If you're looking to eliminate as many phase problems as you can you should really look into using a coincident array like XY.....ORTF / NOS and the like are near coincident arrays and actually use small phase differences as part of forming their typical image
You're pretty much right with your placement, remember that the height and angle of the array will affect the front to back balance so just move it around and up and down until you hear what you want.
If your interface allows it make sure you're recording to a 24 bit file (remember that the sample rate may vary depending on what you're outputting the file to). That way you can afford to leave yourself around 12dB of headroom at the loudest dynamics. It gives you space for transients, but when you process later followed by dithering to a 16 bit file you should find that (most of) the background noise at the lower dynamics is acceptable.
Just try and be careful not to clip he analogue gain stages or you'll have an issue
Oh - incidentally, do you have any recommendations for a cheap OTRF adapter for a microphone stand? At the moment, all I have is standard microphone stands which means I'd have to use two - a bit messy!