Rotel RB-991 power amplifier Measurements

Sidebar 3: Measurements

The measurements of the Rotel RB-991 were made primarily in unbalanced mode, though selected readings were taken in balanced operation. Unless otherwise noted, the results discussed here refer to the unbalanced connection.

Following its preconditioning test, the Rotel RB-991 was very warm, but not unusually so for a high-powered amplifier. Its voltage gain measured 31.8dB unbalanced, 24.9dB balanced. The input impedance measured 36k ohms (45.2k ohms balanced). DC offset was a highish 22.1mV in the left channel, 13.1mV in the right. The RB-991 is noninverting—a positive-going input emerges positive at the output. (Pin 2 of the balanced input is wired as positive.) The unweighted S/N ratio at 1W into 8 ohms was 89dB over a 22Hz–22kHz bandwidth, 83.3dB from 10Hz to 500kHz, and 90.0dB A-weighted. The corresponding balanced S/N figures measured somewhat worse: 83dB, 74.3dB, and 85.5dB, for the same respective conditions.

The RB-991's output impedance varied from 0.04 to 0.05 ohms, the higher value occurring at 20kHz. This should not affect its frequency response with varying loudspeaker loads in any audible way (fig.1). The small dip at 5kHz is the only change visible when the amplifier is driven into our simulated loudspeaker load. The 4 ohm result was the same as the 8 ohm measurement and is not shown; the same is true of the balanced response. The 10kHz squarewave from the RB-991 is shown in fig.2. Apart from the slight rounding of the leading edge found in almost all amplifiers, this is an excellent result, as is the nearly perfect 1kHz squarewave (not shown).

Fig.1 Rotel RB-991, frequency response at (from top to bottom): 1W into 8 ohms, and 2.828V into simulated loudspeaker load (0.5dB/vertical div., right channel dashed).

Fig.2 Rotel RB-991, small-signal 10kHz squarewave into 8 ohms.

At 102dB at 1kHz, reducing through capacitive coupling to 78dB at 20kHz, the unbalanced channel separation (not shown) is more than adequate to render any possible audible effect moot. As with the S/N, the balanced measurements were a little worse than the unbalanced, the crosstalk increasing by 10dB on average.

Fig.3 shows the small-signal THD+noise percentage plotted against frequency. The 2 ohm result shows a slight increase at high frequencies, but this is not unusual, and overall this is a very good result. The balanced THD+noise into 8 and 4 ohms (not shown) is slightly higher than the unbalanced—approximately 0.02% across the entire spectrum. The 1kHz THD+noise waveform at 200W (!) output into 4 ohms is shown in fig.4. The high output was used to capture the waveform because at lower output it was buried in low-level noise. The result is heavily second-harmonic, though higher harmonics are evident.

Fig.3 Rotel RB-991, THD+noise (%) vs frequency at (from top to bottom at 1kHz): 4W into 2 ohms, 2W into 4 ohms, 1W into 8 ohms, and 2.83V into simulated loudspeaker load (right channel dashed).

Fig.4 Rotel RB-991, 1kHz waveform at 2W into 4 ohms (top), distortion and noise waveform with fundamental notched out (bottom, not to scale).

The distortion spectrum resulting from a 50Hz input at 268W into 4 ohms is shown in fig.5. All of the artifacts are below –90dB (0.003%), an excellent result. The 19+20kHz IM spectrum at 214W into 4 ohms (the maximum output with this test signal before visible signs of clipping appear) is plotted in fig.6. The 1kHz intermodulation artifact is at –86.4dB (0.005%), the 18kHz artifact at –75.2dB (0.017%), both excellent results. The 19+20kHz spectrum at 107W into 8 ohms is nearly the same and is not shown.

Fig.5 Rotel RB-991, spectrum of 50Hz sinewave, DC–1kHz, at 268W into 4 ohms (linear frequency scale).

Fig.6 Rotel RB-991, HF intermodulation spectrum, DC–22kHz, 19+20kHz at 214W into 4 ohms (linear frequency scale).

The RB-991's THD+noise vs level curves are shown in fig.7. The amplifier is not rated into 2 ohms; when I tried to run a 2 ohm test anyway to include in fig.8, the amplifier blew its power-supply fuses at just above 400W output. No damage resulted, and replacing the fuses restored normal operation. I made no further attempt to determine the amplifier's output power with a 2 ohm load. The RB-991's discrete clipping on continuous tones is shown in Table 1.

Fig.7 Rotel RB-991, distortion (%) vs continuous output power into (from bottom to top): 8 ohms, 4 ohms.

Fig.8 Rotel RB-991, distortion (%) vs burst output power into 8 ohms (black trace), 4 ohms (red), 2 ohms (blue), and 1 ohm (green).

John Atkinson used the Miller Audio research Amplifier Profiler to test the Rotel's maximum power output on a toneburst signal that more closely approximates a music signal, 10 cycles of 1kHz followed by 40 cycles off. Under these conditions, the modest-looking Rotel proved a powerhouse, delivering 281.5W into 8 ohms, 539W into 4 ohms, 974 W into 2 ohms, and 1615W into 1 ohm. The latter corresponds to an RMS current of 40A by the way! This behavior is shown graphically in fig.8, which reveals the distortion floor to rise in inverse proportion to the load impedance. Nevertheless, even into the punishing 1 ohm load (green trace), the THD remains below –70dB (0.03%) until the 1300W level. That the Rotel should do well on this toneburst test is no surprise, as the Chinese factory uses the Miller Amplifier profiler for production line testing and quality assurance.

The measured results for the RB-991's unbalanced operation slightly exceed those for the balanced mode. This is not, in my experience, particularly unusual, and the amplifier's test-bench performance is excellent in either setup. It is probably best to keep the loudspeaker load above 4 ohms, however, as Rotel recommends.—Thomas J. Norton

Table 1 Rotel RB-991: Clipping

(1% THD+noise at 1kHz)
Both Channels DrivenOne Channel Driven
LoadW (dBW)W (dBW)
8219.2 (23.4)217.8 (23.4)238.7 (23.8)
4349.1 (22.4)345.9 (22.4)416.5 (23.2)
Rotel of America
54 Concord Street
North Reading, MA 01864
(978) 664-3820