. Yet in spite of the tightness In a two-unit loudspeaker one of the difficulties is
of these tolerances, comments have been made that that the high-frequency units avai I able at present
the sound quality varies from specimen to specimen; cannot be operated below apprOXimately 1.5 kHz so
furthermore, some cri ti ci sm has al so been made of that the low-frequency unit must operate in a predic-
the reproduction although it is conceded to be better tabl e manner up to about 2 kHz. I n the past, repro-
than that of any commerci all y avai I abl e loudspeaker. ducible operation of a low-frequency unit above
about 500 Hz was not possible but the situation has
In view of the difficulty in obtaining low- been changed by the advent of the 305 mm pi astic
frequency units of adequate quality and reproduci- cone previously mentioned.
bility, an investigation was started in Research
Department into the possibil ity of producing a It is still difficult, however, to maintain the
thermoplastic cone and these experiments led to the required frequency characteristics away from the
production of the 305 mm unit descri bed el se- axiS of a two-unit design. At 1.5 kHz the wave-
where(1). The listening tests were so successful length of sound is about 220 mm and thus a 305 mm
that in November 1965 itwas decided to commission cone has a diameter considerably larger than a
a new loudspeaker incorporating this unit. It was wavelength. It follows that the radiation will be
clear that by employing a 305 mm unit an appreci- directional at such frequencies and that even when
ably smaller cabinet than that of the LS5/1A would the axi al frequency characteri sti c is made uni form
suffice and it was intended that the new loudspeaker the off-axi s curves wi 11 depart from this condi tion.
2:

On the other hand the high-frequency units, 58 mm An alternative method is to use equalization
in diameter, are small compared with a wavelength, ahead of the power amplifier, but if an excessive
and therefore nearly omnidirectional, up to about degree of equalization is applied, over-loading of
6 kHz. The resulting axial and off-axis characteris- the amplifier will occur first in the bass and once
tics are typified by the curves in Fig. 1. To some again the usable mid-band power wi II be reduced.
extent the difference between the curves can be The quest i on therefore ari ses as to whether the
reduced by fitting a slot in front of the low-frequency programme spectrum is such that it is possi bl e to
unit but as will be shown later, this device is by no . apply equal ization before the amplifier without
means wholly successful in overcoming the trouble. causing overloading in the bass. Experiments were
accordingly designed to explore this possibility
and to determi ne the opti mum shape for the pre-
emphasis curve. It will be seen that, in effect, the
axial curve object of the experiment was to obtain the low-
\ frequency equivalent of the high-frequency pre-
emphasis employed in f.m. broadcasting.

3.2. Experimental detai Is

Various types of programme were examined to
find those which had the highest power levels in
frequency
the bass. Eleven recorded items were finally
Fig. 1 - Typical frequency characteristics of a two- chosen, two of which were organ solos, three were
unit loudspeaker on axis and at 60° from axis light (pop) music and the remainder orchestral
music, the total playing time amounting to about
The use of a three-unit system with crossover 13 minutes; detai Is of the items are given in the
frequencies in the region of 500 Hz and 3 kHz appendix. In all cases the recording was arranged
allows these difficulties to be largely overcome, to peak to 6 on a peak programme meter type PPM/2,
provided a suitable type of middle-frequency unit the peak occuring usually, although not neces-
can be found. There is the extra advantage that sarily, during the excerpt chosen.
with a frequency range restricted to the band from
3 kHz upwards the high-frequency unit will be able The spectrum was examined by means of
to h and I e a larger prog ramme I evel than if it had
octave fi Iters centred on frequenci es rangi ng from
to operate at 1·5 kHz. On the other hand an ad- 1 kHz down to about 50 Hz, the peaks in each band
ditional unit and a more expensive and elaborate of frequenci es bei ng recorded by a peak counter
crossover network are required. reading in steps of 2 dB, due allowance being made
for the insertion loss of the filters. Typical analy-
ses are given in Fig. 2 and the overall peak levels
3. BASS EQUALIZATION
6
3.1. General
4
In practice the axial frequency characteristics (c)' 1\
2
of low-frequency loudspeaker units are not uniform. ,/ 1\ I'~
The reasons for thi s are that in the mi ddl e-frequency o ./,
range the unit becomes directional, concentrating C -2 ./~ ,,/ ~I'\.
the sound energy i ncreasi ngl y in the axi al di rec- o I "J / (a) ,
g -4
tion, whilst at low frequencies over-damping of the
bass resonance takes pi ace, thus produci ng a bass .Bc -6 II /
/

IT(b)
cut; the resulting rise in axial response above E -8 I
the resonance frequency usually amounts to between ~-10
6 and 10 dB. This rise must be equalized electri- I
-12
cally and in past designs, e.g. the type LS3/1A
loudspeaker, it has been carried out in the cross- -14 I

over network, thus enabling a standard amplifier -16

with a uniform response/frequency characteristic -18
to be used. This method involves a considerable 3 5 7 10 2 2 3 5 7 10 3 2
frequency. Hz
loss of power in the mid-band region: for example,
if a 20 watt amplifier is employed and 10 dB of Fig. 2 - Peak levels in octave analysis of
bass equalization is required, only 2 watts are programme
Item (a) Kramer with Dakotas
available to drive the loudspeaker in the mid-band
(b) Mars I
·region. (c) Organ Prelude in G
 3

for the whole range of items is plotted in Fig. 3; TABLE 1

a smoothed curve of the peak spectrum is al so Effect of Bass Equalization on Peak Level of
shown in this figure. It will be noted that the Programme
smoothed curve passes below the pOint plotted for
68 Hz. This point represents a single note from a peak level s on PPM/2. (dB above'S')
bass guitar which stood out considerably above the programme
item circuit condition (see Fig. 4(a))
rest and was therefore ignored in drawi ng the
smoothed curve as it was felt not to be representa- (see append 'IX) no bass boo st circuit (i) circuit (ii) circuit (iii)
tive. a -% -1/2 -112 -%

b -% -% 0 2:..!
10
100 0/0 modulation c 0
e .:!:.l -+2 +3%
-
8U)
6
4
smooth~d curVG
d

e
0

-1'/,
0

-2
0

-2
+1112
-
-1'/,
S /r-.. ~
.5 2
~o lL~ .;- f -2

-lV,
_1';' t'12 ~

g -1 -V2
'tJ
-2
V -+112
h -3 -3 -3 -2'/,
-4
3 5 7 10 2 2 3 5 7 103 2 i -4 -4 -4 -3%
fr~qu~ncy • Hz

j -2 -2 -2 _1112
Fig. 3 - Peak octave analysis of programmes, all
items k 0 +1/1
- -+1 +1
-

7 -
~ '" ~i~
(iii)

C
I L
\
6 R
1\, \ 1\ c
Jl
L
H Cl
R
5 (j) 8,4 3,8 1600- t-
\ 600 600
1\ (jj) 4·0 3·1 1600

1\ , ,1,
4 (j i j) 6'0 2·1 1800 - t-
dB
3

2 1\ "I', ",
l"
" '"
Tb
1 I' r-..... " " ""
~ ...:-~
o
1
3 5 7 10 2 2 3 5 7 lOa 2 3 5 7 10 4 2
frcqu~ncy, Hz

Fig. 4(a) - Response/frequency characteristics of bass-lift circuits

Equal ization was designed for the smoothed

curve and for two similar but progressively more
extreme conditions as shown in Fig. 4(a). The
recordi ngs were then repl ayed through the di fferent
circuits to see by how much the equalization in-
creased the peak level of the complete programme
as read on a PPM/2; the results are given in
Table 1. It will be seen that the level of item c is
increased by 1 dB even by circuit No. (i) and it was
Fig. 4(b) - Circuit used for determination of acceptable
decided to determine whether this degree of over-
distortion with bass-Lift circuits
load at low frequencies would be audible with a
typical amplifier using a considerable degree of
clipping amplifier was adjusted so that a 1 kHz
negat i ve feedback. si gnal of +8 dBm from the source was just cl ipped
A circuit was set up as shown in Fig, 4(b), in at the peaks. The bass -lift circuits were inserted
which the peak clipping is arranged to occur in a in turn ahead of the ampl ifier and the programme
separate ampl ifi er followed by an attenuator wh ich items played through the system, allowance being
feeds a loudspeaker amplifier. The gain of the peak made for the insertion loss of the circuits. It was
4

found that when using circuit No. (iii) of Fig. 4(a) the radiation more nearly omnidirectional in the
distortion was clearly audible on items c and d, horizontal plane. There is, however, a limitation
i.e. the organ passages, none being noticed on the to thi s devi ce: the Hel mhol tz resonator formed by
remainder; when circuit No. (ii) of Fig. 4(a) was the mass reactance of the slit and. the compliance
inserted, distortion was only just detectable on of the ai r enclosed between the si it and the cone
item c and it was therefore concluded that this increases the output to an undesi rabl e extent in the
degree of bass pre-emphasi s is permi ssi bl e. Any region of the resonance frequency, but acts as a
equal i zation requi red in excess of thi s must there- low-pass filter above the resonance, severely
fore be appl ied after the power ampl ifier. reduci ng the output at hi gh frequenci es. The mi ni-
mum slit width which could be employed without
either of these two effects becoming excessive was
4. DESIGN DET AI LS found to be 100 mm and it would appear at first
sight that this width, which amounts to only a third
4.1. Cabinet of a wavelength at 1 kHz, should j,be quite small
enough for thi s purpose.
Experience with the type LS5/1A loudspeaker
had shown that it had an adequate bass range. In the first instance the slit may be regarded
Calculations indicated that a similar range would as a source having uniform sound pressure all over
be obtained with the new 305 mm plastic cone unit its area, but with conditions of radiation inter-
by employing a cabinet of only 0.085m 3 internal medi ate between those for free space and those
capacity, that is 60% of the volume used for the for an infinite baffle and there are three possible
)
LS5/1A. confi gurations whi ch may be regarded as approxi-
mations to these conditions. Of these, a line
Measurements were then made with an experi- source and a circular piston in a baffle may be
mental cabinet to determine the vent resonance shown(2) to have directional patterns given respec~
frequency giving the best combination of power tively by
handl ing capacity and frequency characteristic;
this frequency was found to be 38 Hz, close to • (TTl. )
sin A sin a
that employed for the type LS5/1A. Two types of Ra= _ _ _ __
cabinet were made, one floor-standing and the TTl
other for hanging from the ceiling, corresponding A sin a
to the LS5/1A and the LS5/2A respectively. The
volume and front dimensions of each model were where Ra is the sound pressure radiated at an
the same. angle a between the direction of radiation and the
axis, l is the length of the source and A is the
4.2. Use of SI it wavel ength.

The next factor to be dealt with was the 2TTr

directivity of the units. Fig. 5(a) shows the re-
and 2J 1 (A sin a)
Ra = --:--~---
sponse on the axis and at 60° from it for the un- 2rrr sin a
equalized bass unit in the cabinet. It will be noted A
that there is an appreci abl e di fference between the
two at the hi gher frequenci es. Thi s difference can where r is the radius of the piston and J 1 is a
be reduced by placing a si it in front of the unit; 8essel function of the first order and first kind.
the diffraction from the edges of the slit will make The directional pattern for a piston in the end of a
semi-infini1e pipe is more complicated(3) viz:

4 Ji kr sin a
Ra=---
TT sin2 a [(J i (kr sin a)) 2 + (Yi (kr sin a)) 2]Y2

IRI 2 kr cos a kr 1
x tan- (-J 1 (X)/Y1 (x))dx ]
x
1 - IRI
2 X exp
[ IT
p x
~ [x' - (h sin a) '1[x' + (k~ 'I" 1

where IRI = exp --

2kr skr tan-i (-J d X)/Y. i (X)) d }
-y: x
{ TT x[(kr)2-x2] 2
o
 5

-----
~r h,,(\
5t B /" I-"'"
../ ~

v V\
$/' \!'t
~
- - r~
~".-/ 60·'

~'\/
,...f\ I - \
'"
1 ( \V
/\

/
~ J \ \
\1 \ ~ \
'J ~

0'03 0·05 0'1 0·2 0·5 2 5

fraquancy, kHz
Fig, 5(a) - Response/frequency characteristic of unequalized low-frequency unit
without slit at 0° and 60° to the axis

o· ./\ 1\
5tB
T
V
~ ~-- ....
........
1'\ r-./ \ / \
~' 60·' r\\ ~_/' ~\ IJ~ \ "
~
~.--
.- F-"
\~ ; \V'
V I} H '\,
1 fraQuancy ranga USGd in typa A
I 1
IV \I '\" \
frequency ra~ge
I'
uSed in types IB and C --I ~ -"\..1
0·03 0·05 0·1 0·2 0·5 2 5
fraQuancy. kHz

Fig, 5(b) - Response/frequency characteristic of unequalized low-frequency unit with

100 mm slit at 0° and 60° to the axis

Fig, 6 - Deviation of 60° characteristics from axial

characteristics for differing types of source
(a) Line source (calculated)
(b) Piston source in infinite plane (calculated)
(c) Piston source at end of pipe (calculated)
dB (d) Measured values obtained with slit on low-frequency
unit
(e) Sound pressure concentrated at edges of slit (calcu-
lated)
(f) Measured values taking d as frant of cabinet

0·2
ratio, *
0·3 004 0·5 0·6 07 0·8 o·g
,of slitwidth to wavalangth
lO
 6

and J and Y are real first order Bessel functions of It appears therefore that up to 700 Hz"'the
the first and second kind respectively, according to directivity is largely determined by the width of
the usual notation* and k = 277/ A- the cabi net but that above thi s frequency the width
of the slit plays a large part. That it does not
The calculated response at 60 0 with respect to fully determine the directivity even then is shown
that on the axis is shown in Fig. 6 for these cases. by the fact that the upper part of curve (d) of Fig. 6
As expected it wi II be noted that for si it widths up does not lie in the region of the calculated curves.
to 0·6,\ there is not much difference between them,
Thi s di screpancy is further emphasi zed by the fact
(curves (a), (b) and (c)), and for the pcoposed slit
that in the final design the smaller middle-frequency
width of '\/3 considered at 1 kHz, the mean dif-
unit employs the same width of slit, 100 mm, in the
ference between the axial and 60 0 responses is not
same baffle, yet the deviation of the 60 0 curve from
more than about 1% dB.
the axial curve at 1 kHz is different from that of
I n contrast to thi s the actual frequency charac- the low-frequency unit, the value being 3 dB closer
teristics obtained with a 100 mm slit are shown in to the theoretical figure. Unexpectedly it appears
Fig. 5(b). It may be observed by comparison with therefore as though the size of the unit still affects
Fig. 5(a) that, with the si it, the deviation from the the directional properties in spite of the slit and
axial response is almost unaltered up to about 700 the exact mechanism accounting for the directivity
Hz, although beyond this frequency there is an for the values of.E.. greater than 0.75 is obscure.
appreciable change; furthermore at 1 kHz the devia- ,\
tion with the slit is not 1112 dB as calculated but
nearly 6 dB. The measured deviation is replotted 4.3. Details of Units )
as curve (d) in Fig. 6 and it will be seen that it
does not correspond to any of the three calcul ated 4.3.1. Bass Unit
cases.
As already mentioned, the bass unit employed
This lack of improvement in directivity with is the 305 mm plastiC cone unit described in Report
. the use of a slit was first noticed during the design L-065. A chassi s wi th a more powerful magnet is
of the LS5/1A, when it was found that reducing, now available and an increase in' sensitivity of
below 180 mm, the width of the si it in front of the about 2 dB over the unit described in Report L-065
380 mm cone did not bring about a corresponding
is thus possible. Further experience with the unit
improvement in the off-axis curves.
revealed a slight colouration in the 1.5 kHz region,
and this is accentuated with a later material manu-
One possible explanation which has been
examined is that the distribution of energy across factured as a replacement for the type of Bextrene
the si it is not uniform and the extreme case when formerly used. It is however completely removed
a 11 the energy has been concentrated at the two by painting the cone with a layer 01 polyvinyl
edges has been calculated and is shown in Fig. 6 acetate damping compound known as Plastiflex type
as curve (e). Even under these conditi ons the 1200 P, even though thi s treatment does not cause
directivity is not nearly as great as that experienced any appreci abl e change in the frequency response.
in practice with the low-frequency unit for small (The effect on colouration can easi Iy be demon-
values of d/,\, where d is the width of the si it; strated by applying pink noise (i.e. random noise
furthermore, measurements show that although the with equal power per octave) to the unit in a free-
pressure across the slit is not quite uniform it is field room and making a tape recording of the output
actually higher in the centre by about 2 dB; in before and after pai nt i ng the cone. The two con-
addition the phase change across the slit is also ditions can then be compared sequentially and the
small. improvement obtai ned by'the treatment is evident).
The further possibility arises that re-radiation
from the edges of the cabinet might be responsible In spite of the use of the vent mentioned in
for the directivity. Taking the width of the front Section 4.1 some electrical low-frequency equaliza-
baffle as 350 mm, the actual values obtained for the tion is also necessary. As explained in Section 3,
deviation of the 60 0 curve from the axial for the new it is best to apply this equalization mainly as pre-
values of d/'\ are plotted as crosses in Fig. 6. It emphasi s ahead of the power ampl ifier and to intro-
will be seen that in fact the agreement with the duce the remainder in the crossover network. It is.
theoretical curves is quite good up to a value of d/,\
of 0·75 after which the loudspeaker is less direc- * * At the vent resonance frequ ency the output from the
vent is in quadrature with that from the cone, but as most
tional. This value of d/,\ corresponds to a fre-
of the energy is radi ated from the vent and both sources
quency of about 700 Hz, the frequency above which are very close together, the loudspeaker is omni.direc.
it was observed that the si it has an appreciable lional. Above this frequency the sound radiated from the
effect. vent is rapidly attenuated and the phase difference
between the two outputs becomes zero. The vent there.
fore has little influence on the directivity at any fre.
* In Reference 3 Yl(X) is denoted by N1 (x) throughout. quency.
 7

expected that, as with the LS5/2A loudspeaker, a in Fig. 7, and it will be seen that over the required
further bass lift, amounti ng to about 3 dB at 40 Hz frequency range the two are smooth and nearly
over that required for the floor-standing model, will parallel. Listening tests, however, showed a
be required for the hanging model, and this lift also noticeable colouration in the 1.5 kHz region and
is conveniently applied ahead of the amplifier. It chopped-tone tests were therefore applied. In the
will be seen from curve (ii) of Fig. 4(a) that this region 1.2 kHz to 1.7 kHz these tests revealed
leaves about 4 dB available for the floor-standing three resonances wi th O-factors of the order of 500,
model before the permi ssi bl e amount of pre-emphasi s some 40 dB below the steady-state condition. If in
is exceeded. phase with the steady-state condition, these reson-
ances represent i rregul ari ti es of no more than 0.1 dB
The frequency characteristics of the bass unit on the axial curve and can only therefore be meas-
on the axi s and at 60° from it are those already ured by chopped-tone techniques. It was however
shown in Fig. 5(b). shown that the appl ication of a layer of Plastiflex
type 1200P damping compound to both sides of the
cone reduced the resonances to a marked extent;
4.3.2. Middle-Frequency Units furthermore, the use of pi nk noi se and the recordi ng
technique mentioned for the bass unit demonstrated
No satisfactory commercially-produced mid- a great improvement in the reproduction and the
dle-frequency unit is available but at the time when colouration was reduced to a very low level.
the new loudspeakers were commisSioned, experi-
ments on a 110 mm diameter unit were already The sensitivity of the 110 mm unit is com-
proceeding in Research Department. This unit parable with that of the bass unit described in
used a 25.4 mm voice coil and a flared cone of Report L-065 but there is a growi ng demand for
Bextrene type 237, 004 mm th i ck, together wi th a even greater sound levels from monitoring loud-
surround made of p.v.c. 0.5 mm thick. The bass speakers; whereas the sensitivity of the Iow-fre-
resonance, at about 400 Hz, was well damped, the quency unit could be increased, that of this middle-
intention being to employ this unit over the frequency frequency unit could not, and it was therefore
range 450 Hz to 3·5 kHz. The frequency charac- decided to make a 200 mm diameter unit of increased
teristics on the axis and at 60° from it are shown sensitivity as an alternative design.

5~B
T
/
_. ~

-
~--:: ................ - " " - I~'-~
~

P ___ ~

',-,\J
fl
~
/\0

C7\
/"\.

, A.

L \ ," f'\",600 rv,..,. '"1\

\.'
" \/,.~ v '\. I.-
"-
fraQUatcy ranga usad I .., ' .
01 0·2 0·5 2 5 10 15 20
fraquancy, kHz

Fig. 7 - Response/frequency characteristics of 110 mm dia. middle-frequency unit at 0° and 60° to the axis

5~B
-f r-"", ,"v------
--. ~/'\
~, /\v
If '- f--- 1-/
'-~ R I\.
L "'-,,600

, ..... \ (\..,

L 1\
\J
/-,
\.,
A
\I "
.. 1\".."'-.
0·1
I fraQUa~cy ranga u sad
~'-.
0·2 0·5 2 5 10 15 20
fraquancy, kHz

Fig. 8 - Response/frequency characteristic of 200 mm dia. middle-frequency unit

without slit at 0° and 60° to the axis
8

-
518 ~ '~ ~ / \-
i
i' '--- -- - "'. - '""f. . . .\
hi ~
,\00
I \ 1\

/ \,'\J \ f'v\ \J~

0·2
fraqiancy ronga usad
'1 \J \l'J \/'/1J'p./
0·' 0·5 2 5 10 15 20
frGqucmcy. kHz

Fig. 9 - Response/frequency characteristics of 200 mm dia. middle-frequency runit

with 100 mm slit at 00 and 60 0 to the axis

The cone of the 200 mm unit is made from ment a further model has been produced employing
0.4 mm thick Bextrene type 730 and as with the the same di aphragm, and therefore havi ng si mi I ar
110 mm diameter unit employs a surround of 0.5 mm frequency characteristics, but with a stronger
thick p.v.c. The experience obtained in the design magnet giving an increase in sensitivity of nearly
of the surround of the 305 mm unit was applied to 2 dB.
this unit and in addition a heavily flared cone was
used. The bass resonance frequency in free air is Rola Celestion also make a horn-loaded unit
about 50 Hz but to avoid reaction with the cabinet for the high fidelity market and this was examined
vent resonance the rear of the unit is confined in a but found to be definitely inferior to the type al-
small enclosure. The resulti ng frequency charac- ready in use.
teristics on axis and at 60 0 are shown in Fig. 8;
with this unit the operational frequency range is The K. E. F. uni t type T15 was al so tested and,
400 Hz to 3.5 kHz. It wi II be seen that the axial al though thi s has a more extended axi al frequency
frequency characteristic over this range is smooth, range than has the Rola unit, the corresponding fre-
but that the 60 0 response di verges from it. As quency characteristic is not so smooth and the unit
mentioned in Section 4.2 a slit of 100 mm width is is appreciably more directional at high frequencies.
used to effect an improvement in this respect; the
resulting characteristics are shown in Fig. 9. The The frequency characteristics of the improved
cone was coated on both sides with Plastiflex but unequal ized Rola Celestion unit mounted in the
1200P to reduce slight colouration in the 2 kHz cabinet are shown in Fig. 10 at 00 and 60° to the
regi on and in th i s regard listen i ng tests show that axis.
the reproduction from the coated unit is remarkably
"clean".
5. DESIGN OF COMPLETE LOUDSPEAKERS
4.3.3. High-Frequency Units
5.1. General
As already mentioned, the 58 mm high-
frequency unit employed in the LS5/1A is made by With the units avai labl e three designs were
Rola Celestion. It has a smooth response/frequency possible. Design A was similar to the type LS5/1A
characteristic and has proved to be very repeatable construction and employed the plastic cone 305 mm
in production. At the request of Research Depart- unit and two of the Rola Celestion 58 mm units;

5dB

Fig. 10 - Response/frequency
characteristics of high flux density
Rola Celestion high-frequency unit
at 00 and 60° to the axis

0-5
fraquGncy. kHz
 9

type B used the 305 mm unit for the bass, the 200 11 is not uniform and is rather like that of the
mm unit for the middle frequencies and a single LS5/ 1A in thi s respect.
Rola Celestion 58 mm improved unit for the high
frequencies; Type C was similarto type B but used 5.3. Type B Loudspeaker
the 110 mm unit for the middle-frequency range.
As it was not possibl e to determine from a study of In the type B design the 305 mm plastic-cone
the units which oould g:ive the best reproduction bass unit is employed up to a frequency of 400 Hz.
it was deci ded to bui Id a prototype of each and Above this frequency the 200 mm middle-frequency
carry out final listening tests. unit operates up to 3.5 kHz v.A1ere a change is made
to the Rola Celestion 58 mm improved unit. As
5.2. Type A Loudspeaker already mentioned the bass resonance frequency
of the middle-frequency unit is about 50 Hz and it
The design of the type LS5/1A has been is necessary to enclose the rear to prevent it act-
described in the Technical Instructions and it is ing as a vent at low frequencies. In order to make
sufficient to mention here that the low-frequency use of the sensitivity of the middle- and high-
unit is employed up to about 1.7 kHz. Above this frequency units the high-flux-density version of the
frequency too Rola Celestion high-frequency units low-frequency unit is employed. In this design the
operate in parallel up to approximately 3.5 kHz relative voltages applied to the units are adjusted
above which the output from one is attenuated leav- by means of an auto-transformer placed ahead of
ing one only to cover the remaining part of the the crossover networks; by this method the relative
spectrum. The response/ frequency characteri st i c levels can be adjusted without having to change
of the 305 mm plastic cone unit is smoother than components in the crossover network as was the
that of the 380 mm Goodmans cone used in the case with the LS5/1A. It also has the advantage
LS5/1A and the design of the crossover network that the nominal impedance of the loudspeaker can
is therefore somewhat simpler; a 100 mm slit, be adjusted to any convenient value to suit ampli-
desc-ribed earlier, was fitted over the front of the fiers commercially available. Fig. 12 shows the
305 mm unit. The response/frequency character- response/frequency characteristics in the hori-
istics achieved are shown in Fig. 11 for the hori- zontal plane and Figs. 13 and 14 those in the
zontal plane. The axi al response is smooth but it vertical plane above and below the axis. It will
will be observed that in spite of the 100 mm slit be observed that the curves in Fig. 12 are smooth
the response/frequency characteristic at 60° in Fig. and close together.

~oo

---- "'" f'.....

( ~
.
..::=-~...:...-~
= ---
--'.,:::..----
-......~-.-.,.--
E--' j~-C~
. ............. -..,. ..... , ~
..... --- r-.'M./
.rY~ ~-,
.... ... ~'rl 1
r-- ~

\~'J...-'"
[ I
450 6ci~
'
.....'\\) \.:, \Ir" \
\
~\
~ '\ h I \
5~B

0·03
'. 0·05 0·1 0·2 0'5 2 5 10
\h.,
IT \
u
15
fr<2qu<2ncy, k Hz

Fig. 11 - Response/frequency characteristics of type A loudspeaker in horizontal plane

- 00 2"
00
"";.:; ...
- 1\ J').M~ n
./
V ':":::""":: '="""~...:.::=~-::.
',- - =---~p
,-,~.-.,~

45 0
~,-",;
. ,, ,/ / /
1-"'60 0
"- ~v,\,
v
VA';V' I.. --::~~,r
',,,
',-,r ,
,I \
'\\

/ 11\.",
\ \I
\

"",\i1{i'..
5~B \ t,1l
-. i .
U
0·03 0·05 0·1 0·2 0'5 2 5 10 15
fr<2Qu<2ncy, kHz

Fig. 12 - Response/frequency characteristics of type B loudspeaker in horiz(lntal plane

 10
5.4. Type C Loudspeaker seen that the curves in Fig. 15 are smooth and
except at the highest frequencies very nearly
This design is essentially similar to that of coincident.
Type B but employs the 110 mm di ameter unit for
the middle-frequency range. The lower crossover
frequency in this case is about 450 Hz, the upper 6. LISTENING TESTS
crossover frequency remaining at 3.5 kHz. As the
middle-frequency unit has a bass resonance of about The three prototype loudspeakers were given a
400 Hz the mechanical impedance at low frequencies listening test and compared with a type LSS/1A and
is high and it is not necessary to enclose the a still earlier experimental model known as the
rear. Owing to the lower sensitivity of this middle- R.M.L. which was included because some observers
frequency unit there is no advantage in employing consi dered it to be superior to the LSS/1 A. The
the high-flux-density low-frequency unit and the tests, which were carried out by experienced mem-
lower-flux-density type is therefore used. As with bers of operational and programme staff, incl uded
the type B design an auto-transformer is inserted speech from both dead and reverberant surroundings
ahead of the crossover network. and recorded and live orchestral items, the latter
from Maida Vale 1 studio. For the live music test
The response/frequency characteristics in the the loudspeakers were checked in turn both in the
horizontal plane are shown in Fig. 15. It will be Green Room and in the Sound Control Room both of

V
- JJ.!.... __ ~~ ..........

./
~30" ,
/ , \

,. ,,
5~9
(

:1J

003 005 01 02 05 1 2 5 10 15
fraquancy. kHz

Fig. 13 - Response/frequency characteristics uf type 8 loudspeaker in vertical plane abuve axis

~ ::v:'\
/'"
V - --
--= ---.:..",::::-
~
~~'r-."-:: V""-\;:
~ .. ~":"'
./"'...

...". ~.
. ",-/"/
~JOC:\
~
i ,I ~f 'J "\.\. ,
/ .;'--:'30·
-15"
'\J"
\r v
\
\ 0.1 \
5~9
'~ \!
f
o'03 005 01 02 05 1 2 5 10 15
fraquancy. kHz

Fig. 14 - Response/frequency characteristics of type H loudspeaker in vertical plane below axis

5.19
.'Z ~
O·
-:::.. ......... ....... /'I lA __ I\~ r-v~
/"
V --"''''= -- ~~~2--::" , -
O· 30· ......... ~-;... "--J
\ W... ~- r'~~Y'~~ ~~(y"1
~V "...1 .\ -- \
r -V v· . . . , , ,
\\
\ A
\ I'
JI
003 005 01 02 0·5 1 2 5 10 15
fraquancy. k Hz

Fig. 15 - Response/frequency characteristics uf type C loudspeaker in horizontal plane

 11

which communicate directly with the studio, and in Report L-065; there has been considerable
direct comparisons with the live programme were production experience with the Rola Celestion
thus possible. The quality of reproduction 01 all high-frequency unit. The 200 mm unit was however
three prototypes was judged an improvement on handmade speci all y for thi s prototype and there
that from both the LS5/ 1A and the R.M. L. I t was was no experience of its repeatability'in production.
further agreed by all that the sound qual ity from To speed up acceptance tests a number of pre-
the Type B loudspeaker was outstanding, being producti on model s of the LS5/5 loudspeaker were
better than that from types A and C but that from bui It and advantage was taken of thi s to determi ne
the type C was very slightly coloured by the remains the spread in frequency characteristics likely to be
of the resonances around the 1.5 kHz· regi on previ- obtained in practice.
0us
y mentioned in Section 4.3.2. The wide angle
of radiation of type B in the horizontal plane was
al so favourably commented on. Fi g. 16 shows the spread in the unequal i zed
axial frequency characteristic of six middle-fre-
In view of thi s verdi ct the remai ning measure- quency units measured in the cabinet without the
ments were confined to the type B model. Two rear enclosure; in the figure the curves were
vari ations of thi s design have been constructed; arbitrarily lined up at 750 Hz. It will be seen that
one designated LS5/5, is floor based with a rec- the spread is very small over the operating fre-
tangular cabinet mounted on a plinth, the other, quency range of 400 Hz to 3.5 kHz.
designed for hanging, is lozenge shaped and is
coded LS5/6. In the LS5/6 the vertical positions
of the units are reversed with respect to those of
Fig. 17 shows the spread in axial frequency
the LS5/5, the bass unit being mounted uppermost
characteristics of six complete loudspeakers. It
as in the LS5/2A. This is done in order to keep
should be noted that the trend of the curves is more
the bass unit near to the main reflecting surface in
uniform and the spread is appreciably smaller than
the room, in this case the ceiling.
that to be expected in practice from moving-coil
microphones and even from many electrostatic micro-
7. REPEATABILlTY OFTHE TYPE LS5/5 LOUD- phones. In the past, the monitoring loudspeakers
SPEAKER have been the least predictable link in the studio
Some experi ence of the repeatabl I i ty of the low- chain, but with the introduction of these new loud-
frequency unit has been obtained and was described speakers thi s shoul d no longer be so.

. ,--_ ..

5~B

--
~
-I ~ f'v-.. ...
~~ ~
~
~ ~
!II 1 ~
II v
\h
IfI ,\\
frOqUjnc y ranga: u50:d
"I
0·03 0·05 0·1 02 0·5 1 2 5 10 15
fraquancy, kHz

Fig, 16 - Spread in axial response/frequency characteristIcs of six 200 mm units in large cabinet

c--... ~A /II'\. ;l\;.:r -"l

~
~ --' -::::7 '-'= rYr v l V'"
\
v
7 )
5~B
:i
003 005 0·1 0·2 0·5 1 2 5 10 15
fraquancy. kHz

Fig. 17 - Spread in axial response/frequency characteristics of six LSS/S prototypes

12

3
5~B
T

7
---ro-- --
2 3 4
- r- r-
5 7
I"-

2 3 4 5
ru- ~t"c
7 2
fraquancy, Hz

Fig. 18 - Mean spherica1 response of LSS/S loufj,speaker measured in octave bands

15

10

...- - lD' ".

----
dB 5
~
~
o

4 5 7 10"Z 2 3 4 5 7 tOl 2 3 4 5 7 2
frllquancy, Hz

Fig. 19 - Directivity index of LSS/S loudspeaker measured in octave bands

tap connactlons to ba adjust ad

if nacassary to suit flux dansity
of spaakar units (at tast staged

T1

- -., R1 68 4 I W

input
I
!
4
5
L.F. .
MF.I

H.F.
I
I
i
~
~~m r.
.
I 6 . ~6 M.F. unit
L _______ .J
12

C5
L---~---ill---~---tlt-----------------------~O+

H.F. unit

all componants t. 2 0,.

Fig. 20 - Circuit diagram of crossover network of LS6/S and LSS/6 loudspeakers
 13

100
50 '\. L t--"
. / L -...........

!
~20

.S
015
3,0
5
5 " .........
--- ::-----
~
~
~
- ,..-: -
o
0·03 0·05 0·1 0·2 0·5 1 2 5 10 15
frequency. kHz

Fig. 21 - Modulus of impedance of LSS/5 and LSS/6 loudspeakers

0
fundamantal
-10

-20

-30

-
-40
l\_~_ ..... .•.. ·······l····.. 2nd harmonic
-50
U" '.':::<,. I.~ L·,:::::· ....../ . / ... \ 5th haiJJ10nic 1'.2 d£" .·.t\
dB -60

-70

-80

-90

-100

5 10 15
frllquency. kHz

Fig. 22 - Harmonic distortion of LS5/5 loudspeaker measured at 1 N/m 2 at 1'S m

8. DIRECTIVITY overloading than would the 380 mm unit employed

for the LS51 1A loudspeaker. Fig. 22 shows the
The variation in mean spherical radiated power curves of harmonic distortion measured on the axis
as a function of frequency was measured by the of the compl ete LS5/5 loudspeaker at 1·5 m for a
use of octave bands of noise. It is shown in Fig. sound level of 1 N/m2 and Fig. 23 gives the corres-
18. The corresponding directivity index* is given ponding curves for i ntermodul ation tests; these
in Fig. 19; the variations of both quantities with curves incl ude the effect of the variable impedance
frequency are less than those of the LS5/1A and load on the power amp lifier, and were obtained by
LSS/2A and very much less than those found with special apparatus(4) designed for this purpose.
any other loudspeaker which has been tested.

9. IMPEDANCE AND DISTORTION CHARAC- -20 I----+----j----j---+--j--t-t---i-t-t--t--N

TERISTICS
-30 I----+----j----j---+--I--t-t---i-t-t--t--H
Fig. 20 gives the circuit diagram of the cross-
over network. The inductors in all cases have
Radiometal cores and operate well below the
saturation level. Fig. 21 shows the modulus of the
impedance of the loudspeaker measured on the 25
ohm tapping of the auto-transformer.

Early tests on the 305 mm unit indicated that

it would del iver a higher level of sound without

* The directivity index of a loudspeaker is the logarithm 1rrzquC'lncy, kHz

to base 10 of the ratio of the sound power which would
be radiated if the free-space axial sound pressure were Fig. 23 - Intermodulation distortion of LSS/S loud-
constant over 47T steradians to the octual sound power speaker measured at 1 Nlm'2 at 1·5 m
radiated.
14

The level of the sixth intermodulation product was ment, which provides the usual balanced bridging
too low to measure. It will be seen that the dis- input impedance and also the bass pr~emphasis
tortion levels are quite low even at the lowest fre- circuits, mentioned in Section 4.3.1, which give a
quency at which each unit is used, thus indicating rise of 4. dB at 40 Hz for the LS5/5 and 7 dB at
that they are being operated well within their limits. 40 Hz for the LS5/6.
The distortion curves shown in Fig. 14 of reference
4 were taken on the type LS3/1 loudspeaker at the
11. DIMENSIONS
same sound pressure and comparison with Figs. 22
and 23 shows that the distortion levels of the new The LS5/5 loudspeaker cabi net is approxi-
loudspeaker are appreciably lower than those of the mately 350 mm wide by 430 mm deep by 660 mm
old design in spite of the fact that this used a high, giving an external volume of 0.1 m3 . It is
larger (380 mm) low-frequency unit. mounted on.a plinth, 520 mm high, which houses
the power amplifier. The LS5/6 cabinet is of
i rregul ar shape but has the same vol ume as that of
10. POWER AMPLIFIER the LS5/5; photographs of the two model s are
shown in Figs. 24 and 25 respectively.
A commercially produced transistorised power
ampl ifier is used, capable of supplying 25 watts The weight of the LS5/5 loudspeaker together
into a 25 ohm load. Associated with it is a pre- with the power ampl ifier is 47 kg, that of the LS5/6
amplifier, designed by the BBC Designs Depart- without amp I ifier is 35 kg.

Fig. 24 - Studio monitoring loudspeaker LS5/5 (free-standing version) with and without front cover
 15

Fig. 25 - Studio monitoring loudspeaker LS5/6 (hanging version) with and without front cover

12. CONCLUSIONS 13. REFERENCES

Details have been given of the considerations 1. The design of a low-frequency unit for monitor-
which have led to the design of two new monitoring ing loudspeakers. BBC Research Department
loudspeakers which are suitable for studio or out- Report No. L-065, Serial No. 1966/28.
side broadcast use. The sound qual ity and di rec-
tional properties are appreciably in advance of 2. OLSON, H.F. 1957. Acoustical engineering,
those obtained from the loudspeakers at present in pp. 36 and 44. New York, Van Nostrand.
use and the maximum output level is higher. The
spread in frequency characteristics between develop- 3. LEVINE, H. and SCHWINGER, J. Physical
ment speci mens is extremel y small, and the I evel of review, 1948, 73, No. 4, pp. 383 - 406.
non-linearity distortion is low.
4. HARWOOD, H. D. Apparatus for measurement
of non-I inear distortion as a continuous func-
tion of frequency. BBC Eng. Monograph No.
49, July 1963.
16

APPENDIX

Musical Excerpts used for the Experiment on Bass Equalisation

Item No. Title Type of Music Length of Excerpt

a Gotterdammerung (Wagner) Orchestral 35 sec

b Schwanda (Wei nberger) Orchestral 55 sec

c Prelude in G (Pierne) Organ 1 min 41 sec

d Fiat Lux (Dubois) Organ 1 min 30 sec

e The Gee Men Saturday Cl ub 1 min 41 sec

(Swinger from Seville) (pop)

f Billy J. Kramer with Dakotas Saturday Club 1 min 12 sec

(It's all over now baby blue) (pop)

g Billy J. Kramer with Dakotas Saturday Club 1 min 30 sec

(We're doing fine) (pop)

h Mars from Planets Suite (Hoist) Orchestral 52 sec

i Mars from Planets Suite (Hoist) Orchestral 25 sec

j Jupiter from Planets Suite

(Hol st) Orchestral 51 sec

k Overture: Scapino (Walton) Orchestral 1 min 30 sec

.JMP

Printed by BBC Research Department, Kingswood Warren, Tadworth, Surrey