First, for those readers who don't know what a rotary engine is, or whose first thought when reading the phrase "rotary engine," is "Wankel engine," please go visit Wikipedia's Rotary engine page [link].

Next, take a look at Wikipedia's Radial piston pump page [link], and observe how a rotary engine is very much like an outside impinged radial piston pump -- notably, the part which pushes the pistons is on the inside, and the working space of each piston is on the outside.

This idea is to make an engine which resembles an inside impinged radial piston pump, with the working space of each cylinder located between the cylinder's piston and the (stationary) hollow shaft. The shaft acts as the engine head.

Intake and exhaust are accomplished through ports in the shaft, which the cylinders slide over; similarly, there is a fuel injector and spark plug on the shaft, which the cylinders move past. (Or a fuel injector and glow plug, in the case we want to use a diesel cycle).

Sealing can be done using the same kind of sliding seals and apex seals that are used by Wankel engines.

The simplest design of cycle would of course be a two stroke, since intake and exhaust can be timed entirely by the movement of the cylinders past the ports.

A split cycle two stroke is also quite doable...

By not having valves poking into the cylinder, we have a huge advantage compared to earlier split cycle engines -- namely, the clearance between the compression piston and the ported shaft can be very tiny, thus forcing the entirety of the compressed air charge into the crossover passage, improving volumetric efficiency.

A double rotor version of the split cycle two stroke might be a good idea (with one rotor doing compression and the other expansion), because if the rotors spun in opposite directions from one another, it could minimize unwanted gyroscopic effects... plus it would look cool.

A four stroke version should be possible, though it would be more complicated. If the piston rods are pushed in by a ring to which they are attached, or a circular cam which they slide along, it would require actual valves, not just ports, since each cylinder would need to whole revolutions per cycle. If the pistons are pushed in by a two lobed cam, then simple ports could still work, but the piston motions will be complicated.

It's probably not worthwhile to make a four stroke version of this engine layout.
-- goldbb, Aug 12 2011

Rotary Engines http://en.wikipedia.../wiki/Rotary_engine
[goldbb, Aug 12 2011]

Radial piston pump http://en.wikipedia.../Radial_piston_pump
[goldbb, Aug 12 2011]

The split cycle version http://www.doylerotary.com/
My idea is a /generalization/ of this, since it could just as easily be a regular two stroke, or a counter-rotating split cycle [goldbb, Aug 12 2011]

More prior art http://web.archive....llpistonengine.com/
Whatever happened to the ball piston engine? [goldbb, Aug 12 2011]

From a 'nifty' point of view I like the 'one cylinder head, many cylinders' idea, except it looks like it's adding several cylinders worth of friction, and doubling blow-by.

But regarding split-cycle, if you've a common intake area then you only need two large intake&compression cylinders: one for each half of a revolution.

The ball-piston engine may have suffered the same fate as most cam engines: the power transfer takes place across too small an area. That and the actual 'ball' bit seems to be just for looks: surely an oval piston (inside an oval cylinder of course) with a roller would be much more robust and less frictive.
-- FlyingToaster, Aug 13 2011



random, halfbakery