Consequence of headphones being driven from a high impedance source?

[mike7877]

Adding just a fraction of an ohm in-line with my speakers changes their voicing noticeably. Not so much the bass, but treble - it screws up the crossover. I have to get up to half an ohm or so before the bass starts sounding obviously bloated.

Headphones, though... On a few occasions I've used the RCA outs from a DAC with a 20 ohm output impedance to drive my 250 ohm headphones. I never noticed anything wrong with how they sounded, and with a damping factor of over 12, most things should sound just fine. Then a situation arose where my 250 ohm'ers weren't around, so I had to use my 37 ohm cans... I was expecting a horrible experience, but no - it worked fine! Better than fine, actually...

So then I got curious and plugged my 37 ohm 1440s into a 50 ohm output...
Everything still sounds perfectly fine! Exactly as if it was driven by something less than 0.1 ohm!

Is there something about the way that DAC outputs operate? Do they somehow take the headphones into account with their negative feedback or something? Or does frequency generally not really affect headphones' impedance to the degree of speakers?


Has anyone else driven their headphones from 20 to 50, even 100 ohms? What was your experience?

edit: afterthought: are DAC output stages in any danger of being pushed into over current? What about the resistor? Is it just there for short term protection (like plug/unplug) and not really drop any voltage because most amps are 4,000-50,000 ohm, or do they size the resistor so that there's no danger to the output unless you put a voltage source there...

 

[DVDdoug]

Its usually not the DAC (output), it's usually the headphones (or speakers) which may have varying impedance over the frequency range.

You have a Voltage Divider where the "top" resistance is the source and the "bottom" resistance is the headphones (or speakers). If the headphones are resistive (the same impedance over the full audio range) you get a voltage loss but it's the same over the frequency range so the frequency response isn't messed-up.

If there are bumps-up in impedance the voltage across the headphones goes up, and where there are dips it goes down. The higher the source impedance (relative to the load impedance) the worse the effect.

DAC output stages in any danger of being pushed into over current?

Of course, you shouldn't connect headphones to a line-output...

I've never heard of a headphone-output burning-up from too-low impedance headphones, but there usually is a minimum load spec.

 

[Curupira]

Mismatching impedances in headphones is not as straightforward as it is in speakers.

To complement the other comment above: the impedance rating you see in speakers, in the vast majority of cases, is also not linear, so they will be very affected during the voltage sweep. Meanwhile, in a lot of headphones like, planars, the resistance is linear across the entire Fr, so a impedance mismatch will be minimal and the inneficiency will be heard in volume rather than in tonal quirkness.

That being said, it's worth noting that impedance mismatch is very much a thing in headphones as well. Take a look at the HD650 being fed over 100 ohm at the HP amp output:

Moreover, the effect is much more dramatic in iems with non linear impedance sweeps, below the Truthear x Crinacle red with its 10ohm resistor:

There are also other infamous instances of significant changes in FR, the most notorious one being the Campfire Andrômeda, the one case of cabling mattering in the final sound: it's impedance is so low that even adding a more resistive cable can make the FR sound different.

 

[oleg87]

edit: afterthought: are DAC output stages in any danger of being pushed into over current? What about the resistor? Is it just there for short term protection (like plug/unplug) and not really drop any voltage because most amps are 4,000-50,000 ohm, or do they size the resistor so that there's no danger to the output unless you put a voltage source there...

No output stage is going to be happy being shorted. A series output resistor is a multi-function cover-your-ass design feature, protecting the output from overcurrent, ESD, destabilizing load capacitances...

 

[solderdude]

So then I got curious and plugged my 37 ohm 1440s into a 50 ohm output...
Everything still sounds perfectly fine! Exactly as if it was driven by something less than 0.1 ohm!

SRH1440 driven from 0.1ohm and 120ohm. note: for 50 ohm the tonal balance will be between the green and red line but closer to the red line than the green line.

Whether or not a headphone will sound audibly different from a higher output R is solely dependent on the impedance plot (impedance per frequency)
Some are very linear (not only most planars) but can be everything up to varying substantially.
Also the average impedance has an influence. The higher the headphone impedance is the higher the output resistance needs to be to change the FR.

Note that most integrated amps (incl AVR's) that use the speaker output through a resistor(network) derived from the speaker out can be somewhere between 100 and 500ohm

edit: afterthought: are DAC output stages in any danger of being pushed into over current? What about the resistor? Is it just there for short term protection (like plug/unplug) and not really drop any voltage because most amps are 4,000-50,000 ohm, or do they size the resistor so that there's no danger to the output unless you put a voltage source there...

You have to differentiate between a DAC and a DAC that is also capable of driving headphones directly.

The ones that can also directly drive headphones are current limited at a much higher current than the ones that only have RCA/XLR and are designed to be connected to an amplifier. The latter is usually (well) above 1kohm in load which all opamps can easily drive.
These usually have output resistors above 10ohm or so and is there to ensure a high capacitive load cannot destabilize the (op)amp circuit at the output. It is not intended as current limiter. Most opamps have effective current limiters inside anyway.
or so.

There are just a handful of 'esoteric' DACs that have unusual high output impedances. These usually have a tube or transformer in the output or lack buffer circuits.

 

[usern]

Is there something about the way that DAC outputs operate?

You mean headphone AMPLIFIER outputs?

 

[sergeauckland]

Although I agree that it's best from an engineering point of view to drive headphones from a very low impedance, I have wondered how much difference it actually makes.

I have measured the three headphones I have here, AKG K270P, Koss Pro4AA and some very cheap Sonys. All have an impedance that varies from the nominal by no more than a few %, consequently, there won't be much variation in frequency response if driven from zero ohms, or a higher impedance. The only variation will be a loss of level when driven from higher impedances.

The graphs above show a difference of only around 1-2dB in frequency response, so barely noticeable on programme material. Hardly the Night And Day differences that are often touted.

Since I made the measurements, I've stopped worrying about driving impedance on headphones. It may be noted that the headphone outputs on most integrated amplifiers was a few hundred ohms in series with the headphone sockets, driven from the main loudspeaker outputs. I don't recall many complaints then.

S.

 

[multiformous]

The response of some IEMs can vary greatly with even modest differences in output impedance. For example, here is my Floaudio Bluelover (6BA IEM) as measured with low OI Apple dongle, with a 30Ω inline adapter, and with my old AVR's headphone output (TX-SR706) which has some obscene amount of OI (and no, the receiver doesn't seem to affect the FR of any single DD IEMs I've tried as they typically have flat impedance curves in my experience).

 

[staticV3]

For the impact of output impedance on IEM response, check out this spreadsheet I made a while ago:

Standards for Audio Electronics Measurements?

This is nothing new, Amir has stated for quite some time now that when a company send him a product, he does them the courtesy of giving them the choice to not publish the review if it's not what they expected. Personally I agree, and hopefully the company will take some steps to remedy the...

www.audiosciencereview.com

 

[mc.god]

A good, bad indeed, example of headphones greatly suffering when driven from high impedance output is Sennheiser HD 598/599.
They have large impedance peak at 100 Hz

just where they already have emphasized response, resulting into a really bloated and muddy experience:

I can confirm that they changed a lot for better when I got a good low impedance output to drive them.

 

[usern]

example of headphones greatly suffering when driven from high impedance output is Sennheiser HD 598/599.

My HD 560S also has different tonality when driving from low output impedance Topping or higher output impedance audio interface.

 

[ZolaIII]

Depends on driver type and specific driver. In case of dynamic drivers as most used ones it will have impact in bass area (higher level) but there are and those (even low impedance one's) where it's so small that it's not meaningful.
A database with regular constant amplifier impedance effects measurements (in a Pro reports among other things including loudness and so on):

https://reference-audio-analyzer.pro/en/catalog-reports.php?sp_1=1&tp=1#gsc.tab=0

Have fun and enjoy!

 

[mike7877]

Of course, you shouldn't connect headphones to a line-output...

Yeah, that's what I was doing lol. Not even to 600 ohm headphones? One of my amplifiers has an input impedance of like 1040 ohms, so I thought why not! lol. I guess amplifiers aren't reactive in the way that traditional drivers can be, is that why? I thought one of the reasons they stick resistors there was for overcurrent protection of the output stage and protection from malfunctioning devices. Also limiting current from ground loops

 

[mike7877]

The response of some IEMs can vary greatly with even modest differences in output impedance. For example, here is my Floaudio Bluelover (6BA IEM) as measured with low OI Apple dongle, with a 30Ω inline adapter, and with my old AVR's headphone output (TX-SR706) which has some obscene amount of OI (and no, the receiver doesn't seem to affect the FR of any single DD IEMs I've tried as they typically have flat impedance curves in my experience).

Good to know - is there anything special about the impedance of those headphones? Or is it just the peak and the increasing resistance with frequancy?

Also (unrelated): how does bass sound through those:

 

[multiformous]

Good to know - is there anything special about the impedance of those headphones? Or is it just the peak and the increasing resistance with frequancy?

Also (unrelated): how does bass sound through those:

Well I don't have a measurement of the actual impedance curve of these particular IEMs. Every set will be different. With this set, it just so happens that adding output impedance gives it a brighter tilt overall. So it's the opposite of something like how the Truthear Zero:RED was designed, where adding the impedance adapter 'increases' the bass (in effect). The impedance curve of Zero:RED can be used to 'predict' the effect of amp output impedance. But impedance graphs aren't always available. Some hobbyists measure their IEMs with different impedance adapters sometimes, though.

As for the Floaudio Bluelover: I just absolutely love the tonality of it, from bass to treble. I don't EQ it. For me, it's perfect. No notes. The only thing is that it's a sealed shell, so the you get the pressure feeling in the ear canal.