January 03, 2013, by Pat Brown
A common task in Audio Land is the need to feed a number of inputs from a single signal source. This may include driving a rack of amplifiers, providing feeds to the press, or distributing a signal around a building or campus.
The methods used to accomplish this range from the profoundly simple to quite complex, and the appropriate method must be determined after sizing up the situation.
Impedance matching means that an output is terminated with a “mirror” input impedance (Z). This configuration yields maximum power transfer, and more importantly reduces reflections of the signal voltage from a load back to the source.
In multimedia systems, the matched interface is used for high frequency signals. These include video, antenna and digital interfaces. One drawback of the matched interface is that active or passive splitters must be used if the source must drive multiple inputs.
In a constant voltage interface, an electronic signal source (an output) is expected to develop a signal voltage across a high impedance (an input). The minimum ratio between the source Z and load Z is one order of magnitude (1:10).
This scheme is used universally in the audio industry for passing signals from component to component. One utility of this interface is that it provides the possibility of driving multiple parallel loads from a single source without additional hardware.
The stipulations are:
1. The parallel combination of all loads cannot violate the 1:10 minimum impedance ratio.
2. The path length (interconnecting cable) must be short when compared to the wavelength of the highest frequency component of the signal.
Since the speed of propagation of electricity approaches the speed of light, and audio cables are typically less than a few hundred feet, the second condition is easily met in the vast majority of audio applications.
Radio frequency, digital, and video signal wavelengths are much shorter, and the impedance matched interface must be used in lieu of the constant voltage interface to prevent signal degradation.
Figure 1: Circuit of a single source driving multiple loads. (click to enlarge)
“Y” To The Rescue
Figure 1 shows an equivalent circuit of a single source driving multiple loads. Note that even though the input impedances are not the same, this is a parallel circuit, so all of the inputs have the same voltage impressed across them.
In this example, signal distribution requires a simple “Y” cable connected from the source to the multiple loads. Alternately, the loads can be “daisy chained.” Power amplifiers often provide parallel inputs to facilitate this.
A drawback to Y-cable signal distribution is the lack of isolation between the individual loads and the source.
For instance, a short circuit across any of the inputs will kill the signal to all of the inputs.
For this reason (and others), this method is not recommended for driving loads that lie outside of the equipment rack that houses the source.
In these cases, load isolation can be achieved by using a distribution amplifier (DA).
The DA provides a single high impedance input for the signal from the source, but provides buffered low impedance outputs that can be used to drive remotely located loads.
The load buffering is achieved by using an active stage for each of the DA’s outputs.
A short across any one output is buffered from the other outputs by the active stage (Figure 2).
Yet Another Problem
While the DA solves the loading issue, we’re not out of the woods yet. Another problem that plagues distribution systems results from multiple ground connections between the various components.
Figure 2: Isolation between source and loads. (click to enlarge)
These “shared” ground paths include the AC safety ground, the cable shields, and possibly paths between the equipment chassis via racks, etc. Noise currents will circulate through these “ground loops” and possibly infect the audio signal if this parasitic current finds its way onto a circuit board.
Isolation devices can allow the audio signal to be coupled from an output to an input with no physical wire joining the two circuits, eliminating at least one of the ground loops.
Transformer isolation allows the signal to be coupled via induction (Figure 3).
Figure 3: An isolation transformer. (click to enlarge)
Optical isolation uses pulsed light to couple the signal, but usually requires the signal to be converted to a digital format. The transformer has an advantage in that the signal can remain in analog form.
The irony is that the same mechanism that allows a signal to be coupled between two circuits inductively also allows power supplied fields to be coupled into ground loops (Figure 4).
We’re faced with the classical engineering task of maximizing the effect when it helps us and minimizing it when it is working against us.
Ground loops and power supply radiation form an unintended transformer in a sound system. (click to enlarge)
A Complete Solution
Putting all of these mechanisms to work, an active distribution amplifier with transformer balanced inputs and outputs may be the optimum way of distributing an audio signal to multiple components.
The active stages buffer the inputs from short circuits, and the transformers allow ground loops to be interrupted while allowing the signal to pass, while at the same time providing excellent common-mode rejection.
Many DAs also include level controls for each output, allowing the signal level to be optimized for mic or line level devices. But this approach is not cheap, and may be overkill for simple systems.
The next time you need to distribute an audio signal to multiple inputs, don’t overlook the simplicity of using a properly wired Y-cable to accomplish the task. If the signal needs to extend beyond the rack, a good DA will easily justify the investment.
Pat & Brenda Brown lead SynAudCon, conducting audio seminars and workshops around the world. For more information go to www.synaudcon.com.