importance of output impedance for headphones?

[bentoronto]

I've recently been digging into headphone amplification, particularly amp output impedance.

In the loudspeaker world, people are intent to have low output impedance - say, below 0.5 Ohms - and hence good damping of cone drivers.

But does damping improve headphone sound? That my question.

Sometimes in reading reviews, I think some headphone users glory on tubby undamped bass. And most headphone amps (often IC driven) are happy to have 35 Ohm or bigger resistors sitting in the output leg and rarely tell consumers what the output impedance of their amps is.

With my Audio-Technica ATH-A900 headphones (50 Ohms??), I think I hear a hint of damping when driven by a speaker-like output circuit as compared to a garden-variety headphone amp circuit.

BTW, can anybody tell me the output impedance of the headphone amp in a Behringer Xenyx 1204 mixer? Or where I can get a schematic?

Thanks.
Ben

 

[31697B]

I dont know about that model, but recievers often use 100 Ohms in series from speaker to headphone jack. Mostly, it was just a simple and cheap way to reduce the output level for headphone use. The affect on sound wasnt much of a consideration and headphone makers standardized on 100 Ohm for their designs. But now we have new headphone models and dedicated headphone amps so it is harder to tell what Zo is expected without consulting the mfr.


THx-RNMarsh

 

[scott wurcer]

But does damping improve headphone sound? That my question.

In the opinion of several cell phone manufacturers, yes (to my surprise BTW). Picking the right damping resistor is of great concern.

 

[bentoronto]

I dont know about that model, but recievers often use 100 Ohms in series from speaker to headphone jack. Mostly, it was just a simple and cheap way to reduce the output level for headphone use. The affect on sound wasnt much of a consideration and headphone makers standardized on 100 Ohm for their designs. But now we have new headphone models and dedicated headphone amps so it is harder to tell what Zo is expected without consulting the mfr.


THx-RNMarsh

Loudspeaker amps had those resistors to keep from blowing up headphones and to keep the noise-floor less hearable on the phones.

Worth noting that loudspeaker amps generally have the feedback loop at the tail end, driving down output impedance. But headphone amps I've seen have the protective resistor last in the line and thus a passive resistance.

Yet most boutique headphone amps have IC output transistors and protective resistors in the 25-100 Ohm?? range whether or not that IC really needs protection.

Ben

 

[bentoronto]

In the opinion of several cell phone manufacturers, yes (to my surprise BTW). Picking the right damping resistor is of great concern.

Any details to share?

You are implying something less than max damping might make some manufacturers happy??

Am I wrong to take for granted that headphone motors - like quality loudspeakers - have the moxie to benefit from damping? Or is the physics of closed or open headphones different and hence the need for damping (AKA degenerative feedback) of less importance?

Ben

 

[scott wurcer]

Any details to share?

Either way I don't think there is a consistent story to report. Some of the next generation phones will support a separate "audiophile" signal chain. I think someone mentioned the HTC Harmon Kardon phone here already. Obviously bass is an issue with earbuds so some folks are fiddling with using the resistor to tune it and supplying custom designed earbuds to take advantage.

 

[jcx]

Headroom plots the impedance - lack of axis auto scaling is annoying though

An externally hosted image should be here but it was not working when we last tested it.

the biggest effect is usually the bas mass-spring resonance - visible in some impedance plots - V drive gives the typically plotted frequency response, a series R causes some peaking where the headphone Z rises

Tyll Hertsens has done more measuring of headphones at inner Fidelity: Headphone Data Sheet Downloads | InnerFidelity

some high end iem use multiple drivers, can have multiple impedance bumps over audio

since bass sensitivity is low, bass bumps may not be very audible

...
practically there is little point in extreme accuracy at 10 Hz - several dB, 10-20 % mismatch won't reach audible thresholds that low


we just don't hear, resolve small frequency response differences that low - even if they are easily measured

Clark, David L., "High-Resolution Subjective Testing Using a Double-Blind Comparator", Journal of the Audio Engineering Society, Vol. 30 No. 5, May 1982, pp. 330-338

ABX Amplitude vs. Frequency Matching Criteria
...

headphones are way more linear than most loudspeakers - but there could also be some distortion changes - as you increase output Z some dynamic driver nonlinearities are reduced - really unlikely to be audible if you believe Geddes

 

[scott wurcer]

since bass sensitivity is low, bass bumps may not be very audible

I'm afraid this falls into the customer is always right category for me.

 

[bentoronto]

the biggest effect is usually the bas mass-spring resonance - visible in some impedance plots - V drive gives the typically plotted frequency response, a series R causes some peaking where the headphone Z rises

Tyll Hertsens has done more measuring of headphones at inner Fidelity: Headphone Data Sheet Downloads | InnerFidelity

some high end iem use multiple drivers, can have multiple impedance bumps over audio

since bass sensitivity is low, bass bumps may not be very audible


headphones are way more linear than most loudspeakers - but there could also be some distortion changes - as you increase output Z some dynamic driver nonlinearities are reduced - really unlikely to be audible if you believe Geddes

Headphone acoustics seems much more benign than for loudspeakers (which is pretty close to unsolvable with current devices).

If a user could see an impedance curve (for example, for my A-T A900's) then it should be clear enough if the moving parts are under-damped. If no big resonance peak, no tight coupling and/or no floppy diaphragm and no tubby bass. So no problem having high output impedance amp (like most are).*

Ben
*in preparing this thread, I noticed a sponsor/vendor thread who was giving away pc boards... they had 100 Ohm protective resistors.

 

[jcx]

the old DIN standard is never used today with so many 32 Ohm headphones for DAP running from a single LiIon cell

the high flex cables for headphones may have the wires very close - nF of cable C is possible even in typical length manufacturer supplied headphone cables

some op amps like a isolating impedance for that much C load - but few 10s of Ohms often works
lossy ferrite bead or series R||L as used in audio power amps can give lower audio Z and still isolate the output from the cable C at MHz feedback gain intercept

 

[agdr]

Any details to share?

You are implying something less than max damping might make some manufacturers happy?

I asked the preferred headamp output impedance question a year ago to AKG tech support (Harman) about the AKG K550's. The surprising answer that came back was Zo the same as the headphones, 32R, +/-20R. Not zero ohms. That implies a 50/50 voltage divider with the amp output resistance. No big deal on output swing though given the sensitivity is 114dB/V. I've tried it both ways with Zo = 0.083R and Zo = 32. I really can't say I hear much difference one way or the other, although it is pretty easy to psych myself into thinking I hear the bass being a little "better", whatever better is, with the 32R. All sighted bias.

 

[Kiriakos]

Headphones impedance was somewhat important if you were owning an poor mixing console with a weak headphones output.
Cheap headphones was sounding launder, but most full side headphones was following more strict specifications.

Since 1993 I own my Koss Pro/4X Plus and selected them after testing eight headphones sets of high pricing, just before one week time measured their impedance by my LCR and found to be at 108 Ohm.

 

[bentoronto]

DATA

I ran trustworthy but sort of semi-calibrated impedance curves on my nearby headphones: Audio-Technica A900 (sealed, some mods) and Sony MDR-7 (open, about 30 yrs old but good quality at the time).

Some conclusions:

1. Minor bumps in curves - likely diaphragm resonance
2. Surprisingly precise match between sides
3. No difference when A-T mounted on an elderly human head
4. Since there appears to be not too much undamped resonance with these headphones, I think it is unlikely that a low output impedance would be materially beneficial, esp. in light of substantial DC resistance present. But prolly just enough resonance to lead to Big Arguments among reviewers and different amps and tastes.

The top curve is the impedance of a 83 Ohm resistor and bottom is 39 Ohm.
Orange curve is Sony with peak at maybe 150 Hz; DC resistance about 53 Ohms
Red (lower) curve is A-T with peak at maybe 80 Hz; resistance about 41 Ohms

A loudspeaker (driver) curve would have an Everest style peak at, say, 25 Hz (resonance) and a full-range driver might be going somewhat up from 3 kHz (voice coil inductance).

Ben

 

[coluke]

I've recently been digging into headphone amplification, particularly amp output impedance.

Amp output impedance has a big influence on measured THD on real loads: http://www.diyaudio.com/forums/solid-state/210614-real-load-thd-somewhat-philosophical-question.html; but the question about its influence on perceived THD is still open...

L.

 

[camerabrain]

There were a set of headphones that literally had bass mid and treble cones in them that I had in the late 80's that made amazing bass playback. There was another pair I had that took the bass frequency and split it through a vibrating mechanism so as to bring bass that you could feel.

 

[availlyrics]

In a Stereophile article about headphones, they said of the 120 ohm standard(IEC 61938 from 1996):

“Whoever wrote that must live in a fantasy world.”

sourced from NwAvGuy: Headphone & Amp Impedance

Also, a must read-
"The "0-Ohm" Headphone Amplifier" by John Siau
Benchmark Media Systems, Inc. - The "0-Ohm" Headphone Amplifier

 

[counter culture]

I would suggest that a low output impedance should be the aim.

We've got a highly disorganised scene.

I'm talking here about conventional drivers, not exotics, ESL...

I'm not saying that all headphones currently in manufacture will benefit from a low output impedance in terms of SQ. I'm saying that headphone manufacturers should design with the anticipation of an amplifier with a low output impedance. I'm saying that nobody should be including an output resistor in their amplifier. If you've GOT to have one, build it into the earphone.

For the benefit of the consumer. Unheard of in audio, I know, but...

'Very low' is a simple target, and proposed standards have fallen by the wayside. This way we get a de facto standard, with luck.

 

[mrWagner]

I'm saying that nobody should be including an output resistor in their amplifier.

Unfortunately because of the design of the headphone plug, you short one side with earth when you plug in. Because of safety they have to use some kind of resistance, not to kill the amp. Not always, but there are badly designed plugs/sockets. Or if you use a mono plug (for some reason)

 

[counter culture]

Unfortunately because of the design of the headphone plug, you short one side with earth when you plug in. Because of safety they have to use some kind of resistance, not to kill the amp. Not always, but there are badly designed plugs/sockets. Or if you use a mono plug (for some reason)

You don't have to have a resistor. It's a cheap fix, a poor solution and it's got nothing to do with SQ. Many amplifiers will withstand a continuous SC, the O2, the wire.

 

[mrWagner]

You don't have to have a resistor. It's a cheap fix, a poor solution and it's got nothing to do with SQ. Many amplifiers will withstand a continuous SC, the O2, the wire.

O2 also has resistors in the output, not big, but enough to handle SC.
The wire also has version with fet end. Don't think those like the SC.