Integrated Amp Reviews

For logistical reasons, I measured a different sample of the HiFi Rose RA180—serial number ESL203AB000182—from the one auditioned by Julie Mullins. I performed the measurements with my Audio Precision SYS2722 system, examining the amplifier's performance in both normal mode and in bridged ("BTL") mode, following the instructions in the excellent manual. As Julie Mullins found, the RA180 doesn't recognize 5G Wi-Fi networks, but I was able to connect it to my Netgear NightHawk router's 2.4GHz network. I could control the RA180 with the RoseAMPConnect app on my iPhone, as well as with the remote control and the front-panel controls.

As the RA180's output stages operate in class-D, all the measurements other than frequency and RIAA responses were taken with Audio Precision's auxiliary AUX-0025 passive low-pass filter, which reduces the level of noise above 80kHz and eliminates noise above 200kHz. Without the filter, in normal mode, a very high 1.33V of ultrasonic noise was present at each pair of loudspeaker terminals, with a center frequency of 525kHz. The ultrasonic noise level rose to 1.6V in bridged mode. In both modes, when amplifying an input signal, the RA180 interfered with the reception of a portable FM radio that lives in the test lab.

I don't precondition class-D amplifiers before testing, but the top panel became warm after a few hours, its temperature measuring 104.6°F/40.4°C.

The RA180 preserved absolute polarity, ie, was noninverting, from both its balanced and single-ended line inputs in both normal and bridged modes. The maximum voltage gain at 1kHz into 8 ohms with the line inputs was 26dB, balanced, and 32dB, single-ended, in normal mode, and 6dB higher for both inputs in bridged mode. The balanced input impedance is specified as 44k ohms. I measured 43k ohms at all audio frequencies. The unbalanced input impedance measured 44.3k ohms at 20Hz and 1kHz, dropping inconsequentially to 29k ohms at 20kHz, compared with the specified 47k ohms.

The source impedance in normal mode, including the series resistance of 6' of speaker cable, was a low 0.1 ohms at low and midrange frequencies, rising to 0.4 ohms at the top of the audioband. The variation in the HiFi Rose amplifier's small-signal frequency response with our standard simulated loudspeaker (fig.1, gray trace) was a negligible ±0.1dB. Into 8 ohms (fig.1, blue and red traces), the RA180's response rose above the audioband, reaching +1.2dB at 65kHz, before sharply rolling off, reaching –3dB just above 100kHz. This peak is associated with a degree of overshoot in the RA180's reproduction of a 10kHz squarewave (fig.2), but there was no ringing. Into lower impedances the response starts to roll off in the top audio octave, with the most extreme result being into 2 ohms (fig.1, green trace), where the output at 20kHz is down by 2dB.

The source impedance in bridged mode was higher, at 0.15 ohm at 20Hz and 1kHz, 1 ohm at 20kHz, doubling the variations in response with the simulated loudspeaker (not shown) and increasing the 20kHz rolloff into 2 ohms to –4dB. However, overshoot was now minimal in the bridged mode's reproduction of a 10kHz squarewave into 8 ohms. (In bridged mode, each channel is effectively driving half the connected load impedance, so the RA180's bridged response into 8 ohms was the same as the normal mode's response into 4 ohms; fig.1, cyan and magenta traces.)

Fig.1 was taken with the volume control set to its maximum, and the excellent channel matching was preserved at lower settings. However, a peculiar 1dB boost in the midrange and bass appeared at lower settings of the volume control (fig.3, blue and red traces). This behavior was consistent with balanced and single-ended inputs in both output modes. The cyan and magenta traces in fig.3 show the action of the Subsonic setting. The main outputs roll off rapidly below 100Hz, reaching –9dB at 40Hz. This rolloff seems too extreme for typical listening, so perhaps it could be used when adding a subwoofer to the system. However, there is no means of altering either the high-pass filter's frequency or its slope.

The behavior of the bass and treble tone controls set to their maximum and minimum settings is shown in fig.4. Some overlap in their passbands can be seen in this graph, and the maximum cut or boost is higher than useful, at ±15dB in the treble and ±17.5dB in the bass. The Active Crossover controls adjust the turnover frequency level of the secondary amplifier outputs, which can be used to drive supertweeters. Fig.5 shows the behavior of this output with the frequency set to "0.85" and "2.17." The control's numerical values approximately coincide with the high-pass filter's –3dB frequency.

The RA180's channel separation (not shown) was 80dB in both directions below 2kHz, but the reading, taken with the Audio Precision low-pass filter, was affected by the presence of residual ultrasonic noise. This noise, which had a level of 1.5mV and a center frequency of 45kHz, is equivalent to an unweighted, wideband signal/noise ratio in normal mode (ref. 1W into 8 ohms, measured with the unbalanced inputs shorted to ground) of 66dB (average of the two channels). This ratio improved to 80dB when the measurement bandwidth was restricted to the audioband, and to 83.6dB when A-weighted. Spectral analysis of the low-frequency noisefloor while the HiFi Rose in normal mode drove a 1kHz tone at 1Wpc into 8 ohms with the volume control set to its maximum (fig.6) revealed a higher level of random noise in the left channel (blue trace) than the right (red). Repeating the measurements with the volume control set to –20dB reduced the left channel's noisefloor by 6dB (green trace) but the right channel's floor by 2dB (gray). In bridged mode (not shown), the levels of the low-frequency random noise in both channels increased by approximately 8dB.

HiFi Rose specifies the RA180's maximum power in normal mode as 200Wpc into both 8 and 4 ohms, respectively equivalent to 23dBW and 20dBW. With our usual definition of clipping as being when the THD+noise reaches 1%, the HiFi Rose amplifier with two channels driven in normal mode exceeded its specified powers. It clipped at 290Wpc into 8 ohms (24.6dBW, fig.7) and 400Wpc into 4 ohms (23dBW, fig.8). In bridged mode with again two channels operating, the RA180 went into protection below the actual clipping point, turning off at 460Wpc into 8 ohms (26.6dBW, fig.9) and 322W into 4 ohms (22.1dBW).

The downward slope below 50W in figs.7–9 indicates that the measured THD+N percentage is dominated by noise. (A constant amount of noise becomes a smaller percentage of the measured level as the power increases.) Fig.10 shows how the percentage of THD+N varied with frequency in normal mode at 20V, which is equivalent to 50W into 8 ohms (blue and red traces) and 100W into 4 ohms (green, gray). The THD+N was very low in the bass and midrange, particularly in the right channel, but rose in the top audio octaves. In bridged mode, the performance into 8 ohms was similar to the normal mode's behavior into 4 ohms.

Fig.11 shows the distortion waveform in normal mode with a 1kHz sinewave at 50W into 8 ohms. The presence of bursts of switching noise at the waveform peaks makes interpretation difficult, though it looks as if the second harmonic is dominant. This was confirmed by spectral analysis (fig.12), though the third harmonic was almost as high in level. Repeating the analysis at the same power in bridged mode (fig.13) revealed that all the distortion harmonics were lower in level than they had been in normal mode, the third harmonic in particular.

Intermodulation distortion was low in level. With the RA180 driving an equal mix of 19 and 20kHz tones at 100Wpc peak into 4 ohms in normal mode (fig.14), which will be equivalent to 50Wpc into 8 ohms in bridged mode, the 1kHz difference product lay close to –110dB in both channels (0.0003%), and while the higher-order products at 18 and 21kHz were higher in level, they still lay below –80dB (0.01%).

Turning to the phono input's performance, this preserved absolute polarity in both MM and MC modes. The maximum gain measured at the loudspeaker outputs in normal mode was 78.4dB, MM, and 98.3dB, MC. Bridged mode increased both gains by 6dB. The input impedance, specified as 47k ohms, was 44k ohms at 20Hz and 1kHz, in both MM and MC modes, but dropped to 6.7k ohms at 20kHz, again in both modes.

I examined the phono input's adjustable equalization set to "RIAA" and the output measured at the speaker outputs without the Audio Precision low-pass filter. The error in the RA180's RIAA equalization (fig.15) was very low, with superb channel matching. This graph was taken with the volume control set to its maximum; at lower settings the midrange and bass were boosted by 1dB, as they had been for the line inputs. Assessed with the inputs shorted to ground but the volume control set to its maximum, the phono input's unweighted, wideband S/N ratios in MM mode were 72.3dB ref. 1kHz at 5mV in both channels. Restricting the measurement bandwidth to 22Hz–22kHz increased this ratio to 74.1dB, while switching in an A-weighting filter further increased the ratio to 80.6dB. The ratios in MC mode, ref. 1kHz at 0.5mV, were all 20dB lower, correlating with the 20dB increased gain in this mode, but the HiFi Rose's phono stage is still relatively quiet.

The phono input's overload margins were respectable in both modes, at 13dB at 20Hz and 1kHz, ref. 1kHz at 5mV (MM) and at 0.5V (MC). However, the overload margin at 20kHz was disappointing, at –5dB, again ref. 1kHz at 5mV/0.5mV. In MM mode, an input signal at 20kHz reached 1% THD+N at 28mV, which will be marginal with normal-output moving magnet cartridges. (I performed the overload testing with the volume control set to –20dB to avoid clipping the RA180's output stage.) Harmonic distortion was very low via the phono input, the second and third harmonics with a 1kHz tone at 10mV respectively lying at –94dB (0.002%) and –100dB (0.001%) and all other harmonics at –110dB (0.0003%). And even at a peak level of 25mV, just below the MM mode's overload point, an equal mix of 19 and 20kHz tones produced a 1kHz difference product at –74dB (0.02%), with higher-order intermodulation products at negligible levels.

The HiFi Rose RA180 offers high powers in both its normal and bridged modes, especially into 4 ohms, with low, mainly benign second-harmonic distortion. I remain puzzled, however, by that 1dB boost in the midrange and bass at low volume control settings, by the higher-than-usual level of ultrasonic noise at the loudspeaker outputs, and by the phono input's limited overload margin at the top of the audioband.—John Atkinson