CanJam at RMAF 2014: Audeze Prototype-Z High Impedance Planar Magnetic Headphone

Here's something I bet you didn't see coming: A 1200 Ohm impedance planar-magnetic headphone from Audeze.


Well, it makes a lot of sense to me. Thank about the damping factor with a high output impedance tube amp. Let's say you have a 120 Ohm output impedance amp. With this headphone at 1200 Ohms impedance you'd still have a damping factor of 10. Sheesh, that's cool. I'd love to try that.

Audeze didn't have too much information on these cans yet, but Tara did have a hand-out. I think I'll just reprint the important bits here.


It features a new ultra-high-density voice-coil, a first in planar magnetic design, along with a newly-patented diaphragm. The end result is a whopping 1200 Ohm impedance. The force that's generated on the surface of the diaphragm is directly proportional to the length of the voice-coil and the prototype delivers unmatched dynamics, extreme resolution, precise and detailed sound across the audio band, plus excellent imaging.

With 1200 Ohms impedance the prototype can be directly driven by tube amplifiers without an output matching transformer. That will lead to a new wave of high-power OTL tube headphone amplifier designs. In fact, Audeze is working on an OTL design right now.

The Prototype-Z can be driven with any regular voltage-mode amplifier, but better results can be achieved with high quality current-mode amplifiers like the BAKOON. Current-mode amplifiers work very efficiently into this type of load and will be louder at the same volume setting than any other headphone. Voltage-mode amplifiers will require the volume knob to be higher to generate proper driving power into such high impedance modes.

Finally, we've addressed added comfort with our new carbon fiber headband.

Very cool stuff. It's going to be fun playing around with these cans. Probably have to chat with Pete Millett as well, he'll surely have some ideas about tube topologies for something like this.


tony's picture

She's a nice presenter , I'd certainly have her on the Show Floor presenting or just being there helping .
By the way , why would that kind of high Impedance be useful ? and where are they getting Voltage to drive these things ? ( if they are ) or is it just a spoof to show off the Carbon head band ?
Tony in Michigan

mithrandir39's picture

It's not April fools yet, and I think it would be a little counterproductive to introduce spoof product at a trade/enthusiast show.

mithrandir39's picture

Looks like a possible new Flagship from Audeze! I'm loving my LCD-2 F's.
Btw, Tera is an extremely cool chick; got to meet her at T.H.E. Show this year.
@coolphones- Not that it matters, but she's straight. Don't be a jerk in this forum, you troll.

Tyll Hertsens's picture
This is the third time that guy has registered and trolled.

For readers seeing some of the strange replies here, they were in answer to a troll's post that has been removed.

TheAudioGuild's picture

When did damping factor become relevant to planars in the same way that it is for dynamic drivers? With their flat, virtually resistive impedance, where's the resonance that requires any damping (and damping has everything to do with resonance)?

Have you done any comparative measurements of Audeze 'phones driven by high and low output impedances?

superjawes's picture

You could argue that dmping factor is a good rule of thumb with planars to get goo efficiency between amp and headphone. Even if you don't "need" damping on a resistive load, you get a better pairing efficiency with a high damping factor than you would if the amp and loads were matched.

Again, this is a really weak case I'm making. I think the real benefit of a high impedance planar driver could be better pairing with high impedance amps. That would at least make it interesting for people who built systems around high impedance dynamic headphones instead of having their amps fail hard.

TheAudioGuild's picture

Not sure just what you mean by "if the amp and loads were matched." Do you mean a 50 ohm load being matched by an amp with a 50 output impedance?

superjawes's picture

That was a bit sloppy on my part, but when the impedance of your components and/or lines are the same, it is common to call them "matched" (although that is more of a transmission line applications).

If you have a 50 Ohm output impedance on your amp and a 50 Ohm load, you basically have a voltage divider cutting your voltage in half before getting to the headpones. If you see 5 Volts across your headphones, then your output impedance is "stealing" another 5V.

If you bump up your headphone impedance to 500 Ohms and leave the source alone, you'll see about 9V across your 'phones, and you'll be making much better use of your amp's output capabilities. This might not be necessary, but it also doesn't hurt.

TheAudioGuild's picture

Ok, I'm with you now.

Yeah, outside of transmission lines, voltage sources ideally want to drive an infinite impedance and current sources ideally want to drive a short circuit (zero impedance). Which is why I don't understand why these headphones would be advantageous for current source amplifiers like the BAKOON, except from the misguided perspective of "damping factor."

Their high impedance is anathema to current source drive. That's why the input impedance of the SATRI-LINK (which is their current source drive for line level signals) on the BAKOON is only 3.68 ohms, versus 100k ohm for the normal (voltage source) input.

x838nwy's picture

There's still the mechanical resonance of a circular diaphragm, surely. If it's a system that vabrates, it will have a resonance and a response curve. Which means it's damped to a certain degree and can be critically damped. Or am i missing something here...?

TheAudioGuild's picture

I spoke about that in a post below.

Tyll Hertsens's picture
Interesting point, but damping can be applied to non-resonant systems. Think cruise control for a car and how the PID controller creates a critically damped response to a speed change.

The numbers are there to calculate a damping factor---how important it is I'm not sure. And no I've not measured anything related to it.

TheAudioGuild's picture

No, "damping" is used in the same sense with regard to PID controllers.

From this piece on PID controllers:

"Figure 10.2c illustrates the behavior of a system with a PID controller: the system is oscillatory when no derivative action is used and it becomes more damped as derivative gain is increased."

Damping is a well defined term and I don't see it used outside the context of a resonant system except in the "audiophile" world.

It would be instructive to do such comparative testing. Anecdotally, my LCD-2s perform just fine when driven from my TEAC A-H500 which drives the headphone jack off the speaker outputs through 399 ohm series resistors.

ultrabike's picture

So this is how I understand damping stuff:

In a PID controller, damping is related to the proportional (resistive or kp) gain, integral (capacitor or ki) gain, and perhaps a differential (inductive or kd) gain of a feedback loop. A system can be over, under or critically damped.

Over-damped in the PID controller context means that there is proly too much integral gain and the system will take a long time to settle and start tracking. Under-damped means things are tracking fairly fast and the system will oscillate a bit before settling. Critically damped is what one wants as that achieves steady values faster.

In the second order RLC network context, which have no feedback and bare resemblance to PID in how the step response behaves, it's more a function of the network topology and R, L and C values.

In audio AFAIK its a measure of how dominant the load impedance is versus the amp impedance, and that's about it. If the headphone designer had a close to ideal voltage source in mind for his/her cans then having a high damping factor in the audio context proly make sense.

IMO, same word with different meanings depending on context.

TheAudioGuild's picture

No, it's all the same meaning and in both cases has to to with the behavior of resonance.

"Damping factor" is a dimensionless term that's derived by dividing the nominal load impedance by the amplifier's output impedance. But it's ultimately geared toward the resonant behavior of the loudspeaker.

If you look at the impedance plot of a typical dynamic loudspeaker driver, you'll see a big hump at low frequencies. That represents the driver's natural resonant point and the point where it peaks is the driver's resonant frequency.

If you look at the spec sheet for a raw driver, you'll see a figure called Qts. This describes the resonance in terms of its Q factor. An underdamped resonance will have a higher Q and an overdamped resonance will have a lower Q. A Q of 0.5 is called "critically damped." In the critically damped case, the driver will have a maximally flat response without any ringing or overshoot. Above 0.5 the driver will start to exhibit ringing and overshoot, as well as begin to show a peak in the low frequency response.

The Qts of the driver assumes it is being driven from a low (ideally zero) impedance source and this is what loudspeaker designers generally assume when designing loudspeakers.

As the driving source impedance increases, it affects the Q of the driver. The driver's Qts is derived from two different resonances. Its mechanical resonance (Qms) and its electrical resonance(Qes). Generally Qes is overdamped and Qms is underdamped. As the driving source impedance increases, Qes becomes more and more underdamped and eventually you're left with just the mechanical damping controlling the driver's resonant behavior. And since Qms is typically underdamped, that's the behavior you'll start seeing in the loudspeaker. More ringing and overshoot, and a peaking in the low frequency response.

So "damping factor" is just a convenient means of describing how ideal a voltage source a given amplifier is. However it's not of a whole lot of value because there can be quite a lot of leeway before the amplifier's output impedance has any meaningful effect on the loudspeakers behavior. But that didn't stop manufacturers from turning damping factor into a numbers game marketing war.

Anyway, point being that whether in a PID controller or a loudspeaker, damping is all about resonant behavior. And A loudspeaker driver simplifies to an RLC resonant circuit. On the left side of the impedance hump it behaves inductively. On the right side of the hump it behaves capacitively. And the losses in the system determine the overall behavior of that resonance.

ultrabike's picture

Thanks! I think I understand what you're saying. In all cases we are dealing with the step response of a roughly second order system. Furthermore, the source impedance might factor into Re in the speaker RLC circuit model and change it's characteristics.

That may apply to certain dynamics, but don't think it applies to resistive load drivers. I also think damping seems kind of an overused marketing term... like I wanna 1000000 damping ratio.

TheAudioGuild's picture

Yes, that's exactly how it works. To determine the new Qec, it's Qec' = (Re + Rs)/Re, where Rs is the amplifier's output impedance. Then you can calculate the new Qtc from that by way of Qtc' = (QmcQec)/(Qmc + Qec).

And yeah, "damping factor" is really a completely meaningless term when talking about planars. I'm sure planars must have a resonant frequency somewhere, but it's ither so overdamped or outside the audible range that at least for source impedances of upwards of several hundreds of ohms, it doesn't seem to matter.

However the BAKOON has an output impedance of about 2 meg ohms. That might matter. I don't know. That's why I think it would be instructive to see some comparative measurements.

ultrabike's picture

I had the chance to measure a planar out of the Bakoon and DIDN'T!!! But I did measure an HD600 out of the Bakoon using both current and output ports:


FWIW, it seems the output impedance from the voltage port is 1.5 to 1.75 ohms. Current port is somewhere in heaven of course, so maybe 2M or something.

TheAudioGuild's picture

Yeah, that bass bump is what I would expect driving a dynamic from a current source.

Wonder what they're doing for the voltage amp. See an awful lot of high order harmonics in the spectrum plot.

Gotta be someone near you with a pair of planars. Where a outs are you located?

ultrabike's picture

I'm in Irvine, CA. Don't have the Bakoon and planars anymore though. They were on loan :)

It was good to hear them though.

TheAudioGuild's picture

Oh. Well hell. Guess we'll all just have to bug Tyll to do it. ;-)

NickS's picture

It would be great if they redid the arm mechanisms too (to make them lighter and decrees the clamping force). I love my LCD-2s but they are very uncomfortable on the top of my head and the way they drive my glasses into my skull.

sankar's picture

These are quite efficient. We measure almost 106 dB/1mW (ERP) on the prototypes.

xnor's picture

Damping factor with current drive is (close to) zero, but it may actually work well with headphones that do rely less on electrical damping.

The comparison of voltage vs. current drive is a bit fishy. Pushing a certain amount of current through 1200 ohms will result in the same voltage as doing it the other way around.

Of course you will have to turn up the volume control higher with a voltage amp, but so what?
Lets say the current amp has max current of 150 mA. That is 675 mW into 30 ohms. But this would require a voltage swing of 180V into 1200 ohms!
The current amp will also be voltage limited. About 7V in case of the BAKOON mentioned above. So now we're down to 6 mA into 1200 ohms.
With a normal volume control this means clipping between 9 and 10 o'clock...

ultrabike's picture

Nice new avatar Xnor! :)

Unless I effed up something, into 300 ohms the Bakoon from current out can do about 9.78 Vrms in high gain mode and about 5.57 Vrms in low gain mode. Those are absolute max values with the volume knob all the way up so most likely the amp was struggling.

Into an open I got all kinds of weirdness but I think it managed 11 Vrms, so I would be surprised if it did more than 9 mA into 1200 ohms... still, that's a not too bad 100 mW.

Maybe Audezee should just forget about all these high impedance nonsense and put out an electrostatic can already.

Other than that, sound will definitively change in current out for non resistive loads.

Also, like Tyll said below... efficiency is about 106 dB/mW (, so not a lot of current needed it seems.

xnor's picture

I extracted the number from their specs (1 W into 50 ohms). It might very well be higher with other loads, gain options, ...

Measuring the current-out with a voltmeter (couple of megaohms) should result in crazy clipping, because pushing even 0.1 mA through the "load" would result in crazy high voltages.

If that sensitivity rating is more or less accurate, then any 2+V amp should be fine. This would be one of the easiest to drive headphones on the market: mostly resistive, high impedance, high sensitivity.

ultrabike's picture

I think that 1W into 50 ohms is peak power. The way I can approach that number is as follows:

Consider max voltage into 33 ohms is 3.24 Vrms (in current out). This corresponds to 318 mW (rms) or 636 mW (peak). The corresponding peak current is therefore 139 mA. If the peak current is held constant into 50 ohms (amp is not running out of voltage headroom) then peak power would be (0.139)^2*50 = about 1W (peak).

At some point the amp is proly going to run out of voltage headroom though. That appears to be somewhere around 9 to 10 Vrms or something like that. Also, like you said, into an open I proly got all sorts of crazy clipping and stuff.

The amp also seems to measure better in current out that voltage out (distortion-wise). Specially when the load is highish.

Also note that HD6x0 are in the range of 103 dB/mW, and they do need a bit of voltage. So I guess that for these a practical USB deal might struggle a bit (unless listening fairly quiet) but who knows (I think one needs to convert rms voltages to peak to see where they end up when supply is only 5V single ended).

xnor's picture

But just "power" is by definition the result of Vrms*Irms. Well, I guess there's a bit of marketing in the specs.

If there was some clipping with your 300 ohm and volume all the way up test, then RMS voltage would appear bigger (up to 3 dB as we approach a square wave). Funnily enough, 3 dB is the difference between 7V and 10V, but I'm purely speculating at this point. It could very well do over 9V unclipped with high enough load impedance.

HD6xx do 103 dB/V, which is a bit lower than this new Audeze. You should still be fine with a 2V source and no additional gain in the headphone amp, but the Sennheisers would definitely be harder to drive (lower impedance, sensitivity, non-resistive).

Tyll Hertsens's picture
The thing we don't know is the efficiency of it. Maybe it's so efficient that it doesn't need much voltage. Dunno.

Whatever the case, I appreciate the numerous manufacturers who brought prototypes to RMAF. It keeps things quite interesting.

Edit: Whoops, Sankar posted the efficiency. Have to run some more numbers.

xnor's picture

106 dB @ 1 mW would be quite efficient.
That would mean >110 dB with a standard 2V CD player without any additional gain.

As I wrote above, that would be one of the easiest to drive hi-fi headphones on the market!

Most people seem to believe that high impedance means hard to drive, well, actually it's kinda the opposite and only half of the story.
The ideal headphone would have extremely high impedance and high sensitivity. => You could drive it directly from a line-out without an amp.

The Headphone Viking's picture

Any word on the estimated pricing of it? I generally find Audeze's to be overpriced when it comes to technical performance... They just have an entizing presentation. So if can offer a technically better Audeze without going full Abyss on the price it might be interesting :P
And that headband will too - god knows I need it for my LCD-X

DT48's picture

I find the idea of a high impedance headphone quite intriguing, there was a reason Sennheiser and Beyer opted for that solution in the old days. I'm thinking electric motor force, driver control, and coil/membrane mass.

Modern low impedance headphones gain efficiency at the cost of other qualities. Can't get anything for free by laws of nature.

Good electrostats sound quite fine to me.

cundare's picture

Well, now maybe I won't be selling my Quad II's after all. What to do with 15 watts?