Why Balanced Headphone Amps?

[sbradley02]

 
I was originally going to post this to the desktop amp forum, but then saw this forum which is probably more appropriate.
 
In reading various posts in this site I have run across references to balanced amplifiers and this prompted to me dig a bit further into these and the claims made for them. Certain aspects make sense. Balanced inputs have proven noise rejection advantages hence their widespread adoption in pro audio. Having separate ground runs for the left and right cans makes a certain amount of sense, removing resistance from the common ground connection would seem to have advantages though whether it is worth the hassle of rewiring is less clear.
 
Inside the amplifier itself things get a lot murkier. If you read the technical description of these units, what you are basically getting is a pair of floating ground amplifiers being driven in push pull for each channel. This design is used commonly in car stereo head units, where it allows for double the power on a 13.8VDC supply without going to the expense and space required for a DC to DC converter. You pretty much don't run into them much elsewhere. 
 
Fully differential power amplifier designs employing bipolar power supplies have been around since the 70s (I have even built some from kits). In these designs, separate sets of devices handle the positive and negative portions of the signal so they are inherently balanced in nature. The promoters of balanced headphone amps (referring to the floating ground configuration with two completely separate amplifier sections for positive and negative) make the claim that using a balanced configuration results in twice the maximum swing and slew rate. While this is true, it comes at an expense of twice the parts count, and the same result can be achieved by simply creating a design with the double the supply rail voltage in the first place.
 
These balanced amps obviously aren't using the topology and space to save money as is done in car stereos. These units range from expensive to extremely expensive. So why are the companies doing it? Anyone seen a valid technical reason for this? I have a suspicion that it is simply a way for a company that already has a single supply, non-differential amplifier design to get more swing without going to a higher voltage true differential design, but maybe there is something I am missing.

 

[Steve Eddy]

 
Quote:

Originally Posted by sbradley02 /img/forum/go_quote.gif
 
Certain aspects make sense. Balanced inputs have proven noise rejection advantages hence their widespread adoption in pro audio. Having separate ground runs for the left and right cans makes a certain amount of sense, removing resistance from the common ground connection would seem to have advantages though whether it is worth the hassle of rewiring is less clear.

 
Be aware that a fair number of "balanced" headphone amps are just a pair of single-ended amplifiers bridged together in push-pull. This configuration doesn't offer any common-mode rejection.
 
Quote:

Having separate ground runs for the left and right cans makes a certain amount of sense, removing resistance from the common ground connection would seem to have advantages though whether it is worth the hassle of rewiring is less clear.

 
The crosstalk from the common ground contact in a TRS plug/jack can be significant. However it can be largely ameliorated without going balanced. Simply use a four pin connector and tie the grounds together internally in star ground fashion.
 
Quote:

These balanced amps obviously aren't using the topology and space to save money as is done in car stereos. These units range from expensive to extremely expensive. So why are the companies doing it?

 
In many cases, it could simply be because there are a lot of people out there who simply prefer the sound of them over their single-ended counterparts. There's really no arguing that. One's preference is what ti is.
 
Quote:

Anyone seen a valid technical reason for this?

 
Guess it depends on how you define a "valid technical reason." Greater slew rate is the most common claim, but I don't consider it terribly meaningful as it's rather trivially easy to design a single-ended amp that doesn't slew limit.
 
se
 
 

 

[roadcykler]

Marketing and subjectivity. 

 

[sbradley02]

 
Quote:

 
Be aware that a fair number of "balanced" headphone amps are just a pair of single-ended amplifiers bridged together in push-pull. This configuration doesn't offer any common-mode rejection.
 

Yes, this is what I found described in the one example of manufacturer literature I was able to locate, and what prompted this post. This is the same arrangement commonly used in car stereo head units and isn't truly balanced at all. Even single ended output amplifier designs commonly use a differential input pair, and the design can be made differential to a greater or lesser extent throughout the circuit (with typically a ground referenced output stage) with resultant significant common mode rejection with a balanced input. It has been quite awhile since I have been active in circuit design, please let me know if I have gotten any of the above wrong.
 
Quote:

The crosstalk from the common ground contact in a TRS plug/jack can be significant. However it can be largely ameliorated without going balanced. Simply use a four pin connector and tie the grounds together internally in star ground fashion.

 

This makes sense, with the same level of modification still required. I am unused to this arrangement outside of electrostatic headphones. Does it offer an advantage in practice? You end up with non-standard headphones and amplifier, but it sounds interesting. I imagine a number of people have tried this out in the community - if anyone has done this (without any other modifications) I would be eager to hear what kind of results you got. Sounds like a possible large gain for relatively small effort.
 
Thanks

 

[kwkarth]

Quote:

 
I was originally going to post this to the desktop amp forum, but then saw this forum which is probably more appropriate.
 
In reading various posts in this site I have run across references to balanced amplifiers and this prompted to me dig a bit further into these and the claims made for them. Certain aspects make sense. Balanced inputs have proven noise rejection advantages hence their widespread adoption in pro audio. Having separate ground runs for the left and right cans makes a certain amount of sense, removing resistance from the common ground connection would seem to have advantages though whether it is worth the hassle of rewiring is less clear.
 
Inside the amplifier itself things get a lot murkier. If you read the technical description of these units, what you are basically getting is a pair of floating ground amplifiers being driven in push pull for each channel. This design is used commonly in car stereo head units, where it allows for double the power on a 13.8VDC supply without going to the expense and space required for a DC to DC converter. You pretty much don't run into them much elsewhere. 
 
Fully differential power amplifier designs employing bipolar power supplies have been around since the 70s (I have even built some from kits). In these designs, separate sets of devices handle the positive and negative portions of the signal so they are inherently balanced in nature. The promoters of balanced headphone amps (referring to the floating ground configuration with two completely separate amplifier sections for positive and negative) make the claim that using a balanced configuration results in twice the maximum swing and slew rate. While this is true, it comes at an expense of twice the parts count, and the same result can be achieved by simply creating a design with the double the supply rail voltage in the first place.
 
These balanced amps obviously aren't using the topology and space to save money as is done in car stereos. These units range from expensive to extremely expensive. So why are the companies doing it? Anyone seen a valid technical reason for this? I have a suspicion that it is simply a way for a company that already has a single supply, non-differential amplifier design to get more swing without going to a higher voltage true differential design, but maybe there is something I am missing.

Someplace around here at least a year or more ago, I posted the genesis of balanced headphone amplifiers.  It was a brilliant business decision, allowing the company create a new product of higher performance from bridging the outputs from amplifier modules already on hand in stock without having to completely redesign.  If I can find it, I will provide a link.
 
EDIT:
I think this is the most recent discussion...
http://www.head-fi.org/forum/thread/497902/schiit-asgard-unboxing-and-first-impressions/675#post_6833387
 
It starts around there and rambles on for a number of pages.  Good luck!

 

[sbradley02]

Thanks. The origin of these units is exactly as I suspected, and you and I are on the same page. These amplifiers should not even be called balanced, using the term as it is understood in the industry, instead they should be referred to as bridged or floating ground and they have no inherent design advantages. In the discussion, Steve Eddy touched briefly on the associated practice of eliminating the common ground on the outputs. I wonder if this could be behind the perceived differences.
 
Quote:

Someplace around here at least a year or more ago, I posted the genesis of balanced headphone amplifiers.  It was a brilliant business decision, allowing the company create a new product of higher performance from bridging the outputs from amplifier modules already on hand in stock without having to completely redesign.  If I can find it, I will provide a link.
 
EDIT:
I think this is the most recent discussion...
http://www.head-fi.org/forum/thread/497902/schiit-asgard-unboxing-and-first-impressions/675#post_6833387
 
It starts around there and rambles on for a number of pages.  Good luck!

 

[MrGreen]

Great post Eddy, you covered all that I considered essential to the topic without going into why things are in a huge amount of depth which is easily researchable.

 

[kwkarth]

Quote:

Thanks. The origin of these units is exactly as I suspected, and you and I are on the same page. These amplifiers should not even be called balanced, using the term as it is understood in the industry, instead they should be referred to as bridged or floating ground and they have no inherent design advantages. In the discussion, Steve Eddy touched briefly on the associated practice of eliminating the common ground on the outputs. I wonder if this could be behind the perceived differences.

Yes, I think this thing took on a life of it's own and pretty soon companies started to realize they could sell more gear at higher profits if they made the gear "balanced."
 
If I'm running 400' of cable from one studio to another, I want a balanced line, but to go from DAC to headamp 1m away, it's a wasted effort, and doubly so if I'm driving an unbalanced device like a speaker or headphone.

 

[nikongod]

you guys do know that there is more to life than just what measures well.
 
On that note, there are large points towards balanced amplifiers.

 

[sbradley02]

This has nothing to do with measurements (and in fact a "balanced" amp should measure very similar to a single-ended amp with higher supply rails) but it does has to do with engineering choices. Does a manufacturer choose to double up their amplifier modules to get more swing (and thereby greatly increase the cost) or to redesign it with higher supply rails in the first place, producing a simpler and similar performing amp at a lower cost? I am arguing that the latter is the preferred approach. There should be a valid technical reason for engineering choices made. I started this thread in the Sound Science forum because I am interested in the design.
 
What are the points towards balanced amplifiers, other than removing the common return on the output, which arguably could make a difference and which could be implemented on any amplifier without much rework?
 
Quote:

you guys do know that there is more to life than just what measures well.
 
On that note, there are large points towards balanced amplifiers.

 

[Yoga Flame]

Let's say I have a DAC with balanced outputs. I could directly feed the balanced outputs into a balanced amp. Or I could convert the balanced output into single ended and then feed that into a stereo amp. This adds another component to the signal path. Many DACs have both single ended and balanced outputs, but their single ended outputs are either just half of the balanced output, or are converted from balanced via opamps.
 
So whether it's:
A. Convert balanced to single ended externally after the DAC,
B. Convert balanced to single ended internally within the DAC, or
C. Just drop half of the balanced signal from the DAC
 
It seems less optimal than just amping the full balanced signal into the headphones.
 
Just a thought. I have never tried a balanced amp myself. Yet.

 

[Uncle Erik]

I've never thought that doubling the parts count was a wise choice. It doubles the points of failure, increases heat and power consumption, and the parts need to be more carefully matched.

Also, most of the noise problems seem to come from ground loops and RFI. There are simpler, more direct and more cost-effective ways to handle those problems than doubling up on an amp. Balanced sources and balanced switching of sources simply adds more expense and more opportunities for things to go wrong.

 

[kwkarth]

Quote:

Let's say I have a DAC with balanced outputs. I could directly feed the balanced outputs into a balanced amp. Or I could convert the balanced output into single ended and then feed that into a stereo amp. This adds another component to the signal path. Many DACs have both single ended and balanced outputs, but their single ended outputs are either just half of the balanced output, or are converted from balanced via opamps.
 
So whether it's:
A. Convert balanced to single ended externally after the DAC,
B. Convert balanced to single ended internally within the DAC, or
C. Just drop half of the balanced signal from the DAC
 
It seems less optimal than just amping the full balanced signal into the headphones.
 
Just a thought. I have never tried a balanced amp myself. Yet.

Since headphones aren't balanced devices anyway, somewhere along the line you're going to have to go to single ended anyway.
 
The DAC I use most has a balanced output and a single ended output.  The balanced output is all SS where the SE output is 6922 driven.  Sonically, I much prefer the latter to the former.
 
I've got a balanced amp I'll sell you.
 

 

[nikongod]

Quote:

This has nothing to do with measurements (and in fact a "balanced" amp should measure very similar to a single-ended amp with higher supply rails) but it does has to do with engineering choices. Does a manufacturer choose to double up their amplifier modules to get more swing (and thereby greatly increase the cost) or to redesign it with higher supply rails in the first place, producing a simpler and similar performing amp at a lower cost? I am arguing that the latter is the preferred approach. There should be a valid technical reason for engineering choices made. I started this thread in the Sound Science forum because I am interested in the design.
 
What are the points towards balanced amplifiers, other than removing the common return on the output, which arguably could make a difference and which could be implemented on any amplifier without much rework?

Id disagree with many parts of that.
 
A balanced or bridged amp will only measure like a single ended (drives the load from both side, very few people use the term properly) if the single ended amp measures very well. If the single ended amp has significant even order distortions the bridged and SE versions will measure very differently.
 
This bit:
"its better to design for more single-ended drive than simply double modules"
 
This brings many problems.
What if this design path forces you into using active devices you would rather not? SS guys like to bag on tube guys for tube rolling, but real SS guys transistor roll. There is more to life than >20db of global feedback. 
 
What if your active devices are somehow limited in max voltage swing? OTL tubes run into grid current VERY soon for example.
 
What if the designer isnt using feedback, and there are easily demonstrable changes to sound that he likes?
 
The list could go on and sadly the call for "no balanced amps" is about as absurd as blindly looking to own one (which I would also argue against, and often do) but at some point there are things that can only be found balanced. 
 
There ARE places and solid reasons to own a balanced piece of gear. Im not really a fan of how most MFR's build their balanced gear for various reasons, but to say that there is no reason to do it beyond marketing hype is just silly.

 

[nikongod]

 
Quote:

Since headphones aren't balanced devices anyway, somewhere along the line you're going to have to go to single ended anyway.
 

The last time I quoted something incorect that you wrote you deleted it. It was tangential to the discussion in the LCD2 thread where we were chatting about helmholtz resonators: you said that the DIAMETER of the tube effected the resonant frequency. 
 
Lets see how it goes this time.
 
Speakers are balanced. Its the cables that are not.