[saint.panda]

I've just been talked out of getting a balanced amp but I'm still curious why driving headphone in balanced operation is so often thought to be superior.

Certainly a few of the world's best headphone amps such as the Blockhead or Gilmore Reference drive headphones in balanced operation (headphone outs, not talking about balanced inputs). However, I'm curious whether from a pure technical point of view, balanced operation is better than driving headphone non-balanced - and why. "Slew rate", "shared groud" and "more power" are a few terms that often pop up. What do they mean anyway?

So is it a technical fact that the balanced operation of headphones has advantages, which non-balanced operation cannot overcome under any circumstances? What would these be or is it mainly for psychological reasons?

Same question, slightly different packaging: Assuming the same design (dual mono with separate boards and power supplies), could you build a Blockhead with regular headphone outs that has the same sonic merits of a Blockhead with balanced headphone outs?

It's true that in this special case, the Blockhead also has balanced inputs so does that also play a role here? From reading various post, I am rather convinced that a balanced connection between amp and source doesn't necessarily contribute to a better sound but does a fully balanced system from source to headphones bear certain benefits, which single-ended systems don't?

I am slightly confused in these matters and help would be very much appreciated. Thanks.

 

[wali]

Balanced audio cables are useful when using long runs of low impedance cables especially for microphones. At each end of a balanced cable additional circuitry is needed to perform the noise cancellation function (which means more circuits in the signal path – not desirable, but essential for recording purposes).

Having said that I can only conclude that balanced IC makes sense between two devices with balanced circuits; however, how a balanced cable has anything to do with headphones especially in very short runs is beyond me.

 

[jefemeister]

I also would like to know more about this. Maybe headroom or Gilmore can provide an answer? Is it merely the lack of a common ground point for improved current return?

 

[taylor]

I may be wrong, but this is what I think it is:

Slew rate, afaik, is the time it takes to change the voltage. If your amp has poor slew rate and you are playing a CD where it is very quiet then there's a very loud sound, it will take awhile for the amp to go from outputting nothing to outputting a lot. The faster the slew rate, the better it can adjust to changes in the music.
When running the headphones in balanced, it doubles the slew rate.
I don't know exact figures, but to make it easy say your op-amp or transistor or tube takes 10ms to increase the signal by 1v. If the input jumps from 0v to 5v, it takes 50ms for the amp to compensate.
Now say you take two of these amps and run them in balanced. As the source goes from 0v to 5v, your amps aren't. The red amp is going from 0v to 2.5v, the black amp is going from 0v to -2.5v, so now it only takes 25ms.

Shared Grounds:
Again, not toally certain, but when you are running in balanced, each driver has it's own ground, since the grounds are also connected to amps. I'm not sure what tonal difference it makes, but it's good.

More power:
This also comes from having two amps per channel. If you take an unbalanced 5w amp, it consists of a single 5w amp. If you make it balanced, there are two 5w amps, so twice the power.

*feel free to correct me*

 

[jefemeister]

so the performance benefit would dissapear if you take a standard stereo headamp and merely "fake" the balanced ouput in the output stage? Or would it merely be less dramatic?

 

[rodbac]

From my reading, a "balanced" connection is one in which the same signal is sent over two wires but with opposite polarities (with the knowledge that the two wires will be equally affected by whatever noise is introduced (RFI or EMI/inductive)).

The receiving device (your phones in this case) then must have the circuitry to read the difference between the two signals, eliminating the noise.

So I think taylor and wali are largely correct (it may effectively double the amplitude of the signal- not sure), save for the fact that the headphones (AFAIK) have no idea how to read any difference between the two, making a "balanced cable" useless.

Any EEs here confirm or correct that?

 

[mulveling]

Quote:

Originally Posted by rodbac So I think taylor and wali are largely correct (it may effectively double the amplitude of the signal- not sure), save for the fact that the headphones (AFAIK) have no idea how to read any difference between the two, making a "balanced cable" useless. Any EEs here confirm or correct that?

How could it NOT read the difference? If one end of the driver gets 2V and the other end gets -2V, that's effectively the same as 4V - no problems differentiating that, and no extra circuitry needed.
You sure like to write off anything involving cables as useless

 

[timoteus]

I've posted this quote from Mick Maloney of Supratek once before, but think it is appropriate to do so once more. Make of it what you will.

www.supratek.biz

Quote:

Balanced operation I often get asked about the advantages of balanced operation-my reply is based on the following: Balanced/unbalanced is a conentious subject. It comes from the pro audio world where they need it to balance out the noise pickup from very long cables. True balance involves either using balancing transfomers or turning a Single Ended cicuit (as most preamps are) into a push pull circuit with double the components, circuitry etc- this is the only way to achieve "true" balanced throughout the preamp. There's a lot of hype about it, with a lot of so called balanced operation being psuedo balanced, which is basically having an XLR socket wired to accept a balanced signal and turning it back to unbalanced. Personally I prefer Single Ended operation every time- there's an ease and effortlesss about it, wheras true balanced sounds more electronic to me, especially the push pull types of operation. One area where it is useful is in helping with RF noise in the big cities- it is not an isue for me in my country town isolation, but places like NY benefit from it and in my mind the best way to use balanced is to take it off the output transformer of a preamp, which fortunately we can do with the Supratek pres- it is true and perfect balanced with the minimum of circuitry to degrade the sound. ( there's virtually no circuitry!) One balancing point in a preamp is all that is necessary- the output transformer turns the single ended circuit into a balanced circuit with two phases and the noise common to both channels (hum, RF etc) is greatly reduced. Going for balanced throughout is sonic overkill that does more harm than good IMO- it has been a successful marketing campaign by some companies, but it is marketed to sell amps. The Cortese,Sauvignon, Grange and Cabernet come standard with balanced/unbalanced outputs- there is a significant increase in quietness with the balanced output - but with the effortlessness and spatial presentation retained.

 

[rodbac]

Quote:

Originally Posted by mulveling How could it NOT read the difference? If one end of the driver gets 2V and the other end gets -2V, that's effectively the same as 4V - no problems differentiating that, and no extra circuitry needed. You sure like to write off anything involving cables as useless

Read up on how a "balanced" connection works and what it's intended to do.

It's meant to eliminate noise, not double the signal strength (which I'm not sure is the effect anyway- +2v and -2v adds up to zero volts, not +4v. I'm trying to find some information that discusses it in that context, though).

The signal from both wires needs to be processed so noise can be eliminated. Now, the circuitry to do so may be on the end of the cable itself, but I've never seen this- maybe it does exist. You tell me.

 

[saint.panda]

Great input so far, please keep the ideas coming!

Quote:

Originally Posted by wali Balanced audio cables are useful when using long runs of low impedance cables especially for microphones. At each end of a balanced cable additional circuitry is needed to perform the noise cancellation function (which means more circuits in the signal path – not desirable, but essential for recording purposes). Having said that I can only conclude that balanced IC makes sense between two devices with balanced circuits; however, how a balanced cable has anything to do with headphones especially in very short runs is beyond me.

I was thinking along the same lines regarding balanced connections between amp and source in a home audio environement. It's only added circuitry.

Quote:

Originally Posted by jefemeister Maybe headroom or Gilmore can provide an answer?

That would be great of course.

Quote:

Originally Posted by taylor Slew rate, afaik, is the time it takes to change the voltage. If your amp has poor slew rate and you are playing a CD where it is very quiet then there's a very loud sound, it will take awhile for the amp to go from outputting nothing to outputting a lot. The faster the slew rate, the better it can adjust to changes in the music. When running the headphones in balanced, it doubles the slew rate. I don't know exact figures, but to make it easy say your op-amp or transistor or tube takes 10ms to increase the signal by 1v. If the input jumps from 0v to 5v, it takes 50ms for the amp to compensate. Now say you take two of these amps and run them in balanced. As the source goes from 0v to 5v, your amps aren't. The red amp is going from 0v to 2.5v, the black amp is going from 0v to -2.5v, so now it only takes 25ms. Shared Grounds: Again, not toally certain, but when you are running in balanced, each driver has it's own ground, since the grounds are also connected to amps. I'm not sure what tonal difference it makes, but it's good. More power: This also comes from having two amps per channel. If you take an unbalanced 5w amp, it consists of a single 5w amp. If you make it balanced, there are two 5w amps, so twice the power.

That was very helpful, thank you taylor. Your illustration of power and slew rate make a lot of sense. And you bring up an interesting point, which is that balanced amps have 2 amps per channel making it 4 amps in total. Powerwise this should be of advantage on the paper but I wonder whether 2 amps (dual mono) shouldn't provide sufficient power already. Regarding the slew rate, it would appear to me that doubling the slew rate should lead to more dynamics. Regarding the ground, doesn't a non-balanced design also have discrete grounds?

The question is also whether non-balanced amps can do the same.

Another thing is whether headphones are able to read inverting and non-inverting signals because then the slew rate and voltage swing would indeed be doubled. I found this using the search function:
Quote:

Originally Posted by Mastergill Also, sorry guys, but it's wrong to talk about "balanced headphone". You cannot drive those dynamic phones balanced.

Quote:

Originally Posted by kevin gilmore I completely disagree. As long as the headphones are wired for dual mono you can drive them with a balanced amplifier. Just like many speakers are driven from balanced output amplifiers (krell, pass labs...) Other than the balanced input there is a definite advantage to driving headphones balanced. Twice the slew rate. Balanced rise and fall times which reduce many kinds of distortions due to imbalance in the pnp and npn transistors (or tubes with current sources). Disadvantage to balanced amplifiers is that each side see's half the impedance of the headphones. For tube amplifiers and grado's each side sees only 16 ohms of impedance. Unless you are going to put in monster tubes, many tube amplifiers are going to have a very hard time driving this. Balanced input is something different entirely, and is also highly recommended even if the output is not going to be balanced. Basically for the same reasons. Looking back from the balanced source. If you have serious bucks to spend then buy a pair of the smallest atmasphere amps which are balanced input, balanced output and wire your headphones for dual mono.

Regarding power I found this:
Quote:

Originally Posted by ppl BTW Balancing an Amp will not put out twice the current it will put out half the current with twice the slewrate and 2 times the output voltage resulting in twice to 4 times the power for the same load impedance.

Quote:

Originally Posted by kevin gilmore I disagree. The unbalanced output amp can put out 5 volts rms and do so into a 32 ohm load. The balanced output amp can put out 10 volts rms and also do so into a 32 ohm load. The voltage is double, The current is also double, and the power is 4 times as compared to the unbalanced amp. Think of the balanced amp as 2 unbalanced amps each of which driving half the load i.e. 16 ohms. The slew rate will also be double that of the unbalanced amp.

Quote:

Originally Posted by kevin gilmore quote I would have thought the current remains the same as the two amps and the load are effectively in series. At what voltage? The unbalanced amp can out 5 volts rms into 32 ohms. The balanced amp can put out 10 volts rms into 32 ohms. 5 volts rms into 32 ohms is 156.25 milliamps 10 volts rms into 32 ohms is 312.5 milliamps The unbalanced amp sees all 32 ohms. The balanced amp sees only 16 ohms per side. 5 volts rms into 16 ohms is 312.5 milliamps! Twice the current (at max power) Twice the voltage (at max power) 4 times the power that the headphones see.

Quote:

Originally Posted by ppl Oh! Sorry for the confusion let me restate my previous post. > since we have twice the voltage and the load impedance stays the same then the current will be twice since we have twice the voltage flowing. This i understand. However if the maximum output current safely available was being delivered in the single end mode, then changing the same Amp to the BTL mode would force twice the voltage swing all things being equal. This would in fact require twice the current to flow thought the load in the Instant case this is twice the maximum safe output current. Assuming we were maxed out on current in the single end mode to begin with we will not be able to utilize the increased voltage swing unless our load impedance was increased twice in value causing half as much current to flow for twice the output voltage. The maximum safely available output current is the same for both the Single ended and balanced Amplifier configurations. As Keven pointed out at what voltage! This is the critical factor hear. Ya everything is relative I know.

Quote:

Originally Posted by ppl the increced voltage swing is what is resopnsible for only half the available outout current. If operating below the max output current the current flow will be twice as much as there is twice the voltage swing into the same load impedance. my oversight was not considering what happens below the max output current of the amp.

From a different thread:
Quote:

Originally Posted by morsel Fully balanced headphone amplifiers have 4 channels, require custom 4 wire headphones, and are typically operated in bridged mode, which doubles the output voltage and quadruples the power by using inverting and noninverting pairs, but does not offer symmetrical transfer characteristics due to the difference between inverting and noninverting modes.

I should have done my research beforehand.

 

[mulveling]

Quote:

Originally Posted by rodbac Read up on how a "balanced" connection works and what it's intended to do. It's meant to eliminate noise, not double the signal strength (which I'm not sure is the effect anyway- +2v and -2v adds up to zero volts, not +4v. I'm trying to find some information that discusses it in that context, though).

Read up on arithmetic: 2V - (-2V) = 4V

These aren't sound waves; they don't cancel each other out.
What I imagine is that if you have a noise function X caused by <whatever> that's added into the signal (obviously undesireable), then with balanced operation you will have this hitting the headphone drivers:
(original signal + X) - (inverse signal + X) = original signal - inverse signal = 2 * original signal
Whereas with non-balanced you've just got this hitting the drivers:
original signal + X

Hence the noise reduction in balanced mode? Is this right?

Heh, i didn't read up. I hate reading. I'm just speculating too.

 

[rodbac]

Quote:

Read up on arithmetic: 2V - (-2V) = 4V

Um, who told you or the headphones to use subtraction?

It would be +2v + (-2v) = 0v.

Quote:

These aren't sound waves; they don't cancel each other out.

You're half right- they're electrical signals (not waves at all) and they most certainly will "cancel each other out" with regard to charge.

 

[wali]

The Cardas balanced cable for HD650 uses two XLR connectors but on the driver end it uses the regular unbalanced Sennheiser plugs... Now you don't need to be an EE to understand that you can't freaking transmit a 'balanced' signal using 'two' connector jacks.

All the talk of noise canceling and other stuff is useless here because Sennheiser HD650 has only an unbalanced input (show me one headphone with balanced inputs).

End of discussion for those who believe in common sense.

 

[rodbac]

Quote:

What I imagine is that if you have a noise function X caused by <whatever> that's added into the signal (obviously undesireable), then with balanced operation you will have this hitting the headphone drivers: (original signal + X) - (inverse signal + X) = original signal - inverse signal = 2 * original signal Whereas with non-balanced you've just got this hitting the drivers: original signal + X

Wait...

And I think Wali is right.

 

[mulveling]

Quote:

Originally Posted by rodbac Um, who told you or the headphones to use subtraction? It would be +2v + (-2v) = 0v.

Well, we're measuring the difference in voltage between the postive and negative ends of the voice coil - which is why I decided to use subtraction, lol

Heh, this thread needs salvation in the form of a Kevin Gilmore, or some other kind of miracle.