Mastersound 300 B S.E. integrated amplifier Measurements
To my alarm, the Mastersound B S.E. arrived with something rattling around inside the box. It was the tubes! In my opinion, the packaging supplied with this heavy, delicate, and expensive amplifier is nowhere near robust enough to suffer the slings and arrows of outrageous UPS shipping. To add insult to injury, the tube cage can be removed only by unscrewing four internal Phillips-head bolts; screwdrivers long enough to reach these bolts are too thick to go through the slots in the cage, and screwdrivers that will go through the slots are too short. Fortunately, my local hardware store had the necessary 10" by 3/16" screwdriver. My pocket $5 lighter, I found that, with some forcing, the blade of my new driver would go through the tube-cage slots.
I reinstalled the tubes. All appeared to be working properly, though with single-ended tube amplifiers, the word properly acquires a somewhat different meaning from the usual, as you'll see.
The Mastersound 300 B S.E. has two output transformer taps: 4 ohms and 8 ohms. I performed a full set of tests using both sets of output taps, but discuss only a representative selection here. Tested through the normal line-level inputs, the maximum gain into 8 ohms was slightly different for each channel, at 40dB (8 ohm tap) and 38.8dB (4 ohm tap) for the left channel, but 40.5dB and 39.5dB, respectively, for the right channel. Tested via the Direct input, which bypasses the preamp circuit, the left-channel gain was 23.3dB (8 ohm tap) and 21.6dB (4 ohm tap), both figures measured into 8 ohms. The preamp therefore appeared to apply a maximum gain of 16.7dB, which is a little on the high side.
While the regular inputs preserved absolute polarity, the Direct input inverted polarity. The input impedance for the preamp section was a little lower than specified, at 67k ohms across the band for both channels, but the shortfall will be insignificant. The Direct connection had a much lower input impedance, however, ranging from 24k ohms at 20Hz to 13.3k ohms at 20kHz. While the output impedance was high compared with a solid-state design, it was not as high as I was expecting, and was well matched between the two channels, both factors a tribute to the construction of Mastersound's output transformers. From the 8 ohm tap, the output impedance was 3.5 ohms in the midrange, rising to 4 ohms at the top of the audioband; from the 4 ohm tap, the impedance ranged from 1.9 to 3.8 ohms.
The variation in frequency response that results from the Ohm's Law interaction between this source impedance and the impedance of the loudspeaker is large enough to be audible, even with the 4 ohm tap. The green trace in fig.1, for example, is the Mastersound's response with its 4 ohm tap driving our standard simulated loudspeaker. The response variations are ±1dB. Just as important, the lower frequencies into resistive loads become increasingly shelved down with decreasing load impedance, though the fact that the response peaks at the loudspeaker's bass resonant frequency might subjectively compensate for this. I note that Art Dudley did feel that the Mastersound had quite a wide bandwidth for a single-ended design.
At the other end of the spectrum, its response does extend almost 20kHz with higher impedance loads. However, the rolloff is disturbed by an ultrasonic response peak, which can be seen in fig.1, centered on 62kHz. This was almost absent from the set of responses taken from the 8 ohm tap (fig.2), but does result in a small amount of overshoot with a 1kHz squarewave (fig.3). The sloping tops and bottoms of this waveform correlate with the shelved-down lows seen in the frequency-response graphs. Peculiarly, it isn't as clearly seen in the 10kHz squarewave response taken from the 8 ohm tap (fig.4), but is fully developed in the 4-ohm-tap result (fig.5). Channel separation (not shown) was only moderate: 53dB in both directions across most of the audioband. The unweighted wideband signal/noise ratio, taken in the worst case—with the input shorted but the volume control at its maximum—was quite good, at 74dB ref. 1W/8 ohms/8 ohm tap. A-weighting improved this figure to 82dB.
Fig.1 Mastersound 300 B S.E., 4 ohm tap, frequency response at 2.83V into: simulated loudspeaker load (green), 8 (red), 4 (blue), 2 (magenta) ohms (1dB/vertical div.).
Fig.2 Mastersound 300 B S.E., 8 ohm tap, frequency response at 2.83V into: simulated loudspeaker load (green), 8 (red), 4 (blue), 2 (magenta) ohms (1dB/vertical div.).
Fig.3 Mastersound 300 B S.E., 8 ohm tap, small-signal 1kHz squarewave into 8 ohms.
Fig.4 Mastersound 300 B S.E., 8 ohm tap, small-signal 10kHz squarewave into 8 ohms.
Fig.5 Mastersound 300 B S.E., 4 ohm tap, small-signal 10kHz squarewave into 8 ohms.
The Mastersound 300 B S.E.'s maximum output power is specified as 12W, equivalent to 10.8dBW into 8 ohms. However, the graphs of output power against the THD+noise percentage (fig.6, 8 ohm tap; fig.7, 4 ohm tap) show that the amplifier reaches this power at 4% THD only when the load impedance is twice the nominal output transformer tap; ie, 8 ohm tap into 16 ohms, 4 ohm tap into 8 ohms. Into lower impedances, the THD rises to alarming (and audible) levels much above 100mW. However, in the best case—the 4 ohm tap driving 16 ohms—the distortion is quite low below 100mW.
Fig.6 Mastersound 300 B S.E., 8 ohm tap, distortion (%) vs 1kHz continuous output power into (from bottom to top at 1W): 16, 8, 4, 2 ohms.
Fig.7 Mastersound 300 B S.E., 4 ohm tap, distortion (%) vs 1kHz continuous output power into (from bottom to top at 1W): 16, 8, 4, 2 ohms.
The level of distortion doesn't change appreciably with frequency (fig.8, 8 ohm tap; fig.9, 4 ohm tap), but again rises considerably with decreasing load impedance, especially from the 8 ohm tap. The two channels behaved very similarly, however. As J. Gordon Holt first wrote almost 40 years ago, what matters subjectively with distortion is not so much the absolute level but the spectrum. The Mastersound generates what looks like almost pure second-harmonic distortion (fig.10); ie, each musical note is accompanied by another note exactly an octave higher in pitch. FFT analysis reveals some third-, fourth-, and fifth-harmonic content (fig.11), but this is much lower in level. This graph also shows that while some low-frequency random noise is present, AC supply–related spuriae are commendably low in level for this kind of amplifier. At low frequencies (fig.12), the second harmonic remains the highest in level, meaning that the output transformer is free from saturation, at least at moderate powers. The 120Hz power-supply harmonic can be seen at –76dB, which is low for this kind of design.
Fig.8 Mastersound 300 B S.E., 8 ohm tap, THD+N (%) vs frequency at 2.83V into: 16 (blue), 8 (red), 4 (magenta), 2 (green) ohms.
Fig.9 Mastersound 300 B S.E., 4 ohm tap, THD+N (%) vs frequency at 2.83V into: 16 (blue), 8 (red), 4 (magenta), 2 (green) ohms.
Fig.10 Mastersound 300 B S.E., 4 ohm tap, 1kHz waveform at 1W into 8 ohms (top), 0.874% THD+N; distortion and noise waveform with fundamental notched out (bottom, not to scale).
Fig.11 Mastersound 300 B S.E., 4 ohm tap, spectrum of 1kHz sinewave, DC–1kHz, at 1W into 8 ohms (linear frequency scale, left channel blue, right channel red).
Fig.12 Mastersound 300 B S.E., 4 ohm tap, spectrum of 50Hz sinewave, DC–1kHz, at 1W into 8 ohms (linear frequency scale, left channel blue, right channel red).
Finally, the Mastersound 300 B S.E. performed surprisingly well on the demanding high-frequency intermodulation test, even at a power level just below visible clipping on the oscilloscope screen (fig.13). The 1kHz difference component lay at –40dB (1%); I have seen worse performance from push-pull designs.
Fig.13 Mastersound 300 B S.E., 4 ohm tap, HF intermodulation spectrum, DC–24kHz, 19+20kHz at 6W peak into 8 ohms (linear frequency scale, left channel blue, right channel red).
Its output transformers are well engineered, but the Mastersound 300 B S.E.'s measured performance can't escape its single-ended provenance: both its response variations and its level of distortion are large enough to have audible consequences. You get the lowest level of distortion from its 4 ohm tap, but the smoothest ultrasonic rolloff from its 8 ohm tap. And neither tap works too well with impedances below its nominal value. The question is therefore begged: Did AD like the amplifier because of what it did right despite what it did wrong, or because of what it did wrong? I have to say that I have no idea.—John Atkinson