In pro audio balanced connections are used almost exclusively, and with good reason. Using balanced connections is very effective in reducing and eliminating pickup and ground loops in the audio.
The most common balanced connections are in the form of XLR cables and ¼” TRS (tip-ring-sleeve) cables, and spade lugs are sometimes used.
Let’s take a look what a balanced connection is and at the physics behind why it works so well. (Balanced cables are used in other areas besides audio, so it applies equally to those situations too.).
A balanced connection will have two signal cables, an audio ground, and a shield. The shield in an XLR cable does not carry any audio information. The shield is connected at each end to the equipment or chassis ground (which is generally connected to the earth ground; many studios, especially ones in a building structure have a special grounding system, and other building codes to follow, but that is beyond our scope here).
The shield is effective against electric fields (and I won’t go very deep into any of these explanations, the physics involved can be found with Google), since the electric field inside the shield is zero (or approaches zero). The shield is made of a foil or braided wire or both.
Sometimes inexpensive cables will have a shield that is not braided, but only a layer of wire wrapped around the core. This can lead to an opening in the shield, especially at a bending point. The result is generally not severe for short runs, but on long runs, well, sometimes you get what you pay for, and at the very least you pay for the amount of copper in your cables.
Another thing to watch out for is a balanced cable with a drain wire running under a foil shield. The drain wire reduces the cable resistance somewhat, but since it will tend to be closer to one of the signal wires, it can cause an unbalanced pickup on the cable, which spoils the magic of the balanced cable.
So there, I said it… Sometimes there is a real, tangible, scientifically valid reason to choose one cable over another… Now back to the magic… er, physics of balanced cables.
The two signal cables are also twisted together. This is a common noise reduction technique called a “twisted pair.” Cables can pick up currents from outside electric and magnetic fields, and field strength is reduced as distance from the source is increased. If one wire is closer to the source than another wire, then the closer wire will be affected more. A twisted pair causes each wire to “take turns” as to which is closer to the source. In the end it tends to average out and both wires will pick up the same current/voltage. It is a physical fact that wires will pick up noise, and so we use this to our advantage as you will see…
The two signal cables in an XLR connection carry the same signal, but 180° out of phase. There are a few ways that the source equipment can do this but that is not important right now… what is important is what happens on the other end of the cable, at the receiving end. For simplicity, let’s say that the voltage one of the signal wires is 1Vrms and the other signal wire is -1Vrms (same magnitude, opposite direction, compared to the signal ground).
The equipment at the receiving end will have a differential amplifier (isolation transformers work too). For simplicity let’s say that it is a unity gain differential amplifier. That means it multiplies the difference of the incoming signal by one. So, (1Vrms - (-1Vrms)) * 1 = 2Vrms.
Now any interference that gets injected into the signal wires, as we saw a few paragraphs back, will be the same on both wires, both in magnitude and direction. Engineers have a fancy term for this. We call it “common mode” noise. Let’s say that each wire picked up 1Vrms of random noise on it way. Now watch what happens when we put it into our differential amplifier. (1Vrms – 1Vrms)*1 = 0V. The noise is gone!
Now if we mix signal and noise on the same cable let’s see what happens. ((1Vsignal + 1Vnoise) – (-1Vsignal + 1Vnoise)) * 1= 2Vsignal. In the real world the difference may not be exactly zero but a noise reduction of 40-60dB is not uncommon, so if the noise floor was halfway decent to begin with in the unbalanced case, the pickup with a balanced system will be vanishingly low.
Side note: instead of getting 1Vrms from an unbalanced signal, there is a 2Vrms signal at the receiving end, an apparent 6dB “gain” without the noise penalty. Some equipment employs this method some does not.
The other part of the magic of an XLR connection is this… separate signal ground and cable shield. Any currents due to pickup or ground loops are routed harmlessly in the shield, and stay out of the audio signal ground! Not so in a balanced ¼” TRS.
Now look at the ground loop HACDBJ in this sketch with a balanced XLR connection. Any currents that are picked up in the loop have no part at all in the audio signal. And any noise injected into the twisted pair cable is cancelled at the differential amplifier in DUT 2. Additionally, if (or when, as is generally the case) the amplifier stages in DUT 1 have power supply noise injected (as in an op amp… assuming that the two amplifiers are part of the same package and injected equally) there’s a good chance that it will be common mode as well, or nearly common mode, and therefore rejected at DUT 2 (at least to the level of CMRR at the receiving end).
What is CMRR? It is the Common Mode Rejection Ratio, expressed in dB, and for a good audio quality op amp can be 100dB, which is probably better than the cable, and that’s the point. How much rejection do you need? Well that depends on how much noise is picked by the cable. In a professional installation when you’re running hundreds of feet of mic cable or line level signals, you need all you can get.