What's the point of balanced inputs?

[smoguzbenjamin]

Hi guys,

I've been accepted for my school's technical department, and they've given me block diagrams of the mixer to figure out how the thing works. However, on every balanced input, it just flips the phase of one signal and feeds the two in-phase signals together. So I'm wondering 'why invert one signal in the first place?' Is there some reason I don't understand, or is this just a funky mixer that doesn't use the 180 degree phase shift?

cheers
Ben

[David]

Ben:

Congratulations on your being accepted.  What you're describing is a low-impedance connection scheme that's called XLR here in the States.  Most microphones now use this technique.  Actually, what you have is an inverted signal and a non-inverted signal in the cable.  This is standard operating procedure for mixers, recording and sound reinforcement equipment.  Your mixer is normal.

[smoguzbenjamin]

Yeah that's the name, XLR, that's it. Okay, it's normal, that's good. So why would you want the non-inverted AND the inverted signal in the cable? Is there a special reason for this?

[David]

Yeah that's the name, XLR, that's it. Okay, it's normal, that's good. So why would you want the non-inverted AND the inverted signal in the cable? Is there a special reason for this?

There is.  I don't know what it is, though.  If you need the specifics, I think it's Google time!

[smoguzbenjamin]

Thanks David

[Samuel]

i believe the balanced scheme is a method for isolating line distortion across long cable runs. If you have a phase inverted version of your signal transmitted simultaneously you can add them together at the input to complete cancel the signal (identical signals that are 180 degrees out of phase can be added together to produce nothing, since since the original will be positive where the flipped signal is negative, and vice versa).  When you cancel the signal in this way, what you are left with will be any signal artifacts introduced along the length of the cable (radio interference, etc). This can in turn then be phase flipped and added back to the original signal to cancel out this noise, giving you a noise-reduced version of your original.

I think this is the reasoning, but I could be wrong.

[smoguzbenjamin]

You know, looking at the block diagrams, that could be exactly what it's doing... Not sure thought, the block just says "180degree phase shift" and the next block says "noise cancellation" so that might be it.

could it also be that in a guitar cable that ground is working as negative as well, so that the ground cable is doing 2 tasks?

[Samuel]

Nope. Guitar cables are definitely unbalanced. That's why XLR mic cables are three connector cables, and guitar cables are only two.

edit:Besides which a connector couldn't be both a steady "0V" reference and an AC signal...

[smoguzbenjamin]

That was what I meant.   All my words are coming out wrong today

[Samuel]

To clarify, after rereading my first post - the reason you are left with just noise after phase cancelling your two incoming signals is that any noise introduced along the cable run will affect the two signals in the same direction, despite the fact that the signals themselves are inversions of one another. Thus, the noise is not phase cancelled at the receiving end, and is thereby left behind to be eliminated...

[Peter Snow]

Well the way I understand it is that you take the original signal (+), invert it (-) and send both signals down 2 separate cables (plus a third ground/shield cable).  

The idea is that any noise that is picked up on the cable run is picked up equally by both wires. But note that the noise signal is + on both cables.  So now one cable carries "+signal+noise" and the other cable carries "-signal+noise".

When the signals reach their destination, the signal on the negative cable (along with the noise) is inverted again.  This gives "+signal-noise".  This is then added to the +signal+noise on the other cable.  Adding the - noise to the +noise cancels out the noise and adding the former -signal (now +) to the existing +signal produces 2x the original +signal.  This may be further attenuated back to it's original level if that is important.

Someone else may be able to clarify this further and clean up my terminology a bit.

Peter

[Peter Snow]

Ahh, Samuel beat me to it and a much clearer explanation too!

[Samuel]

my explanation may have been clearer, but I think the nuts and bolts of yours is probably closer to fact (one inversion stage instead of my unnecessary two). I understand the theory but to be honest I've never really examined a schematic application of the principle.

[Johan]

...or for a guitar analogy of balancing..think "humbucker"...the two coils pick up the same noise out of phase and the signal in phase and then when they are combined the noise is canceled and the signal is doubled in strength..same idea, applied a little different but still the same...

Johan

[Peter Snowberg]

The noise canceling is also why twisted pair wire is used for high speed data connections like EtherNet.

In both cases the wire consists of a signal and an out of phase anti-signal.

The ground connection isn't really part of the signal path; only a reference to the "ground" potential.

[bioroids]

As a recording engineer I can tell you Samuel's and Peter's explanation are right, that's how it's done.

I think there are also guitars (Rickenbacker?) that provide a balanced output (if you use a 3 conductor cable), but never heard of a stompbox with a balanced input... that would be nice when playing near radio stations!

Luck!

Miguel

[roknjohn]

It is based on a differential amplifier. In other(my) words, the difference between the two signals (the "+" and the "-") is what ends up getting amplified. Anything that isnt different (induced noise) doesn't get amplified (in theory anyway).

This is just my way of explaining what Peter and Samuel have already stated, but may be easily visualised. Check out "differential amplifier" on Google or an electronics source.


John

[roknjohn]

Oh yeah,
I forgot to send my Congratulations as well.


John

[smoguzbenjamin]

Thanks guys, it all makes sense now. Sorry that I could't respond earlier but I've been busy for school. Thanks again

[Mike Burgundy]

Noise elimination through cancellation, yes.
The electrical implemenatation of this in the mixer will be very simple. An opamp has both inverting (-) and non-unverting (+) inputs.
Feed on signal to - and the other to +,  plus is sent through, - is inverted and then sent through, leaving you with the original signal, nice and strong, with no hum, clicks or other nasties.
The send bit involves an extra opamp section that inverts (simply use the - input) and as a whole amplifies 1:1.