The often talked about, but rarely accomplished...

13B based 4 Rotor

How many times have you heard someone say that they were going to build their own 4 rotor engine, and that they heard it was as easy as stacking the rotor sections of a coupla engines together?

Well here's a 4 Rotor I built, but it's the result of a couple years of brainstorming, much frustration, and several thousand cans of RC Cola. The following is a brief summary describeing what worked for me.

The engine design uses modified 13B eccentric shafts, and the two engines are joined by a custom made 3/8" thick aluminum adapter plate.

Removed from the front of the rear engine are the front cover, oil pump, thrust bearing, counterweight, water pump, etc. Pretty much wiped clean with the exception of the stationary gear and modified rear ecc. shaft. The 3/8" plate bolts to the rear engine first, as internal bolts are needed to seal passages exposed when the front cover is eliminated from the rear engine.

A round donut shaped pilot ring was machined, which lightly pressed into the recess in the exposed front stationary gear on the rear engine, and into the seal groove of the rear stationary gear of the front engine. This pilot ring served to accurately align the front and rear engine's eccentric shaft bores, while allowing the e-shaft coupling to pass thru it's center. O-ring grooves in the pilot ring seal the front engines' rear stationary gear to the adapter plate, preventing oil leaks from inside the front engine's rear housing.

The 2 lower exposed (outside bottom) tension bolt holes are bored out and the stock bolts replaced with very long bolts that go thru both engines, to pre-tension the lower part of the engine, as I used only front and rear motor-plates, and no mid-plate support.

At the heart of the design is the coupling, which has to take a lot of abuse, yet remain small enough to fit thru a main bearing to allow assembly. The coupling design requires welding up the eccentric shafts (or building new shafts from scratch), which lengthens the full diameter portion of the shaft adjacent to the main bearing journals, so that the coupling can be machined into them. It is a tapered/pinned design that requires a draw bolt passing thru the bored out center of the front ecc. shaft, threaded into the male half of the coupling (which is machined into the rear e-shaft), to draw the two ecc. shaft tapers together. The tapered portion was about 1-1/2" long, with a series of (12) 3/16" dia. x 2" long aluminum pins arranged in a circle about the taper parting line. As the pin holes are bored parallel to the e-shaft centerline, the parting line passes thru the pin at an angle, with 1/4" of each end of the pins firmly rooted in each of the shafts. The drawbolt thru the bored out center of the front e-shaft pulls the coupling together, and allows the front engines' thrust bearing to work for both shafts.
By using the tapered/pinned design, the loads on the coupling are effectively spread out over a much larger cone shaped area. Less stress concentration, than what would have been present with any other type of useable coupling we could dream up, was an added bonus. All this in such a compact and relatively inexpensive design.

Because the design removes what would have been the 2 middle counterweights, the two engines were indexed the same (#1 rotors for each fired at the same time). This puts the two middle eccentrics 180 degrees (in rotation) from one another, thus canceling the need for center counterweights to maintain balance.

For the oil pump, I use the big oil pump on the front of the front engine, which feeds both engines, and modify the drive ratio. By machining off and welding the toothed part of another oil pump sprocket onto the eccentric shaft sprocket, it is possible to drive the oil pump at twice speed (1:1 instead of 1/2 speed), which roughly doubles the volume. The special drive chain required is made from parts of 2 chains put together, as I could not find a source for the odd pitch chain.

A separator/de-airiation plate was made that fit across the entire bottom of the assembled 4 rotor, which helps seal the area between the engines for the full length, baffled and gated custom built oil pan. A special larger pickup tube is made and the oil circuits "ported" to ease the demands on the pump.

The water pump housing is modified so that the output can not flow into the front engine from the front housing, but instead is re-directed into an external distribution manifold. From here, the coolant is split and directed into the top of the center housing of each engine. The modified rotor housing coolant flow is into each center housing from the top, clockwise (from the front) around to the top on the right side, cooling the combustion side of the engine first, then to exit ports that return the coolant to the radiator. No thermostat is used, only center housing outlet restrictors to regulate coolant flow rates and engine temperature.

For the ignition system, my 1st design fired only 1 plug per rotor, and allowed using the normal point type distributor. Points were adjusted to fire leading/trailing at the same time, with leading plug wires going to the front engine, and trailing wires to the rear. A more modern ignition would be a direct fire using double posted "waste spark" coils, one post for the front engine, one post for the rear. 4 coils would be used.

Starting the beast was harder using 1 plug per rotor, so a 24v starting system was used. The added voltage increased cranking speed, but the starter would overheat at anything over 20 secs.(melted the solder in the brushes). A system that uses 18 volts cranks almost as fast, as is not nearly as hard on the starter. If you use both plugs per rotor, the 18v system will work for you.

During the duration of the project, which included quite a few nites racing spread out over a year and a half, the couplings experienced no damage at all. The draw bolt, which I felt would be the first external indicator of failure, maintained it's assembly torque with no sign of any loosening. When the engine was finally torn down (after an oil line was torn off during a race), the coupling and all it's components looked like new. As proof of this, when the engine was assembled, the coupling pins and taper was coated with loctite, which I felt would fill any minor discrepencies, and improve the strength of the coupling. Upon diassembly, the shafts were quite hard to seperate, as the loctite bond was still intact.

For the type of racing that the engine was built for, eliminating as much rotating mass as possible is a much sought after goal. The effects on accelleration can be very surprising. By joining the two engines as I did (using only 1/2 the normally required flywheel and counterweight mass), the engine rev'ed much faster than a single engine by its self. Although the exhaust note might be more pleasant if the engines were phased 90 degrees from one another for an even firing order, the reduced mass and synchronized peak torque pulses make for less stress on the coupling, and the reduced rotational inertia over that of two separate engines far outweigh any performance advantages of the smoother sounding 90 degree firing order.

The Outlaw Dirt Cars were one of the few places that the 4 rotor car was allowed to run, so that's where we ran it. In the form shown here, output was around 550hp. Even at this low power level, the engine held it's own at times against the $25,000. aluminum blocked V-8s it ran against. Turbos were not allowed, so plain 'ol low octane pump gas was used. Race gas seemed to hurt power and make the headers glow. Thought was given to increasing the compression ratio by filling in the depressions on the rotor flanks, either by metal spraying them or bolting on some pieces of metal, but it was never done. We had planned to slip in some nitro, which would have helped with the relatively low compression ratio, but never got around to it.

Both the engine and the car were specially designed & fabricated by myself, and burned up a couple years worth of free, and increasingly not so free, time. Eventually, growing demands on my time forced me to cut back on the long hours the racing program consumed. After a while, I found myself dreaming of a more conventional power plant. One that I could actually BUY parts for, without having to spend endless hours modifying before I could actually use them.

After some damage from a ripped off oil cooler line, the 4 rotor was replaced with a 700hp alky burn'n small block Chevy, which proved more than competitive, and consumed way less time.

Was it fun? doing over 100mph on a short dirt track, tires blazing, rubbing fenders with the competition, all with a 5' high steel wall sometimes less than inches away? fun just doesn't describe it.

The series only lasted a season longer before being cancelled, so the car was eventually sold. Last I heard, it was somewhere in Oregon.


an addition featuring a description of the coupling making process, and a few of the other key components, including some dimensions taken from some old notes from the build in the early 1990's (before we had a computer or digital camera)...

CLICK HERE to check out the old scratch paper notes!!!


From here can seen the aluminum adapter plate between the two engines, which features the lifting eye that you see sticking out of the top/center of the engine. The external cooling plumbing can also be seen entering and leaving the center housings, as can the dual oil filter setup, which serves as a distribution manifold for supplying oil to the rear engine. The huge airbox filters out the dirt clods using a reinforced furnace filter (hey, I was on a budget). A Chevy power steering pump assists a Ford power rack & pinion steering gear. Unseen behind the firewall is a custom made planetary "quick steer" box, which speeds things up quite a bit, resulting in 1 turn lock to lock steering. No need to take your hands off the wheel, no matter how out of shape you get!

The car we built for the 4 rotor

Of the few rules the class had, one was that the front-end had to be stock appearing. We said the nose looked like a 1st gen RX-7. The tech inspector bought it.

Stock car racing on dirt is a contact sport, and body panel replacement is just part of the game. Big reason dirt Stock car bodys are made of flat sheet aluminum? - you get tired of building fancy panels.

The transmission is a hand fabricated unit, featuring only direct drive and neutral. A small multi-disc motorcycle clutch in the countershaft is used only to get the car rolling. Engaging dog ring on the mainshaft allows power to pass directly thru, bypassing the clutch. An aluminum flex plate, mounted on two bearings and one-way clutch, supports the ring gear. The one way clutch allows the starter to start the engine, but after the engine starts, it is no longer required to spin the flex plate. The result is the car could start and drive onto the track as per the rules, but didn't have to pay the penalty of the extra rotating mass of a conventional multi-disc clutch and flywheel. They never did figure out that one.

From here, you can see what can be a somewhat puzzling rear suspension. The basic design is a 4 link, in conjunction with an adjustable watts linkage, but here's where the dirt cars can get different. The 4 link brackets are not welded to the axle tubes, but allowed to rotate on greasable sleeves. This allows the rear axle to rotate in each sleeve independantly, eliminating the binds from their linkage. The axle housing rotation is controlled by a torque control rod attached to the aluminum plates atop the Quick Change centersection. The other end of the torque rod is attached to a spring and shock controlled "reaction arm", anchored to the chassis just to the right of the steering wheel. Barely visable is the blue specially made minature coil-over shock that controls the reaction arm. The reaction arm controls the wind-up of the rear axle, providing a degree of cushioning to help maintain traction over the rough and sometimes slick dirt track surfaces. The initial angle of the rod produces chassis lift to plant the rear tires on accelleration, but as the miniature coil-over is compressed, the angle levels out, reducing the vertical bind that an angled rod might otherwise have on the suspension travel. It works in reverse during de-celleration, controlling the initial pitch when the rod is closer to horizontal, then gradually increasing in angle to add squat to the rear.
The brake calipers are also mounted on floaters (the RR control rod can seen just inside the RR tire). These allow the individual brake reaction forces to be seperately tuned, which can really help when you are at the limits of your traction on a rough dirt track.
This all worked quite well, helping to get the most out of the sissy 14" wide hard tires we were forced to run by the rules (hey, the Sprint Cars were lighter and got to run soft 18" rubber). The suspension and steering featured 35 rod ends (not counting the throttle & brake balance bar linkages), each of which had to be cleaned and lubricated every week to keep things working freely.
Because of the low torque/hi-RPM characteristics of the rotary engine, we had to use the steepest change gears we could find. Typical rear axle ratio used was 8.50 & up.


Although the scrub radius of this geometry is huge, it can actually be benifical on a dirt track, producing an automatic jacking effect that changes the rear tire loading as the steering wheel is moved from side to side. The spindle and control arm geometry was all special to compliment the Ford steering rack.

Here's another unusual Dirt Late Model we built back in the '80s...

This one was built to beat an engine setback rule. The entire BB Mopar engine was reversed in the chassis, placing it's water pump, distributor, timing gears and oil pump on the "rear" of the engine, and only a lightweight ring gear and starter on the front. The net effect was like an additional 4" of engine setback. The mandated setback was 10-1/2" maximum from the upper balljoint centerline to the center line of the #1 sparkplug hole. Some argued that the #1 sparkplug was now at the rear of the engine, but luckily for us, the stock BB Mopar intake manifold is reversable (no water passages) and has it's runners numbered. We bolted it on and clearly pointed out the factory numbering sequence, with #1 being up front. Nobody said we had to keep the stock firing order, so we marked it as 1563487 instead of 18436572 because of the changes.
As a bonus, the reversed engine gave us a reverse-rotation drivetrain. The quick-change rear had it's centersection modified to allow flipping the diff carrier over, placing the ring gear on the other side of the pinion, allowing the rear tires to spin in the right direction.

Here's a closer look at the little home-made transmission that's used in the backwards engine car, same one that was later used in the 4 Rotor car. Tha aluminum cover on the right of the pic is the hydraulic cyl that was used to apply the motorcycle clutch that was located in the countershaft. The long rod extending from the right side is the spring loaded shift rod, which is basically moving a dog ring slider to engage direct, bypassing the countershaft clutch for racing. The input shaft was driven off of a splined hub that was bolted to the BB Mopar's balancer. The Balancer got 2 more "keys" added between it and the crank snout to prevent shearing. A small fabricated bellhousing mounted the transmission to the front (or rear?) of the engine in the picture above, but it was originally mounted back 18" farther to the chassis and used a really long input shaft.