Melos SHA-Gold headphone amplifier Measurements
The Melos's measurements presented here were taken from the front-panel headphone jack unless otherwise stated—since the SHA-Gold was reviewed primarily as a headphone amp. Selected measurements, however, were also made from the back-panel preamp outputs.
The Melos was non-inverting, both from its headphone jack and its main outputs. (Using its balanced output, pin 2 was wired to be positive, pin 3 negative.) Its input impedance was 82k ohms. Its output impedance was 0.53 ohms at the headphone jack, 21 ohms from output 2.
The S/N ratio measured 81dB from the front jack, 73.6dB from output 2, and 67dB from output 1 (all 22Hz–22kHz, unweighted, ref. 1V). DC offset measured 2mV, left, and 8.9mV, right from the headphone jack, 7mV/5.9mV from output 2, and 2.1mV/2.2mV, from output 1. There was considerable fluctuation in the readings, due to very–low-frequency noise, as is typical in tube designs (the figures given were the measured maximums).
The tracking of the Melos's volume control was good, the left and right outputs had a maximum of 0.39dB deviation at half rotation of the volume control. Voltage gain at the maximum setting of the control was 18.6dB at the headphone jack, 19.2dB at the balanced output, 13.2dB at output 2, and –2.3dB at output 1 (the "passive" output).
The frequency response of the SHA-Gold into a high impedance load is shown in fig.1. The 10kHz squarewave response is shown in fig.2. The latter has good rise time and the typical slight rounding at the top leading-edge corners associated with the ultrasonic rolloff seen in fig.1. At the bottom leading-edge corners, the slight notching seen is not a measurement or printing artifact, it was also visible on an analog oscilloscope trace. I have no ready explanation for it, other than a slight oscillation that only occurs in the negative portion of the signal. It is also just barely visible in the 1kHz squarewave response (not shown), which also has a very slight overshoot at the top leading-edge corners, but good risetime and no visible ringing.
Fig.1 Melos SHA-Gold, frequency response at 1V into 100k ohms (0.5dB/vertical div.).
Fig.2 Melos SHA-Gold, small-signal 10kHz squarewave into 100k ohms.
The SHA-Gold's variation of channel separation with frequency is shown in fig.3. It is only fair as preamps go, but the crosstalk is well-matched between channels. The somewhat lower-than-average separation, particularly at high frequencies, should not be of audible significance.
Fig.3 Melos SHA-Gold, crosstalk (from bottom to top at 100Hz): L–R, R–L, output 2; L–R, R–L, headphone jack (10dB/vertical div.).
Fig.4 plots the variation in THD+noise with output voltage of the Melos SHA-Gold into 40 and 150 ohm loads (headphone output) and a high impedance load (output 2, unbalanced). The headphone output voltages should be more than sufficient to drive any typical headphones to extremely high levels. The line-stage output clips at 2.5V, lower than expected, but sufficient to drive any power amplifier to full level.
Fig.4 Melos SHA-Gold, distortion (%) vs output voltage into (from bottom to top at 2V): 100k ohms, output 2; 150 ohms, headphone jack; and 40 ohms, headphone jack.
The Melos's THD+noise level is plotted against frequency in fig.5, for outputs of 3V and 300mV into 40 ohms. As expected from fig.4, the 3V distortion is moderately high, but this level is well above that likely to be required by most headphones. The THD+noise vs frequency plot for an output level of 1V is also shown for output 2 (unbalanced, left channel). This is suitably low. The THD waveform at 250mV into 40 ohms (1kHz) is shown in fig.6. It is heavily second-harmonic. As the output level increases to 2V (not shown) higher harmonics begin to appear, but the second-harmonic component remains dominant.
Fig.5 Melos SHA-Gold, THD+noise vs frequency at (from top to bottom at 1kHz): 3V into 40 ohms, headphone jack; 300mV into 40 ohms, headphone jack; and 1V into 100k ohms, output 2.
Fig.6 Audio Alchemy HPA, 1kHz waveform at 250mV into 40 ohms (top); distortion and noise waveform with fundamental notched out (bottom, not to scale).
Driving the Melos with a frequency of 50Hz at an output of 300mV into 40 ohms results in the distortion spectrum shown in fig.7. Power supply products predominate (120Hz, 240Hz, 360Hz, etc). The largest artifact of the 50Hz input signal is the second-harmonic (100Hz) at –75dB (about 0.017%). Fig.8 shows the output intermodulation spectrum for an input of 19+20kHz at 2.17V into 40 ohms, just prior to visible clipping of this waveform. The difference component artifact at 1kHz is the highest in level, at –42.5dB or about 0.8%. This is a satisfactory, though not particularly notable, result.
Fig.7 Melos SHA-Gold, spectrum of 50Hz sinewave, DC–1kHz, at 300mV into 40 ohms (linear frequency scale). Note that the second harmonic at 100Hz is the highest in level but that the spectrum is dominated by an AC supply harmonic at 120Hz.
Fig.8 Melos SHA-Gold, HF intermodulation spectrum, DC–22kHz, 19+20kHz at 2.17V into 40 ohms (linear frequency scale).
The test-bench measurements of the SHA-Gold were reasonable, though it did not perform as well on an objective basis as the other headphone amplifiers reviewed by WP in this issue. The technical results suggest that headphones with impedances above 40 ohms are best suited for use with the Melos. Certainly, WP did not have anything negative to say about the preamp's subjective performance.—Thomas J. Norton