One of the most widely misunderstood concepts in audio is signal transmission between equipment, specifically balanced and differential signaling. We all know that a balanced signal is one where two conductors each carry a signal at opposite polarity, right? If you’ve answered yes, you’ve fallen victim to a widely accepted internet lie. Don’t worry, they got me once too.
So let’s clear this up!
Balanced signaling is two conductors (with an optional shield) that have equal impedance to ground.
Differential signaling is two conductors (with an optional shield) transmitting the same signal at opposite polarity.
Though these two signaling techniques are often used together, it’s important to understand they are two different concepts with distinct advantages. In audio, a balanced signal line will often utilize differential signaling but not always. A balanced line may have one conductor with an audio signal, and the other with no signal (0V) or each conductor may carry the same signal at opposite polarity (differential). The only requirement for a balanced line is that both conductors have the same impedance to ground.
Usually, the two conductors will be wrapped in a shield to provide additional protection from unwanted radio frequency interference (RFI). This shield, which is usually tied to chassis ground at one or both ends, is the first line of defense against noise. Any noise voltage induced in the shield will be shorted to chassis ground. Furthermore, the balanced conductors will be twisted. Having a shielded balanced signal in a twisted pair provides the best EMI (50/60hZ) and RFI rejection. However, if the noise reaches the wires internal conductors it will induce a voltage in these internal conductors.
What happens when noise reaches these internal conductors? Let’s start with an unbalanced system that consists of an audio signal in one conductor and a conductor tied directly to ground. This ground reference conductor will have a near zero impedance to ground (dependent on the wire and its length) and the signal conductor will have a larger impedance to ground. Noise induced in the ground reference will be shorted, however some of the voltage will be induced in the signal conductor. You are left with a conductor carrying your original signal and a noise signal that are referenced to the ground conductor. How do you separate the noise from the signal? You can’t really.
By replacing this unbalanced line with a balanced line where both conductors have an equal impedance to ground, and noise that penetrates past the shield will induce the same voltage in each conductor. Now we can use a differential input (not to be confused with differential signal) to remove this added noise. The differential input senses a signal by measuring the voltage difference between conductors. Signal voltages common to both conductors (the noise) are ignored. So in an impedanced balanced line that has a 1V signal on one conductor and no signal on the other (0V), the input would pass a 1V signal ( the difference between the two conductors). This is an example of a balanced signal that is NOT differential. If the same impedance balanced line also had 0.25V of noise induced in each conductor, the input would reject that signal as there is no voltage difference between the two. In the real world, no circuit can reject all of the noise. The ability of the circuit to reject these signals is called Common Mode Rejection (CMR) and is usually expressed in dB as a ratio (CMRR).
Now if we take the same 1V signal and place it on the second conductor in opposite polarity, we’ll have a differential signal. At their peak one conductor will have a 1V signal and the other will have a -1V signal so the input will see a 2V difference between the two. Furthermore, utilizing differential signaling on the balanced line gives us better control of the reference voltage (“cold” side) and produces a 6dB larger (double) output signal improving signal-to-noise ratio.
If you want to get a more detail understanding, check these links!