Specifications

TEST RESULTS
26 | Australian Hi-Fi
er er
Manley Laboratories Stingray Valve Integrated Amplifier
Test Results
Newport Test Laboratories usually meas-
ures the power output of an amplifier
at either clipping, if no other waveform
distortion is visible, or at 1.0% THD+N if
waveform distortion occurs prior to clip-
ping. This works well for solid-state ampli-
fiers which have very little distortion, but
not with valve amplifiers, which usually
have fairly high average distortion levels
and a variety of different waveform non-
linearities as they near their maximum
power output. Reconciling measured out-
put with claimed output is further compli-
cated by the fact that most valve amplifier
manufacturers state their maximum power
output when THD+N is equal to 3.0% or
even more. In the case of the Stingray, the
reporting of power output is further com-
plicated by Manley Laboratory’s decision
to optimise the output into a ‘non-stand-
ard’ 5Ω load. After some discussion about
whether to measure the Stingray’s power
output into a 5Ω load, it was decided that
in the interests of uniformity, in being able
to compare with other amplifiers, the out-
put would be measured into the standard
load resistances of 8Ω, 4Ω and 2Ω.
As you can see from the tabulated
results, power output was remarkably
uniform across the different loads. If
Manley had instead opted for a standard
load, one might have expected maximum
output into that load, and much less into
the other two. Generally, the Stingray
delivered 30-watts of power into all load
impedances at 1kHz, reducing to 20-
watts at 20kHz, when using 1.0% THD
as the ‘bar’. This 30-watt figure is 1.2dB
lower than specification, though Manley
uses a higher THD figure (1.5%) which
effectively ‘lowers’ the bar. Indeed the
Stingray can deliver 50-watts of power
into 8Ω, but only at a frequency of 1kHz,
and only with a significantly distorted
waveform (around 10% THD). The
amplifier could not sustain this power
output at either 20Hz or 20kHz into 8Ω,
where output fell to just 20-watts.
The Stingray’s frequency response
into 8Ω and 4Ω non-inductive loads
was moderately extended, stretching
from 10Hz to 53kHz –1dB (normalised,
this would be 10Hz to 53kHz ±0.5dB)
and from 4.7Hz to 72kHz –3dB. The
accompanying graph shows the frequency
response between 5Hz and 30kHz into
an 8Ω resistive load (black trace), a 4Ω
resistive load (blue trace) and into a
load similar to that of a two-way bass
reflex loudspeaker (red trace). The shift
in level between the 8Ω and 4Ω traces
is the direct result of the Stingray’s low
output impedance, measured as 3.8Ω at
20Hz, 3.6Ω at 1kHz and 3.8Ω at 20kHz.
The effect of this varying impedance on
the response into a simulated speaker is
evident on the graph, with the response
swinging almost 2dB, from a –1.5dB low
at 4.5kHz to a +0.5dB peak at 20kHz.
Channel separation measured 68dB
at 20Hz and 69dB at 1kHz, both of
which are excellent results, reducing to
52dB at 20kHz (and 49dB at 30kHz). The
Stingray does not invert polarity and
input impedance was high, averaging
around 47kΩ for midpoint settings of the
volume control and dropping slightly
at extreme settings. Channel balance
was 0.4029dB, which is pretty good for
what is essentially a dual mono design
approach. Channel phase was very good
at 1kHz, but more than 8 degrees in error
at 20Hz and nearly 5 degrees at 20kHz.
To put this into perspective, the error at
20kHz is on a par with the phase error in
most bit-stream CD players.
Signal-to-noise ratios were adequate,
with the weighted noise below 1-watt
coming in at 79dB. There wasn’t the
expected improvement when using rated
output as reference, with the A-weighted
signal-to-noise ratio increasing by just
1dB, to 80dB. THD+N at one watt was
ManleylabHFSep06.indd 26 22/01/2009 1:56:11 PM