A question about true balanced differential design signal paths.

[Chris J]

The thing to understand here is that "balanced" is an overloaded term, such that it has two different (but similar) meanings that are often confused or mixed.

Balanced in pro audio and signal transmission means that there are 3 lines: positive, inverted, and ground.  This allows the positive signal to be added to the inverse of the inverted signal to cancel out line noise.  It also implies, but does not require, a +4 standard for signal level, which gives an extra 12 or so decibels of audio over the consumer unbalanced standard, which should result in a better signal to background noise ratio for the same content.

Balanced in headphone wiring means removing the common ground between the two channels, and replacing it with separate grounds that carry the inverted signal.  This is believed to decrease channel crosstalk, and therefore increase stereo separation.  Whether that belief is factual is dependent on the two signals being compared.  One thing that is generally true is that balanced amplifiers use high end designs, and generally sound really good, though it is arguable whether any of this goodness is actually due to having balanced outputs.

(My apologies to people in this thread who already know this stuff.  I thought having a clarification may be useful for some.)

Balanced is the same signal in pro audio as it is when driving a headphone: there is a non-inverted and inverted signal, both referenced to ground.

However, in proaudio when you are sending a balanced signal to a balanced input, At the receiving end, the inverted signal is subtracted from the non-inverted signal.
This creates a signal which is 6 dB higher than the non-inverted signal alone, in addition since you are subtracting the inverted from the non-inverted, then you are also rejecting the noise picked up by the two signal leads.

In headphones, the headphones themselves are not rejecting noise, they are just receiving a 6 dB hotter signal. You could argue thus actually does nothing to improve the signal but create twice as much distortion and double the output impedance of the head amp.

 

[Steve Eddy]

Balanced in pro audio and signal transmission means that there are 3 lines: positive, inverted, and ground.  This allows the positive signal to be added to the inverse of the inverted signal to cancel out line noise.

That's one of the most common misconceptions about balanced interfaces. The rejection of common mode noise has absolutely nothing to do with signal polarity. Keep in mind that a balanced interface will reject common mode noise even when there is no signal.

Balanced interfaces rely on the balance of imedances between each line and ground for the rejection of common mode noise. Also, there's no requirement of a ground connection. A balanced interface only needs two wires. And the current standard for balanced cabling uses pin 1 as shield, not ground.

http://pin1problem.com/

se

 

[akiroz]

Quote:

Balanced is the same signal in pro audio as it is when driving a headphone: there is a non-inverted and inverted signal, both referenced to ground.

However, in proaudio when you are sending a balanced signal to a balanced input, At the receiving end, the inverted signal is subtracted from the non-inverted signal.
This creates a signal which is 6 dB higher than the non-inverted signal alone, in addition since you are subtracting the inverted from the non-inverted, then you are also rejecting the noise picked up by the two signal leads.

In headphones, the headphones themselves are not rejecting noise, they are just receiving a 6 dB hotter signal. You could argue thus actually does nothing to improve the signal but create twice as much distortion and double the output impedance of the head amp.

 
Yep, that's called common-mode rejection.
 
I guess balanced driving would have the advantage of more output power when you're using a very low voltage supply, since the peak output voltage is doubled. It also eliminates DC offset, power-on/off transient and the need for a "fake" ground for single supply systems. So it may have slight advantages in portable battery powered systems? 
 
But yeah, each amp would appear to be driving a load half the impedance of the headphone's rated impedance which would increase distortion.

 

[akiroz]

Quote:

That's one of the most common misconceptions about balanced interfaces. The rejection of common mode noise has absolutely nothing to do with signal polarity. Keep in mind that a balanced interface will reject common mode noise even when there is no signal.

Balanced interfaces rely on the balance of imedances between each line and ground for the rejection of common mode noise. Also, there's no requirement of a ground connection. A balanced interface only needs two wires. And the current standard for balanced cabling uses pin 1 as shield, not ground.

http://pin1problem.com/

se

The ground is used as voltage reference. Without the ground, the signal would electrically be "floating".

 

[Chris J]

That's one of the most common misconceptions about balanced interfaces. The rejection of common mode noise has absolutely nothing to do with signal polarity. Keep in mind that a balanced interface will reject common mode noise even when there is no signal.
Balanced interfaces rely on the balance of imedances between each line and ground for the rejection of common mode noise. Also, there's no requirement of a ground connection. A balanced interface only needs two wires. And the current standard for balanced cabling uses pin 1 as shield, not ground.
http://pin1problem.com/
se

Yep, that's it.
The only thing the differential line driver really does is create a balanced output impedance.
Having said that, you may as well just put a non - inverted signal on one line and nothing on the other line. Or put a non- inverted signal on one line and an inverted signal on the other line. It doesn't matter either way.
It is also important to use a balanced cable, i.e balanced WRT ground.
The third thing you need is a balanced (or differential) receiver at the other end of the line to reject the induced noise, i.e. the noise picked up by the balanced cable.

 

[Chris J]

Yep, that's called common-mode rejection.

I guess balanced driving would have the advantage of more output power when you're using a very low voltage supply, since the peak output voltage is doubled. It also eliminates DC offset, power-on/off transient and the need for a "fake" ground for single supply systems. So it may have slight advantages in portable battery powered systems? 

But yeah, each amp would appear to be driving a load half the impedance of the headphone's rated impedance which would increase distortion.

you can still get DC offset from a balanced head amp.
You can still get power on/off transients
but you do get twice as much output voltage from the same power supply.

You get increased distortion because you are driving the phones with two outputs, instead of one output and ground.

 

[akiroz]

Quote:

you can still get DC offset from a balanced head amp.
You can still get power on/off transients
but you do get twice as much output voltage from the same power supply.

You get increased distortion because you are driving the phones with two outputs, instead of one output and ground.

 
True, but it will be significantly less than single ended driving.
Given a theoretical situation where the 2 amps driving the transducer is 100% identical, then there will be no DC offset or power on/off transients.
 
When you look at the "fake ground" that some portable system use, it's basically a third channel with the input connected to a half supply voltage reference so you will still get the distortion.

 

[Chris J]

True, but it will be significantly less than single ended driving.
Given a theoretical situation where the 2 amps driving the transducer is 100% identical, then there will be no DC offset or power on/off transients.

When you look at the "fake ground" that some portable system use, it's basically a third channel with the input connected to a half supply voltage reference so you will still get the distortion.

I don't really disagree!
I'm just pointing out that now you need two identical amps! LOL!

As for the fake ground, may as well add a fourth channel and create a pair of balanced line drivers instead of a stupid fake ground.

 

[Steve Eddy]

Quote:

The ground is used as voltage reference. Without the ground, the signal would electrically be "floating".

 
So?
 
se

 

[akiroz]

Quote:

Quote:

The ground is used as voltage reference. Without the ground, the signal would electrically be "floating".

 
So?
 
se

 
I actually do not know. LOL
With my current knowledge in this field I don't see a problem with not having a ground reference but It's just common practice to have one.
 
btw, what's with the "se"

 

[Chris J]

So?

se

Hmmmm, I think I'll try and find some pictures of Pippa Middleton's butt on the internet, there's gotta be some on the Internet somewhere.........

 

[Steve Eddy]

Quote:

 
I actually do not know. LOL
With my current knowledge in this field I don't see a problem with not having a ground reference but It's just common practice to have one.

 
Having a ground reference is the single biggest source of noise. Read the Pin 1 problem link I posted. Pin 1 should be shield, not ground. It's those who ground pin 1 that create problems for others.
 
Quote:

btw, what's with the "se"

 
My initials.
 
se

 

[Chris J]

I actually do not know. LOL
With my current knowledge in this field I don't see a problem with not having a ground reference but It's just common practice to have one.

btw, what's with the "se"

The differential line driver and differential line receiver should both be grounded at the same potential, but not via the balanced signal cable.

Couldn't find any good pictures on the 'net.......

 

[akiroz]

Quote:

 
Having a ground reference is the single biggest source of noise. Read the Pin 1 problem link I posted. Pin 1 should be shield, not ground. It's those who ground pin 1 that create problems for others.
 

\Thanks for the link.
 
The signal still has to be grounded, but a good grounding trace or plane is needed. So they use the chassis (earth ground) instead of the signal ground.
My question now is if grounding the signal with the "earth" ground create ground loop problems?

 

[Steve Eddy]

Quote:

 
The signal still has to be grounded, but a good grounding trace or plane is needed. So they use the chassis (earth ground) instead of the signal ground.
 

 
You only need a local ground reference. And if course if you're using transformers, not even that. The interface is completely isolated.
 
Quote:

My question now is if grounding the signal with the "earth" ground create ground loop problems?

 
That's the single most common cause of ground loop problems.
 
se