When you convert you use all sensels info, so it is actually 72 sensel info gathered. For producing an in-camera b&w, you'd still need some sort of color array in order to be able to balance hues, since silicon photosites won't respond evenly to different colors. For maker to provide filtering options color capture is required.
Then think about it in present conversion: you have GRgBG... info, then you use that to produce monochrome set of information. Given that most scene areas are not balanced equally in color terms, sensels will collect different amounts of photons. Even if one is blown, the other carry info. That means that actually the b&w DR is much higher than the full color one, likely 1-2stops higher. This is very obvious if you do b&w conversions, areas that have one or two channels clipped still can be used if others or 3rd are not, so common in skies and red yellow flowers, or skintones.
So you control it yourself, not the camera's maker.