barnabas: I am well aware of the advantages of balanced interconnections and their ability to fight inductively and capacitively coupled interferance.

Such interconnections are usually not used on consumer gear because (a) interconnections are short (though 5.1 setups are making that less true), (b) they add expense. Nevertheless, their value in combating interferance is well known.

That said, soundhound's observations regarding a relative increase of odd order harmonic distortion bears investigating: if the benefits of balanced drive are not there (because of a short run), and there are greater relative disadvantages because of the nature of balanced output stages, it might not be worth it.

As for running a balanced output to an unbalanced input (via simple XLR to RCA), you may be defeating the purpose, as the termination could unbalance the circuit (depending on the output stage topology). Furthermore, IIRC, balanced signals are referenced to a higher level than unbalanced ones, so you might need some attenuation (this may be built into some XLR to RCA adapters).

I hadn't considered the value of a balanced run to a sub amp: induced 60 Hz rejection, though that makes some sense. My sub, using a Hypex HS200 amp, does not offer balanced inputs :-( I have found that a 30 foot run of decent interconnect cable connected to signal (and chassis) ground at both ends carrying a single-ended sugnal to my sub amp does not pick up audable 60 Hz interferance, though. I did have a ground loop problem at one point, but running the three grounded power cord from the sub amp through a GFCI, and lifting the ground from the GFCI worked. (This is much safer than simply lifting the safety ground as any current inbalance caused by a ground fault (i.e. through a body) will trip the GFCI.)

Given runs of up to 65 feet, I am definately biased in favour of running line level signals to amps located near speakers, rather than speaker level signals and dealing with resistive losses: the BG speakers present nominal 4 ohm loads. If I have to run line-level signals that long, I'd prefer driving a balanced line.

So, the question remains: how bad is the odd order harmonic distortion from the balanced output of the 990? Is it as bad as soundhound suggests? I am not entirely convinced that (a) it is significantly worse than the unbalanced output and (b) the reduction of even order harmonic distortion from balanced output stages somehow masks what is present less.

I don't deny that the odd harmonic distortion from a balanced output stage might be worse than from an unbalanced one. The questions are "Is it noticible?" in the context of the rest of the signal quality produced by a $1k component, and "Is it the lesser evil when considering a 65 foot signal run?"

As I already have a decent stereo amp for the L+R mains (though I am considering upgrading to a better one, with balanced as well as unbalanced inputs), and the run from the 990 to that amp and speakers is the shortest of all, it isn't an issue for the most critical speakers when listening to two channel music. So, some minimal inaudable distortion to the surround speakers might be tolerable, even if the signal is theoretically "worse" by some measure.

I suspect however, that the benefits offered by a better physical connector usually used for balanced interconnects over an RCA phono plug probably exceed any increased odd harmonic distortion.

BTW, I hear that someone is offering to mod M200s for balanced operation. Does anyone know if that's correct, and at what cost?