[hagtech]

But if the two terminal source is already balanced ... The balance is already there

Yes, but it is easily destroyed.  As soon as you connect that source (say, a cart) to a receiver, everything goes out the window.  The receiver ends up defining the mode of operation, whether single-ended or balanced.

Consider: The secondary of an output transformer and the primary of an input transformer.

Excellent example, Steve.  This situation works best when only one of the center taps is grounded.  That prevents ground loops via power cords, etc.  In your example, you connect center taps, which is fine.  You are correct that no current will flow in the center conductor.  It can be disconnected and you still preserve the balance, no change in performance.  However, this again ignores the big picture.  What did we miss?  That damn third terminal again.  It is still there, although in a minimal fashion.  The connection is via leakage.  The insulation in the transformers and wiring is imperfect.  But it exists.  It may be 1000M ohm from primary to secondary, or to ground, yet it is there.  It keeps the common mode voltage of the system in check.  Sometimes.  If the insulation is really good, then the cable can pick up static charge over time.  Without any common mode connection, the coils float, and are at the whim of external corruptive influences.  It is possible that the static charge build-up is so great that the voltage difference ends up breaking down the insulation, causing an arc.  That is, unfortunately, the real world we live in.  In this example, the leakage path is actually needed to keep the common mode voltage within reasonable limits. 

Another good example is the telephone system.  It appears at first glance to be a two-wire system.  Tip and Ring.  But that is only with blinders on.  In the big picture, there is an Earth connection too.  No currents are supposed to flow through Earth, but it does provide a handy common mode reference.  It is only connected at the CO.  Now if the balance between Tip and Earth and Ring and Earth gets unbalanced, guess what?  We get currents through the Earth, which is not good.  Hence, maintaining balance is again crucial for performance.

A Simple Alternative

This is a clever connection that accommodates a 3-wire system.  But it's not really a balanced transmitter because there is no signal on pin 3.  Unless!  Unless you receive with a center-tapped transformer, which then back drives the cable so that the signals are indeed balanced.  That is why this is so clever.

jh

[Steve Eddy]

Yes, but it is easily destroyed.  As soon as you connect that source (say, a cart) to a receiver, everything goes out the window.  The receiver ends up defining the mode of operation, whether single-ended or balanced.

I don't see how the input of the receiver changes the symmetry of the coil and therefore its balance or the behavior of the coil for that matter.

Excellent example, Steve.  This situation works best when only one of the center taps is grounded.  That prevents ground loops via power cords, etc.

There's no need to ground any of the center taps for that. The transformers by their nature provide galvanic isolation between the two components and therefore break any ground loops that would be there otherwise.

In your example, you connect center taps, which is fine.  You are correct that no current will flow in the center conductor.  It can be disconnected and you still preserve the balance, no change in performance.  However, this again ignores the big picture.  What did we miss?  That damn third terminal again.  It is still there, although in a minimal fashion.  The connection is via leakage.  The insulation in the transformers and wiring is imperfect.  But it exists.  It may be 1000M ohm from primary to secondary, or to ground, yet it is there.  It keeps the common mode voltage of the system in check.  Sometimes.  If the insulation is really good, then the cable can pick up static charge over time.  Without any common mode connection, the coils float, and are at the whim of external corruptive influences.  It is possible that the static charge build-up is so great that the voltage difference ends up breaking down the insulation, causing an arc.  That is, unfortunately, the real world we live in.  In this example, the leakage path is actually needed to keep the common mode voltage within reasonable limits.

I dunno, Jim. You seem to really be stretching here to keep that third terminal in there.

Let's say we do connect the two center taps. How does that amoeliorate any static charge (assuming there's one to begin with)? The static charge would still be there. Where else can it go?

Another good example is the telephone system.  It appears at first glance to be a two-wire system.  Tip and Ring.  But that is only with blinders on.  In the big picture, there is an Earth connection too.  No currents are supposed to flow through Earth, but it does provide a handy common mode reference.  It is only connected at the CO.  Now if the balance between Tip and Earth and Ring and Earth gets unbalanced, guess what?  We get currents through the Earth, which is not good.  Hence, maintaining balance is again crucial for performance.

Not sure how you're defining "Earth" here. Do you mean a literal connection to the literal earth?

This is a clever connection that accommodates a 3-wire system.  But it's not really a balanced transmitter because there is no signal on pin 3.

It is indeed balanced. At least in the context we've been discussing here. Balance in this context has absolutely nothing to do with whether or not there's a signal on pin 3 and everything to do with a balance of impedances, which are required to achieve good common-mode noise rejection. And common-mode noise rejection performance hasn't anything to do with whether or not there's a signal on pin 3.

Unless!  Unless you receive with a center-tapped transformer, which then back drives the cable so that the signals are indeed balanced.  That is why this is so clever.

First, why a center-tapped transformer? What are you going to do with the center tap? If you connect it to the source's ground (i.e. the node between the two 100k resistors) you're right back where you started, with no signal on pin 3.

Second, there's no need for there to be a signal on pin 3. Common-mode noise rejection hasn't anything to do with signal balance. It's all about impedance balance. And the circuit as shown provides just that.

se

[hagtech]

Let's say we do connect the two center taps. How does that amoeliorate any static charge?

No, no.  The connection has to be to ground or some other reference voltage that defines the common mode operation.  And you only do it to one side.  If the cable floats, you can get into trouble.  You have to have at least some leakage path.  Really, it's not a stretch.  Same diagram, just that the R1 and R2 values get big.

Do you mean a literal connection to the literal earth?

Yep.  Same thing with power transmission.  Same with the old telegraph system, which of course predates telephone.  They all connect to Earth using copper rods and salts.  This is how one gets the phantom connection (a common mode signalling technique).

everything to do with a balance of impedances

I'm ok with that.  Well put.

What are you going to do with the center tap?

The idea was to tie the center-tap of the receiving transformer to the return (opamp ground, actually).  The signal on pin 2 is then forced onto pin 3 because of autoformer action, but of opposite polarity.  In this way we achieve both impedance and signal balance.

jh

[Steve Eddy]

No, no.  The connection has to be to ground or some other reference voltage that defines the common mode operation.  And you only do it to one side.  If the cable floats, you can get into trouble.  You have to have at least some leakage path.  Really, it's not a stretch.  Same diagram, just that the R1 and R2 values get big.

The context of my comment was with regard to an output transformer feeding an input transformer. Yeah, if you're driving straight into say an opamp circuit, the inputs need to reference its internal ground.

Yep.  Same thing with power transmission.  Same with the old telegraph system, which of course predates telephone.  They all connect to Earth using copper rods and salts.  This is how one gets the phantom connection (a common mode signalling technique).

Far as I'm aware, the literal earth connection in those systems is simply for lightning protection. A bit different scenario from a couple of interconnected audio components.

The idea was to tie the center-tap of the receiving transformer to the return (opamp ground, actually).  The signal on pin 2 is then forced onto pin 3 because of autoformer action, but of opposite polarity.  In this way we achieve both impedance and signal balance.

Hmmm. Off the top of my head, I'm not seeing it work that way. I'm seeing virutally all of the signal current flowing through the top of the transformer and right back to the source ground through the connection to the center tap.

Now, if you just fed the output of that "alternative" balanced output into a regular input transformer with no center tap, then you would have both lines driven.

se

[Bill Epstein]

If it's the phono stage I think it is, ... Even using the adapters supplied by the manufacturer you should be able to ascertain for yourself whether the balanced input is offering superior sound to the RCA's (but keep in mind-once again if it's the stage I'm thinking it is-that you will have to adjust for impedance using the RCA's where you do not with the balanced in). ...

I think you're right about the phono stage you think it is and it's a completely different sound with the XLR adapters and none of the adjustments needed for the single-ended connection.

As for the discourse on how Samuel F.B. Morse was grounded, Jim and Steve, glad I could help