Ray Samuels Audio Emmeline The Hornet headphone amplifier Measurements
Wes Phillips enthused over Ray Samuels Audio's Emmeline The Hornet portable headphone amplifier ($350) in our early-June eNewsletter, as did Sam Tellig in "Sam's Space" in September, and again in this issue. I thought it worthwhile, therefore, to complement those positive impressions with some measurements. (I remembered to fully charge the Hornet's internal 9V battery before starting the testing.)
The Hornet's maximum voltage gains, with its volume control set to its maximum, were 5.6dB, 12.7dB, and 20.8dB with the rear-panel switch set to its Low, Medium, and High positions, respectively. (There was some crackling when the volume control was operated, which suggests the presence of a DC voltage on the potentiometer wiper.) The Hornet preserved absolute polarity (ie, was noninverting) in all three gain conditions. The input impedance was a fairly high 23k ohms at low and midrange frequencies, dropping to 11k ohms at 20kHz. These figures were not affected by the volume-control position or by the gain setting. The output impedance was extremely low, at well below 1 ohm.
The Hornet had a very wide bandwidth, being flat from 10Hz to 100kHz under all conditions, and was down just 0.22dB at 200kHz, as can be seen in fig.1, taken with the volume control at its maximum. The channel matching was superb. Turning the volume control down to 3 o'clock—the unity-gain position in the Low Gain condition—introduced an inconsequential channel imbalance of 0.1dB. Channel separation was good, at 90dB L–R and 98dB R–L at 1kHz, though it decreased at higher frequencies due to the usual capacitive coupling (fig.2).
Fig.1 Ray Samuels Audio Emmeline The Hornet, Low Gain, volume control at maximum, frequency response at 1V into (from top to bottom at 2kHz): 100k, 150 ohms (0.5dB/vertical div., right channel dashed).
Fig.2 Ray Samuels Audio Emmeline The Hornet, High Gain, channel separation, L–R solid, R–L dashed (10dB/vertical div.).
With the Hornet set to Low Gain, its unweighted, wideband signal/noise ratio was a superb 98.7dB (ref. 1V, with the input shorted but the volume control at its maximum), this improving to 116.7dB when an A-weighting filter was switched into circuit. These figures worsened by about 7dB in the Medium Gain setting, and by 14dB in the High Gain setting, but are still excellent in absolute terms. DC offset was low in the Low Gain mode, at 1mV in both channels, but rose to 2.7mV in Medium Gain and to 7.2mV in High Gain. All figures should be low enough not to cause problems.
Distortion levels were also extremely low. The sawtooth nature of the traces in fig.3, for example, which plots the THD+noise percentage against output voltage at the Low Gain setting, indicates that the actual distortion lies at the limit of the Audio Precision system's resolving power. This graph also indicates that the Emmeline Hornet will deliver just over 3.5V at 1% THD into loads ranging from 150 ohms to 100k ohms. Only into the 30 ohms typical of Grado and Ultimate Ears headphones does the maximum output decrease, and then only by a little. The maximum level was the same in all three Gain settings, but with higher THD+N percentages below actual waveform clipping in the Medium and High Gain modes, due to the slightly higher noise floors. This can also be seen in the graph of THD+N percentage against frequency (fig.4), taken at 1V into loads of 100k ohms and 150 ohms. Other than a slight rise in THD above 10kHz in the High Gain setting into the lower impedance, the distortion is vanishingly small.
Fig.3 Ray Samuels Audio Emmeline The Hornet, Low Gain, THD+noise (%)vs 1kHz output voltage into (from right to left at 1% THD): 100k, 600, 150, 30 ohms.
Fig.4 Ray Samuels Audio Emmeline The Hornet, THD+N (%)vs frequency at 1V into (from bottom to top): Low Gain into 100k, 150 ohms; Medium Gain into 100k, 150 ohms; High Gain into 100k, 150 ohms (right channel dashed).
This low distortion is shown in different form in fig.5, which plots the amplifier's output spectrum at 1V into 8k ohms in Low Gain mode. The only harmonic rising above the residual level of the signal generator's output is the second, at –108dB (0.0005%), left, and –105dB (0.0006%), right. Intermodulation distortion was also very low (fig.6), with the second-order difference component resulting from an equal mix of 19kHz and 20kHz tones at 1V into 8k ohms lying at –96dB (0.0015%).
Fig.5 Ray Samuels Audio Emmeline The Hornet, Low Gain, spectrum of 1kHz sinewave, DC–10kHz, at 1V into 8k ohms (linear frequency scale).
Fig.6 Ray Samuels Audio Emmeline The Hornet, Low Gain, HF intermodulation spectrum, DC–24kHz, 19+20kHz at 1V into 8k ohms (linear frequency scale).
This is superb measured performance in absolute terms. However, to get the maximum performance, Hornet owners should use the lowest Gain setting that gives acceptable levels with their preferred headphones. For example, I use Low Gain with the Ultimate Ears UE-5 and UE-10 in-ear headphones, which gives me more than enough output to compete with the noise of the subway on my daily commute.—John Atkinson