Balanced lines are typically – and quite incorrectly – explained as follows. A signal is split into two equal but antiphase parts in a balanced line driver. These signals are connected to each leg of a pair cable and eventually arrive at a balanced receiver. This device inverts one leg of the signal. In doing so it adds the antiphase signals together to reproduce the original one and also cancels out any in-phase interference signals that may have been induced in the cable en route. Almost all of this is wrong!
The term phase should always be replaced by polarity, since phase implies shifting the signal in time rather than merely inverting it. But balanced lines are balanced irrespective of whether any signal is present and do not require symmetrical signals to operate correctly.
IEC Standard 60268-3 explains the truth –
“A balanced line is a two-conductor circuit in which both conductors … have the same impedance with respect to ground and to all other conductors”.
If the impedance is the same for each leg then the voltage induced across it by interference will be the same, so there will be no difference in voltage between the two: there will be no “differential” interference voltage. Therefore balanced lines are distinguished by their ability to ignore interference and by nothing else.
There is no requirement for each leg to be driven, let alone symmetrically, and polarity inversion is not always used in the receiver – a simple transformer, which is the classic balanced input, cannot invert one signal leg. What is universal in receivers is the ability to sense the line differentially (differential mode) – the difference in voltage between each leg – since this is the distinction between the wanted signal and most interference, which will be common (common mode) to each leg.
Balanced lines are, in effect, extended Wheatstone bridges and the driver, cable and receiver all have a role in maintaining the balance condition. An excellent balanced receiver loses all its beneficial characteristics if connected to an unbalanced cable or a poor driver. Poor tolerance resistors or bad contacts at any one point can have severe consequences in damaging the ability of the entire circuit to reject common-mode interference even though this may not be immediately obvious in the handling of the wanted, differential signal. Under some broadcast conditions a 1 ohm common-mode impedance difference can reduce interference rejection by 15-20 dB.
As with so many audio matters balance will vary across the frequency range. Good transformers can maintain their common-mode rejection across a very wide bandwidth but many electronic input circuits lose this ability at higher frequencies and can be very poor at rejecting RF interference.
Most (though not all) electronic inputs cannot “float” – the maximum common-mode voltage is limited to the voltage swing of the line receiver – and large levels of interference may restrict the undistorted dynamic range of programme. Transformer-connected systems can float since there is no direct connection between the two windings but this can also mean that dangerous voltages could exist between the input and the output – approach with care.