acoustic - CARVER site!

Transcription

acoustic - CARVER site!
DAVE BERRIMAN of
PRACTICAL HI.FI
REVIEWS
ACOUSTIC
RESEARCH
ARgO
LOUDSPEAKERS
,r,,,?
would
ments introduced in the brilliant AR9. Scaled to fit handsomely in most listening
repetitive. lrrstead l'll highlight
those areas which differ. For
backgrouno rnformation refer
to my Audiolab review on the
AR9s in the February issue.
HE AR9O
- a four-way,
system
I ftoor-standing
I incorporating refine-
rooms. " Not my words. I hasten
to add. but AR's, from a Stateside promotional leaflet on the
new AR9O.
Allowing
for their copy-
writer's enthusiastic superla-
tives it does just about sum up
the AR90s. They are a sort of
shrunken version of the ARg
(which I reviewed in the
Februaryl 979 issue). The same
tweeter, upper midrange dome
and lower-midrange units are
used in the AR90s. The rather
less voluminous
measure
1.1
AR90s
02mm high
x
368mm wide x 386mm deep
1433/a
x
litres
(3.1
1
41/2
x
151% o-inc hes),
givrng a volume of some 90
8
cubic feet)
and
weigh a mere 37kg (82 lb) nowhere near so intimidating
in domestic surroundings as
the ARS but nearly as back-
dislocating Fortunatelyfor
become rath.er
lnside the AR90s
Drive units of the AR90s
differ in the bais where two
eight-ohm 25Omm (ten-inch)
paper-coned drive units are
incorporated. These are connected in parallel to give the
equivalent of one larger fourohm unit. and the drivers are
mounted on the floor to the
sides of the enclosure. As for
the
AR90s.
this form
of
mounting eliminates response
crevices due to the floor and
walls which it used to boost
bass output by as much
as
9dB. Unlike the ARgs, there is
no need for a tuned circurl 1a
flatten response and rarse the
electrical impedance. To
achieve the correct low-fre-
quency response. the drivers
are weighted with lead.
once in position than man'/
Internal acoustic damping rs
greater than the AB9s and the
large monitor loudspeakers on
stands due to their reasonablylow centre of gravity.
claimed by AR to be at 0.7. a
little higher than the AR9 Of
sqfety, they are more stable
Not only is the size
and
weight lower than the ARgs.
'but
also the cost.
Since the AR90 is
a
develop-
ment of the AR9, I won't go
over their every feature which
overall O
of the system
is
course, the crossover for these
two drive units in the AR90
is
totally different .from that in the
AR9s. AR tell me that the filter
secttons on the other units
remain unchanged. though the
loudspeakers both sound and
measure differently in the
region covered by
these
recessed w th the MA3sl
and
(seer,'r-
ful . s rghtly dull sound wh ch
was not thought strictly
cymbals were sharper
rngly emphasrsed slight
y)
Kick drum w th the ARg0s was
crisper. havrng greater punch.
solldity and definrtion.
Lookrng back over my
ong nal notes on the ARgs
there s absolutely no doubt
that the AR9s produced a less
cnsp bass end compared to the
MA3s
So. the AR90s
obvrously have a ba3s performance, which out performs the
AR9s in terms of punch.
solidity and defrnition The
difference was clearly
apparent to me even wthout
making direct comparrsons.
but at least double-checktno
rn
thrs way does ensure that
rrve rs.
Aesthetically, the AB90s are
more pleasrng simply because
of their smaller size. So.
the
and
AR90 is smaller, lighter
cheaper than the ARg with
which it shares a design philosophy and a numberofcomponents. On the other hand it's
still a large, heavy and
expens,ve loudspea<er .n
context of the Brittsh market
and so deserves a critical
assessment. How wouid it
sou nd
?
AR90s in use
As usual. the AR90s
were
instantly inserted into mV
resrdent system which now
rrcher tonal balance
slightly less apparent openness. Bass seemed solid
-
more predo.nrnant dnd cr,sper
than the MA3s. Mid-band of
the AB90s was -nore convtnL-
ing
reproducing the nch
fruity- sound of the saxophone,
for rnstance, more naturally.
Treble of the AR90s was
smooth yet slightly dull. which
resulted in a slight y muff ed
effect. Reproducrng
massed
choral works, voices yra the
AR90s seemed less'blurred
rouno rhe eciges though the
individuai voice parts were
about as easy to separate wlth
both loudspeakers.
Stereo image of the ARgOs
included a Mission 774 arm,
was not
amp ifier,
cartridge,
Denon AU32O transformer.
listenrng inlo the sound stage
Sansui AUs17
Denon DLl03
Linn Sondek turntab e
and
Pioneer TX9500 tuner.
Using the AR9Os wth the
controls al setto 0dB (wh ch is
maximum on all drve unrts)
compared them to a pair of
I
Monitor Audio MA3 Series
11s. The AR9Os made the
I
have not been mistaken.
and
pin potnt in nature hut
revealed that. like the AR gs. the
AR90s p,oouce a conv ning y coherent type of stereo
rmage. which rs not an
exaggeration of life but
a
repreSentdt,Oa Of it especrally when reproducrno
we I recorded orchestral
works and sma ll musica
qroups.
MA3s seem tunnelly and nasa
by comparison
closed-in
- clearer
with a superficially
Usrng a studro-produced
close-miked track (Joan
if
less tonai ly realistrc sound. The
Ar
AR9Os. on the other hand.
produced a sound which was
away from the boxes with a
Joan's
voice stayed well tn focus (not
pushed back as vra the ABgs).
Now her voice seemed slightly
matrad,ng. Opportunityt
was interesting that
Fig 1: Sinewave frequency response measured at one metre on the
tweeter axis. lRead off the O-sOdB scale for all graphs- lnput equa.ls 2.83
volts rms for all curves except Fig 7)
ti
smooth
treble have been retained. but
plucked sounds. snares
d
stereo lmage and
Background revei'b was
easier to hear wrth the M43s.
Going
back to
my
comparrsons between
the
AR90s and the MA3s: overall,
the AB9Os on each track
played produced a less
coloured more convtncng
sound albert with a comparatrve lack of upper mid,/treble
the AR90s possess a similar
accurate. The music. however.
was getli'lq lhrolgh we
Obv olslv. wttn encoJ,agtng
I
results like th s. further
listening agarnsta neweT, more
sophisticated and more
n\pens;ve oudsoearer desrgr
was necessary Those closest
to the AR90s rn price available
fcr comparison were KEF
105s, which are being
used
increasingly as a reference
these days.
Round two
ARgOs
v KEF 1 05sChanoing the auxilary
equipment to Ouad ll valve
power amplifiers in order to
reveal subtle differences rr
texture and fine detai moTe
clearly, and using the
Meridian 1O3 transistorised
pre-amplifrer with built-in
Supex movrng-coil module.
the tests conttnued. Later on
I
used the Trevor Lees
Positive Vibrations valve
pre amplrfier and pre-pre,
clarity. Piano reproduction was
amplif rer.
AB90s. conveyrng the power of
the rn5trLrn-ont and al.owrrg
one to hear the individua
notes as they are struck even
wher- trght y brnL hed l
o{ low power. they were not
d'rven tnto c,opt1g oLrtrg
exceptionally good with the
compiex passages.
It may lnitiaily seem unf air to
set the AB9Os agatnst
the
MA3s.
much less-expensive
(which in context of their price
still feel have much to offer n
terms of musical analys s) but
readers of the AF9 review w
remember that the AR9s fared
I
I
leSS we i,. T y md 1 c, ', u S'l
be,r-g a puch ng Ddc" ot \ o p.
both male and female on var-
Though the Ouads are only
these tests qnd provided sound
levels which were quite
adequate,f rol ear-shalterrl[j
with the KEF 105s. Later on,
tests at a higher level wrth the
Sansui AU51 7 conf irmed my
findings revealed with
only more powerfu. but was
ious types of programme
matelal. Playing the same
also more reiaxed
materia through theAR90s,
speaker's stride.
Treb e of the
as
reported on here. revealed that
thrs charactelstic had
been
much reduced or eliminated. In
add
ition. bass rea lrsm
has
been rmproved in the AB90s.
Posrtrve features audrble rn the
AR9s, such as the r upper
bass,/lower mid-band detai
rea'strc yet
J1i1-press,\e
,
the
Ouad 1 1 s.
Startrng off w th the AB90s
set w th all dr ve un t sw tches
at OdB as before. the ARgOs
revealed again a rrcher sound
with a f ul er deep bass delivery.
Organ. for nstance was not
freer as if
- the oudit were easrly within
1
05s
was
brighter resultrng in a thinner
rathe" ttzzy sou'lo he'e.'
Reproducing programme
malerrd w th some treo p
distorlron the 1O5s wcre mo,e
irritating due to their brighter
balance and though strrngs
were less
brg
ht with
Fig 2: Sinewave response at onb metrc measured mid-way between
tweeter and mid axis, upper-mid and lower-mid controls at+3d8, tweeter
at odB
the
AR90s, the delicate feathery.
rosined texture of
massed
strings was better revealed by
the latter. There was a certatn
degree o' nasal coio'at o1 ,r
the KEFs as opposed to the
AR's
mro
baeo
Cotoratron
Overa , the 1 O5s lacked body
by comparison. lt was felt that
be both subJectrvely
1
KEF
O5s
Testrng for coloration now
that the responses were mo.re
al ke (we also measured the
KEF 1O5s to further confirm
th s) proved nlerest rq.
upper bass,/m d-band. Bass of
the KEFs proved to be a tri{ e
flacc,d ano sporgey. rack'ng ir-
lnitia ly, w th the contro s at
zero. it was thought that the
i 05s were less coloured. but
now which ever way I carr ed
out the companson one loud
speaker parr made the other
pair sound colouredl
punch and attack.
Stereo image of the 1O5s
was excellent srmi ar to the
AB90s with good ooherence,
seemed less coloured tonally
on applause whlle the ARg0s
still had a coloration in the up
the 'body of the AR90s was a
little over done with a slight
excess somewhere rn the
naturalness and depth
Superficially. like the MA3s.
the 105s are clearer but the
105's mid-band is slrghtly
deader and muff ed and I think
that it's th s effect coupled with
their bass charactelst c whrch
makes the 1O5s seem less
to lrsten to than the
AR9Os or MA3s
Even so. the warmer and
duller balance of the AB90s
stiil struck me as berng less
excrting
accurate than the 105.c. At this
point lstarted to experiment
with the controls, reaiising that
the predominance cou
d
possibly be compensated for
By switching the
2OOmm
(eight-inch) lower m drange
unit's control to -3dB. the
balance much improved and
became closer to the KEFs.
Subsequent .measurements
revea ed
un
KEFs were nasal but
pe, bass/.ower-mrd regron.
thouqh now this was much
reduced due to the swltch
The latter was about
3dB
down, obviously grving rrse to
my comments about a du
hr gh-f requency performance.
New control settings
All further listenrng tests
105s were less coloured on
applause, i//a the AR90s the
individual handclaps
stayed
more separate and were easrer
to differentiate between
sl ghtly over-full mid-band and
a sublective reduction in h gh-
Irequency orrrlness'es
J,1
r1g r-
and
Male voice reproduced by
the AR9Os still had a slighty
husky character. but was
ludged to be rather thin and
unmasculine via ttse 105s.
with the AR90s: whether l was
bass guitar rhythms. complex
piano passages, massed
was
that while the KEFs made male
voice seem more feminrne and
strings. the wrnd noises of
a
f lute, the rosinysound of a solo
violrn or the punch and impact
of a kick drum. the AB90s
Paradoxrcal
y, the
the
KEFs
produced a seehrngly c earer
view of the scene in which the
musicians perform. whereas
the AR90s produced a
ess
c ear vrew (which could even
be interpreted as clouded or
distanced on first hearing) but
reveal the nature of the instru
ments, their acoustic textures
upper midrange swrtches both
the 3dB posit on as th s
Thrs s the overal character of
the AR90s and sums up their
balance which was thought to
agarnst the 105s. However.
main areas of improvement
were now a reductlon In the
Neither ts correct of course so
rt's a waste of time debatrng the
point.
I
and how they were oeing
played, more convrncingly.
resulted rn a frequency
new control settings. but many
comments wil duplrcate those
in the or ginal compar sons
This effect was echoed
throughout the audio band
were done with the ower and
set to
I cou d go over the sublective results in deta I with the
may be slightly predominant.
Cur ously. though, whrle the
rnstruments more clear y. wtth
greater prec sion and definrt on than the KEFs.
bass'."l d reg;on do,vt no'p tn
line with the tweeter's output.
oud-
parameter.
range
3dB
y to the
However. if there is one
critrcrsm it is that the bass
perforrrance whrre drarr.atr-
musical characteristrcs of the
c cuTVe
further, br nging the upper
part cular
speakers when so set.
ferent ways
- you couldn t put
them in a ranking order on thrs
seemed to reproduce
mrdrange unrt to
these atter comments apply
extended treble performance
s
f attened the anecho
their upper and lcwer
mrdrange switches to -3dB
and I should poinl out that
ln a nutshell, both speakers
are equally uncoloured rn dif-
a
and also switching the upper-
the
AR90s d d sound much better'
when set to what I r-egard as
the r flat position. that rs with
a smooth yet analytrcal
th
t's operating
coloratrons. However.
posrtrons adopted.
that th s produced
flatter characteristic !n
dnve
The
response characteristics and
more
natural and closer to the
Capabrl,t eS better than any 'n-
depth analysis of frequency
My marn observation
nrolected femalo \orcp
comparatively well, the ARs
prolected male voice better
and made female vorces sound
more mascul ne. Thats abcut
as c ose as I can get in words
describrng the tonaL
drfferences on voice It also
concludes mV sublecttve
analysrs of the AB90s.
How they measured
Measuring the A890s at one
metre on the midrange ax s
revealed a large dip of 1 5dB at
7kHz. reminding me of the
25dB dip measured at
this
positron with the ARgs.
Again, it's a cancellatton effect
i
n the crossover reg on
between upper mid-band and
tweeter which rs ellmtnated
when measurrng etther on the
tweeter axrs or midway
between the tweeter and
N
Fig 3: Sinewave response at one metre on tweeter axis with tweeter
control at OdB. -3dB and -6d8.
upper-mrdrange axes.
The best response was
chtarr eo rr 1.e ralle r-oi Iral
an'J s 3 'eai,rt.'redsJrt'-rg
point srnce rt coincrdes alrnost
y with my ear height
when lstening on a ow
exact
sofa. ln contrast w th Ihe ABgs
the m dr-ange axrs more or less
co rcrded w trt llp lstel 1g
axis. hence resu ting n a 25dB
dip at listenrng
he
ght
as
opposed to the f at respcnse
w th the Af?9Os This rnay n
part expla n theARg0's livelier.
,eSS resl'd naC sn Ja(l pa.l arlv whcr reorca,. 'rg \o .?.
I don't thrnk t s the who e
I
story. for e'ren olacrng the
sanre dnve rnrts and crossover
nto t\^/o drfferently drmens oned enclosures of similar
internal vol-me can esr, t tn d
different sound. Ihis s Cue to
drfferences in d ffract on
patlerns from the front baffle
and di{fering air and panel
resonances of the enclosure
Looking at Frg /, which
shows the response at one
metre on the tweeter axls.
reveals a narrow dip arounC
1.9kHz flanked by two humps
in the ;esponse wlth a treb e
output about 3dB below the
mean
m
d band level. This
undoubtedlv accounts for rny
commenls about treble dull
ness. All rnsrst that
theV
measlrre the AR9OS and AB9s
rn TOoms and though I concur
with thrs phiiosophv rn the lowfrequency reqion where walls
,rvill augment the response. at
rrid and h gh frequencies the
absorptron characteristtcs of
the room v,, I affect the sound
JJst,l: 'r a .r'rrdr
Sltratro
t'oom
ldeally, bringing up
tweeler. output would
the
ha're
flattened tne resoonse. butthts
was not possrble, so the upper
and lower r.rdrange controls
were both set at -3dB
!-esponse. it a so sol,nd.s f latler
and less coloured iike this Frgs
3, 4 and 5 shcw the e{fect of
the controls. Note that output
hv
Fig 4: Sinewave response at one metre on tweeter axis'with upper-mid
control at OdB, -3dB and -6d8.
as
descrbred to give the f atter
curve shown tn Fig 2 wh ch
w'as measured between the
upper mrd and tweeter axes.
Note how much flatter rs the
can only be reduced
Orive units:
Tu,o 25Omm (1O-inch)
acoustic suspension
woofers.
2OOmm (8-inch)
acoustic suspension
lower-midrange driver.
38mm (1% inch) liquid
cooled dome uppermidrange driver.
19mm (% inch) liquid'
cooled dome treble unit
Crossover !requencies:
2@Hz, 1 .2k9z, 7kHz.
lmpedance:
4 ohms nominal.
3.2 ohms minimum.
Controls
Three three-position
switches for lolvermidrange, upper
midrange and
h igh-range
from
maximum by -3dB and -6dB
Power rsquirements:
50 watts minimum per
channel, recommended.
System low-Ilequency
response:
-3dB at 32 Hz
for all apart irom the bass
unrts. for which there rs no
Power handling:
May be used with ampliflers
capable o{ del ivering 300 watts
continuous power per channel
being driven to clipping no
more than ten per cent of the
time, using normal source
material (speech and musici.
Enclosure
Volume: 90 litres
(3.1 8 cubic feer)
Dimensions:
1 102 x 368 x 386mm
rs
Crossovei network:
Hal{ section LC networks on
woofers and lower and upper
midrange units. Network of
upper-midrange unit also
includes an impedance
equalizing circuit. Full section
network on high-range. All
capacitors are computer-grade
bi-polar electrolytics. All
chokes are air-core would with
17 AWG solid conductor
deep (4378 x 141/2
inches).
x
15316
Weight: ln cartoni 46kg.
(1O1 rb).
Unpacked: 37k9.
(82
rb).
DIN specitications:
Frequency range: 30kHz
lmpedance: 4 ohms
Sensitivityi 8 watts
Nominal
power
handlingi 1 50 watts
Maximum
power
handling: 250 watts
control of course.
The ofr-arrs perforrnance
shown dotted tn Frg .6
sdperrmposed over the o1 alts
curve. Note how the treble
response s well maintained
and that the outputs of the
indivldual units remain well
integrated even off axis. Note
too, how the dip at lust below
2kHz vanrshes off axis wh ch
indicates that this rs most likely
simply due to cabinet diffrac
tion. an effect whrch I would
have thought, should
have
been elim nated or reduced by
AR's acoustrc blanket
the
layer of felt over the front-of the
policy of mounling the drive
unrts as close to the floor and
e to give
bass
operating range. ln
an
b
augmentatron over their
anechoic room the absence of
these boundarres results rn a
l.ll.i--i]= rllJ i l- l-..
;if ='
r TtErll
liiri{i
#-+i+81
L++if
il
ll:f+:{H-€
1,
'i
l .l..,.i
-r;J r ii1,r,i I
rlrt
=
=F
l
I
.!
I
;-
_l il
made. one at 0dB and one at
my"flat' setting. The
curves
have not been reproduced due
to lack of space. ln both cases
there were dips of fractionally
over three ohms at between
TOHz to 14ONz and ar 1k\z.
With the units fully in. a dip of
low-frequency peak shows
a
Thus the AF90s lust scrape
intoAR'sspectf cat onof afour
ohm nominal impedance wrth
a minrmum of 3 2 ohms. but as
with the ABSs you will need to
be careful about using
dmpr I er of s._t",crenl
qLa
an
ty Io
look ng atthe curves Menta ly.
add, 6dB to 9dB or so progress vely to the anechoic cLrrve to
passages
grve a rough tdea of
the
.-UUir-.
t'lc
in-room performance below
iviu!!
ruvt.
anechoiC Toom is not
D rsto rtion
Drstortron curves weTe taken
current cielivery on
loud
Summing up
Conc uo,rg .rv slbJect ve
and oblective look at the
Afr90s t s piarn tnat lm more
enthus astic about these oudspeakers than the AR9s They
offer advantages tn terms of
price and size, whrle still
re'na n rg
pxpels,ve.
Comparing to a state-of-theart Brtrsh design ike the KEF
1
05 reveals, on the other hand,
that the AR90s can do more
than hold their own.
The KEFs set a high standard
pr ce wh ch many
fol owing table (see Ftg 7 for
the curves).
at therr
Second harmonic:
4OHz 5.0 per cent
achieve, so l'm tn no way
50Hz 3.5 per cent
AR90s need to be better than
the 105s to iustifv their extra
desrgners aspire to but farl to
denigrat
140H2 0.8 per cent
1.SkHz 0.71 per cent
Elsewhere: below O.56 per cent
Third harmonic:
4OHz 1.8 per cent
ng the 1 05s.
The
cost and I believe that f the
reproduction of musica
rnstrunrents. with their tonal
and rhytl'n-ic cha.aC.{errqt .5 s
TOAHz O.11 per cent
used as a criterion. then the
Elsewhere: below 0.56 per cent
AR90s do lustrfy their expense
and size. I do feel, thouoh. that
they perform at therr best when
adjusted for the r flattest
anectro c 'esponse d1d rT rs
Distortron at low fre
quencies rs not qutte so low as
wth the AB9s (2 5 per
cent
second harmonic at 40Hz and
50Hz) but s st I very lorv, nodouot a doo oy tne o.ra,-dl\e,
approach. Dont forget also
that n a room the funda
menta w lbe increased at low
freqencres due to the bass
driver's ciose proxrmity to the
,loor and wolis. lhls reducinQ
the distort on percentage by
more than a half.
FigT: Distonion: Above is response for gOdB spl at one metre, measured
mid-way between upper mid and tweeter axes. lJpper and lower mid
units both at -3d8. Solid line beicw is second harmonic and dashes
represent third-harmonrc dislort ion.
Finally comes rmpedance.
Two measurements were
avord cl pping during heavy
ower
n the
-3dB position w th the AR90s
producing the statutory 9OdB
at one metre at 1 kHz Frgures
were excellent as shown n the
'at odB.
sensttivity.
of bass and thrs
should be allowed for when
wrth the upper and
midrange un t controls
Fig 6: Sineutave response at one metre mid-way between tweeter and
upper mid axes. Upper mid and lower mid units both set to -3d8, tweeter
AB90 is oi above average
severe lack
completely re iable anyway.
Fig 5: Sinewave response at one metre on tweeter axis with lower
midrange at OdB, -3dB and -6d8.
sensitivity figure is quoted by
AB but these indrcate that the
well damped bass resonance
at lust below 40H2.
wall as'poss
ll+l
before,
sensitrvity was reduced slightly
to 82 sdBA and 868 lin No
be
l
l.ti
flattest reponse, as
alarmed by the rapid bass fa
off below 150H2. This ,s qurte
simply due to AR;s del berate
consequence. Don',t
i].t,i.,
and
B6dB lrn. (unweighted) with
cortrols rn the 0dB pos l,o1
Wrth the controls set for the
lust over three ohms also
appeared at 2kNz. The
characterstic infinite baffie
loudspeaker. Still. it is of llttle
Manuracturer: Teledyne Acoustic Research,'lO American Drive, Norwood,
Ma O2O52 USA.
High Street, Houghton Regis, Dunstable, Bedtordshire, England.
to be B4dBA (weighted)
Sens rtrvrty
Sensrtivity was measured
with oink norse and was foLtnd
whe- used ,r' t-,s condrt,or
that lbase my overall high
assessment. Differing tastes
and svstem
characteristtcs
may, on the other hand sug-
gest alternatrr,e setlrngs
rr-
some rnstances.
How,ever, putting tonal
analyses aside the AR9Os are
loudspeakers io' entov 1g
music. lf that s your arm and
your pocket's deeP enough l'd
recon'mend so.re selous
listen,'lg r1 lhe,- directior