I tried a search on this but could not find in simple terms what does it mean when you are talking about "BALANCED" audio? Thankyou.

Balanced is referring to how the audio signal is carried over the cables. A balanced signal requires three wires instead of two. Two of the three wires carry the signal 180 degrees out of phase with each other (the third is a common ground). Along the way, noise may be introduced into the signal. When the signal reaches its destination the two wires are flipped 180 degrees again putting them back in phase with each other BUT out of phase with any noise that may have been introduced. So balanced cables can run for very long lengths (100's of feet) without introducing any noise.

Unbalanced uses just two wires and any noise that gets introduced can be heard at the other end. These two wires can act like a radio antenna and pick up all sorts of interference. Unbalanced cables should only be run for short lengths. 10-15 feet.

~jr

Thanks John,
That was an easy to understand explanation. If I remember correctly, XLR mics have three prongs so this automatically makes all XLR mics balanced...is this correct. Also, is the "on-camera" mics built into a camcorder balanced since it is built in and no cables involved? Thanks again John.

The noise benefit, nicely descrbied above, is call "common mode rejection". That is, the noise is common to both of the balanced lines, and is rejected by the 180º cancelation characteristic.

Yes, Having an XLR adapter or cable is a good indicator of being a balanced audio device. On-camera mics don’t need balancing because the wires are so short.

~jr

Having a 3 pin XLR connector does not necessarily mean that the circuit is balanced. As an example of this I have cables (purchased - I didn't have to make them up) with RCA connectors on one end and three pin XLR's on the other that I use to connect the XL-2 audio output to a mixer input. Obviously only 2 of the pins on the XLR are used in this case and that is fine. It is quite possible to connect a single ended (unbalanced) circuit to a balanced one in this fashion. Note that while what I'm describing does work it is better to use a "balun" (balanced-to-unbalanced circuit: a simple transformer) in applications that require this.

Having a 3 pin XLR connector does not necessarily mean that the circuit is balanced. Thanks for pointing that out. That’s why I said it was a “good indicator” and not that it “means” they are balanced. I guess my choice of words was too subtle and I should have been more specific. XLR is just an adapter, you can wire it any way you want. The only way to be sure that a device is balanced, is to check the specs.

~jr

Balanced means the two signal conductors have the same impeadance to ground. This eliminates hum and buzz. The shield is for RF protection. You can have a balanced output without driving both signal lines but both lines need to have the same impedance to ground.

Sam

Hope you all don't mind me butting in here for a moment, but to echo Steve's reply: John Rofrano, your post (#2) is the BEST explanation I have ever heard regarding balanced/unbalanced cables. Finally, someone made it understandable--and easy to remember--for "regular folks" like me! Thank you!!

John,
I agree. Your explanation of balanced cables is great and easy to understand. Could you do the same with output impedance? What does one need to know about that value when looking at mics? Any advice would be much appreciated.

Ramdas

Impedance, as the word itself suggests, is the degree to which a circuit resists the flow of current. If 1 volt is applied to a circuit the amount of current which will flow is 1/impedance (note that this is a simplification as impedance equals resistance + j*reactance where j is the square root of -1 but reactance should be small in audio circuits and if you are saying "Wait a minute, there is no such thing as the square root of a negative number" you'll understand why we don't want to talk about this aspect of impedance any further here). What is important with microphones and especially loudspeakers is that if two circuits are interconnected the maximum power is transferred between them when their impedances are the same. If they are connected by a transmission line (cable) that cable must have the same impedance as the source and load. Thus a microphone with 50 ohms output impedance will deliver the maximum possible amount of power to a mixer if the mixer has 50 ohm input impedance and the connection is made via a cable with 50 ohms characteristic impedance. It would also be acceptable to use a center tapped transformer to change the single ended 50 ohm impedance of the microphone to 600 ohms balanced, run through a 600 ohm transmission line, then use another transformer to go back to 50 ohms single ended and connect that to a 50 ohm input mixer. Thus the important thing is matching the impedances of the items being connected and the cables connecting them. In audio this is not so terribly important with short cable runs but with long ones (really long) mismatches may have an effect on the frequency response of the system. The problem is that the power which is not delivered to the load is reflected back down the cable. The inefficiency associated with power reflection in a microphone circuit is not significant because it is easily recovered by amplification. The power loss with a loudspeaker system is. In either case the waves reflected back down the cable add with the ones headed towards the load and it is this that causes the frequency response effects.

Impedance matching is very important at higher frequencies (video, sync lines, SDI, FireWire...) where the interconnecting cables can be an appreciable fraction of a wavelength or indeed several wavelengths long. Impedance mismatches will result in appreciable signal distortion in such cases.

Impedance is the measure of signal resistance. Less resistance (lower impedance) is better because it means the signal can travel longer before the signal deteriorates (usually in the high frequencies first).

The only thing you really need to remember is that, in general, a low impedance microphone should be connected to an input with the same or higher impedance. That means you should check the impedance of equipment you’re going to plug the mic into as well. If a microphone is connected to an input with lower impedance, there will be a loss of signal strength because the mic has more resistance than the equipment it is being plugged into. (i.e., don’t plug a high impedance mic into low impedance equipment) In this case you should use a line-matching transformer.

Bottom line: lower is better, matched with the equipment is perfect.

~jr

John's post #2 in this thread is a convenient way to think of this subject and except for the idea of "flipping the wires 180 degees" is correct in that it is the way XLR cable is used to carry a balanced mono signal and cancel interference without the need for active circuits. I've used that explanation myself in an XLR FAQ at camcorderinfo.com.

But as far as the true definition of balanced conductors, Sam Gates's is the one that's correct. All that's required to carry a balanced signal is that both conductor have the same impedance to ground. Electronic circuits (differential amplifiers) can do the noise cancellation at the receiving end.

John,

You are confusing impedance with loss. A signal will travel further in a high impedance transmission line with less loss than a lower impedance line with more loss. Telephone lines, for example, are 600 Ohms. A signal travels with less attenuation in RG/8 than it does in RG/58 even though both are 50 Ohms. The former is less lossy (because it's bigger). Given that things are kept matched it doesn't matter what the system impedance is. If things are not matched and actual power transferred is not important (as would be the case with a mic) it is generally a good idea to have the source at a low impedance. The fraction of the source voltage appearing at the load is Rl/(Rl + Rs) with Rl being the load impedance (assumed to be resistive) and Rs the source impedance (also assumed to be resistive). This also assumes the interconnection is very short relative to a wavelength. If the source impedance could be set to 0 all the voltage from the source would appear at the load. An "ideal voltage source" is a source with 0 output impedance (there is no such thing in the real world but it can be approximated closely).

OTOH you can very well plug a high impedance mic into a low impedance device if you are not concerned with frequency response distortions (shouldn't be an issue with short cables) and if the low impedance device has enough gain to overcome the voltage division loss (and low enough noise relative to the dropped incoming voltage). In fact this might be a good way to deal with a "hot" mic.

It is always best to stay matched!

Balance actually means what it says - that if the voltage on one signal wire exactly balances (equals) the voltage on the other signal wire (with respect to ground in both cases) they will cancel or in other words that the circuit described as balanced subtracts the voltage on wire 1 from the voltage on wire 2. The subtration may be done with transformers or differential amplifiers - it doesn't matter and it doesn't matter what the impedances are though they are, naturally, usually equal in both sides. Thus if a cable picks up voltage Vi on both signal lines from an unwanted source the receiver produces Vi - Vi = 0 at its output. In fact some of Vi makes it through. The ratio of the amount that makes it through to Vi is called the "common mode rejection" and is a measure of how good the balance is.

The way to get a signal through this system is to put Vs on one wire and -Vs on the other. The receiver output is then Vs - (-Vs) = 2Vs. To get -Vs from an AC signal Vs one inverts the phase 180 degrees. That's what John was talking about in his first post on this subject. To subtract two (analog) signals from one another one flips the phase of the one and combines it with the other. That's what he meant when he talked about flipping 180 degrees again.

A.J. and John, thanx much for all the info. Now I have to figure out how to use it in my situation. In short, I have a PD 170, and I am considering either a Azden SGM-2X (impedance = 680 ohms) or an Audio Techinca AT897 (200 ohms). I will be doing documentary filming in India, and some of the footgtage will include village music. That is primarily why I want another mic and wonder if the AT is necessary for that, considering budget constraints. I called B&H to see if I could get a tech to help out, but they were too busy. Also, cannot get the discount on the Audio Techinica without an AES #. Any suggestions, help? Thanx again for all the help.

Ramdas
rlamb@hawaii.rr.com

According to its spec sheet the input impedance for the XLR inputs is 10,000 ohms. The signal drop from the 200 ohm mike would be 0.2 dB and from the 680 ohm .6 dB i.e. less than a dB in either case and you'd never notice the difference so I don't really think impedance needs to be a factor in your choice.

Impedance
1. Voltage is the "electrical pressure" that makes current want to flow, very analogous to fluids driven by pressure. This analogy is often used in electronics courses.
2. Resistance is the simple opposition to current flow due to the chemical and physical nature of the conductor, some of which conduct better than others. It does not depend upon frequency.
3. Impedance is also the opposition to current flow, but it takes into account the phenomena of inductive reactance and capacitive reactance that occur when a voltage and the resulting current alternate in polarity (AC signals). They do depend upon frequency. Lower frequencies, like audio frequencies give higher capacitive reactance, and lower inductive reactance.
4. Large inductances give high inductive reactance, but small capacitances give high capacitive reactance. Inductance and capacitance values both tend to be very small in conductors, so capacitive reactance is more of a factor at audio frequencies in conductors than inductive reactance.
5. Inductive and capactive reactance are arithmetic opposites that cancel each other out by subtraction, so the bigger of the two dominates in any circuit.
6. Bottom line: Impedance is the square root of the sum of the squares of resistance and reactance.

Balanced circuits
1. Signals require two conductors.
2. Balanced inputs, outputs and cables are ones where the impedance of each conductor with respect to some reference, usually ground, are equal.
3. When used for a mono signal, 3-conductor XLR cable the sending and receiving ends have both of the internal conductors at the same impedance with respect to ground. THAT is what makes it a cable being used in balanced mode.
4. The sending device using balanced mono XLR cable normally applies the voltage between the two conductors, so they are 180° out of phase with each other if you look at their voltages with respect to ground at any instant. But that is NOT the definition of balanced. It is an elegant way of cancelling any induced noise without the need for active electronic circuits to do that at the receiving end.

Impedance matching
1. It is true that to transfer the most power at connections that the impedances of both sides must be equal.
2. However, in audio circuits, particulary in connecting microphones to inputs, often the goal IS NOT TO TRANSFER MAXIMUM POWER but to apply maximum VOLTAGE to the input. To make this happen, the impedance of the input should be MUCH HIGHER than that of the microphone. This is because the input amplifiers are driven by voltage, not power, and voltage divsion in a circuit is proportional to the impedances of the devices in the circuit. The amplifiers get any additional power they need pumped in from their own power supply.
3. The thumb rule is that it is desireable that the microphone's output impedance be 1/10 or less of the input impedance.

According to its spec sheet the input impedance for the XLR inputs is 10,000 ohms. The signal drop from the 200 ohm mike would be 0.2 dB and from the 680 ohm .6 dB i.e. less than a dB in either case and you'd never notice the difference so I don't really think impedance needs to be a factor in your choice.

A.J. That is exactly what I needed to know. Much appreciation.
Ramdas

I don't want to turn this into EE 101 but a couple of points since you seem interested:

1. Voltage is a measure of potential difference i.e. the amount of work that is done on a unit charge in moving it from one point in a circuit to another.
2. Resistance is the reciprocal of conductivity which is the current that flows per unit of potential difference. Resistance does depend on frequency. At higher frequencies currents flow selectively near the surfaces of conductors. This is not an appreciable effect at audio frequencies. The amount of current which flows in a circuit in phase with the applied voltage is given by the ratio of the potential difference (voltage) to the resistance.
3. Capacitance is the amount of electrical charge stored per unit of potential difference between two isolated conductors. If an AC current is applied across a capacitance current will flow but that current will lead the applied voltage by 90 degrees. The amount of current that flows is the ratio of the applied voltage to 1/j*2*pi*frequency(in Hz)*capacitance(in farads). Thus -j/2*pi*f*C is called the capacitive reactance which decreases as capacitance or frequency increases. Electrically -j represents the leading phase and the fact that no energy is dissipated in a capacitor (except in real ones whose leads have resistance). Mathematically it means that the current is imaginary. Inductance is a measure of the magnetic charge stored per unit of current flow in a conductor. If a voltage is applied across a conductor with very small resistance the current which flows is proportional to the applied voltage divided by j*2*pi*frequency*inductance (Henrys). This current lags the applied voltage by 90 degrees. Thus j*2*pi*f*L is the inductive reactance which increases with inductance and frequency. j idicates lag. No energy is dissipated in an inductor (except real ones whose conductors have resistance). Total current is V/(R - j/2*pi*f*C + j*2*pi*f*L).

While capacitive reactances may be large at low audio frequencies they usually shunt circuits. Thus the bigger they are the less effect they have. It is when they capacitance becomes large enough that the capacitive reactance approaches the impedance of the source that capacitance loads a circuit appreciably. Consider a twisted pair with 10 picofarads per foot. No problem for a few feet but a mile long twisted pair will exhibit inter conductor capacitance of a fraction of .05 microfarad.

Whether inductive or capacitive reactance dominates depends on whether they are in series or parallel. In a parallel circuit the smaller sets the current flow. In a series circuit it is the larger.

What is most important here is that we are talking about transmission lines (cables) which exhibit a characteristic impedance whose value is (in an ideal line) the square root of the ratio of the inductance per unit length to the capacitance to unit length. What this means is that the input end of a long cable looks like a pure resistor with value given by sqrt(L/C) which is not a function of frequency (as long as the look is quicker than the time it takes a singnal to propagate out to the end and get back) unless the end is terminated in the characteristic impedance.

Now what is true is that a short mismatched transmission line looks like a shunt capacitance and that may be what you were trying to get at. Match it properly and it looks like a pure resistance. This is another argument for low source impedance if the line is mismatched.

Balanced Circuits

Here are a couple of definitions of a balanced circuit:

"A circuit having its conductors electrically symmetrical with respect to a reference potential plane (e.g., ground). The potentials between the two sides and ground are equal and of opposite sign. For example, a horizontal two wire line may be a balanced line." ( US Patent Office)

"a circuit in which two branches are electrically alike and symmetrical with respect to a common reference point, usually ground." (IEEE definition)

I suppose you could argue that equal impedances are required for the branches to be "electrically alike" but consider an op-amp with 100 ohm resistors connected to both (+) and (-) inputs, a 1000 ohm feedback resistor and a 1900 ohm resistor from (+) to ground. The input impedance on both legs is then 100 ohms but the gain on the inverting input is 10 while the gain on the non inverting leg is 19. The common mode (noise) rejection would be dismal. I'd say the circuit was not balanced.

Good job Fred!
Here is a link that explains it balanced and unbalanced.
http://www.jensen-transformers.com/an/an003.pdf
You are correct about the impedance also. In audio we are looking for maximum voltage transfer so the input impedance of a device should be at least ten times the output impdance of the device feeding it.

Sam

A.J., I have an MSEE from the University of Connecticut so, I'm biased toward the IEEE definition. I once wrote a FORTRAN program to emulate the Smith Chart for transmission lines just for fun. I just try to keep it at level here useful to most curious readers.

FORTRAN? Smith Chart? You've been at this as long as I have!

I was afraid I might be dating myself :>)