Mark Levinson No.383 integrated amplifier Measurements
A one-hour preconditioning warmup at one-third full power exerts maximal thermal stress on an amplifier with a class-B output stage, and after that hour the No.383 was indeed warm to the touch—but not so hot that I couldn't keep my hand on its heatsink.
The amplifier didn't invert absolute polarity. (The XLR inputs appear to be wired with pin 2 hot.) The input impedance was 19k ohms balanced and 10k ohms unbalanced at 1kHz, both lower than the specified 100k ohms. The tape output source impedance was a low 50 ohms. With the line-stage gain set to "6dB," which was how LG had it set up, and the volume control to its maximum "73.2" setting, the overall voltage gain into 8 ohms was 33.6dB, suggesting a power-amplifier gain of 27.6dB. The power-amplifier output impedance was very low at approximately 0.055 ohms across most of the audioband, this rising to a still negligible 0.065 ohms at 20kHz.
Because of this very low source impedance, there was virtually no change in frequency response as the load varied (fig.1). The response was flat from 10Hz to 20kHz, with then a 1dB drop at 90kHz and a 3dB drop at 180kHz. As a result, the amplifier's reproduction of a small-signal 10kHz squarewave was virtually perfect (fig.2). Channel separation (fig.3) was superb, at better than 110dB below 500Hz. There was a smooth increase in crosstalk at higher frequencies due to capacitive coupling. Nevertheless, channel separation was still around an excellent 80dB at 20kHz.
Fig.1 Mark Levinson No.383, frequency response at 1W into 8 ohms, 2W into 4 ohms, and 2.83V into dummy loudspeaker load (5dB/vertical div.).
Fig.2 Mark Levinson No.383, small-signal 10kHz squarewave into 8 ohms.
Fig.3 Mark Levinson No.383, crosstalk (10dB/vertical div., L-R dashed).