Sidebar 3: Measurements
I measured one of the Audionet Max monoblocks (serial no. 9.16.20) with my Audio Precision SYS2722 system (see the January 2008 "As We See It"). Before doing any testing of a power amplifier, I precondition it by running it at one-third its specified power into 8 ohms for 60 minutes with a 1kHz tone: This power level results in the highest thermal stress on the output devices of an amplifier having a class-B or -AB output stage. The Max turned itself off after 42 minutes, its front panel displaying the message "SEVERE ERROR OVERHEATING." The Audionet's top panel at that point was fairly hot, at 117.3°F (47.4°C).
What I found interesting about the amplifier's behavior was that its THD+noise, which was extraordinarily low with the amplifier cold, at 0.0008%, stayed close to that level for almost the entire 42 minutes, but then rose to around 0.03% (fig.1). If I then reduced the level of the input signal so that the output power was lower than one-third power, the THD+N remained around 0.03%. If I didn't reduce the input level, the amplifier turned itself off several seconds after the increase in THD. I let the amplifier cool down for a couple of hours, then resumed testing. The Max's protection circuits kicked in several times during the measuring when I was operating it at high powers, particularly at high frequencies, giving me the message "SEVERE ERROR HIGH FREQUENCY." Each time, I let the amplifier cool down before continuing the testing.
The output impedance was very low, at 0.045 ohm (including the speaker cable) at 20Hz and 1kHz, rising slightly to 0.07 ohm at 20kHz. The response with our standard simulated loudspeaker varied by just ±0.05dB (fig.2, gray trace). The Max offered a wide small-signal bandwidth, with the response into 8 ohms (blue trace) down by just over 1dB at 200kHz. The Audionet's reproduction of a 10kHz squarewave into 8 ohms (fig.3) thus featured very short risetimes, and no overshoot or ringing was visible.
Fig.1 Audionet Max, THD+N (%0 vs time (seconds).
The voltage gain into 8 ohms for the balanced input was a low 23.4dB. Unusually, the gain via the unbalanced inputs was 6dB higher. Both inputs preserved absolute polarity, the XLR jack being wired with pin 2 hot. The input impedances were close to the specified values, at 35k ohms unbalanced and 3k ohms balanced. The latter is very low; if this amplifier is used with tubed preamps, which typically have a high output impedance at low frequencies, the bass will sound rolled off.
Fig.2 Audionet Max, frequency response at 2.83V into: simulated loudspeaker load (gray), 8 ohms (blue), 4 ohms (magenta), 2 ohms (red) (0.5dB/vertical div.).
Fig.3 Audionet Max, small-signal, 10kHz squarewave into 8 ohms.
Not only does the Audionet Max have very low distortion—it is also very quiet. Its unweighted, wideband signal/noise ratio, ref. 1W into 8 ohms and taken with the input shorted to ground, was a superb 85.4dB, which increased to 98.5dB when the measurement was A-weighted. Fig.4 reveals that the only spuriae present in the output were at the 60Hz supply frequency and its odd-order harmonics, but these are all at very low levels.
Fig.4 Audionet Max, spectrum of 1kHz sinewave, DC–1kHz, at 1W into 8 ohms (linear frequency scale).
The Audionet Max exceeded its specified output power of 400W into 8 ohms (26dBW), clipping at 450W into that load (26.5dBW, fig.5). Note that in this graph, the actual distortion lies beneath the noise at powers below 30W, indicated by the upward slope of the trace as the power decreases. Fig.6 reveals that the amplifier also exceeded its specified power of 700W into 4 ohms (25.4dBW), clipping at 750W (25.75dBW). The Max clipped at the specified 1100W into 2 ohms (24.4dBW), though the AC wall voltage had dropped from 123V to 118.5V at the clip point.
Fig.5 Audionet Max, distortion (%) vs 1kHz continuous output power into 8 ohms.
Fig.6 Audionet Max, distortion (%) vs 1kHz continuous output power into 4 ohms.
Fig.7 plots the THD+noise percentage against frequency at a level, 20V, that is equivalent to 50W into 8 ohms and 100W into 4 ohms, and where I could be sure I was looking at actual distortion rather than noise. Even so, the THD+N percentage remains very low, with only slight rises in the top octave and into the lower impedance. I haven't shown the THD+N plotted against frequency into 2 ohms at this level (equivalent to 200W), as despite repeated attempts, the amplifier's protection circuit cut in and displayed the "SEVERE ERROR HIGH FREQUENCY" message. At low frequencies and high powers into low impedances, the highest-level distortion harmonic was the third (fig.8), though with a 1kHz tone at the same output voltage into 8 ohms, the second harmonic was the highest in level (fig.9). Intermodulation distortion was also extremely low (fig.10), even at a level close to visual waveform clipping on an oscilloscope, where the 1kHz difference product rose from –120dB (0.0001%) to –96dB (0.0015%, fig.11).
Fig.7 Audionet Max, THD+N (%) vs frequency at 20V into: 8 ohms (blue), 4 ohms (magenta).
Fig.8 Audionet Max, spectrum of 50Hz sinewave, DC–1kHz, at 200W into 4 ohms (linear frequency scale).
Fig.9 Audionet Max, spectrum of 1kHz sinewave, DC–10kHz, at 100W into 8 ohms (linear frequency scale).
Fig.10 Audionet Max, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 100W peak into 8 ohms (linear frequency scale).
Fig.11 Audionet Max, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 300W peak into 8 ohms (linear frequency scale).
As I am coming to expect from expensive German amplifiers, Audionet's Max performed very well on the test bench, offering high powers with very low levels of distortion and noise. It doesn't quite have enough heatsink capacity for its power-output capability, but I suspect that that shortcoming will not be a problem in normal use, especially given its very effective protection circuitry. (Despite the abuse I gave it during testing, the Max didn't break.) My only reservation concerned its very low balanced input impedance, but this won't be an issue with modern solid-state preamps and digital source components.—John Atkinson
