Integrated Amp Reviews

To examine the performance of the A25's DAC, I used the coaxial and optical S/ PDIF inputs, both of which accepted data sampled at rates up to 192kHz, and USB-C data sourced from Roon on my iPad mini. (My MacBook Pro, which I usually use to test devices with USB ports, doesn't have a USB-C port, and the USB A–C adapter cables I had to hand had the wrong gender USB-A connector.)

The ARCAM's digital inputs preserved absolute polarity from all the outputs. With the volume control set to the maximum, the A25's output level into 8 ohms at the speaker outputs with a 1kHz tone at –25dBFS was 19.92V, which is just 3dB below the clipping voltage. With a 1kHz tone at –20dBFS, the output level was 4.725V at the headphone output and 972mV from the Preamplifier output. As is often the case with integrated amplifiers, the A25's DAC offers much higher gain than is necessary.

The Menu/Settings button on the ARCAM A25's front panel and on the remote control doesn't offer a choice of reconstruction filters. Fig.10 shows the A25's impulse response with 44.1kHz S/PDIF data. The filter is a short minimum-phase type. However, the impulse response with USB data (fig.11) was different. This response is typical of the ESS DAC chip's hybrid filter (footnote 1), with most of the ringing following the single high sample.

The magenta and red traces in fig.12 show the USB filter's ultrasonic rolloff with data sampled at 44.1kHz. They reach full stop-band attenuation at half the sample rate (this indicated by the vertical green line), with the aliased image at 25kHz of a full-scale tone at 19.1kHz (cyan, blue) suppressed by 100dB. The harmonics associated with the 19.1kHz tone all lie at or below –85dB. With the short S/PDIF filter (fig.13), full stop-band attenuation isn't reached until 28kHz, with the aliased image at 25kHz suppressed by just 30dB. The frequency response with 44.1kHz data and the S/PDIF filter starts to roll off above 15kHz, reaching –4dB at 20kHz (fig.14, green and gray traces).

An increase in bit depth from 16 to 24, with dithered data representing a 1kHz tone at –90dBFS, dropped the A25's noisefloor by 13dB (fig.15). This implies a resolution of around 18 bits. When I played undithered data representing a tone at exactly –90.31dBFS, the waveform was symmetrical, with negligible DC offset, and the three DC voltage levels described by the data were clearly defined (fig.16). With undithered 24-bit data (fig.17), the A25 output a somewhat noisy sinewave.

As it had with the analog inputs, the ARCAM's digital inputs produced very low levels of harmonic distortion. Intermodulation distortion with an equal mix of 19 and 20kHz tones, each at –6dBFS, was also very low, though as expected from fig.13, the aliased images of the primary tones were only suppressed by 20–30dB with TosLink data (fig.18).

Fig.19 shows the spectrum of the A25's output when it was fed high-level 16-bit J-Test data via TosLink. Most of the odd-order harmonics of the undithered low-frequency, LSB-level squarewave lie close to the correct level, though the central spike that represents the high-level tone at one-quarter the sample rate (Fs/4) is surrounded by low-level sidebands of unknown origin at ±68Hz and ±441Hz. Repeating this analysis with 24-bit J-Test data unmasked another pair of sidebands at ±1.368Hz (fig.20). The ARCAM behaved better on this test with USB data (figs.21 & 22), though the sidebands at ±68Hz are still present.

To examine the performance of the A25's MM-compatible phono input, I connected a wire from one of the Audio Precsion's ground terminals to the grounding lug on the ARCAM's rear panel. This input preserved absolute polarity in all three modes at all three output types. The input impedance is specified as 47k ohms; I measured 45k ohms at 20Hz and 1kHz, and 38.1k ohms at 20kHz.

The maximum gain at 1kHz was a very high 83.9dB from the loudspeaker output. It was 65.67dB from the headphone output and 52.5dB from the Preamplifier outputs. To avoid overdriving the amplifier's output stage, I performed all the subsequent phono stage measurements at the headphone output, which mutes the speaker output.

The phono input's RIAA correction (fig.23) was extremely accurate and well matched between the channels. The lowest frequencies were rolled off, reaching –6dB at 11Hz. Channel separation was 80dB or greater across the audioband. The wideband, unweighted S/N ratio, ref. 1kHz at 5mV, assessed with the inputs shorted to ground and with the volume control set to –20dB, was a good 65.3dB in both channels. Restricting the measurement bandwidth to 22Hz–22kHz increased the ratio by 12dB, while inserting an A-weighting filter resulted in a ratio of 81.0dB.

To be sure I wasn't prematurely clipping the headphone output, I examined the phono input's overload margins with the volume corol set to –40dB. The margins were impressively high from 20Hz to 20kHz, at 24dB ref. 1kHz at 5mV. The second harmonic was the highest in level in the phono input's distortion signature (fig.24), but at –100dB (0.001%) was inconsequential. The levels of the intermodulation products with an equal mix of 19 and 20kHz tones were equally low.

Overall, the ARCAM A25's measured performance was excellent. It offers extraordinarily low distortion from its line, phono, and digital inputs, and low noise provided the volume control is not used above a setting of –12dB or so.—John Atkinson

---

Footnote 1: See, for example, fig.4 here.