Impedances and transmission line effects...

[Mark Hennessy]

I would point out that in the past I've seen a short (<50m) run of twisted pair driven from a 'balanced' source that was anything but - and of about 1000 Ohms output impedance - suffer from notable treble loss.

The driver was an op-amp with 1000 Ohms in series with its output in the hot side of the feed: the cold side was a 1000 Ohm resistor in series with a 100uF capacitor to earth.

With RG58 at 58pF/m - let's call it 60pF as we're engineers, not scientists - then 50m would be 3nF. With the total Zout being 2k, the -3dB point is ~26kHz. There would be some attenuation at lower frequencies, of course, but that doesn't seem like much to me. Measurably, certainly, and should be attended to in an installation, but would it leap out as being a big problem unless you were doing an AB comparison?

Regarding your point about balance, I'd say that actually, it was nearly perfectly balanced.

One of the biggest misconceptions that I have to clear up for folk at work is that balanced operation is nothing whatsoever to do with the voltages on the two lines. Rather, it's about having identical impedances to earth, so that an interfering signal affects both lines equally, so that the differential input stage receiving the signal is able to subtract the interference from the wanted signal.

I said "nearly" because as well as the resistance to earth, the capacitance ideally needs to match as well. Unlikely with a coax, naturally...

Of course, if you do drive the source in true antiphase, you get an extra 6dB headroom for a given set of PSU rails. So that's an obvious win, but it isn't essential by any means.