This review page is supported in part by the sponsors whose ad banners are displayed below |
 |
 |
|
 |
First a brief disclaimer. If power were the EF6's main or only contribution to this game, the obvious question tantalizing cheapskates (whether some old receiver or brand-new speaker integrated—say a Dayens Ampino—couldn't be converted for pennies on HifiMan's dollar) falls outside this assignment. It requires nothing but a jerry rigged converter from binding posts to 4-pin XLR to find out.
Three provisos apply. 1/ not every circuit meant for 4-8Ω loudspeakers is happy or stable into a 50Ω or higher headphone load. 2/ whatever the power rating into 8Ω should be down about 25% or more into an HE-6. 3/ the noise floor of such an amp might be too high for headphones. But if $1.600 has you gasp, certainly experiment. You might already own something which can do the heavy headphone lifting to your satisfaction. With low-Z cans like HifiMan's it's certainly a far better prospect than with 600Ω brutalizers of speaker amps. In case of success, you might go fancy and build yourself a proper cable like this Stefan AudioArt leash for the K-1000 terminated in Cardas Rhodium spades. End of disclaimer.
|
|
|
 |
Beneath the sheets. I asked Fang Bian to explain his circuit in as much detail as he was comfortable.
"Input impedance is 100K. Power output is 5 watts into 50 ohms, 3 watts into 100 ohms. The pre-out impedance is 600 ohms at its max 2V (the preamp circuit is separate from the headphone output and uses an OPA627 input and OPA2604 output buffer). Overall circuit gain is 25dB. The NFB switch injects 10dB of feedback and lowers the output impedance from 1.2R to 0.66R.
The amplifier topology is a 3-stage circuit with an OPA627AP input followed by a Toshiba 2SK246GR 2SJ103GR Jfet buffer, a B716/D756 voltage gain stage and 3 pairs of Toshiba 2SK214 2SJ77 output Mosfets. The attenuator stops equate to values of -∞dB, -60dB, -55, -50.1, -45.2, -40.2, -37.9, -35.6, -33.2, -30.9, -28.6, -26.2, -23.9, -21.6, -19.3, -17, -14.6, -12.3, -10, -7.6, -5.3, -3 and 0dB."
|
 |
Sub covers in place
|
 |
Sub covers removed
|
 |
The stepped attenuator
|
|
|
"The volume control uses DIP resistors instead of SMT types. At the same price level SMT resistors are smoother in operation but sonically inferior. If we used a smoother attenuator with less physical resistance but maintained the present sound quality, its raw cost would easily exceed $150. This would significantly impact the final retail price. About the toggle switch in the back, even though it says 'gain' it really is a negative feedback switch. But most headphone customers don't know what that is. Since feedback reduces gain, we've labeled it 'gain' to avoid confusion. And yes, the XLR and 6.3mm outputs may be used simultaneously without much effect on the sound quality. The 6.3mm socket has a series resistor to protect normal-efficiency headphones. The output stage is parallel single-ended class A biased at 0.3A per channel with an idle power draw of 58 watts."
|
 |
One channel's output stage
|
 |
Three output Mosfets per channel
|
 |
Voltage regulators
|
|
|
|
 |
Buffer JFets
|
 |
Input selector switches
|
 |
Relay-switched input socket board
|
 |
Final closeup of attenuator
|
 |
This tour of the EF6's innards demonstrated Chinese workmanship not quite on par with Prima Luna or Mystère. Enclosure cosmetics were far more agrarian than for example the slick casings US competitor Woo Audio sources from the PRC for their DAC and transport models [left, $1199/ea.].
As the makers of my standby comparison models, Burson and Linnenberg too enjoyed eye candy and tactile advantages. Those tend to affect general buyer perception. Ultimately it's sonics of course which are our primary concern. But it's fair to say that for a presumed statement effort the EF6 plays it a bit coy on sexiness. That also went for the really clunky attenuator and equally stiff input selectors. How would Being Earnest translate at the ears then? |
|
|
 |
  |
 |
|