All engines with ports (and no reed valves) suffer from flow reversion when operated outside the engine's power band. The Aixro rotary engine is no exception. To get the engine to breath at higher rpms, the port timing must be such that intake manifold air flow will "reverse" at lower rpms. You can see this affect on the dynamometer when operating the engine at high throttle openings and low engine speeds. You literally get a cloud of fuel vapor forming around the carburetor inlet as the fuel/air mixture is first blown out the carburetor then sucked back in.
Carburetors do not care which direction the air is flowing and will tend to provide fuel when air is moving in and when it is moving out of the engine. Thus the carburetor provides far too much fuel when inlet air flow reversion occurs. The benefit of fuel injection is that it does not know which direction the air is going and thus does not care. It simply knows the throttle angle and engine speed from which it provides a pre programmed fuel pulse. So, where the carbureted rotary suffers from overly rich mixture in the 3,000 to 6,000 rpm range, the fuel injected version has perfect mixture.
Once we were able to control mixture at lower rpms, we learned that the engine was capable of providing significantly more torque; on the order of 25% more from 3000 to 6000 rpm. We accomplished this by manually opening the peripheral port on the engine at these lower speeds. Normally, the vacuum diaphragm operated peripheral port will not open below 6000 rpm. The downside to this approach is that the driver must be judicious in throttle actuation at lower rpms as you can easily stall airflow by slamming the throttle open too quickly. We are constantly working with accelerator pump effect and different linkage relationships between the side and peripheral ports to address this issue.
Once the engine was running on gasoline, we turned our attention to main bearing failure resulting from high transfer port temperatures. Charged cooled rotaries where the intake mixture travels through the center of the rotor and acts to cool the rotor can suffer from high bearing temperatures as horsepower is increased. In this case, we were recording transfer port temperatures in excess of 290 degF! The most logical solution was to turn to methanol for fueling the engine as it acts as a terrific cooling agent. Engines require almost twice the volume of methanol as they do gasoline so twice the fluid is "boiling" off the rotor and, in turn, cooling it. We left our oil to fuel ratio the same so we affectively doubled the amount of oil lubricating the engine as well. In addition, we allowed for trailing throttle fueling of the engine so that fuel continues to cool the engine when you lift at the end of a straight. The results were almost a 100 degF drop in transfer port temperatures which equated to a denser charge reaching the combustion chamber resulting in a dramatic improvement in power throughout the operating range.