Wallpaper photo of Convair LACES Aerospaceplane model (ca. 1961).Many thanks Circle-5, and as follows my contribution about the matter as a part of Scott Lowther's APR old number ([size=78%]http://www.up-ship.com/eAPR/ev2n5.htm[/size])
Thank you Archipeppe. I believe your drawing represents an earlier, development version of the Convair Aerospaceplane, with separate air intakes for the LACE oxygen separators (on three quadrants, behind the flight deck). It appears that Convair engineers were later satisfied with diverting enough air from the six hybrid engine intakes to perform that function. The fuselage intakes then became unnecessary. I've seen a model of your Aerospaceplane variant, but I do not have it, unfortunately.
No word on how the exposed ventral intakes would withstand atmospheric re-entry -- perhaps they were solid, unobtainium castings.
Since the pictures you posted (thanks for that circle-5, VERY much!) show a shock cone inlet,
I would expect that the cone would have the capability to close off the inlet when they were done
breating air and for reentry. Plus it was probably a moving spike inlet anyway for more efficient
compression. And as far as cooling, since LH2 is used for the LACE to cool the air and liquify it,
I would expect LH2 was probably also employed to cool the structure there.
As far as the fuselage based or engine inlet integrated air liquifier, I would think that the most
flexible approach would be for an inlet that could breathe non or low-cooled air, more deeply cooled air
(but not liquified) to increase the compression ratio for a turbojet or certain cycles of a variable cycle
turbofan, or to cool the air all the way to liquifying it, for the LACE (Liquid Air Cycle Engine).
Such a propulsion system would have great flexibility if such an inlet could feed a turbojet or
variable cycle turbofan, or a rocket.