Sidebar 3: Measurements
I measured the two-channel version of the Emotiva XPA Gen3 with my Audio Precision SYS2722 system (see the January 2008 "As We See It"). Before performing any tests, I ran it at 100Wpc into 8 ohms for an hour. (With a class-AB or -B amplifier, one-third full power is when the heat dissipation in the output devices is at its maximum.) At the end of that time, the louvers on the top panel above the two circuit boards were hot, at 134.4°F (56.9°C), though the rest of the exterior case was cooler. This may be an inexpensive amplifier, but it has adequate heatsinking for its power rating, though this might be due to the fact that the case can hold up to seven amplification modules rather than the two of our review sample.
The gain at the loudspeaker terminals for both balanced and unbalanced inputs was 29.4dB, and while the output preserved absolute polarity (ie, was non-inverting) for the unbalanced input, the balanced input inverted polarity, suggesting that the XLR jacks are wired with pin 2 cold, the opposite of the modern convention. The balanced input impedance was moderately high, at 27.5k ohms; the unbalanced input impedance was 14.5k ohms at low and middle frequencies, dropping to 9.5k ohms at the top of the audioband.
The output impedance, including the series resistance of the speaker cables I used, was very low at 20Hz and 1kHz, at 0.09 ohm, and rose only slightly at 20kHz, to 0.125 ohm. As a result, the response with our standard simulated loudspeaker varied by just ±0.1dB (fig.1, gray trace). Fig.1 shows that the response into 8 ohms (blue and red traces) starts to roll off at the top of the audioband, with the output reaching –0.3dB at 20kHz and –3dB at 70kHz. The response rolled off only slightly earlier into lower impedances, reaching –0.5dB at 20kHz into 2 ohms (green trace). The Emotiva's reproduction of a 10kHz squarewave was very good (fig.2), with no overshoot or ringing.
Specified as delivering 300Wpc into 8 ohms (24.8dBW) or 550Wpc into 4 ohms (24.4dBW), the XPA Gen3 clipped at 315Wpc into 8 ohms (25dBW) (fig.4), and at 520Wpc into 4 ohms (24.15dBW) (fig.5), both readings taken with both channels driven. Looking at how the percentage of THD+noise varied with frequency at a moderately high level (20V, equivalent to 50W into 8 ohms and 100W into 4 ohms), the THD+N remained low at low and middle frequencies, and was even slightly lower in the low treble (fig.6). However, as can be seen in this graph, the percentage starts to rise in the mid-treble before climbing precipitously above 15kHz, exceeding 1% at 20kHz. All I can think of to explain this behavior is that the circuit has limited open-loop bandwidth, meaning that there is insufficient corrective negative feedback available at the top of the audioband.
Emotiva's XPA Gen3 amplifier offers high powers with low levels of noise and distortion, at least at frequencies below 10kHz, at a very affordable price. But the amplifier's problems at the top of the audioband bother me. Predicting the subjective effect of this objective behavior is probably an exercise in futility, but I do wonder if it correlates with Herb's finding that the Emotiva sounded "hard and 100% masculine" through two of the speakers he tried with it.—John Atkinson
Fig.1 Emotiva XPA Gen3, frequency response at 2.83V into: simulated loudspeaker load (gray), 8 ohms (left channel blue, right red), 4 ohms (left cyan, right magenta), 2 ohms (green) (0.5dB/vertical div.).
Fig.2 Emotiva XPA Gen3, small-signal, 10kHz squarewave into 8 ohms.
Channel separation (not shown) was good rather than great, at 75dB or so from 20Hz to 5kHz, decreasing slightly to 67dB at the top of the audioband. The unweighted, wideband signal/noise ratio, taken with the unbalanced inputs shorted to ground, was good at 70.5dB left and 69.3dB right, both ratios ref. 1W into 8 ohms. Limiting the measurement bandwidth to the audioband improved these respective ratios to a superb 95 and 93.9dB, while an A-weighting filter offered further improvement, to 99 and 98dB. The spectrum of the XPA Gen3's low-frequency noise floor while it drove a 1kHz tone at 1W into 8 ohms is shown in fig.3. Other than some very low-level spuriae at 60Hz and its odd-order harmonics in the left channel, the XPA's noise floor is very clean.
Fig.3 Emotiva XPA Gen3, spectrum of 1kHz sinewave, DC–1kHz, at 1W into 8 ohms (linear frequency scale).
Fig.4 Emotiva XPA Gen3, distortion (%) vs 1kHz continuous output power into 8 ohms.
Fig.5 Emotiva XPA Gen3, distortion (%) vs 1kHz continuous output power into 4 ohms.
Fig.6 Emotiva XPA Gen3, THD+N (%) vs frequency at 20V into: 8 ohms (left channel blue, right red), 4 ohms (left cyan, right magenta).
At lower frequencies and moderate powers, the distortion signature is predominantly the third harmonic (fig.7). Though they can't be seen in the graph, there are small spikes surrounding each positive and negative waveform peak; ie, four per cycle, that could be seen on an analog 'scope. I have no idea where these spikes come from, unless they are related to the class-H voltage rail switching. At high powers, the third harmonic is joined by higher-order odd harmonics, as well as by the second and fourth harmonics (fig.8), though these are all very low in level.
Fig.7 Emotiva XPA Gen3, 1kHz waveform at 50W into 8 ohms, 0.0082% THD+N (top); distortion and noise waveform with fundamental notched out (bottom, not to scale).
Fig.8 Emotiva XPA Gen3, spectrum of 50Hz sinewave, DC–1kHz, at 200W into 4 ohms (linear frequency scale).
I usually test an amplifier for intermodulation distortion by driving it at a high power with an equal mix of 19 and 20kHz tones, which is very much the worst case. Given the dramatic rise in THD at the top of the audioband in fig.6, I was not surprised to find that the Emotiva XPA Gen3 did very poorly on this test. Fig.9 reveals a picket fence of intermodulation products, with the second-order difference product at 1kHz lying at –33dB (2%) ref. the waveform's peak level. Reducing the level to 1W (fig.10) eliminated the difference product, but many higher-order intermodulation products are still present, the highest in level lying at –66dB (0.06%). I repeated the 100W test with an equal mix of tones and octave lower (fig.11); though high-order intermodulation products can still be seen, these all now lie at or below –80dB (0.01%).
Fig.9 Emotiva XPA Gen3, HF intermodulation spectrum, DC–24kHz, 19+20kHz at 100W peak into 8 ohms (linear frequency scale).
Fig.10 Emotiva XPA Gen3, HF intermodulation spectrum, DC–24kHz, 19+20kHz at 1W peak into 8 ohms (linear frequency scale).
Fig.11 Emotiva XPA Gen3, HF intermodulation spectrum, DC–24kHz, 9+10kHz at 100W peak into 8 ohms (linear frequency scale).
