HeadRoom BlockHead headphone amplifier Measurements
We covered the behavior of the HeadRoom Process in some depth in earlier reviews—see my review of the HeadRoom Supreme in January 1994 and Wes Phillips' review of the HeadRoom Max in February 1997. Suffice it to say that this proprietary circuitry adds equalized and time-delayed crosstalk to make the experience of listening to normal amplitude-encoded stereo recordings on headphones less artificial. Fig.1, taken from my Supreme review, shows the Process's different frequency responses when the channels are in phase compared with when they are out of phase.
Fig.1 HeadRoom Supreme, frequency response at 1V into 100k ohms with Process on and the two channels in phase (top trace at 100Hz) and out of phase (bottom at 100Hz). (Right channel dashed, 2dB/vertical div.)
As the high frequencies can tend to sound a little recessed when the Process is engaged, the HeadRoom amplifiers feature switchable treble EQ, called "Filter." The BlockHead has a three-position Filter switch; the amplifier's responses with the two EQ positions selected are shown in fig.2. Both result in a moderate peak of +3.5dB in the treble, but the Brighter setting extends lower in frequency. Note the 0.6dB difference in level between the two channels in this graph, which was taken with the Gain switch in its centered, minimum position. As can be seen from fig.3, which shows the BlockHead's frequency response with the Filter and Process switches set to Off and the Gain switch in its three positions (4.04dB, 11.75dB, and 15.72dB, all into 100k ohms), this interchannel level imbalance decreases as the gain is increased. At the highest gain setting, the imbalance is just 0.1dB. The response didn't change when the test load was reduced from the unrealistically high 100k ohms to the 150 ohms more typical of moving-coil headphones, by the way.
Fig.2 HeadRoom BlockHead, frequency response at 1V into 100k ohms with Filter set to Bright and Brighter positions (right channel dashed, 1dB/vertical div.).
Fig.3 HeadRoom BlockHead, frequency response into 600 ohms with Gain switch set to (from top to bottom at 200kHz): High, Medium, Low (right channel dashed, 0.5dB/vertical div.).
More significant, note in fig.3 that the BlockHead's ultrasonic response also changes with the Gain setting. In the lowest Gain setting, the output is a sensible -0.25dB at 20kHz. With the middle Gain setting selected, the -0.25dB frequency increases to 70kHz, but with the highest Gain setting, the response actually peaks slightly above 200kHz. This is reflected in the 10kHz squarewave responses in the highest and lowest Gain settings (figs. 4 and 5), the former showing a slight overshoot on the leading edges of the wave, the latter the expected slower risetime.
Fig.4 HeadRoom BlockHead, High Gain, small-signal 10kHz squarewave into 600 ohms.
Fig.5 HeadRoom BlockHead, Low Gain, small-signal 10kHz squarewave into 600 ohms.
The BlockHead's input impedance was very high, which is good. It looked as if the input impedance was well over 200k ohms at 1kHz, and just below 200k ohms at the edges of the audioband. However, the change in output level as the Audio Precision's generator source impedance was changed from 600 to 50 ohms was just a couple of millivolts, which means that the experimental error in this figure is large. Nevertheless, the BlockHead will not load down the source component in any way. The unit appeared to be noninverting, with pin 2 of the XLR plugs connected as "hot." Front-panel switches can be used to invert polarity. The output impedance was less than 1 ohm across the band, the measured result mainly reflecting the series impedance of the cables used.
With the BlockHead's dual-mono construction, I wasn't surprised to find that channel separation was buried beneath the noise floor in the bass and midrange. This did decrease to a still good 67dB at 20kHz, perhaps due to capacitive coupling across the Process switches. Not only was the noise low—around 106dB ref. 1V output—but distortion was very low as well, as can be seen from fig.6. This graph was plotted at a high 6.67V output to raise the distortion products out of the noise. Nevertheless, the THD hovers at the 0.001% level (-100dB) below 3kHz and rises slightly only in the top audio octaves, due to the circuitry's decreasing gain-bandwidth product in the region.
Fig.6 HeadRoom BlockHead, Low Gain, THD+N (%) vs frequency at 6.17V into (from bottom to top at 1kHz): 100k ohms, 600 ohms, 150 ohms (right channel dashed).
As well as the absolute level of distortion, its spectrum matters. Fig.7 shows that all the distortion harmonics are very low in level, even with the BlockHead driving a 50Hz tone at 6.17V into 150 ohms. The second, third, fourth, and sixth harmonics are all very low and the highest in level was the fifth, at -94dB (0.002%). This is at odds both with the fig.6 results and with measured data I was supplied by HeadRoom's chief designer, Daniel Bartlett. It's possible, therefore, that this result is incorrect. Nevertheless, the BlockHead's linearity seems beyond reproach.
Fig.7 HeadRoom BlockHead, Low Gain, spectrum of 50Hz sinewave, DC-1kHz, at 6.17V into 150 ohms (linear frequency scale).
Finally, fig.8 shows how the percentage of distortion and noise present in the BlockHead's output changes as its output voltage increases. Ignoring the little sawtooth jiggle in the traces at 11V, which is due to the Audio Precision System One's gain-ranging, the sloped-down shape of the traces below 10V or so indicates that the measurement is dominated by noise. The amplifier does hard clip at higher voltages, but as these voltages are so high—even into 150 ohms, the HeadRoom will put out 12V, way above what's needed to turn typical headphones into molten slag—that users will not get anywhere close to clipping.—John Atkinson
Fig.8 HeadRoom BlockHead, Low Gain, distortion (%) vs output voltage into (from bottom to top): 100k ohms, 600 ohms, 150 ohms.