Headphone & Amp Impedance Questions? Find the answers here!

[KeithPhantom]

I always read that voltage is what matters with high impedance headphones?

Both current and potential are important for high and low impedance headphones. You also have to factor out sensitivity/efficiency (related but not the same). High impedance headphones are less sensitive to imperfection in the amplifying and source stages, also are less of a load to an amplifier (it is easier to supply voltage than current). Actually, low impedance headphones are more of a load to an amplifier, but it happens that portable gear has really high efficiency/sensitivity values which let them reach higher volume at a single mW (and being low sensitivity they don't need as much voltage). I also have to add that you need way more power to power lower sensitivity headphones, almost regardless of the impedance presented.

 

[dazzerfong]

High impedance, high sensitivity: low voltage, low current
High impedance, low sensitivity: high voltage, low current
Low impedance, high sensitivity: low voltage, high current
Low impedance, low sensitivity: high voltage, high current

Power requirements are determined from your sensitivity/efficiency. The distribution between voltage and current is determined by your impedance. That's why whenever people say high impedance headphones need more power, I roll my eyes.

 

[Chris Kaoss]

Yes.
But it's like computer specs, there's a "minimum requirements", which drives them "well", and "recommended ones", which will have a small impact.

 

[KeithPhantom]

"recommended ones", which will have a small impact.

If you're able to reproduce the highest peaks of your chosen SPL within a power value, that's enough power for the transducer.

 

[jotaku]

OK so maybe I haven't spelunk' d the forums enough (but I think this is the right thread for this discussion this time), but I have never found an answer that even remotely satisfied me concerning this topic.

So I have a degree in electrical engineering. One of the classes I took was about transmission lines and the importance of matching your output and your load impedances. The grossly simplified version being the following:

• When you match impedance from your source (amp; ZS) and your load (headphone/speaker; ZL) you are just fine. (According to audio literature, the reactive part of impedance is ignored i.e. the inductive/capacitive part of impedance. So we only care about the resistance i.e. Ohms. Don't know how true this is. I have my doubts.)
• When you have an impedance mismatch you have signal reflections in your transmission lines that cause destructive interference.

I was wondering why no one talks about this when talking about equipment synergy. Is it a non-factor? Does Dampening Factor matter more? (DF = ZL/ZS) Or is it a large part of it and people don't realize it?

It seems like when people talk about amps that work well with certain headphones, upon further research these amps have an output impedance closer to the impedance of the those headphones (less reflections/destructive interference). Or maybe it is the result of some magic DF ratio, but I didn't look at that closely.

As a slight tangent, signal reflections affect certain frequencies differently; supporting the claim that certain amp/headphone pairings improve certain frequency responses, like improved bass response or extension.

Am I missing something fundamental here? Oversimplifying things?

Just wanting to get my head screwed on straight.

 

[castleofargh]

OK so maybe I haven't spelunk' d the forums enough (but I think this is the right thread for this discussion this time), but I have never found an answer that even remotely satisfied me concerning this topic.

So I have a degree in electrical engineering. One of the classes I took was about transmission lines and the importance of matching your output and your load impedances. The grossly simplified version being the following:

• When you match impedance from your source (amp; ZS) and your load (headphone/speaker; ZL) you are just fine. (According to audio literature, the reactive part of impedance is ignored i.e. the inductive/capacitive part of impedance. So we only care about the resistance i.e. Ohms. Don't know how true this is. I have my doubts.)
• When you have an impedance mismatch you have signal reflections in your transmission lines that cause destructive interference.

I was wondering why no one talks about this when talking about equipment synergy. Is it a non-factor? Does Dampening Factor matter more? (DF = ZL/ZS) Or is it a large part of it and people don't realize it?

It seems like when people talk about amps that work well with certain headphones, upon further research these amps have an output impedance closer to the impedance of the those headphones (less reflections/destructive interference). Or maybe it is the result of some magic DF ratio, but I didn't look at that closely.

As a slight tangent, signal reflections affect certain frequencies differently; supporting the claim that certain amp/headphone pairings improve certain frequency responses, like improved bass response or extension.

Am I missing something fundamental here? Oversimplifying things?

Just wanting to get my head screwed on straight.

I'm going to do the captain obvious part. With impedance bridging design we don't need a different amplifier anytime we use a headphone with a different impedance.

About transmission line. We're dealing with short cables and most of all, audio frequencies. Digital interconnects go at higher frequencies and do use impedance matching(even for the cables). It's just not necessary at audio freqs. As far as I understand that's really the key point here.

About your potential correlation between sound preference and impedance matching. The scenario of 2 amps with vastly different impedance outputs, usually carries many variables changing along with the amps. In practice, maybe one amp would be 0.3ohm SS amp with a ton of negative feedback, while the other amp would most likely be some old tube amp design if we're looking for 100ohm or more. That makes observation not so conclusive about one single cause. And that's assuming the listeners even bothered to properly volume match the amps, and used blind testing. As that almost never happens, the existing testimonies should be taken with a grain of salt.

But there is no doubt that the impedance change at the amp could lead to audible changes with some headphones. Starting with good old frequency response. We can expect the amplitude in the subs to change a little. Also the typical impedance bump at the main resonant frequency of the transducer, would boost that freq with the high impedance amp. Then is the total damping of the transducer that might perhaps be audible in some cases, but I haven't seen much in term of demonstrating an audible impact on headphones. I've seen something regarding fairly massive sub-woofers once, but it's likely that the impact is much smaller for tiny very light transducers with an already high mechanical damping. Whereas the frequency response change from change in the impedance ratio, is something that will easily be audible as soon as it reaches big enough amplitude. And that impact has been confirmed several times. You just have to add resistors in series between the amp and the headphone so get frequency response change.
So until I see supporting evidence for the rest, I'm tempted to assume that FR variations, and just 2 amps being 2 different amps, are the lead causes for perceived changes. That and of course, audiophiles hearing a lot of stuff with their eyes and wallet.

 

[71 dB]

So I have a degree in electrical engineering. One of the classes I took was about transmission lines and the importance of matching your output and your load impedances. The grossly simplified version being the following:

• When you match impedance from your source (amp; ZS) and your load (headphone/speaker; ZL) you are just fine. (According to audio literature, the reactive part of impedance is ignored i.e. the inductive/capacitive part of impedance. So we only care about the resistance i.e. Ohms. Don't know how true this is. I have my doubts.)
• When you have an impedance mismatch you have signal reflections in your transmission lines that cause destructive interference.

I was wondering why no one talks about this when talking about equipment synergy. Is it a non-factor? Does Dampening Factor matter more? (DF = ZL/ZS) Or is it a large part of it and people don't realize it?

It seems like when people talk about amps that work well with certain headphones, upon further research these amps have an output impedance closer to the impedance of the those headphones (less reflections/destructive interference). Or maybe it is the result of some magic DF ratio, but I didn't look at that closely.

As a slight tangent, signal reflections affect certain frequencies differently; supporting the claim that certain amp/headphone pairings improve certain frequency responses, like improved bass response or extension.

Am I missing something fundamental here? Oversimplifying things?

Just wanting to get my head screwed on straight.

Impedance matching in transmission lines applies only when the transmission line is so long compared to the wavelength of the signal, that the voltage level at the ends of that transmission line differ from each other. At audio frequencies this means the transmission lines (signal speed about 2/3 of the light speed or 200.000 km/s) must be kilometers/miles long. This is why audio cables don't require (thank God!) impedance matching. The frequencies are low enough. The reflections are nearly identical with the signal and destructive interference doesn't really happen. At 20 kHz, just one full cycle of electrical signal in an audio cable would be 10 km (6 miles) long! Having cable length of 1 % of one cycle makes the reflections are nearly identical. That's 100 m (~300 feet) of cable and that's for 20 kHz, which older people can't even hear!

Damping factor does matter, because it affects sound quality. The higher damping factor the more faithful the resulting sound is to the original signal, but one can of course reduce damping factor by increasing source resistance to have less faithful, but perhaps more relaxed sound.

Some headphones are mechanically damped so much they don't really need electrical damping meaning the damping factor can be poor. Typically these headphones also have very flat impedance curve (impedance is almost the same on all frequencies). This means high source impedance won't affect the frequency responce much. Some other headphones have very little mechanical damping and need electrical damping (low source impedance). Typically this kind of headphones have a very curvy impedance curve and a huge bump on the resonance frequency of the driver (around 100 Hz or so). If these headphones are driven from sources with high output impedance the result is not only inadequate damping, but even more seriously a poor frequency response with a bump on the driver element resonance frequency.

 

[jotaku]

Impedance matching in transmission lines applies only when the transmission line is so long compared to the wavelength of the signal, that the voltage level at the ends of that transmission line differ from each other. At audio frequencies this means the transmission lines (signal speed about 2/3 of the light speed or 200.000 km/s) must be kilometers/miles long. This is why audio cables don't require (thank God!) impedance matching. The frequencies are low enough. The reflections are nearly identical with the signal and destructive interference doesn't really happen. At 20 kHz, just one full cycle of electrical signal in an audio cable would be 10 km (6 miles) long! Having cable length of 1 % of one cycle makes the reflections are nearly identical. That's 100 m (~300 feet) of cable and that's for 20 kHz, which older people can't even hear!

Damping factor does matter, because it affects sound quality. The higher damping factor the more faithful the resulting sound is to the original signal, but one can of course reduce damping factor by increasing source resistance to have less faithful, but perhaps more relaxed sound.

Some headphones are mechanically damped so much they don't really need electrical damping meaning the damping factor can be poor. Typically these headphones also have very flat impedance curve (impedance is almost the same on all frequencies). This means high source impedance won't affect the frequency responce much. Some other headphones have very little mechanical damping and need electrical damping (low source impedance). Typically this kind of headphones have a very curvy impedance curve and a huge bump on the resonance frequency of the driver (around 100 Hz or so). If these headphones are driven from sources with high output impedance the result is not only inadequate damping, but even more seriously a poor frequency response with a bump on the driver element resonance frequency.

Thanks everyone for contributing to answer my question. I just found this article from sound on sound that makes a lot of sense coupled with everything people have said.
https://www.soundonsound.com/techniques/understanding-impedance

 

[redrol]

I read that a general rule of thumb is the driver impedance should be around 8x the amplifier impedance. Sounds reasonable from what I know. Is this a decent idea?

 

[71 dB]

I read that a general rule of thumb is the driver impedance should be around 8x the amplifier impedance. Sounds reasonable from what I know. Is this a decent idea?

Actually this is not a very good rule of thumb, because it all comes down to how much mechanical damping the driver has got on its own. Drivers with a lot of mechanical damping work well even when the driver impedance is less than 8x the amp output impedance while some headphones have for sensitivity reasons so little mechanical damping that even 8x ratio of impedances isn't enough. I think in the past headphones where less sensitive in general (not designed to work with portable devices) and they had generally enough mechanical damping for the 8x rule to make sense, but the World has changed and headphones of today can be quite sensitive (have very little mechanical damping meaning they require electrical damping which means very low amp output impedance). For this reason the 8x rule should be updated and I suggest to make it a 20x rule if we insist on having one. However, I think a better way to deal with this amp output impedance issue it to divide headphones into two categories:

1. High mechanical damping models (measured impedance curve is very flat) => Amp output impedance can be pretty high.
2. Low mechanical damping models (measured impedance curve is not flat) => Amp output impedance should be very low.

Even when mechanical damping is high, the voice coil inductance causes the impedance to rise at highest frequences. In order to avoid frequency response error bigger than 0.5 dB, the output impedance Zout of the amp should be at most:

Zout ≤ Zmax × Zmin × (𝛃 - 1) ÷ (Zmax - 𝛃 × Zmin),

where Zmax and Zmin are the maximum and minimum impedance values and 𝛃 = 10^(0.5/20) ≈ 1.06 is defined by the maximum allowed frequency response error. If we plot the maximum allowed amp output impedance as a function of Zmax/Zmin ratio, we get a graph like this:

 

[redrol]

OH wow. I dunno why I never get notifications for these threads. THANK YOU @71 dB

I fully understand now.

 

[71 dB]

OH wow. I dunno why I never get notifications for these threads. THANK YOU @71 dB

I fully understand now.

You are welcome @redrol
I'm glad I could help.

 

[redrol]

Yeah, I am now over here doing the maths on my expensive low impedance IEMs with wacky impedance curves to match. Much appreicated!! I am telling all my buddies because this is more important than DACs and other stupid things people spend money on.

 

[Chris Kaoss]

Actually this is not a very good rule of thumb, because it all comes down to how much mechanical damping the driver has got on its own. Drivers with a lot of mechanical damping work well even when the driver impedance is less than 8x the amp output impedance while some headphones have for sensitivity reasons so little mechanical damping that even 8x ratio of impedances isn't enough. I think in the past headphones where less sensitive in general (not designed to work with portable devices) and they had generally enough mechanical damping for the 8x rule to make sense, but the World has changed and headphones of today can be quite sensitive (have very little mechanical damping meaning they require electrical damping which means very low amp output impedance). For this reason the 8x rule should be updated and I suggest to make it a 20x rule if we insist on having one. However, I think a better way to deal with this amp output impedance issue it to divide headphones into two categories:

1. High mechanical damping models (measured impedance curve is very flat) => Amp output impedance can be pretty high.
2. Low mechanical damping models (measured impedance curve is not flat) => Amp output impedance should be very low.

Even when mechanical damping is high, the voice coil inductance causes the impedance to rise at highest frequences. In order to avoid frequency response error bigger than 0.5 dB, the output impedance Zout of the amp should be at most:

Zout ≤ Zmax × Zmin × (𝛃 - 1) ÷ (Zmax - 𝛃 × Zmin),

where Zmax and Zmin are the maximum and minimum impedance values and 𝛃 = 10^(0.5/20) ≈ 1.06 is defined by the maximum allowed frequency response error. If we plot the maximum allowed amp output impedance as a function of Zmax/Zmin ratio, we get a graph like this:

Thanks for heading up with and i do agree with.

Nonetheless, i'm struggling to apply this to the almost flat imp line (curve) of a planar.

Did someone encounter an unusual behaviour with a planar on a higher imp out amp?
Regarding to my Era-1, which is flat on 21 ohm from bottom to top.

Sadly, i don't have an amp with higher imp out to test it by myself.

Thanks guys.

 

[71 dB]

Thanks for heading up with and i do agree with.

Nonetheless, i'm struggling to apply this to the almost flat imp line (curve) of a planar.

Did someone encounter an unusual behaviour with a planar on a higher imp out amp?
Regarding to my Era-1, which is flat on 21 ohm from bottom to top.

Sadly, i don't have an amp with higher imp out to test it by myself.

Thanks guys.

If you mean by struggling to apply this using the equation for the largest allowed output impedance then one needs to notice that it doesn't "work", if the variation of impedance is less than ~6 %. It will give NEGATIVE output impedance for the amp. This is not a problem at all. All it means is the headphone impedance curve is so flat it doesn't matter how high the amp output impedance is, the frequency response error will always be less than 0.5 dB. The only problem with high amp output impedance in these cases such as your 21 Ω cans is that only a small portion of the amp voltage is over the headphone, for example if the amp output impedance was 100 Ω, only ~17 %* of the amp voltage would be over the headphone. This means almost 12 dB attenuation** compared to an amp with 10 Ω output impedance given same amp voltage.

Hopefully this helps in your struggles.
________________________________
* 21 Ω ÷ (100 Ω + 21 Ω) = 0.173355... ≈ 17 %
** 20 × log10( (21 Ω + 10 Ω) ÷ (21 Ω + 100 Ω) ) = 20 × log10(31 ÷ 121) ≈ 11.8 dB