Solid State Power Amp Reviews

Sidebar 3: Measurements

Although the ELAC Alchemy DPA-2 is a stereo amplifier, Jason Victor Serinus auditioned it as a pair of monoblocks. (In mono mode, the signal is fed to either input and the output taken from the left channel's red binding post and the right channel's black binding post.) I therefore performed a complete set of measurements with a DPA-2 in mono mode then repeated some of the tests in stereo mode. As the DPA-2 has an output stage operating in class-D, I inserted an Audio Precision auxiliary AUX-0025 passive low-pass filter between the test load and my Audio Precision SYS2722 system. This filter eliminates RF noise that could drive the SYS2722's input circuitry into slew-rate limiting. I used it for all the tests other than frequency response.

Though the Alchemy DPA-2 has both balanced and single-ended inputs, I performed most of the measurements using the balanced inputs, which the manual says are preferred. With the amplifier in stereo mode, I measured a low voltage gain of 18.7dB into 8 ohms with both types of inputs and the gain set to Normal. High gain mode increased the gain by a useful 5.9dB. In mono mode, the DPA-2's gain into 8 ohms was 24.6dB and 30.4dB on the Normal and High gain settings. The amplifier preserved absolute polarity (ie, was noninverting) with both inputs in both stereo and mono modes. The XLR jack is wired with pin 2 hot. The input impedance is specified as 12k ohms single-ended and 19k ohms, balanced. I measured 11.9k ohms at 20Hz and 1kHz for the single-ended inputs, dropping inconsequentially to 10.7k ohms at 20kHz. The balanced input impedance was 18k ohms from 20Hz to 20kHz, ie, 9k ohms per phase.

The output impedance in stereo mode is specified as 0.03 ohms at 1kHz. I measured a very low 0.078 ohms at 20Hz and 1kHz rising to 0.095 ohms at 20kHz. (These figures include the series impedance of a 6', spaced-pair speaker cable.) The output impedance was a little higher in mono mode, as there are now two output stages in series. I measured 0.105 ohms at 20Hz and 1kHz, and 0.14 ohms at 20kHz. Even so, the modulation of the ELAC's frequency response driving our standard simulated loudspeaker was minuscule, at ±0.05dB (fig.1, gray trace). The small-signal bandwidth was restricted by the low-pass filter between the amplifier's class-D stage and its output terminals. Into 8 ohms (fig.1, blue trace), the ultrasonic rolloff reached –3dB at 50kHz, a little more into lower impedances. The rolloff lengthened the risetimes of a 10kHz squarewave (fig.2). There is a critically damped overshoot on the tops and bottoms of the waveform, but there is no ringing.

Without the auxiliary low-pass filter, 191mV of ultrasonic noise was present at the DPA-2's output terminals. With the AP filter and the DPA-2 set to stereo mode and normal gain, the unweighted, wideband signal/noise ratio, taken with the single-ended inputs shorted to ground, was an excellent 82dB ref. 1W into 8 ohms. It improved to 95.4dB when I restricted the measurement to the audioband and to 98.1dB with an A-weighting filter in circuit. In high-gain stereo mode, these ratios were all around 3dB lower; they were down another 3dB or so in mono mode. Nevertheless, this is a quiet amplifier. Spectral analysis of the low-frequency noise floor in mono mode (fig.3) revealed some AC-supply–related spuriae at 60Hz and its odd-order harmonics, though these all lie below –100dB ref. 1W into 8 ohms. The red trace in fig.3 is the spectrum with the amplifier set to normal gain. High gain (blue trace) increased the random noise components by around 3dB.

The Alchemy DPA-2's maximum power is specified at 325Wpc into 4 ohms in stereo mode (22.1dBW ref. 1W into 8 ohms). With clipping defined as when the THD+noise in the output reaches 1%, the ELAC's clipping power in this mode with one channel driven was 339W into 4 ohms (22.3dBW, fig.4). In stereo mode, the amplifier clipped at 202W into 8 ohms (23.05dBW, not shown). As expected, considerably more power was available in mono mode. At 1% THD+N, the DPA-1 delivered 590W into 8 ohms (27.7dBW, fig.5) and 550W into 4 ohms (24.4dBW, fig.6). Though the 8 ohm power is a little lower—0.3dB—than the specified power of 625W (28.0dBW), the AC power line had dropped to 118.5V with the amplifier clipping into 8 ohms.

Figs.4–6 show that the Alchemy DPA-2 had very low distortion at powers below a few tens of watts. I examined how the ELAC's THD+N varied with frequency at 20V (equivalent to 50W into 8 ohms, 100W into 4 ohms, and 200W into 2 ohms).3 The distortion into 8 ohms was close to 0.01% (fig.7, blue trace), though it rose at high frequencies and into 4 ohms (magenta trace) and 2 ohms (red trace).

The ELAC amplifier's distortion signature in mono mode into 8 ohms was primarily third harmonic in nature (fig.8). Though some second, fifth, and seventh harmonics are present (fig.9), these are all at or below –90dB (0.003%). At the same output voltage into 4 ohms, which is equivalent to 200W, the second and higher odd-order harmonics rose in level (fig.10), though the third harmonic dropped a little. With an equal mix of 19 and 20kHz tones and the signal peaking at 200W into 4 ohms (fig.11), the decrease in the circuit's linearity at very high frequencies resulted in some higher-order products, the highest in level of which, at 21kHz and 18kHz, each lay at –70dB (0.03%). The difference product at 1kHz, even at this very high power, lay at –96dB (0.0015%).

The ELAC Alchemy DPA-2 amplifier is a powerhouse, especially in mono mode.—John Atkinson

---

Footnote 1: There was one oddity I noted when I measured the DPA-2. I have a portable FM radio in the test lab that I keep tuned to WNYC, New York City's NPR station. When I powered up the DPA-2 and fed it an audio signal, I could hear a tone at the frequency of the test signal, overlaid with fizzy noise, reproduced by the radio's speaker. This was very noticeable when I swept the test signal down from 20kHz to 20Hz to generate the traces in fig.8. I have only encountered this RF interference with one other modern class-D amplifier, the Spec RPA-W7EX that Ken Micallef reviewed in March 2016.