I am working on a project that takes in a single line-level audio channel and outputs 2 balanced/unbalanced +/- outputs from that channel. I have read a number of the excellent published articles online on circuit topologies for balanced/unbalanced inputs/outputs and have been testing the circuit shown in the partial schematic attached. I have tested the output on both a decent quality sony boombox (unbalanced RCA inputs) and a high end QCS amplifier (balanced 1/4"), and have discovered a number of quirks I've been trying to troubleshoot.
1) When connecting to either the QCS or Sony, as long as the audio signal (-) output is at the same potential as GND in the schematic, everything works fine, the sound is top notch. When connecting the Sony with the audio signal (-) as shown in the schematic, I would expect noise since the output is not (probably) at same potential with the Sony signal GND ... this is exactly what I get (a high pitched buzz). However, I would expect that the QCS, having balanced inputs (active I assume) wouldn't care about the actual (-) potential (I measured around 1mV DC, 5mV ACrms). However I get the exact same result with both, a high pitched buzzing noise). Anyone know why that might be?
2) Turning on the project device results in a light pop heard from the speakers, presumably from the initial DC bias current flowing from the cap. Is there a way to reduce or eliminate this? Would a larger output resistor help? (I'd prefer to keep output impedance at 100Ohm if possible.)
3) Without power to my project device, the sound still comes through the opamp follower (via an audio transformer) and is output (albeit very distorted with no DC bias) to the speakers. I wasn't expecting sound output without opamp power ... is there a common/easy way to prevent this?
4) When the device is working unbalanced with excellent sound, it still does add a minimal amount to the noise floor (white noise/hiss). It's not really significant, but what is the best way to minimize this generally?
Thanks for any input.