The Basics of How It All Started
The rotary engine, properly called a Wankel engine, is an internal combustion engine that uses a triangular rotor spinning inside a specially shaped housing. There are no pistons. There are no crankshafts doing up-and-down work. The rotor traces an figure-eight path called a epitrochoid, and the housing is roughly oval-shaped, known as a pericon. The three sides of the rotor act as moving boundaries between combustion chambers, completing intake, compression, power, and exhaust in a single rotating motion. That is the basic mechanical truth of it. The whole thing was patented by Felix Wankel in the late 1950s, and the first working prototype ran in 1957. He had been experimenting with rotary valves on standard piston engines since the 1920s, and the insight that a rotating triangle could replace reciprocating parts came from that earlier frustration. NSU Motorenwerke AG in Germany licensed the Wankel patent and built the first production car with one: the 1964 NSU Spider, powered by a single-rotor 491cc unit making about 40 horsepower. It was light, compact, and smooth, but the apex seals wore out quickly. Mazda, then a relatively small manufacturer, saw the potential for a high-revving, compact powerplant that could fit in their small cars. They spent years solving the sealing problem and the housing distortion issue before launching the Cosmo Sport in 1967 with a twin-rotor 982cc engine. TheRX-2, RX-3, and RX-7 followed through the 1970s and 80s. The rotary reached its commercial peak with the RX-7, which is still celebrated today. But fuel economy and emissions concerns in the 1990s, combined with the inherent difficulty of meeting exhaust standards with side-port scavenging, led Mazda to eventually drop it from passenger cars. They have kept it alive in the RX-8 until 2012 and still use it in some niche applications and motorsport. I have spent more time than I care to admit pulling apart Wankel housings and measuring wear patterns, and here is the thing most people miss. The classic textbook explanation says the rotor orbits while also rotating, completing one power stroke per rotor side per revolution of the eccentric shaft. That is correct, but it is only half the story. The real complexity is in how the exhaust ports are timed and positioned on the housing, not on the rotor. The side-port exhaust layout that Mazda used for most of the RX-7's life meant that exhaust gas evacuation was never clean. You get short-circuiting where fresh charge escapes right alongside the exhaust, and you get trapping of burnt gas near the housing walls. This is why rotary engines always ran richer than their displacement would suggest. It is not an oiling quirk or a tuning preference. It is a fundamental scavenging limitation of the side-port design.
The counter-intuitive part that nobody tells beginners is that a worn rotary engine can sometimes run better than a new one. When the housing bore goes slightly out of round from heat cycling, and the apex seals begin to wear past their initial break-in window, the clearance characteristics change in a way that actually improves airflow at certain RPM ranges. This is one reason why a tired RX-7 might feel more responsive after a complete rebuild that restores everything to factory-spec clearances. The engine feels sluggish because the fresh sealing surfaces create tighter combustion chamber boundaries than the seasoned ones did. I learned this the hard way when I rebuilt a 13B-REW for a customer who complained about poor mid-range torque. Everything measured within spec, compression was good, timing was correct, and it just felt flat. The fix was essentially running the engine hard for a break-in period rather than trying to force it into a richer tune. The seals bedded in properly and the power came back after about 800 kilometers of normal driving. The cooling system in a rotary is another area where the literature gets sloppy. Because there are no valves, the thermal load is distributed differently across the housing. The area around the spark plugs, called the crown, runs significantly hotter than the exhaust side. This means the cooling passages in the housing are not symmetrically designed. The inlet side gets more coolant flow because it needs it less, and the exhaust side gets restricted flow because the metal there is thinner to allow the exhaust ports to exist. If you are machining or resurfacing a housing, you cannot treat it like a flat cylinder head. The warpage limits are tight, and the mating surface must be checked with a straightedge and feeler gauge across the entire oval, not just at the edges. A housing that is even slightly distorted will throw off the gear train timing because the eccentric shaft bearings sit directly in the housing bore. I once found a case where a shop had machined the housing face flat but left the bearing bores out of alignment by a fraction of a millimeter, causing premature seal failure within a thousand miles. The engine looked fine on paper. One detail that deserves more attention is the gear ratio between the rotor and the eccentric shaft. It is a 3:1 reduction, meaning the rotor rotates once for every three revolutions of the output shaft. This is mechanically elegant because it means each rotor side fires once every three shaft revolutions, matching the four-stroke cycle across the three rotor sides. But it also means that any imbalance in rotor mass creates a centrifugal force that varies at the rotational frequency of the shaft, not at half that frequency like in a piston engine. The balancing shafts in a Mazda rotary are designed for this specific frequency, and replacing them with incorrectly specced parts is a common mistake in rebuilds. I have seen aftermarket balancer assemblies that looked correct but had the weights positioned for a different engine entirely, causing vibration that destroyed main bearings within a few thousand kilometers.
The 13B-REW twin-turbo version introduced in 1992 added water-to-air intercoolers and sequential turbocharging, which made the engine even more complex. The primary turbo spooled early for throttle response, and the secondary turbo engaged at higher RPM for peak power. The plumbing for the intercooler system ran through the engine block itself, which created a hotspot that accelerated wear on the adjacent apex seal. This is one reason why the later 13B-REW engines had shorter service lives than the earlier 13B-PEI versions. The thermal management of the intercooler circuit was never adequate for sustained high boost, and the housing crown near the secondary turbo outlet consistently showed excessive carbon buildup and seal degradation. If you are looking at acquiring or maintaining a rotary-powered vehicle, the realistic assessment is straightforward. These engines are lightweight, smooth, and capable of very high specific output, but they require more frequent seal replacement than a comparable piston engine requires top-end work. A full rebuild with new rotors, seals, side seals, oven seals, and apex seal retainers, done correctly, runs about as expensive as a piston engine big-job overhaul, but the parts are less available and the machining requirements are more specialized. The 12A and 13B families share many components, which helps with parts sourcing, but the 18B tri-rotor from the MX-6 Rotary is a different animal entirely with unique bearing sizes and a longer case. There is no real alternative to finding a specialist if you want the job done right. A general mechanic who has only worked on piston engines will likely miss the critical details about gear timing, housing finish, and seal orientation that determine whether a rebuild lasts or falls apart quickly.
Get the Full Details
