IT’S AES/EBU FOR YOU!

The advent of digital audio transmission standards began for me with Denon CD cartridge players. These were the first devices to cross my path that had an XLR connector for digital output. And so began the transition to digital audio standards. And there have been a few surprises along the way.

The first thing to know about digital audio wiring is that the various common formats available, whether they use balanced shielded wire, or coaxial cable, or fibre-optic cable, are all very similar, and it’s usually quite easy to adapt from one to another. The second thing to know is that 99% of all problems are related to impedance mismatches. The high bit-rates involved make digital audio look and act more like RF than audio, and as a result if you think of the signal as an RF carrier, you’ll intuitively stay out of much trouble.

Okay, first the good news: in true digital fashion, this digital audio signal will not pick up hum, or impair its frequency response, or get audibly distorted by traveling around the radio station. The bad news is that the inevitable degradations are largely undetectable until they reach the equally inevitable digital cliff, at which time operation becomes flaky and unreliable. And nobody wants that!

The main differences between “digital twisted pair” and the regular analogue product are found in the characteristic impedance of the wire, and the capacitance of tip and ring to ground (are we allowed to still call the conductors tip and ring?). Our normal shielded twisted pair 22 AWG wire has a typical but generally unspecified impedance of 40 to 80 ohms. The AES/EBU specification for digital cable allows for 88 to 132 ohms, with the ideal impedance being 110 ohms. While you can generally get away with using old familiar wiring for short jumpers, if your signal is going farther than, say, 15 metres or so, you’re going to need to use “digital” wiring.

As a consequence of the higher impedance and desired lower capacitance, you’ll find that the wires tend to be smaller (26-24 AWG) and hence more fragile, and the insulation, being foam-based, is thicker, softer and tougher to strip off. Take care not to crush the wire, as that insulation will compress easily, and the conductor spacing is a critical factor in maintaining the specified impedance. The AES/EBU standard calls for the use of shielded cable, but the common mode noise spec is so loose, that really the shielding is not needed. All of which is moot, because when you’re shopping for digital wire, shielded is what you’re going to find. And it will be expensive. Since you’re paying for shielding anyway, you should look for a cable that has braid shielding. Foil alone is most effective at shielding below 1 MHz, and our digital signals are going way above that!

One thing to bear in mind is that even though you’re spending the big bucks on that special wire, your transmission lengths are still limited to 300 m or so. The exact distance depends upon your bit-rate. Your signal can travel much further at 75 ohms using coaxial cables, but you’ll need balun transformers to impedance-match and unbalance the signal unless your equipment already has unbalanced I/O. Since TV stations are generally running all sorts of precision 75 ohm cable around anyway for video, this option is quite popular in TV-land. Although special “digital audio coax” is available and of course recommended, it’s difficult to find much wrong with using a precision “analog video coax” for digital audio. Digital video transmission, with its bandwidth requirement up into the multi-GHz, is of course another story. But it’s always okay to use a “digital” cable to carry analogue signals.