Final Laboratory Music-4 phono preamplifier, Music-5 line preamplifier, & Music-6 power amplifier Measurements part 2
Fig.7 Final Music-5, gain trim at 12:00, spectrum of 1kHz sinewave, DC-10kHz, at 1V into 8k ohms (linear frequency scale, 10dB/vertical div.).
Fig.8 Final Music-5, gain trim at maximum, spectrum of 1kHz sinewave, DC-22kHz, at 10V into 100k ohms (linear frequency scale, 10dB/vertical div.).
Fig.9 Final Music-5, gain trim at maximum, HF intermodulation spectrum, DC-24kHz, 19+20kHz at 1V into 1k ohm (linear frequency scale).
With its trim controls set to give maximum gain, the Music-5 is fairly straightforward in its measured performance. However, if the trim controls are set to anything other than their maximum settings, the preamp's behavior becomes very dependent on the partnering power amplifier's input impedance.
The Music-6 power amplifier's variable negative feedback control could have made measuring it a time-consuming exercise in complexity. I therefore simplified the task by performing complete sets of measurements on the amplifier with the control set to its two extreme positions, on the assumption that the performance with the control set somewhere in the middle can be inferred. The battery voltage for the measurements started at ±25.5V, which is when I measured the Music-6's output power, but had drifted down to ±24V by the end of the tests. I don't believe this had a significant effect on the results.
The input impedance at 1kHz measured a very high 470k ohms—just as well, given the Music-5 preamplifier's high source impedance. The Music-6 didn't invert signal polarity. With the output unloaded, the amplifier's voltage gain depended, as expected, on the setting of the feedback control, varying from 25.4dB (maximum feedback) to 31.4dB (minimum). To my surprise, when the Music-6 was loaded with 8 ohms, the gain was the same in the two conditions, at 24.3dB. Rapid operation of the feedback control introduced a varying DC offset as the amplifier's internal operating conditions changed; this could reach as high as 70mV.
With maximum feedback, the Music-6's output impedance was moderate across the audioband, at 0.76 ohm. But with minimum feedback, the output impedance both rose to the highest I have ever encountered—15 ohms!—and displayed a relationship with the load impedance. As a result, there will be an extraordinarily high modulation of the amplifier's frequency response by the varying impedance of the loudspeaker.
This can be seen in figs. 10 and 11, which show the response in the two feedback conditions. With maximum feedback (fig.10), the response variation with our simulated speaker load is around ±0.5dB and the high-frequency output is 3dB down at a very high 156kHz, well above the audioband. With minimum feedback (fig.11; note different vertical scale), there is now almost 10 times the variation in response with the dummy load, and the ultrasonic bandwidth is modulated by the load impedance, with a peak appearing above 100kHz. Interestingly, though this results in an overshoot on the leading edges of a 10kHz squarewave, this overshoot didn't disappear as I had expected from fig.10 when the feedback was increased (fig.12).
Fig.10 Final Music-6, maximum feedback, frequency response at (from top to bottom at 2kHz): 2.83V into dummy loudspeaker load, 1W into 8 ohms, 2W into 4 ohms, 4W into 2 ohms (0.5dB/vertical div., right channel dashed).
Fig.11 Final Music-6, minimum feedback, frequency response at (from top to bottom at 2kHz): 2.83V into dummy loudspeaker load, 1W into 8 ohms, 2W into 4 ohms, 4W into 2 ohms (2dB/vertical div., right channel dashed).
Fig.12 Final Music-6, minimum feedback, small-signal 10kHz squarewave into 8 ohms.
Channel separation (not shown) was poor in the L-R direction, at 56dB over most of the audioband, but was about 10dB better in the other direction. The A-weighted S/N ratio—an excellent 98dB ref. 2.83V into 8 ohms—was not affected by the negative-feedback control, but the wideband ratio was, measuring 93dB with maximum feedback, 85dB with minimum, which correlates with the greater potential gain in the latter state.
To my surprise, given that negative feedback is generally used to improve a circuit's linearity, not only were there only small differences between the manner in which the Music-6's small-signal THD varied with frequency with maximum and minimum negative feedback, but the amplifier produced more high-frequency distortion with the maximum feedback. This is graphically shown in figs. 13 and 14 (maximum and minimum feedback, respectively). Alarming behavior!
Fig.13 Final Music-6, maximum feedback, THD+noise vs frequency at (from top to bottom): 2W into 4 ohms, 1W into 8 ohms.
Fig.14 Final Music-6, minimum feedback, THD+noise vs frequency at (from top to bottom at 5kHz): 2.83V into dummy loudspeaker load, 4W into 2 ohms, 2W into 4 ohms, 1W into 8 ohms.