Integrated Amp Reviews

I performed a full set of measurements on the Rega Brio, using my Audio Precision SYS2722 system (see the January 2008 "As We See It"). Via its line inputs, the Brio's maximum voltage gain into 8 ohms was typical for an integrated amplifier, at 39.55dB. The amplifier inverted absolute polarity for line input signals at both the speaker terminals and headphone jack, and its input impedance at 20Hz and 1kHz was, at 41k ohms, fairly close to the specified 47k ohms. Though the line input impedance dropped to 36k ohms at 20kHz, this will be inconsequential.

The headphone output impedance was a high 107 ohms from 20Hz to 20kHz, so will not be optimal with low-impedance headphones. However, the output impedance at the speaker terminals was very low, measuring just 0.06 ohm at low and middle frequencies, rising to a still-low 0.15 ohm at the top of the audioband. As a result, the change in response with our standard simulated loudspeaker, due to the Ohm's law interaction between this impedance and that of the load, was minimal (fig.1, gray trace), and there was very little difference between the response into 8 ohms (blue, red) and those into lower impedances. However, this graph reveals the Brio to have a relatively limited small-signal bandwidth, the –3dB point lying at 41kHz. This results in increased risetimes in the Rega's reproduction of a 10kHz squarewave (fig.2), although, commendably, no overshoot or ringing can be seen. The traces in fig.1 were taken with the volume control set to its maximum; at lower settings of the control, there are no changes either in the response or in the superb channel matching.

Channel separation was good, though it was greater from right to left than in the other direction, those crosstalks respectively lying at –78 and –65dB. The wideband, unweighted signal/noise ratio, taken with the line inputs shorted to ground but the volume control set to its maximum—the worst possible case—was okay, at 65.5dB ref. 1W into 8 ohms. Restricting the measurement bandwidth to 22Hz–22kHz increased the ratio to 69.35dB, while switching an A-weighting filter into circuit resulted in further improvement, to 72.2dB. (All figures are the average of the two channels.) Fig.3 indicates that the measured S/N ratios are due to both random noise and spuriae at the power-supply–related frequency of 60Hz and its harmonics.

Specified as offering maximum powers of 50Wpc into 8 ohms (17dBW) and 73Wpc into 4 ohms (15.9dBW), both channels driven, the Brio actually delivered slightly more power at clipping, which we define as when the level of THD+noise in the amplifier's output reaches 1%. Figs. 4 and 5 show how the percentage of THD+N changes with output power into, respectively, 8 and 4 ohms. You can see that while the Brio offers very low THD+N, it does clip quite sharply, reaching 1% THD+N at 52Wpc into 8 ohms (17.2dBW) and 80Wpc into 4 ohms (16dBW).

To be certain I was looking at actual THD rather than noise, I plotted (fig.6) the way the THD+N percentage changed with frequency at a fairly high level, 8.9V, which is equivalent to 10W into 8 ohms and 20W into 4 ohms. Even so, the percentage at low and middle frequencies was very low—much lower than what I found with the original Brio, which Wes Phillips reviewed in 1998. But as with that early Brio, the THD+N increased in the top octaves, presumably due to the circuit having limited open-circuit loop gain. This means that the amount of corrective negative feedback available decreases as the frequency rises, and to a greater extent in the right channel (red and magenta traces) than in the left (blue, cyan).

The distortion itself is heavily second-harmonic in nature (fig.7), which would be benign even if it were a lot higher than it is in the Rega's output (fig.8). However, this graph reveals that some higher-order harmonics are present, particularly in the left channel (blue trace), where they're almost as high in level as the second—if you can refer to anything at or below –94dB (0.002%) as "high." Despite the decreasing linearity in the top audio octave, the Brio did relatively well when I fed it an equal mix of 19 and 20kHz tones at a level just below visible waveform clipping into 4 ohms (fig.9). When I repeated this test at 1W into 8 ohms (fig.10), the difference product at 1kHz lay at just –84dB (0.006%).

The Brio's phono input offered 39.9dB of gain measured at the Tape Out jacks, and 79.6dB at the speaker terminals with the volume control set to its maximum. The input impedance was 44k ohms at low and middle frequencies, dropping to 36k ohms at 20kHz. Both the gain and impedance are appropriate for moving-magnet cartridges. The phono input inverted polarity at the speaker terminals but preserved polarity at the Tape Out jacks.

The RIAA equalization (fig.11) was superbly accurate, with a slight downward slope in the upper bass and the two channels very closely matched. Channel separation was also excellent, and the Brio's phono input was very quiet, the unweighted, wideband S/N ratio (ref. 1kHz at 5mV) measuring 68.8dB. This improved to 79.9dB with an A-weighting filter in circuit.

Phono overload margins were superb, at 27dB at 20Hz and 1kHz, and though the margin decreased to 24.5dB at 20kHz, this is still excellent. I had to raise the input level at 1kHz to 21mV to see what distortion harmonics were present; even at this level, only the second and third harmonics could be seen (fig.12), and these respectively lay at –106 and –114dB (0.0005% and 0.0002%). Intermodulation distortion was also extraordinarily low (fig.13).

Rega's Brio is a well-sorted little amplifier, as the Brits say, and I was especially impressed by its moving-magnet phono stage. But it runs hot: after 60 minutes of driving one-third its clipping power into 8 ohms, the temperature of its side panels was 135.6°F (57.6°C).—John Atkinson