The aforementioned controls are on the

top surface of the amplifier chassis, as is the

rotary power switch plus four green LEDs
and four access holes for bias adjustment.
Also atop the chassis, just up from the rear
panel, are four binding posts for matching
VAC P A 1 6 0 loads of 1 to 2, 2 to 4, or 4 to 8 ohms—
ranges chosen to let you select a tap to
M O N O A M P L IF IE R match a speaker’s nominal or minimum
impedance. (VAC suggests putting a 4-ohm
speaker on the “4-8” tap for maximum
power.) A small, two-position toggle switch
can be used to bypass the chassis connec­
tion to the a.c. third-wire ground via a par­
alleled resistor and capacitor. The bypass
breaks d.c. ground loops that could cause
hum and noise. On the rear panel are a
combination fuse-holder and IEC power-
cord socket, plus RCA and XLR connectors
for unbalanced and balanced signals.
Taking off the bottom plate of the all­
alum inum chassis reveals an interesting
wiring technique not often seen in audio
products. A large fiberglass board, about Vs
inch thick, covers the internal area of the
chassis and is spaced away from the chassis
by perhaps another Vs inch. This board car­
ries all the top-m ounted chassis com po­
nents except the transformers, but does not
connect them. All connections are hard­
wired, via turret terminals staked to this
board. (The only actual p.c. board is small
and holds the bias pots and LED-indicator
circuitry; it is mounted to the main board.)
Teflon-insulated wire is used; the general
appearance and neatness of the wiring are
n 1990, the Valve Amplification Com­ are “6 L 6 /K T 6 6 ,” “E L 3 4 /K T 7 7 ,” a n d nothing to write home about, but VAC says

I
pany introduced its PA series of tube “6550/KT88.” (That last position will also this wiring arrangement is deliberate, to
power amplifiers. These amps struck accommodate KT90s, according to VAC.) avoid the stray couplings that can occur
me as truly beautiful, and, presumably, with wires paralleled in neat bundles. Parts
they were well executed. I am also con­ quality looked to be appropriate, with a
siderably impressed with company THANKS TO ITS VARIETY num ber of metal-film resistors and by­
president Kevin Hayes’ knowledge of vacu­ OF USER ADJUSTMENTS, passed specialty film capacitors in the signal
um-tube design. path. The main chassis piece is bent to form
The PA160, one of the newest products
THE PA160 SHOULD
the top and sides. Front and rear panels are
from VAC, has a variety of user adjustments SATISFY A WIDE RANGE separate pieces, bolted to corner brackets
that should enable it to satisfy a wide range OF MUSICAL TASTES. that tie everything together. Half-inch lips
of musical tastes. For starters, there is a at the bottom of the main chassis piece are
three-position rotary switch for triode, Ul­ lined with Pemm nuts for bolting on the
tra-Linear, or pentode (beam-power) out­ The amps are normally supplied with Gold­ bottom plate.
Photos: Michael Groen

put-tube operation. Another three-position en Dragon KT88s, from China, which VAC
switch selects bias range and B+ voltage for feels possess a nice combination of image Circuit Description
three families of output tubes, and a six-po­ size, tonal weight, and openness. With the A 6SN7 octal-based dual triode, known
sition switch allows you to vary the amount supplied KT88s, power output is a healthy to be a linear and good-sounding tube, is
of negative feedback. Output tubes associ­ 160 watts in pentode mode and a very us­ used in both the first and second stages of
ated with the three bias/B+ switch positions able 80 watts or so in triode mode. the PA160. The first stage is configured as
what is known as a floating actuality, this amounts MX LABS LIUELCdBr' v* nOQ(Ki)
+5
paraphase phase inverter. In to a stage that provides
this arrangem ent, the input one o u tp u t phase d i­
signal is applied to the grid of rectly and couples that
a triode acting as a common- output to another stage,
cathode amplifier, which pro­ which has 100% invert­
vides one phase of the stage’s ing feedback, to provide
output signal. This output is the other output phase.
resistively coupled to a point From a circuit purist’s -20
10 100 IK 10k 200K
that I shall call “point x.” The standpoint, this scheme FREQUENCY - Hz
other triode’s plate, also resis­ has the disadvantage
that the second generat­ Fig. 1— Frequency response
tively coupled to point x, pro­
at " 4 - 8 " output tap.
vides the other phase of the ed phase goes through
stage’s output. A coupling ca­ one more tube than the BHX LABS LIUKL(dlr)
mi nUCQ(Ki)

pacitor from point x drives the first generated phase.

grid of the second triode. In Further, the output im­
pedance of the two phases is likely
to be quite different.
SPEC S The two output phases of the
phase-inverter first stage are ca-
Power O u tp u t into 4 O hms: W ith pacitively coupled into the grids of
KT88 o u tp u t tubes, 160 w atts in the second stage. This stage acts as
pentode mode, 152 watts in U ltra- a short-tailed push-pull amplifier
with cathodes returning to ground FREQUENCY - Hz
Linear mode, or 79 watts in triode
mode; with other tubes, in pentode through a common, unbypassed Fig. 2— Frequency response
mode, 125 watts with KT77s, 150 cathode resistor; the plate outputs vs. feedback setting, 4-ohm
watts with EL34s, or 100 watts with of this second stage are capacitor- load on "4 - 8 " tap.
6L6GCS. coupled to the output tube grids.
THD: At full rated power, less than According to VAC, the modest dif­
2%; at 10 watts, typically 0.23% at 1 ferential function o f this stage
kHz, less than 1% from 20 Hz to 20 helps to offset any imbalances
kHz. from the paraphase inverter.
Pow er B a n d w id th , U ltra -L in e a r All four output tubes have po­
Mode, with KT88s: 12 Hz to 42 kHz, tentiometers to provide adjustable
+0, -0.5 dB; 8.5 Hz to 75 kHz, +0, negative grid bias in two ranges,
-3 d B . as co n tro lled by th e tu b e -ty p e
Frequency Response at 1 Watt: 5 Hz switch. Each output tube’s cath­
to 48 kHz, +0, -0.5 dB; 3 Hz to 82 ode returns to ground through a
kHz, +0, -3 dB. 10-ohm sampling resistor. In an
S/N: Greater than 88 dB. arrangement used successfully in a Fig. 3 — Square-wave
O utput Noise: Less than 1 mV. number of other tube power am­ response fo r 10 kHz into
S en sitiv ity : 0,34-V in p u t for full plifiers, the voltage drop across 8 ohms (top), 10 kHz into
o u tp u t, w ith KT88 tubes, U ltra- each cathode sampling resistor is 8 ohms paralleled by 2 pF
Linear mode, and feedback setting compared, in a quad comparator, (middle), and 4 0 H z into
to a reference that represents the 8 ohms (bottom).
“D.”
Power C o n su m p tio n : At idle, 288 correct cathode current. The out­
puts of each comparator drive the appro­ Two power transformers are used in the
watts; at full power, 528 watts.
Dimensions: 1134 in. W x 2014 in. D x priate LED indicator. Correct bias for each PA 160. One is for the high-voltage supply,
814 in. H (29.9 cm x 52.1 cm x 21.6 output tube is obtained at the adjustment and the other provides tube heater power,
point where the LED just goes out. The bias voltage, and power for the ou tp u t
cm).
overall negative feedback, which is taken tubes’ bias-setting circuitry.
Weight: 53 lbs. (24 kg) each.
from the “4-8” output tap, goes to the feed­ The output of the high-voltage supply
Price: $4,990 per pair.
C o m p an y A ddress: 164 S arasota back switch. At this switch, adjustable feed­ transform er is rectified by a full-wave
back resistors are selected and applied to bridge; it is then applied to four 330-pF,
Center Blvd,, Sarasota, Fla. 34240.
For literature, circle No. 95 the cathode of the signal input tube of the 45G-V filter capacitors in a series-parallel
first stage. arrangement. This arrangement yields the
 M LABS (x) <m w w u rri UUH(U) main B+, and is bypassed with an­ set to position “A” (no feedback) and to
other 330-pF, 450-V capacitor. In a position “D.”
similar fashion, the supply to the Frequency response in the Ultra-Linear
first stage is derived from the sec­ mode— as a function of open-circuit, 8-,
ond stage’s supply point through a and 4-ohm loading on the “4-8” tap— is
resistor and a final 330-pF, 450-V plotted in Fig. 1. As can be seen, response is
bypass capacitor. well behaved to beyond 100 kHz. The spac­
One aspect of the PA160’s circuit ing between the curves is uniform (suggest­
bothered me. The balanced input, ing uniform output impedance with fre­
which I assumed was balanced be­ quency); the amount of spacing suggests a
cause of the XLR input connector, is fairly high output impedance (no surprise
actually not balanced at all. Pin 2 of for tube amplifiers, especially those with
OUTPUT - WATTS
the XLR is tied to the hot pin of the low negative feedback). Figure 2 shows how
Fig. 4 — TH D + N at 1 kH z,
unbalanced RCA input connector. gain and frequency response change with
and SM PTE-IM distortion,
vs. power.
Pin 1 is tied to circuit common. Pin different feedback settings. For the pentode
3, which would be the negative mode (not shown), the open- and closed-
m ia h b m u ) «• m q a ti) phase of an incoming balanced sig­ loop gains were higher; the curves were
nal, is not utilized at all and is mere­ thus more widely spread between maxi­
ly term inated w ith a resistor to mum and minimum gain, and the curves
ground. Blasphemy! While VAC for higher amounts of feedback had some
never explicitly claims this is a bal­ high-frequency peaking. In the triode mode
anced input (the owner’s manual (also not shown), the gains were the lowest
only says it’s “for compatibility with and the curves were more closely spaced
balanced . . . sources”), I think that than in Fig. 2.
they should clearly and explicitly Rise- and fall-times for an output of ±5
state in the manual what was done. V into 8 ohms on the “4-8” tap were 4.3 pS.
According to Kevin Hayes, however, Figure 3 is a ’scope photo of square-wave
FREQUENCY - Hz
“The XLR jack merely provides a response. The top trace, for 10 kHz into an
Fig. 5— TH D + N vs. place to plug a [balanced] source in 8-ohm load on the “4-8” tap, is fast and well
frequency. without a clumsy adaptor, and pro­ damped. In the middle trace, with a 2-pF
vides a balanced load to each phase capacitor paralleled across the 8-ohm load,
I IMS 00 - Ota)
of the source.” the damping is superb, with rise- and fall-
times of perhaps 8 to 10 pS. This amp
Measurements would do well with electrostatic speaker
Since there were so many opera­ loads. Tilt, or pulse droop, at 40 Hz (bottom
tive combinations available with the trace) is quite low, indicating extended re­
PA160s, I selected the following as sponse below the audio band. Square-wave
the nominal condition for my test­
ing: Ultra-Linear operation, feed­
back at position “D,” the supplied
THE SQUARE-WAVE TEST
KT88 output tubes, and loading on
the “4-8” output taps. SHOWS THE PA 160
The voltage gain of the PA 160 WOULD DO WELL WITH
1 -kHz test signal. amplifiers is somewhat higher than
ELECTROSTATIC LOADS.
normal. This can cause such system
equivalent of a 330-pF, 900-V capacitor, problems as audible noise from preamp
which can withstand the operating B+ of output stages or preamp volume controls
over 500 V. The secondary voltage applied restricted to just the first few degrees of response in the pentode m ode did have
to the rectifier and filter system is con­ travel. The manufacturer realizes this, and a some overshoot and ringing at the higher
trolled by the three-position tube-type se­ modification will be available in the near feedback settings, whereas the responses for
lector. The h ig h est voltage is for the future to lower the overall gain if needed. triode mode were a bit slower than seen in
“6550/KT88” position, with successively (Kevin Hayes says that, in many instances, the photo but otherwise very well behaved.
lower voltages applied to the “EL34/KT77” higher overall gain can result in superior The THD + N at 1 kHz, and SMPTE-IM
and “6L6/KT66” positions. sound quality.) Voltage gain and related in­ distortion, are shown in Fig. 4 as a function
Supply voltage to the second stage is de­ put sensitivity are presented in Table I for of power output and load on the “4-8” tap.
veloped through a resistor connected to the the three operating modes, with feedback The curves for the triode and pentode
 frequency. In pentode mode, the PA160 had Q uicksilver Audio pream p or my own
a damping factor of about 2 and was simi­ preamp (a tube phono stage with a pas­
larly flat with frequency. In triode mode, sive selector and volum e control). For
the damping factor was the highest, at 4.2, CDs, I used C ounterpoint’s DA-11A and
but fell off slightly above 2 to 3 kHz. PS Audio’s Lambda transports to drive a
Dynamic power in the U ltra-Linear Sonic Frontiers SFD-2, a Resolution Audio
mode was about 116 watts into an 8-ohm Reference 20, and other (experim ental)
load on the “4-8” tap. With a 4-ohm load, D/A converters. Additional signal sources
the equivalent sine-
wave power at the vi­ TaDle I— Gain and IHF sensitivity. Position “A” is for operation
sual onset of clipping with no feedback.
was 171 watts at the
b e g in n in g o f th e Gain, IHF Sensitivity,
burst and 162 watts dB mV
at its end. For 2-ohm LEFT RIGHT LEFT RIGHT
loading, power at the Feedback at “D”
beginning and end of Triode Mode 36.3 36.2 42.1 43.7
THE POWERFUL the burst was 182 and Ultra-Linear Mode 39.2 39.1 31.1 31.4
AND WELL-DAMPED BASS 163 watts, respective­ Pentode Mode 42.) 42.0 22.2 22.4
ly. With 4 ohms as a Feedback at “A”
SHOWS THAT THE PA 160s
reference load and Triode Mode 39.8 39.6 28.9 29.5
CAN DEFINITELY the m a n u fa ctu re r’s Ultra-Linear Mode 44.4 44.3 17.0 17.3
KICK BUTT! spec of 152 watts, the Pentode Mode 50.9 50.7 8.0 8.2
dynam ic headroom
worked out to 1.0 dB
modes were, surprisingly, very similar in at the beginning of
magnitude and shape to the Ultra-Linear the burst and 0.55 at its Table II— Output noise levels.
curves shown, save that the clipping point end. Clipping power
was lower in triode mode. In Fig. 5, THD + (steady-state power at Output N oise. mV
N versus frequency and power for 4-ohm th e visual o n se t of Mode and Bandwidtn LEFT RIGHT
loading on the “4-8” tap, we see some un­ c lip p in g ) was 115 PENTODE MODE
usual behavior below about 1 kHz. This is watts for 8 ohms, 158 Widi Band 3.17 2.36
evident as a distortion increase that is most watts for 4 ohms, and 22 Hz to 22 kHz 2.88 2.14
noticeable at lower power levels. It turns 145 watts for 2 ohms. 400 Hz to 22 kHz 0.243 0.175
out that there is some 120-Hz power-sup- The readings for a.c. A-We ghted 0.430 0.341
ply ripple modulating the distortion. Figure line current were 3.0, ULTRA-LINEAR MODE
6 is a spectrum of the harmonic-distortion 3.9, and 4.8 amperes Wideband 2.26 1.63
residue for a 1-kHz signal at 10 watts into 4 for the same loads. At 22 Hz to 22 kHz 2.19 1.54
ohms on the “4-8” output tap. While the idle, a.c. line draw was 400 Hz to 22 kHz 0.183 0.128
actual harmonics fall off with order in a de­ 2.4 amperes. With the A-Wei ghted 0.323 0.242
sirably rapid manner, you can see the afore­ 4-ohm load as a refer­ TRIODE MODE
mentioned 120-Hz sidebands on each side ence, clipping head- Wideband 1.77 1.32
of the gap where the 1-kHz fundamental room was 0.34 dB. 22 Hz to 22 kHz 1.74 1.26
frequency has been nulled out. 400 Hz to 22 kHz 0.123 0.103
Amplifier output noise, as a function of Use and A-Weighted 0.245 0.199
bandwidth and operating mode, is listed in Listening Tests
Table II. For wideband measurements and R e c o r d - p la y in g
those taken over the range from 22 Hz to 22 equipment used in my
kHz, the readings are mostly hum, a mix­ system during the re­ Table III— IHF S/N ratios.
ture of 60 and 120 Hz. This amount of hum view period included
is a bit much and, as I later found, was an Oracle turntable S/N, dB
sometimes audible when I was near the fitted with a Well Tem­ LEFT RIGHT
speakers. pered Arm and Stan­ Pentode Mode 76.4 78.4
Dam ping factor in the U ltra-Linear to n ’s 98IHZS mov Ultra-Linear Mode 78.8 81.3
mode on the “4-8” tap was about 3, referred ing-m agnet pickup, Triode Mode 81.3 83.1
to an 8-ohm load, and was very flat with fe e d in g e ith e r a
were Nakamichi’s ST-7 FM tuner and 250 that was slightly more up front than is THIEL, continued from page 61
cassette recorder and a Technics open-reel my usual preference. However, the sound close to 20 Hz yet never lost definition and
recorder. Preamplifiers included a DGX Au­ was very realistic, w ith great dynam ics control or the ability to move the room at
dio DDP-1, a Quicksilver Audio model, and dimensional properties. Bass was pow­ loud volumes. Its sheer bass power and
Forssell tube line drivers, a First erful and well dam ped. The deep bass extension were not quite up to
Sound II passive unit, and the PA160s can m ost definitely the super-subwoofer level, but I have yet to
passive selector and attenuator kick butt! A m bient sounds, hear a subwoofer more musically realistic
sections of my preamp. Power su ch as a u d ie n c e an d p e r ­ and accurate. Upper bass and midrange
amplifiers on hand for compar­ fo rm e r noises, were rendered were excellent in virtually every respect. I
ison were a C row n M acro with great clarity. detected no significant tonal variations not
Reference, Quicksilver Audio I then tried a new set o f attributable to my listening room. Detail,
M135s, and a pair of Cary Au­ Shuguang Chinese EL34 tubes, transparency, and timbre in this frequency
dio Designs CAD-805 single- from a set of 12 that Kevin range were as good as anything I’ve heard,
ended triode tube amps. The Hayes had kindly given me at a but it did require the right amp to get the
speakers hooked up in my ref­ CES f o r m y Q u ic k s ilv e r proper warmth.
erence system were B & W 801 Ml35s. After letting these out­ The upper midrange was really excellent!
M atrix Series 3s, augm ented put tubes break in for a while, I It was musically natural; it had the air and
from 20 to 50 Hz by went through the o p ­ resolution that were once the province of
my tw o subw oofer era tin g m odes and electrostatics, com bined with a musical
systems, each using a quickly settled in on “life” just short of the best ribbon speakers.
JBL 1400Nd driver in
THE SOUND OF THE PA 160s
the triode mode, with The treble was sweet and natural, with out­
a 5-cubic-foot ported IS VERY REALISTIC, no feedback. These standing resolution; it was fun to listen to,
enclosure. WITH GREAT DYNAMICS amps sounded won­ not simply an exercise in technology.
The owner’s m anu­ derful this way! The Only a few of the larger (and more ex­
AND DIMENSIONALITY.
al for the PA160s is sound was more re­ pensive) ribbon/subwoofer combinations
very inform ative. It fined than with the have better dynamics than the CS7. This
covers the expected KT88s, more like my Thiel speaker could handle both low-level
sound, reliability, and various types and Quicksilver M135s. Bass had excellent and large-scale dynamics with excellent de­
manufacturers of tubes. The manual en­ heft and punch, and resolution, space, and finition and without shifts in sound charac­
courages you to experiment with the amp’s delicacy were of a high order. ter and timbre. The CS7 handled rapid m u­
various settings to get the best sound in In using these amplifiers, I did have some sical transients equally well. I got every
your system. One item deserves special problems with their extra-high gain. For in­ nuance of the differences in transient re­
m ention: VAC indicates that the “initial stance, when I used my preamp, which is ac­ sponse, and only a few ribbon drivers
break-in tim e of high-resolution equip­ tive only in its phono stage, I was frequently sound “faster.”
ment is infuriatingly long. The PA160 will operating at the lowest two positions of its The CS7’s transparency, harmonic natu­
continue to season for at least 200 hours.” passive volume control. One position was ralness, and detail really demonstrated how
When I first tried out these amps, I left too low in volume, and the next one up was good dynamic speakers have become: Both
them in the factory-set mode of Ultra-Lin­ too high. Using a preamp with line-stage accurate and musical. The soundstage was
ear operation, with the feedback switch at gain would be most problematic in this re­ stable over a wide listening area. Width and
position “D.” Initially, I thought the amps gard. Also, any line-stage o u tput noise depth were very faithful to the content of
produced a gutsy rendition of the lower might be audible. In fact, I could hear some my reference recordings; there were no sur­
midrange in louder orchestral passages, but hiss coming from the speakers when using prises, shifts, contractions, or exaggera­
there was a noticeable glare and irritation the Forssell balanced line stage with the tions. Soundstage detail was better than
that bothered me. After loaning the units to Sonic Frontiers D/A converter. I could also that of any other dynamic speaker I have
a friend and beating up on them in the lab, hear some 120-Hz amplifier hum from the heard to date, and the focus was much clos­
I returned them to my system for more lis­ speakers, mostly in the pentode mode. The er to that of a point source than in any oth­
tening. Soon afterward, I got the feeling that bias a d ju s tm e n t d id d r ift w ith tim e er large dynamic speaker I know of. The
these puppies were beginning to break in and a.c. line voltage, and I found myself imaging focus was possibly the best in both
and mellow out a bit. The glare and irrita­ doing a lot of tweaking of the bias pots— width and depth I have heard from any
tion were now greatly reduced, and I wasn’t m ore so with the KT88s than with the cone speaker, and was stable over a relative­
even close to 200 hours of operation. EL34s. But, hey, what do you expect? I am a ly wide listening area.
Experimenting with the various operat­ tweak, and the VAC amps are really bias- The CS7 is definitely a reference-quality
ing modes and amounts of feedback, I set­ stable enough. speaker whose only major drawbacks are
tled on the pentode mode, with feedback in In conclusion, I think the PA160s are price and amplifier sensitivity. It helps de­
the “D” or “E” position. Now 1 was getting damn good amplifiers, and 1 definitely rec­ fine the current state of the art in high-end
somewhere! These amps had a presentation ommend a serious listen. A sound. A real pleasure to review! A