An engineering dream

In looking at the Itala as a “Forgotten Make” in the May issue I touched on how this car was one of the only successful ones to use rotary (instead of poppet or sleeve) valves and then only prior to the First World War and for a limited number of models. For, you see, from that time up to the Cross and Aspin engines with such valves in the late 1930s, rotary valves have never been able to oust the poppet valve in universal usage They remain an engineer’s dream, a rather unhappy lost cause. . .

The reason why Itala had some success with their rotary valves was due to a number of reasons. In the first place, they used vertical rotaries which each served two cylinders of an engine, and thus they ran at one-quarter crankshaft speed. Their design enabled efficient water cooling of the rotating element to be used. Finally, the standard of manufacture of Itala cars was very high; indeed, one authority has said that “for superfine manufacture and design the Fabbrica Automobili Itala SA of Turin has been a household word in the motor-car world for over a generation and the standard of reliability achieved for their products in the mountainous country of Italy, if equalled by others at the period, was excelled by none”.

So the Itala rotary valve had everything going for it! If at first it is difficult to see how one valve rotating at 1/4-engine speed was made to serve two cylinders of a four-stroke engine, this was achieved by having double-ported rotors feeding a single port in each cylinder, the exhaust gas passing through the top of the valve-rotor, fresh mixture through the bottom of the rotor. Two piston rings prevented gas leakage between the two parts of the rotor. (I confess I do not fully understand it, but we will leave it at that.) A simple pressure-balancing device ensured that excessive friction on the valve due to pressure on the firing stroke was obviated, although the valve rotors had to be a good running-fit in their casings; exactly how good was never disclosed. A centrifugal water pump fed coolant through the valves, a feature exclusive to the Itala system, which it was claimed gave substantially the same fit or clearance to the rotor when hot as when cold, due to the external and internal cooling of the rotors.

The drive to these vertical valves was from a flexible helical or skew gear on the engine crankshaft, acting as a vibration damper, that meshed with a similar gear on a side shaft carrying two worm gears, one left-handed, the other right-handed, which rotated in opposite directions, the two rotors serving two engine cylinders. As mentioned in my Itala article, fail-safe Oldham-type couplings connected the valve spindles to their drives; although this was described as no evidence of weakness, only of foresight and sound engineering practice, the fact is that these couplings did break on occasions, preventing damage should a valve tighten up. The Itala rotary valve, engineered by Alberto Balloco, was first seen in Britain at the 1911 Olympia Show, on the 35 hp 105 x 150 mm car; at the same time a 25 hp 90 x 130 mm engine and a 50 hp 127 x 150 mm engine with these valves was on exhaustive test in Italy.

Experts said that these rotary valves were well thought out in every detail, and beautifully constructed. Itala used them for racing and their total retirement from the 1913 French Grand Prix was not, apparently, through any fault of the valves, while at Brooklands a private owner, Robertson Shersby-Harvie won many races with his 8345 cc rotary-valved Itala, which was timed over the half-mile at 119 1/2 mph and was good enough afterwards to be used as a touring saloon for some years. Yet after the war the Itala rotary valve engines were dropped. The reason may well have been, as was the opinion at the time, that these engines would never be part of an inexpensive car, as they were costly to make, because to overcome every known technical defect that might have intruded with these valves, the highest grade materials and workmanship were employed in their making, calling for lavish expenditure, and the use of too many parts, for this to be an economic proposition unless a large market could be obtained. So the most promising design in this pursuit of greater reliability, quieter-running and more durability than the then-current poppet-valve engines were thought to give, faded out. (Incidentally, in America, at that time, a system similar to that used by Itala was employed in Russell engines used in farm tractors and lighting-plants, but this dual vertical rotary valve was dismissed as an example of expending a vast amount of attention and ingenuity to overcome possible adverse characteristics in practice, rather than designing out any such possibilities.)

Another well-known car-maker who sought to use rotary valves was Alexandre Darracq et Cie of Billancourt, Paris, who also introduced an engine of this type at the 1911 Olympia Motor Show. The venture was a dismal failure; after only a few cars had been sold with this Henriod power-unit they were withdrawn and production of them ceased. In this case the simpler scheme was adopted of having the rotor alongside the cylinder block and running at half crankshaft speed. A scooped-out section of this long rotor, one for each cylinder of the four-cylinder 95 x 140 mm engine, provided the passage for the inlet and exhaust gases. This method had two serious shortcomings. One was that, as in a simple two-stroke engine, some exhaust gas mingled with the incoming charge on the exhaust stroke of the piston. The other was that in the same way the rotor-scoop could not prevent some of the new charge being lost to the exhaust at the end of the induction stroke. Timing could to some extent cure the first defect but not the subsequent gas loss, so low volumetric efficiency and a high fuel consumption resulted. Moreover, timing to get the best effect was complicated, because to avoid pressure on the firing stroke being transmitted to the valve rotor it was masked by the piston for one-seventh of the stroke.

It seems that the Darracq engine also suffered from poor lubrication of the rotor, heavy frictional losses over the whole of it augmented by pressure from within the engine cylinders, and temperature distortions leading to early wear, and thus leakage, from the valve. So Darracq quickly returned to poppets. . . These were by no means the only sans soupape engines being built at this period. For example, the gunmakers, Horsley & Sons Ltd of York, converted a four-cylinder 102 x 127 mm engine in 1913 to run with water-cooled disc valves in the cylinder heads, which developed over 42 bhp at 1400 rpm. Between the wars Minerva of Antwerp experimented with a rotary valve invented by M Bournonville, a Belgian residing in New York. Perhaps Minerva was tired of paying royalties on the Knight double-sleeve-valve engine they were using, or sought a means of obtaining similar quiet running without the costly complexity and lubrication difficulties associated with the American Knight power-unit. Bournonville used a cylindrical rotary valve running the length of the cylinder head, running at only one-sixth engine speed, which possessed three pockets per cylinder. The sequence was thus: port one open on the inlet stroke, ports two and three closed for the compression and firing strokes, port two partially open for the exhaust stroke, port two open for the exhaust stroke, port three partially open on the next inlet stroke, the valve rotating anti-clockwise, opposite to the crankshaft’s rotation.

Thus the new gas cooled the exhaust port, helped by the low rotor speed, 666 rpm at 4000 engine rpm. A special elastic bearing of wedge type made it almost impossible for the valve to seize, and valve lubrication was carefully carried out, one drop of oil being fed by four leads to the wedges and the valve every 150 rpm of the engine, by a distributor on the valve shaft. The valve could be easily dismantled, using ordinary tools, in less time than it took to withdraw a poppet valve. It sounded just the job! The Minerva engineers did bench and road tests with two converted engines, an 80 x 112 mm four-cylinder with big ports and the then high compression-ratio of 5.9 to 1, and a six-cylinder with a 5.2 to 1 c r. The smaller engine gave 60 bhp at 4000 rpm, the six-cylinder 34 bhp at 3000 rpm.

One rotary-valve Minerva was run on the road for eight days and nights non-stop, at 4000 rpm whenever possible, and then for a day behind another car over very dusty roads. It was reported that it was practically impossible to seize-up the rotary valve, pistons, big-ends and main bearings seizing before defects showed in the valve, and that quietness and flexibility of running were outstanding. Fuel thirst was that of a normal engine, oil consumption rather less than that achieved with Minerva-Knight sleeve-valve cars. The experimental engines had fixed heads but the Bournonville valve was easily adaptable to a detachable head. In fact, it is an interesting reflection on the difference between maintenance then and now that it was stated that, although the valve-lubricating distributor was unlikely to cause trouble, if an aperture was cut in the car’s dashboard it should be possible to withdraw it in five minutes!

The Boumonville engine was easier to make than a sleeve-valve engine too, which could have been what Minerva was looking for. Yet they didn’t adopt it. One possible problem was that, should the engine back-fire and momentarily rotate the wrong way, the valve-wedge would be drawn inwards and might lock-up the rotor. Or was it felt that correct adjustment of the wedge’s springs was too complicated to be left to ordinary mechanics; or maybe the balance between friction, heat dispersion and gas leakage in this department were critical? Or perhaps the complexity of driving the valve was against it? Whatever, Minerva continued to be a Silent Knight to the end of its days.

In Britain R C Cross commenced experiments with rotary valves in 1922. He obtained some remarkable results from motorcycle engines, although the introduction of sodium-cooled poppet exhaust valves rather reduced his high c r claims. His later engines incorporated “controlled loading” of the overhead rotary valve which has been called “an achievement comparable with the Mitchell thrust-block, if of less universal application”. Cross engines also had patented scraper lubrication and port edge sealing. One such engine went into Halford’s racing HRG. F M Aspin’s rotary valve, a rotating conical sleeve in the cylinder head, was equally acclaimed, after plug overheating and seizure of the valve before oil reached it had been overcome. Both engines were made in various sizes and for different purposes and Aspin claimed that a 249 cc motorcycle engine with a 14 to 1 c r would run at 14,000 rpm for long periods and after over 700 hours showed no undue wear.

Yet in spite of the seeming crudity of poppet valves land reciprocating pistons) both remain in 99% of car engines to this day, thanks to metallurgical science, and rotary valves in this context are as dead as the proverbial Dodo. With the complexities of ensuring enough lubricant to satisfy the rotors without fouling the plugs and of preventing gas leaks where piston ring techniques hardly apply, it is easy to see why. For low-speed engines there was more success. For example, the National Gas & Oil Engine Co of Ashton-under-Lyme has gone on record as stating that some of their rotary valve engines dating back to 1895 were still being serviced in 1935, and a 1903 Crossley vertical rotary-valve gas-engine, used for driving machine-tools in a private workshop, would toil away hour after hour when it was quite old; if it overheated, all that was necessary was to clean the valve’s “keep” with paraffin, apply a fresh supply of oil, and reassemble, although adjustment of the “keep” spring was a delicate operation. W B