Thanks for the images - can I check two thing firsts. Has this been positioned based on the R3 Prusa extruder or the current 0.6 Bear extruder. I ask becuase it looks like the nozzle is too far towards the back of the heated bed. This is going by the 'ROB_Channel_middle.JPG' - the 3 'dividers/roof support' pillars should converge at the X0 and Y0 which on my model is the centre of the nozzle (using Greg's CAD model). I'd also say that the nozzle is also too low - again based on Greg's CAD the lowest part of the duct should be 2.5mm up from Z0 (in an ideal nozzle placement world).
That being said, I can't argue that I need to make some more room for the side inlets and if I'm reading it correctly the centre duct could do without the roof support.
Would it help if I sent you an X_T (Parasolid) of the duct, or even just the internal chambers (they were a Boolean subtraction from the main body)?