Power output with 1kHz test signal
- 8-ohm load at 1% THD: 107W
- 8-ohm load at 10% THD: 129W
- 4-ohm load at 1% THD: 102W
- 4-ohm load at 10% THD: 135W
General
The Audio Research Reference 110 is a medium-power stereo
push-pull tube-based power amplifier utilizing two pairs of 6550 output tubes per channel.
This unit is most unusual in vacuum-tube power-amplifier design in that it has no phase
inverter and depends on receiving a balanced input signal for the output stage push-pull
action. Pity the person who might use single-ended RCA female-to-XLR male adapters with
pin 3 (minus phase) grounded, as is usual. I did that in my initial testing and found out
that the amp responded as a single-ended amplifier putting out about 20 watts per channel
at 10% distortion! Needless to say, when driven with a balanced input signal, the
amplifier behaved properly. Checking with Audio Research, I found out that all of the
Reference-series tube power amplifiers were designed this way. I guess there is something
to be said for circuit simplification by not having a phase inverter if it is known that
the amp will always be driven by a balanced preamp, which the Audio Research Reference 3
is.
Chart 1 shows the frequency response of the amp with
varying loads. The output impedance, as judged by the closeness of spacing between the
curves of open-circuit, 8-ohm, and 4-ohm loading, is reasonably and acceptably low for a
tube power amplifier. With the NHT dummy speaker load, the variation is better than
+/-1dB. One thing to note, though: There is a noticeable gentle roll-off in the audio
range starting at about 500Hz with the 4-ohm loading. This is also present with the 8-ohm
loading to a lesser degree and not quite noticeable in the chart with the usual scale
factor that I use. This, for a perfectly flat speaker load, would have the potentially
audible effect of a softening of the highs or, to put it another way, give a little more
weight to the lower midrange – generally a good thing, as many amplifiers are a
little too thin in this region.
Chart 2 illustrates how total harmonic distortion plus
noise vs. power varies for 1kHz and SMPTE IM test signals and amplifier output load. Of
interest, this amplifier’s general distortion characteristic is typical of many tube
amplifiers and is one of continuously increasing distortion with increasing power. As is
the case with most power amplifiers, this one puts out more power, but not much more, into
a 4-ohm load.
Total harmonic distortion plus noise as a function of
frequency at four different power levels is plotted in Chart 3. The amount of rise in
distortion at high frequencies is admirably low, although there is the typical rise in
distortion at low frequencies due to the characteristics of the output transformers.
Damping factor vs. frequency is shown in Chart 4. This
curve has an unusual shape and, in fact, looks somewhat like the shape of the 4-ohm
frequency response. It makes me wonder if this was a deliberate design aspect of the
Reference 110.
A spectrum of the harmonic distortion and noise residue of
a 10W 1kHz test signal is plotted in Chart 5. The principal signal harmonics are the
second and third with all higher harmonics some 40dB or so lower. AC-line harmonics are
low, although there are some line-harmonic-related sidebands around the nulled-out 1kHz
test-signal fundamental -- something I have seen in quite a few of the power amplifiers
that I have tested, but not all of them.
![[SoundStage!]](../../titles/sslogo3.gif)
Home Audio 
