It's late, and I don't have all my references handy, but IIRC it doesn't matter whether the detector is on the ceiling, on the wall, or down by the floor. The molecular weight of CO is 28, which is the same as nitrogen (N2; composing ~78% of the earth's atmosphere) and within spitting distance of oxygen (O2; MW=32, ~21% of the earth's atmosphere), unlike CO2 which has a MW of 44 and a density therefore of 1.51 versus 0.967 for CO. CO mixes freely with room temperature air, and since it isn't generated (usually) by an open air fire, it isn't driven as a plume to the ceiling as smoke would be. Therefore, for a constant but relatively small volume of generation per unit time, the CO concentration will be the same throughout the enclosed space. For low voltage combination smoke/CO detectors, no closer spacing than that required for smokes alone is called for, so I carefully assume that the standard 30 foot spacing is acceptable. I'd call the manufacturer before committing a design to paper however. Silent Knight says use 30 foot spacing for their IDP-FIRE-CO product.