Europe's New Wankel

From Autocar Magazine, Week ending 8th November 1973

A look at Wankel engine history and the latest unit to appear, the Comotor Bi-rotor announced for the Citroen GS





ROTARY ENGINES go back in history a long way before Dr Felix Wankel's obsession with them before and after the last war. If you admit to the basic similarity between engines and pumps (the action of many engines can be reversed to form pumps), then an Italian engineer called Ramelli was the first in 1588, with his invention of the vane-type pump. Pappenheim, a German, followed this 48 years later with the gear-type pump, which did away with Ramelli’s slide valves but suffered from leaky seals, water or gas tightness being the weak point of all rotary designs.

In 1799 a colleague of James Watt adapted Pappenheim's gear pump into a rotary steam engine. Despite wooden scrapers in the tips of the teeth, there was a lot of leakage and a poor efficiency. Sixty years later an Englishman called Jones simplified the gear pump down to two double-lobed intermeshing rotors, this configuration later being adopted in the famous Rootes supercharger.

One of the first rotary engine patents was filed by an American called Cooley in 1901. His invention featured an epicycloid with internal meshing- gears and an enveloping casing with three fixed seals. In 1908 Umpleby, an Englishman, converted Cooley’s rotary steam engine into an internal combustion engine, but the gas sealing problems prevented the design from being successful.

Inventors throughout modern history have repeatedly been fired by enthusiasm for rotary engines, the attraction generally being smoothness of operation from the elimination of reciprocating parts and a high specific output for the weight and bulk, cylinders being arranged in a much more compact layout than the conventional side-by-side cylinder configuration of a reciprocating engine. Apart from the often impossibility of actually making some of the parts, in the main it has been gas sealing that has caused the designs to fail. Even when combustion was sustained, power outputs were very low due to poor combustion chamber shapes and the thermal problems of putting the cylinders so close together.

Born in the Black Forest in 1903, Felix Wankel acquired a keen interest in rotary engines during the five years he spent working for a scientific publishing house in Heidelberg. At the age of 21 he set up his own workshops and made models of rotary piston engines. Very soon he established that the weakness of all designs was the gas sealing and he set about beating this problem with a programme of intensive development. With the co-operation of BMW, he designed a practical unit in 1934 and in 1936 he received support from the German Air Ministry to set up a research organization. In 1942 the laboratories were expanded to undertake work on rotary-valve aircraft engines, but in 1945 the Allied Forces occupying Germany destroyed Wankel's whole operation.

Nothing daunted, Wankel immediately started work again in his Black Forest home and in 1951 his first contract with NSU was signed. In 1954 the discovery was made that a four-stroke cycle could be performed by the rotation of a three-lobed rotor in an epitrochoidal casing and suddenly the door to success was open. Two years later, an NSU motorcycle with a rotary-piston supercharger wiped the board in competition and took several world speed records on the Great Salt Lake.

In 1957 the first Wankel engine ran. This early type DKM was extremely complicated with both rotor and housing rotating. From a capacity equivalent to only 250 cc, 29 bhp was developed at 17,000 rpm. The next step was to make the rotor trace an orbital path and keep the casing stationary and the engine in this form (known as the KKM) successfully ran in 1958. Endurance trials the following year proved the reliability and in January 1960 at a German engineering convention in Munich, Wankel told the world about his revolutionary new engine.

Already the Curtiss—Wright Corporation in the USA had a licence to develop the Wankel engine, and enquiries poured in from all over the world. By the beginning of 1961, Toyo Kogyo and Yanmar Diesel in Japan had taken out licensing agreements, to be followed by Daimler-Benz, Alfa Romeo, Rolls-Royce, Porsche, Nissan, General Motors, Toyota, Ford of Cologne, BSA, Yamaha, Kawasaki and American Motors, to name only the automotive companies in the total list of 25 licensees. In 1967 Citroen set up a joint company with NSU, named Comotor, to manufacture and market rotary-piston engines, and in 1971 the Anglo-Rhodesian mining company, Lonrho Ltd, acquired Wankel's interest in the original design and thus became the recipient of a share of all licence fees.

European Development

The first Wankel engine to go into series production was a 300 cc unit developing 18 bhp at 6,000 rpm. lt was used as the power unit for a small water scooter controlled by a ski-mounted driver. From October 1962 onwards a total of 3,000 were manufactured.

ln 1959 the Wankel engine made history by powering a car for the first time. Two NSU Prinz minicars were fitted with a 500 cc unit developing 44 bhp at 9,000 rpm and sent out on long-term endurance testing. These cars helped convince the NSU management that the engine had potential, but it was still greeted with a lot of scepticism, all previous rotary engines having been considerable white elephants.

A year later and an engine specifically designed for fitting to a car was ready, six Prinz Sport coupés being built as prototypes. There followed a long period of development and an increasing undercurrent of unrest amongst NSU shareholders about the considerable financial commitment. This precipitated the premature launch of a production car, the NSU Spider in September 1964, powered by a single—rotor Wankel engine of 996 cc developing 50 bhp at 6,000 rpm. Like the early experimental engines, this power unit suffered from very poor bottom-end torque and a useful range of only 4,000 to 6,000 rpm, top end performance being limited by the seal life at that time. For this reason and the fact that there had to be a premium price on a low volume output, the installation chosen was a convertible version of the Sport Prinz, the idea being that sports car drivers expected engines with these characteristics. No more than 15 per day were produced, the total built until production ceased in 1968 being only 3,200.

A lot of valuable experience was gained in service, however, often at the expense and inconvenience of the customer, although NSU have always been very quick to replace Wankel engines free of charge. We ran one of these Spiders as a staff car and suffered excessive seal wear and a cracked rotor casing.

From single rotor engines work progressed to double rotors and more, Toyo Kogyo and Mercedes eventually getting as far as four rotors experimentally. NSU's great landmark was the launch of the twin—rotor Ro80 in 1967. For the first time in Europe there was a new car designed around the Wankel engine, the compact size of the power unit allowing a low aerodynamic nose and front-wheel drive. To compensate for the still-evident deficiency in bottom—end torque, a semi—automatic transmission was employed, with manual shifting and a torque convertor. The Ro80 engine had exactly the same internal dimensions as the Spider unit, two rotors being ganged together in series.

As early as 1964, Citroen, who have never been afraid of complication or advanced engineering, established an association with NSU, the aims being to develop a joint rotary-engined car. There can be little doubt that some of the Citroen thoughts on passenger car layout went into the Ro80.

In 1967 Comotor was established in Luxembourg as the Compagnie Europeenne de Construction Automobile, with objectives more specifically towards the manufacturing and marketing of rotary engines. Two years later a large factory site was acquired in the Saar and in June this year the first phase was opened with an initial work force of 200 and an output of 30 engines per day. Eventually the plant will be four times its present size and capable of producing between 5,000 and 6,000 engines per day.

Since the original formation of Comotor, NSU have been taken over by Volkswagen, so there is now a potential of Wankel—powered cars from NSU, Citroen, Audi and Volkswagen. Of these companies, Citroen are most in need of a new engine range, especially for their big D—range of saloons. To gain experience of the Wankel engine in production, they first embarked on a massive proving trial with single rotor units using components from the Ro80 engine. Originally it was planned that 500 experimental cars based on the Ami 8 and called the M.35 would be placed in the hands of typical high—mileage French drivers for endurance testing. Later the number was reduced to 350 and in fact between 1971 and 1972 only 260 were delivered. The M.35 was a fastback coupe using mainly Ami 8 body panels but also fitted with prototype hydropneumatic suspension which was later introduced with the GS saloon.

Logically speaking the M.35 should have led directly to a twin—rotor saloon, but Citroen were nervous about the Wankel and wanted more time to prove its reliability. Consequently the GS was launched with a flat—four reciprocating piston engine while the M.35 prototypes continued to build up a total mileage of more than 18 million miles between them.

At Last, the Bi-rotor

Development wheels turn slowly when new factories and new designs are involved, but at this year's Paris Salon the twin—rotor Citroen saloon was launched officially, even though the model will not go on sale until next March. As expected, the engine is a double version of the M.35 unit, which brings it back to being very like the one in the Ro80. Equivalent swept volume is 1,990 c.c. and from a compression ratio of 9 to 1 107 bhp (DIN) at 6,500 rpm is produced. This is 8 bhp less than the Ro80 (115 at 5,500) but the shape of the torque curve has been improved dramatically. As the graph shows, the Citroen engine's torque rises rapidly to reach 90 |b.ft. by about 1,600 rpm and is then flat (apart from a slight peak of 100 lb.ft. at just over 3,000 rpm) all the way to 5,000 before it tails slowly off. On the Ro80 engine the torque rises slowly to a peak of 117 lb.ft. at 4,500 rpm and then falls off sharply.

To increase the life of the vital rotor tip seals, Citroen use sintered iron rubbing against a nickel—silicon liner to the alloy trochoid casing. Both the rotor and its side seals are made from cast steel. To minimize the danger of over revving (rubbing speed is the critical factor for seal life), a warning buzzer is fitted which is triggered at 6,800 rpm.

The installation of the Comotor engine in the GS saloon is so neat that it must have been envisaged before the reciprocating piston engined version was conceived. Mounted transversely, the engine drives a torque converter and manual three-speed transmission in line with the rotor shafts, before the final drive which is ahead of and slightly below the gearbox. A unique arrangement is a cross-shaft with bearing on the opposite side to permit equal length drive shafts. To make the installation even more compact, the air injection pump for the anti—pol|ution equipment is mounted in front of the gearbox and driven by a thin shaft connected to the back of a jockey pulley at the other end of the engine. This jockey is driven by twin vee-belts off the rotor shaft nose, as is the water pump. A single vee-belt drives the alternator off a pulley on the air pump.

As well as air injection into a reactor, emissions are controlled by electronic ignition timing which is varied according to gear lever position, oil temperature, engine speed, air intake temperature and induction vacuum.

According to the terms of their licence, Comotor can develop Wankel engines with power outputs between 40 and 200 bhp for land vehicles. This more than covers all the Citroen needs, even the Maserati engine in the SM developing only 178 bhp. In its present stage of development, the Comotor engine would be eminently suitable for most of the D-range of cars, with the exception of the DS 23. By the time Citroen are ready with a replacement large car (1975 is my calculated guess), Comotor could easily have developed a three—rotor unit with around 150 bhp. The attraction of Wankel engineering, apart from the simplicity and small size, is that the ganging up of modules is feasible and relatively easy. For this reason I think Comotor will stick to the 67mm rotor width that has been a feature of NSU Wankels since the Spider unit designed in 1962. All the combustion development has been based on this rotor width with the radial dimension to its tip of 100mm and an eccentricity of 14mm.— However, the terms of Wankel licences stipulate an exchange of technical information with licencees and that other producer of Wankel car engines, Toyo Kogyo, have wider and bigger rotors in their Mazda RX range, so Comotor could get to 130 bhp another way.

Whatever turns up in 1975, Citroen seem to have committed themselves to a big stake in the Wankel engine and l doubt if we shall see them design another four, or even a six cylinder reciprocating engine. From what we know, the GS Bi—rotor will sell for a premium price in France next year, partly to throttle the demand and partly to allow for the possibility of trouble in service. Despite the complicated manufacturing process, however, Comotor have the advantage of starting manufacture from scratch in a new factory, so eventually the unit cost per engine must come down to the level of conventional engines. By then, Citroen believe, the few remaining snags will have been eliminated.

Rolls Royce make a Wankel

From Autocar Magazine, Week ending 17th December 1970.



Rolls-Royce Make A Wankel

Original two-stage diesel design

By J. R. Daniels, BSc

Alone among British firms, Rolls-Royce have shown an abiding interest in the Wankel principle. After six years of research. details are revealed which shows the originality of their work and its promise for the future

WE have long deplored the fact that Britain's motor industry has chosen, by and large, to ignore the Wankel engine. By following a policy of masterly inactivity they may be saving development effort and leaving themselves the option of taking up the rotary engine if it becomes really attractive; but this sort of approach is at present tightly controlled by the licensing arrangements centred on NSU. The only two British licence-holders to date are Perkins and Rolls-Royce, and only Rolls have done any serious development work. For a long time their approach was shrouded in secrecy, except for whispers that their thinking was highly original and intended primarily for application to military vehicles.

Details of their six—year development programme, however, have recently been given in a paper delivered to the Institution of Mechanical Engineers by F. Feller, C Eng, MIMechE. The story is something of a classic, starting with basic research on modified NSU engines and culminating in the design of a 350 bhp engine for military vehicle use. Even at this stage, Rolls-Royce regard the story as only half-written; much more development will be needed before the projected engine becomes a full production unit, even for the Army, while commercial prospects are even farther off.

The basic Rolls requirement was for an engine of small size and low fuel consumption. Existing standards were set by the opposed—piston two—stroke, and it was decided in 1964 by the Military Vehicles Engineering Establishment that it would be worth trying the Wankel engine as a diesel, for the sake of economy and the ability to operate on a wide range of fuels.

In fact, the Wankel in its familiar form makes a poor diesel. Its geometry makes it difficult to obtain a high enough compression ratio, and the long, thin combustion chamber has a poor surface—to—volume ratio (which results in high heat losses and poor combustion). Changing the geometry to push up the compression ratio results in a much bigger engine and an even worse surface-to-volume ratio, and so another solution was sought. The obvious answer was to pre—compress the ingoing air, by means of a Roots blower, a turbocharger, or a positive-displacement unit. Since the Wankel is itself a positive-displacement machine, the most elegant solution was to run two Wankels in series, as it were, with the first merely serving as a compressor for the second, in which combustion would take place. The first unit would then complete the thermodynamic cycle by acting also as the expansion stage for the exhaust gases.

The compressor stage must actually be larger than the combustion stage, since the initial compression depends directly on the relative size of the two units. In effect, one chamber of the compressor feeds air into the much smaller chamber of the engine proper. The basic compression ratio of the combustion stage is thus multiplied by the ratio of the displacements of the two units.

Physically, the compressor does not have to be very much bigger than the engine to give a multiplication of two or three. Together, the two-stage engine turns out to be little bigger than the equivalent single—stage one (since, as already explained, the latter must be made much bigger to obtain the same compression ratio). At the same time, the two-stage engine has a much better——i.e. lower surface—to-volume ratio.

Work was thus concentrated on the two-stage layout, with a three-rotor design as a back-up. This had separate rotors of simpler design to serve as the inlet compression and exhaust expansion stages. In the event, this alternative was not needed, but after a period in engineering limbo it is now being studied as an advanced exercise by the Royal Military College of Science.

Basic studies

Aside from the design of the two-stage engine, a great deal of basic research has been done on both combustion and on apex seal design. The combustion research resulted from the decision to use direct fuel injection rather than to` have a pre-combustion chamber in the wall of the rotor housing; there were. several design disadvantages, including leakage past the tip seals and difficult starting, which overruled the possible advantages of pre—combustion.

Combustion conditions in a Wankel are very different from those in a conventional reciprocating engine. Instead of the nice, stable column of air in the centre of the combustion chamber, and the very low piston speed around top dead centre when the fuel injection takes place, the chamber and its charge of air are travelling past the injector very fast indeed. These conditions are not necessarily worse for combustion; it was just that the engineers were working in a field where very little was known, compared with the extensive work which had been done on the reciprocating engine. Such work as had been done on the Wankel related to the petrol-burning, spark—ignition engine.

Using a small NSU Wankel engine as a test bed, 30 different combustion chamber shapes were tried, along with six different fuel injector positions. Some of the combinations were incompatible, but even so, over 100 different arrangements were tried before the best chamber shape was arrived at. In this design, the fuel is injected into a relatively wide recess carefully shaped to induce air swirl. Air for the combustion process comes from above the trailing half of the rotor, and squirts into the area through a narrow delivery channel.

It must be emphasized that this research applies to the Wankel in its diesel form, and it remains to be seen how much of it may be applied to the four-stroke petrol engine.

Seal research

While the combustion test work was going on, the little Wankel engine had to be modified to withstand diesel operating conditions. An early change was made from carbon to steel tip or seals, and it was found necessary to make new seal springs out of Nimonic 90 nickel alloy.

Early studies showed signs of misfiring which were soon traced to seal misbehaviour. Since the seals cannot be a perfect fit in their rotor-tip slots, they were tilting and jamming. The solution here (also adopted by NSU before the appearance of the Ro80) was to machine slots in the leading face of the seal to admit high—pressure gas underneath it and force it upwards into contact with the rotor housing. Later, a tendency of the seals to lose contact with the rotor housing as they passed into a lower-pressure area was overcome by recessing the trailing edge of the slots. This resulted in a 30 per cent improvement in low-speed fuel consumption. A further development was the use of a stepped apex seal, retaining the advantages of the recessed slot while also reducing the seal mass.

Again, these studies were carried out with the diesel engine in mind; but it would be surprising if designers of passenger-car Wankels did not take them into account in future.

Engine developments

The first Rolls—Royce development engine was the R1 which was conceived purely as a research tool. With a compressor stage of 1,126 c.c., and a combustion stage of 500 c.c., it produced over 50 bhp and achieved specific fuel consumptions of better than O.5lb/bhp/hour. Among other things, it was used to develop the best inter—porting arrangement between the two stages.

The R2 engine was the alternative three-stage layout, built but not investigated in detail. R3 refers to a combustion stage only, which is being used as a basic unit to build up a range of engines; it has a displacement of 1,216 c.c., and has produced 180 bhp at 4,500 rpm under test conditions.

The remaining engine of which details may be given is the 2—R6. This is a military engine formed of two banks of a two—stage engine. Each high pressure (combustion) stage has a displacement of 1,265 c.c., and is fed by a low-pressure stage of 3,250 c.c. The design power is 350 bhp at 4,500 rpm, for a weight of 939lb-a spectacular power-to-weight ratio for a diesel.

Rolls—Royce emphasize that this engine is not even running as yet, and that it will be some years before it sees even military service. But it is encouraging to see at least one British firm indulging in advanced and original Wankel research.