Solid State Power Amp Reviews

Since he wrote about them in his March 2009 "Music in the Round" column, Kalman Rubinson has been using as his reference power amplifiers Bel Canto's e.One Ref1000 Mk.2 class-D monoblocks ($5990/pair). Like the original Ref1000, which KR wrote about in November 2006, the Mk.2 (now called Ref1000M) uses Bang & Olufsen's 1000ASP ICEpower modules, but with improvements in mechanical damping, parts selection, and filtering of the residual switching noise present at the amplifier's output. With the Bel Cantos driving his B&W 802D speakers, Kal found the amplifiers sounded "even livelier and more transparent than their predecessors, with no vestige of HF grain to mar the awesome illusion....Compared with nonswitching amps such as Classé's Omicron or CA-3200, the Bel Canto e.Ones could seem somewhat 'literal' in terms of their tightly defined two-channel soundstage," he wrote, "but that disappeared in multichannel use. In fact, even in stereo, I could make a case for the Bel Cantos being more truthful than the Classés, if less luxurious."

As part of Stereophile's continuing program of publishing measurements Follow-Ups of sonically promising products that have been reviewed by the magazine's regular columnists, I subjected the Ref1000M to my usual battery of tests. I mainly used Stereophile's loan sample of the top-of-the-line Audio Precision SYS2722 system (see the January 2008 "As We See It" and www.ap.com); for some tests, I also used my vintage Audio Precision System One Dual Domain. To minimize the effect of the HF switching noise on the measurements, I used an Audio Precision passive low-pass filter ahead of each analyzer. As usual, I preconditioned the amplifier (serial no. RI-149) by running it at one-third power into 8 ohms for an hour before performing any measurements. Like any other amplifier that uses a switching output stage, which is very efficient at converting electrical power from the wall into loudspeaker-driving current, the Bel Canto did not get hot under these conditions.

The input impedance was to specification at low and middle frequencies, at 200k ohms balanced and 100k ohms unbalanced. These figures dropped to 160k and 58k ohms at 20kHz, respectively, but these are still usefully high input impedances. Both inputs preserved absolute polarity; ie, were non-inverting. The amplifier's output impedance was <0.1 ohm in the audioband, rising to 0.35 ohm at 20kHz, due to the effect of its series low-pass filter ahead of the output terminals. The effect of this filter can be seen in fig.1, which shows the amplifier's small-signal frequency response: the output into 8 ohms rolls off rapidly above 20kHz (blue trace) and with our standard simulated loudspeaker (black), and above 10kHz into 4 ohms (magenta) and 2 ohms (red). There is some peaking evident into 2 ohms (the lowest impedance the amplifier is rated into), but even with 8 ohms, squarewaves show a well-damped overshoot (figs.2 & 3). (These graphs were taken with the AP filter; without the filter, ultrasonic noise with a center frequency of 110kHz and an amplitude of 1.14V obscured the waveform tops and bottoms.)

Fig.1 Bel Canto e.One Ref1000M, frequency response at 2.83V into: simulated loudspeaker load (gray), 8 ohms (blue), 4 ohms (magenta), 2 ohms (red). (0.25dB/vertical div.)

Fig.2 Bel Canto e.One Ref1000M, small-signal 1kHz squarewave into 8 ohms.

Fig.3 Bel Canto e.One Ref1000M, small-signal 10kHz squarewave into 8 ohms.

Bel Canto rates the Ref1000's maximum power at >500W into 8 ohms and >1000W into 4 ohms. I measured maximum powers of 600W into 8 ohms and 1200W into 4 ohms (27.8dBW). With continuous drive, the amplifier turned itself off at 1kW into 2 ohms (24dBW), which was well below clipping (fig.4). Despite its minuscule size and weight, this is a very powerful amplifier. Distortion was very low in the midband and bass, particularly into high impedances, but rose in the treble (fig.5). Though the rise in THD+noise was not to an alarmingly high level into 8 ohms, this graph confirms that the Ref1000 will be more comfortable with higher-impedance speakers, where the residual distortion at relatively low powers was primarily lower-order harmonics (fig.6). Higher-order harmonics appeared into low impedances and at higher powers (fig.7), due to discontinuities at the signal's zero-crossing points. These harmonics all lie at relatively low levels, however.

Fig.4 Bel Canto e.One Ref1000M, THD+N (%)vs output power into (form bottom to top): 8 ohms, 4 ohms, 2 ohms.

Fig.5 Bel Canto e.One Ref1000M, THD+N (%)vs frequency at 13.7V into: 8 ohms (blue), 4 ohms (magenta), 2 ohms (red).

Fig.6 Bel Canto e.One Ref1000M, 1kHz waveform at 6W into 8 ohms (top), 0.008% THD+N; distortion and noise waveform with fundamental notched out (bottom, not to scale).

Fig.7 Bel Canto e.One Ref1000M, spectrum of 50Hz sinewave, DC–10kHz, at 215W into 4 ohms (linear frequency scale).

Testing for intermodulation products with an equal mix of 19 and 20kHz tones proved problematic, because there appeared to be a "history effect" at levels much above 20V. At first, the REF100 Mk.2 reproduced the signal without any problem, but after a short while, the waveform squared and remained clipped until the drive signal was reduced. Fig.8 shows the spectrum of the Bel Canto's output at the highest continuous level it would deliver into 4 ohms with this demanding signal. While the difference product at 1kHz is very low, the higher-order components at 18 and 21kHz lie just below –60dB (0.1%).

Fig.8 Bel Canto e.One Ref1000M, HF intermodulation spectrum, DC–24kHz, 19+20kHz at 50W peak into 4 ohms (linear frequency scale).

This latest version of Bel Canto's e.One Ref1000M monoblock offers very high power from its diminutive chassis. It does work best with higher-impedance loudspeakers, however, and at lower frequencies. Fortunately, music rarely includes enough high top-octave energy to reveal the amplifier's unhappiness in this region.—John Atkinson