My Theory on Balanced vs Single Ended for High/Low Impedance Headphones
Aug 1, 2012 at 4:42 AM Thread Starter Post #1 of 92

TwoEars

500+ Head-Fier
Joined
Mar 16, 2011
Posts
980
Likes
223
My Theory on Balanced vs Single Ended for High/Low Impedance Headphones ####################################################
 
 
Edit & Disclaimer: This post contains errors that shows that I'm no EE, what follows is a discussion on the topic.
 
 
Basically the idea goes like this:
#####################
 
1) Balanced might improve channel separation and imaging in theory, but in practice the channel separation for singel ended already has a 90dB noise floor or more. So balanced shoudn't really improve imaging relative to SE in practice (balanced might improve imaging but if so it is d/t other factors which we'll discuss later).
 
2) Balanced has a definitive disadvantage in that you need twice the amount of components and the push/pull has to match perfectly timing wise. So there is actually a much greater risk of messing up the imaging with balanced and making it powerful but "fuzzy".
 
3) Since balanced has twice the internal components it is usally better to go for a really good SE amplifier than a mediocre balanced for the same amount of money. This beeing said there will be a big difference between high impedance headphones with well dampending membranes (HD650, T1, HD800) and low impedance headphone (LCD2/3, HifiMan).
 
 
If the headphones have high impedance:
###########################
 
The headphone membranes now act as a resistor on the signal and prevent the electrical current from flowing freely towards ground, the more resistans the slower the respons from the membranes. It will take a beefy amp that doesn't care about resistance to make a pair of HD650, HD800 or T1's come alive.
 
A balanced amp can really help here since instead of just "pushing" on the coil and membrane the amp is now pulling on the headphone coil and membrane as well. This should really help hard to drive headphone come alive and offer better control, deeper bass and faster respons. This is also where the percieved improved imaging comes from - and not the increased channel separation as mentioned in "1)" above.
 
 
If the headphone have low impedance:
##########################
 
The headphone membrane still act as a resistor but to a much lesser extent, what is needed now to drive the headphones is not primarily an amp that is "beefy" (and ignores resistance) but rather one that can deliver lots of current into low loads. A balanced amp is not going to help that much here since amp circuits usually are designed for a certain maximum effienct throughput.
 
And it doesn't help to have two circuits connected in tandem in a push/pull configuration if both are running out of throughput at the same time. Instead what is needed is the ability to push lots and lots of current into low loads without running out of juice. Which is also why the best matches for HE-6 tends to be tube amp that can deliever a lot of current or even solid state speaker amps.
 
 
So in summary:
##########
 
1) I believe that balanced can offer significant sonic improvements to hard to drive headphones with high resistance such as the HD650/HD800/T1.
 
2) I do however not believe that balanced should offer any noticiable (or justifiable) upgrade for low impedance headphones (Hifiman & LCD2-3 in particular). Instead for the same money what you need is the best SE amp you can get that delievers a lot of current and can do it cleanly.
 
 
What do you think? Does it reflect your listening experience? Any electrical engineers / circuit designers among us who can offer technical feedback and/or corrections?
 
Discuss!
smily_headphones1.gif

 
Aug 1, 2012 at 7:43 AM Post #2 of 92
Quote:
Originally Posted by TwoEars /img/forum/go_quote.gif
 
The headphone membranes now act as a resistor on the signal and prevent the electrical current from flowing freely towards ground, the more resistans the slower the respons from the membranes.

Beyerdynamic argues exactly the opposite in their FAQ.
 
 
Quote:
It will take a beefy amp that doesn't care about resistance to make a pair of HD650, HD800 or T1's come alive.

Define 'beefy'. Generally speaking, an amp cares less about high resistance loads than low ones.
 
 
Quote:
A balanced amp can really help here since instead of just "pushing" on the coil and membrane the amp is now pulling on the headphone coil and membrane as well. This should really help hard to drive headphone come alive and offer better control, deeper bass and faster respons. This is also where the percieved improved imaging comes from - and not the increased channel separation as mentioned in "1)" above.

A headphone driver doesn't see +X volts on one side and 0 volts (ground) on the other, it only sees the potential difference. Also, there's no 'pushing' and the output is AC.
 
What offers better control over the headphone driver is low output impedance and such 'bridged' amps seem to have a higher output impedance, ergo less control.
 
 
Quote:
The headphone membrane still act as a resistor but to a much lesser extent, what is needed now to drive the headphones is not primarily an amp that is "beefy" (and ignores resistance) but rather one that can deliver lots of current into low loads.

Whether an amp needs to be 'beefy' or not depends on the sensitivity of the headphones and I'd argue that most headphones need only tiny amounts of current or power.
 
 
Quote:
A balanced amp is not going to help that much here since amp circuits usually are designed for a certain maximum effienct throughput.

I don't understand this, care to elaborate?
 
 
Quote:
Which is also why the best matches for HE-6 tends to be tube amp that can deliever a lot of current or even solid state speaker amps.

Again, I think you've got it backwards.
 
 
Quote:
So in summary:
##########
 
1) I believe that balanced can offer significant sonic improvements to hard to drive headphones with high resistance such as the HD650/HD800/T1.

Define 'hard to drive'. It seems you think high impedance means hard to drive?
 
Aug 1, 2012 at 10:30 AM Post #3 of 92
You seem quite muddled, so as xnor has done a good job of picking through your post I'm just going to throw out some stuff which might clear things up.
 
The power needed to get to a given volume is determined by the sensitivity of the headphones (given as how loud per power, i.e. dB/mW).
How much of this "power" is voltage and how much current may be estimated by ohms law (for recent orthos with flat impedance/frequency this is a very good approximation, for "humpy" relationships it probably isn't that great(?))
 
Historically, tube gear were better voltage sources, solid state gear better current sources. This is less the case these days, most things will drive most other things. Electrostats an obvious exception where tube is still essential.
 
A headphone is only universally "hard to drive" if it's low sensitivity (HE-6, K1000) as it consumes a great deal more power than sensitive headphones (think 20-40 times as much power). This means more voltage and more current, one (or both!) of which might be lacking in a given amp.
 
For higher sensitivity phones, the current/voltage mix depends on the impedance. High impedance needs more volts/less current, low impedance needs less volts/more current.
 
Solid state sources these days might typically do say 2 Vrms, and I was expecting here to write about how high impedance headphones wouldn't get the voltage swing they need, however it seems that for a few headphone specs I just tried in my little excel calculator, a 2 Vrms signal will get them similarly loud (T1=110dB @ 1.95V/3.2mA/600 ohm, LCD2=109dB @ 1.95V/32.4mA/50 ohm, HD700=111dB @ 2V/13.2mA/150 ohm). I'd argue that many solid state sources these days have the power to drive common headphones. Of course as above, 2 Vrms makes far less noise through a pair of HE-6, only 102 dB (40 mA current).
 
A lot of what you've written is very heuristic. I agree with points 1-3. However you should realise it's not just voltage swing which makes a "beefy" amp, but how much current it can deliver, which is largely due to the impedance (low impedance = more current). Speakers are often 8 ohms, and IEMs typically tens of ohms, and these are usually the hardest things to drive, consuming far more power than high impedance drivers (though you don't usually see this in IEMs as they're very sensitive).
 
Measurements taken with a small load will usually be much worse (in distortion etc.) then with a large load. High current = hot hot hot!
 
Aug 1, 2012 at 6:09 PM Post #4 of 92
Quote:
 
 
2) Balanced has a definitive disadvantage in that you need twice the amount of components and the push/pull has to match perfectly timing wise. So there is actually a much greater risk of messing up the imaging with balanced and making it powerful but "fuzzy".
 
3) Since balanced has twice the internal components it is usally better to go for a really good SE amplifier than a mediocre balanced for the same amount of money. This beeing said there will be a big difference between high impedance headphones with well dampending membranes (HD650, T1, HD800) and low impedance headphone (LCD2/3, HifiMan).
 
 
If the headphones have high impedance:
###########################
 
The headphone membranes now act as a resistor on the signal and prevent the electrical current from flowing freely towards ground, the more resistans the slower the respons from the membranes. It will take a beefy amp that doesn't care about resistance to make a pair of HD650, HD800 or T1's come alive.
 
A balanced amp can really help here since instead of just "pushing" on the coil and membrane the amp is now pulling on the headphone coil and membrane as well. This should really help hard to drive headphone come alive and offer better control, deeper bass and faster respons. This is also where the percieved improved imaging comes from - and not the increased channel separation as mentioned in "1)" above.
 
 
If the headphone have low impedance:
##########################
 
The headphone membrane still act as a resistor but to a much lesser extent, what is needed now to drive the headphones is not primarily an amp that is "beefy" (and ignores resistance) but rather one that can deliver lots of current into low loads. A balanced amp is not going to help that much here since amp circuits usually are designed for a certain maximum effienct throughput.
 
And it doesn't help to have two circuits connected in tandem in a push/pull configuration if both are running out of throughput at the same time. Instead what is needed is the ability to push lots and lots of current into low loads without running out of juice. Which is also why the best matches for HE-6 tends to be tube amp that can deliever a lot of current or even solid state speaker amps.
 
 
So in summary:
##########
 
1) I believe that balanced can offer significant sonic improvements to hard to drive headphones with high resistance such as the HD650/HD800/T1.
 
2) I do however not believe that balanced should offer any noticiable (or justifiable) upgrade for low impedance headphones (Hifiman & LCD2-3 in particular). Instead for the same money what you need is the best SE amp you can get that delievers a lot of current and can do it cleanly.
 
 
What do you think? Does it reflect your listening experience? Any electrical engineers / circuit designers among us who can offer technical feedback and/or corrections?
 
Discuss!
smily_headphones1.gif

 
I'm an EE, in few words or less...............
 
You are correct, balanced will have more components, therefore higher cost, more complicated than SE.
And if the two halves don't balance properly, the results may be MORE distortion.
But both the inverting amp and non inverting amp will have the same propagation delay, som timing between the inverting and non-inverting outputs will NOt be scrwed up.
 
High impedance cans vs. low impedance cans:
A good headphone amp will not really care whether it is driving high or low impedance cans, i.e this does not really relate to the speed of the amp.
This is an oversimplification, but it will probably do, in the context of this discussion.
The impedance is usually a combination of resistance, inductance and capacitance.
 
A balanced amp will (all else being equal) have twice as much output impedance as a SE amp. Not a good thing.
 
Try reading this:  http://www.head-fi.org/a/headphone-impedance
 
"Beefy" always relates to Voltage, Current, Power and Impedance.
 
Aug 2, 2012 at 7:27 AM Post #5 of 92
Thanks a lot guys, great to have such knowledgeable people here at head-fi
smile.gif

 
I obvioulsy still have a lot to learn but I will try to read up on the subject a bit.
 
I think it's interesting stuff but how electricity actually works is actually very tricky when you get down to the gritty details!!!
 
For instance I had no idea that headphones respond faster with high impedance (but it kind of makes sense now that I think about it). Or that amps care less about high resistance than a low one (but again it kind of makes sense now that I think about it).
 
May I ask what would happen in the extreme case that an headphone had an "infinite resistance" - there would be no current flowing through but there would be a voltage differance. Would the membrane still move??? My have always been "no" since you need current moving through the coil in order to create a magnetic field - and thus move the membrane. Or will the membrane still move even without current?
 
That's why I thought that higher resistance meant slower current, slower build up of the electrical field and slower moving membranes. But I could be completely wrong here.
 
Might I also ask in general terms if you think there is anything in my general conclusion? That SE would be better for low impedance and balanced would be better for low impedance - or is that complete heuristic BS?
 
Aug 3, 2012 at 8:39 AM Post #6 of 92
Quote:
Thanks a lot guys, great to have such knowledgeable people here at head-fi
smile.gif

 
I obvioulsy still have a lot to learn but I will try to read up on the subject a bit.
 
I think it's interesting stuff but how electricity actually works is actually very tricky when you get down to the gritty details!!!
 
For instance I had no idea that headphones respond faster with high impedance (but it kind of makes sense now that I think about it). Or that amps care less about high resistance than a low one (but again it kind of makes sense now that I think about it).
 
May I ask what would happen in the extreme case that an headphone had an "infinite resistance" - there would be no current flowing through but there would be a voltage differance. Would the membrane still move??? My have always been "no" since you need current moving through the coil in order to create a magnetic field - and thus move the membrane. Or will the membrane still move even without current?
 
That's why I thought that higher resistance meant slower current, slower build up of the electrical field and slower moving membranes. But I could be completely wrong here.
 
Might I also ask in general terms if you think there is anything in my general conclusion? That SE would be better for low impedance and balanced would be better for low impedance - or is that complete heuristic BS?

 
Despite what a lot of folks seem to think, this stuff ain't easy to understand.
 
Simplified version:
 
Headphones do not respond faster electrically because they have a high impedance.
They often respond faster, or more accurately because high impedance drivers are often lighter than low impedance drivers.
It has nothing to do with the actual electrical impedance.
 
Assuming equal electrical efficiency:
Low impedance loads  require a bit less voltage, a bit more current
High impedance loads require a bit MORE voltage, a bit LESS current.
 
If an amp can output enough voltage and current, then (in practice) it does not matter to the amp if the load is low impedance or high impedance.
For example, Bryston's new headphone amp has enough voltage and current output to drive almost any real world headphone.  There are several other makes and models which can do this too.
Infinite impedance: if voltage is not applied to a load, then current does not flow, and if current does not flow:
power =  voltage X currentt.
If no power is drawn from the amp, then no work is done and the diagphram will not move, and the air will not be moved to generate sound waves.
 
Bryston claims their new balanced headphone amp sounds the same whether you use it balanced output or SE output.
 
Cheers,
Chris
biggrin.gif

 
Aug 3, 2012 at 11:47 AM Post #7 of 92
Quote:
 
Despite what a lot of folks seem to think, this stuff ain't easy to understand.
 
Simplified version:
 
Headphones do not respond faster electrically because they have a high impedance.
They often respond faster, or more accurately because high impedance drivers are often lighter than low impedance drivers.
It has nothing to do with the actual electrical impedance.
 
Assuming equal electrical efficiency:
Low impedance loads  require a bit less voltage, a bit more current
High impedance loads require a bit MORE voltage, a bit LESS current.
 
If an amp can output enough voltage and current, then (in practice) it does not matter to the amp if the load is low impedance or high impedance.
For example, Bryston's new headphone amp has enough voltage and current output to drive almost any real world headphone.  There are several other makes and models which can do this too.
Infinite impedance: if voltage is not applied to a load, then current does not flow, and if current does not flow:
power =  voltage X currentt.
If no power is drawn from the amp, then no work is done and the diagphram will not move, and the air will not be moved to generate sound waves.
 
Bryston claims their new balanced headphone amp sounds the same whether you use it balanced output or SE output.
 
Cheers,
Chris
biggrin.gif

 
This is starting to make more sense now.
 
But how about "powerful amps" vs "not so powerful amps" driving the same load, is a more powerful always better? Will more power make the amp/headphone faster and give it a more "powerful sound" for a lack of better term? I know there are issues with heat and noise floors but other than that would you say that more power always is better?
 
And is the only real benefit of balanced drive that you have twice the voltage difference and hence twice the power? (if things about channel separation are ignored)
 
I.e would a 10W SE amp be nearly identical in sound and performance to two 5W amps run in balanced mode (for a total of 10W also) - if you get what I'm trying to say. Total output power is the same but one is SE the other balanced.
 
I guess the short question is: what would an EE buy? Would he buy into balanced for home use?
 
Aug 3, 2012 at 12:39 PM Post #8 of 92
Quote:
 
This is starting to make more sense now.
 
But how about "powerful amps" vs "not so powerful amps" driving the same load, is a more powerful always better? Will more power make the amp/headphone faster and give it a more "powerful sound" for a lack of better term? I know there are issues with heat and noise floors but other than that would you say that more power always is better?
 
And is the only real benefit of balanced drive that you have twice the voltage difference and hence twice the power? (if things about channel separation are ignored)
 
I.e would a 10W SE amp be nearly identical in sound and performance to two 5W amps run in balanced mode (for a total of 10W also) - if you get what I'm trying to say. Total output power is the same but one is SE the other balanced.
 
I guess the short question is: what would an EE buy? Would he buy into balanced for home use?

 
You are better off having too much power (within reason) than not enough power, or just enough power.
 
For me, the ideal headphone amp would (among other things) have enough voltage and current available to drive a headphone of any impedance and any efficiency to any reasonable volume.
 
Twice as much voltage is actually four times as much power, assuming the amp can provide enough current. You could argue that is the only advantage when driving headphones.
 
The whole balanced vs. SE thing is something that has been hotly debated on Head Fi.
You can find a lot of folks who will tell you their headphones sounded much better after rewiring to balanced operation.
Personally don't really see why a headphone or speaker amp with a balanced output would sound better than a SE output amp, unless it just comes down to having more power available.
 
However, Balanced pre-amp to balanced power amp (or balanced source to balanced pre-amp) has some advantages for noise immunity.
Personally, I think it is more important to run balanced low level signal lines for noise immunity.
 
Aug 3, 2012 at 5:27 PM Post #9 of 92
Quote:
I guess the short question is: what would an EE buy? Would he buy into balanced for home use?

 
An EE would tend to answer any question like this with an answer like this:
 
"It Depends"!
LOL!
 
Aug 3, 2012 at 6:08 PM Post #10 of 92
Quote:
 
But how about "powerful amps" vs "not so powerful amps" driving the same load, is a more powerful always better? Will more power make the amp/headphone faster and give it a more "powerful sound" for a lack of better term? I know there are issues with heat and noise floors but other than that would you say that more power always is better?

 
Hopefully I'm not going to bore you with details you already know, but it's important to get this straight.
 
The power delivered by the amplifier depends on the volume you set (and the headphones, specifically its impedance).  More power delivered to the headphones mean they vibrate the airwaves more and you get a louder sound.  Unless you're using insensitive headphones or listen pretty loudly, you're generally not getting over 10mW or so delivered to the headphones.  Yes, a lot of people with amps capable of 3W and so on, are using a tiny fraction of that most of the time.  
 
We've been discussing the limits of headphone amplifiers, which are based on the internal design and components used.  Unless you're actually running up against these limits by operating them at a level higher than they can handle, it doesn't really matter what they are.  Unless you're hitting these limits, this does not determine the sound quality of the amplifier.  Let's say you have 32 ohm headphones, amp A can supply up to 50 mW into 32 ohms, and amp B can supply up to 5 W into 32 ohms.  If your listening only requires 1 mW (enough to blow your socks off with some headphones, seriously), then who cares about the limits?  Based on the design of amps A and B, A could sound more "powerful" (lively? sharp? whatever kind of descriptors you want to use) than B.  Or not.  When most people are talking about a "powerful" sound, they're talking about sonic characteristics that often don't have to do with the amp's limitations.
 
As a sanity check of the output power levels you need, check here:
http://www.innerfidelity.com/headphone-data-sheet-downloads
 
Those pdfs list the output power required to reach 90 dB SPL.  Around 85 dB or so (average, not peak) is starting about the point where health/government agencies think it is starting to be an occupational hazard for long-term exposure.  But if you're listening to music with very quiet and very loud passages, maybe you could need a lot more for the loudest bits.  If you want the headphones to reach 110 dB SPL, for example, then multiply the power level for 90 dB by 100.  e.g. Senn HD 650 needs 0.13 mW for 90 dB SPL, so 13 mW for 110 dB SPL.
 
I think it would be rare for headphones to need balanced drive just for the extra power, unless you listen really loud and are looking at HiFiMAN HE-6 or something like that.  By the way, a FiiO E5/E6 is an example of something that can supply a little over 50 mW into 32 ohms—but less into higher impedances because of the voltage limitation, which is an issue with plenty of popular headphones, however.
 
Balanced has a couple small other advantages and disadvantages, which have been mentioned.
 
Aug 4, 2012 at 2:06 AM Post #11 of 92
Quote:
 
Despite what a lot of folks seem to think, this stuff ain't easy to understand.
 
Simplified version:
 
Headphones do not respond faster electrically because they have a high impedance.
They often respond faster, or more accurately because high impedance drivers are often lighter than low impedance drivers.
It has nothing to do with the actual electrical impedance.
 
Assuming equal electrical efficiency:
Low impedance loads  require a bit less voltage, a bit more current
High impedance loads require a bit MORE voltage, a bit LESS current.
 

Lots of good corrective posts here and it's nice to see those that know what's going on aren't all gaga about balanced when common mode rejection is of little import. Something not yet mentioned is that a balanced amp will have double the output impedance from the same circuit design as a single ended amp.
 
  I may be misinterpreting but I don't get the bolded bit. Why would a high impedance driver be lighter? All equal, they may even be heavier as the coil may have more windings.
 
Aug 4, 2012 at 4:25 AM Post #12 of 92
Quote:
Lots of good corrective posts here and it's nice to see those that know what's going on aren't all gaga about balanced when common mode rejection is of little import. Something not yet mentioned is that a balanced amp will have double the output impedance from the same circuit design as a single ended amp.
 
  I may be misinterpreting but I don't get the bolded bit. Why would a high impedance driver be lighter? All equal, they may even be heavier as the coil may have more windings.

 
Read the posts again, the increased output impedance was already measured.
 
Beyer's guidance is that their high impedance 'phones are lighter and more responsive.
Hmmmmm.......maybe we shouldn't generalize.
If you like the sound of a 'phone, enjoy them and don't agonize over impedance.
Apparently, the higher impedance wire is thinner and therefore the complere winding is lighter.
 
Aug 4, 2012 at 9:02 AM Post #13 of 92
Thanks a lot to everyone for their great comments and explanations!
 
I was reading up on induction a bit here:
 
http://en.wikipedia.org/wiki/Inductor
 
http://hyperphysics.phy-astr.gsu.edu/hbase/magnetic/indcur.html
 
The really interesting piece is this: "The effect of an inductor in a circuit is to oppose changes in current through it by developing a voltage across it proportional to the rate of change of the current."
 
Basically the the faster the rate of change of current in the wire is the more resistance will be seen.
 
Could balanced drive have any advantages here? That they are less sensitive to resistance imposed by induction in the drive coil since they have a knack for delivering twice the voltage of a SE circuit?
 
Aug 4, 2012 at 10:43 AM Post #14 of 92
Quote:
 
Read the posts again, the increased output impedance was already measured.
 
Beyer's guidance is that their high impedance 'phones are lighter and more responsive.
Hmmmmm.......maybe we shouldn't generalize.
If you like the sound of a 'phone, enjoy them and don't agonize over impedance.
Apparently, the higher impedance wire is thinner and therefore the complere winding is lighter.

Woops, I now see where you stated that. I'm 100% on board with your views on balanced and even question it's need in low level portions for home use with shielding and short runs. I've seen some interestng things done with PSs and DACs for common mode rejection of artifacts but they've never been my favorites.
 
 The question could asked. Why not use the thinner wire on the low impedance version? Heat from pulling more current?
 
Anyway, thanks for a 'balanced' view on this.
bigsmile_face.gif

 
Aug 4, 2012 at 10:50 AM Post #15 of 92
Quote:
Originally Posted by goodvibes /img/forum/go_quote.gif
 
The question could asked. Why not use the thinner wire on the low impedance version? Heat from pulling more current?

 
Because thinner wire results in higher resistance and you need a certain number of windings.
 

Users who are viewing this thread

Back
Top