T.H.E. Show Newport 2014: Questyle Current Mode Headphone Amplifiers

Bruce Ball was very helpful giving me the basic rundown on the products, but unfortunately he was unable to help me out with the technical details of Questyle's use of current mode amplification. Fortunately the CMR800R Current Mode Amplifier's web page does go into it in a good bit of detail—the English is a little rough, but I'm always very happy to see manufacturers give a detailed technical explanation of a product's inner workings.

Negative feedback in an amplifier is a dirty word for some audiophiles. In solid-state voltage amplifier designs, transistor capacitance causes a short delay from input to output of active devices. When the output of an amplifier circuit is fed back to an input stage to set the gain and stabilize the circuit (negative feedback) the accumulated delay of the circuit is unaccounted for. This difference in timing between the incoming signal and the feedback signal causes transient intermodulation distortion (TIMD), which is one of the common sources for the "transistory" sound (glary, strident) of some solid-state amps—it's most commonly a significant problem in older amps; topologies and devices have come a long way to make solid-state amps sound better these days.

Most amplifiers operate as voltage amplifiers. Questyle's approach is to take a voltage signal input and immediately turn it into a current signal and then amplify it in a current mode circuit, which allows topologies that dramatically reduce circuit capacitance and therefor dramatically increase circuit speed. Since the signal sees much less delay through this current mode amplifier negative feedback can be employed with much less IMD and TIMD, and because negative feedback is used the amplifier can be very linear, stable, and low distortion.

Questyle's website can be found here. I'll let Bruce Ball take you through the line in the video.

COMMENTS
John Grandberg's picture

The little Q192 that Bruce shows at the end was reviewed here http://www.innerfidelity.com/content/tight-and-tasty-questyle-q192-dache... not too long ago. It's a very satisfying all-in-one.

xnor's picture

Current output: nope. The manufacturer explicitly states that the input and output both operate in voltage mode. So "current mode" is maybe a bit misleading at first sight.
(Current output is not a good idea with most drivers anyway, since it eliminates electronic damping - after an impulsive sound stops the driver may electrically continue to ring freely.)
Sure, internally they may do some forth and back conversion, but what for? (see below)

Negative feedback: yes, possibly plenty of it which is good anyway. The schematic posted by the manufacturer shows a feedback loop around everything.

Delays: are no problem in headphone amplifiers. We're dealing with low frequencies here and in a properly designed amp the delay is not unaccounted for, but dealt with by proper compensation.
They have so many components in there that I strongly doubt that you get less delay from such a circuit.

TIMD: is not caused by delay but by slew rate limiting. Again, this is no problem in headphone amplifiers. As nwavguy demonstrated you don't need a high slew rate for headphones and even 30 cent opamps easily surpass common requirements.

Tyll Hertsens's picture
From what I could glen from their description on their web page, they state something about the capacitance being much less in current amplifier topologies. Do you know anything about this?
xnor's picture

They mention a lot of stuff on their web page.

For example they say that conventional amps have "greatly slacked" high frequency response. Then they show a FR chart of their amp that shows -0.3 dB at 200 kHz. Now if we look at an O2, which uses a "rather slow" (3.6V/us) op amp, with the proper compensation we're down roughly 0.5 dB at 200 kHz with 2x gain.
If you need more speed (and have more money) you can get something like an LME49990 (22V/us).

Then they also mention "650KHz wide frequency", whatever that means. An LME49990 has a GBW of 110 MHz.
Anyway, who cares? No headphone rig is going to make use of even a 100 kHz range, especially not full-scale signals.

They also mention TIMD, but I'd rather not go there because it's just FUD.

They mention that current mode transmission is used in "high speed communication and video processing fields" ... and quite rightly so. We're talking about high frequency and long range transmission here, not audio.

So,

since they use voltage on the input, the output and in the feedback loop I have no idea what they're actually doing and why it should make any difference. And judging from their graphs it doesn't. Looks like a fine amp though.

X