Originally Posted by
Fourteenbore
As an aside, I wonder what sudden piece of lateral thinking made someone think it good idea to bolt the crankshaft to the aeroplane and the propeller to the crankcase.
It was actually quite a good solution to the cooling problem. Rotary engines are self-cooling (that's rather the point) as well as being lighter through not having a flywheel. As specific power outputs increased cooling became a major concern (especially where total-loss splash lubrication was used). Primitive metallurgy and immature understanding of where the heat would accumulate, coupled to VERY primitive understanding of the aerodynamics of cooling ducts, severely limited the maximum power, but a rotary engine addressed most of these problems by putting constant wind over the cylinders. Too see just how far this got I suggest looking up the history of the ABC Dragonfly radial, whose copper-plated cylinder head finds would glow cherry red at high power settings. Liquid-cooled engines worked much better, but were heavier, more complex and had more components to fail.
As the science improved the introduction of better metals, better fin layouts (more focus on the heads where the heat is than the cylinders where it isn't), recirculating oil lubrication with oil coolers and MUCH better understanding of how convection cooling works led to better cowl/duct designs, innovations like the Townend Ring and the NACA Long-Chord cowl, made high power radials more practicable and thus obsoleted the rotary.
The heavy flywheels needed by pre-rotary engines became less necessary as decent cam profiles and ignition curves were developed, together with the general move to higher rpm (for more power). And of course the use of bigger, heavier and geared propellers contributed as well.
PDR