Origins of the British Parallel Engine

[Admin]

Kevin Cameron has been writing about motorcycles for nearly 50 years, first for <em>Cycle magazine</em> and, since 1992, for <em>Cycle World</em>. (Robert Martin/)

Now that many motorcycle makers have returned to building parallel twins for their lower cost, good performance, and market versatility, it’s time to review the circumstances in which the first really successful British twin came to market. That engine was the 1937 Triumph 500cc “Speed Twin,” designed by Edward Turner.

The classic air-cooled British parallel twin was built on a spheroidal aluminum crankcase containing a bolted-together two-bearing crankshaft with central flywheel and 360-degree crankpin phasing. Connecting rods were forged aluminum with plain pressure-oiled big-ends. On the left was an enclosed primary chain drive connected to a separate four-speed gearbox. On the right was the timing case, with its distinctive polished alloy cover, containing the gear drive for the camshafts and magneto, plus a twin plunger oil pump. Valve operation was by pushrods and rockers. This design was not only a fine compromise between function and price, it proved reliable.

The accepted synopsis is that in 1936, Turner saw that the Triumph motorcycle company, newly separated from its parent automaker and sold to wheeler-dealer Jack Sangster, desperately needed a uniquely appealing new product that would stand out from Britain’s long family of thudding singles. Turner was a man with a keen understanding of public taste, even if egotistical and skeptical of engineering analysis.

To “speak” a fresh language to the buyer, this product had to be a twin, not just a hotter single. To avoid the cost of a new chassis, it had to fit where existing singles fit. That ruled out a V-twin. To make modern power, it could not have side valves; but because the problems of overhead cam were as yet unsolved, it could not be OHC. These criteria, so the story goes, defined the classic air-cooled OHV parallel twin.

Triumph’s 1937 Speed Twin was the first really successful parallel twin. (London Motorcycle Museum/)

After Turner made a sales success of Speed Twin, every manufacturer had to offer one; BSA, Norton, Royal Enfield, Ariel, AJS-Matchless. The British parallel twin came to define an entire epoch of motorcycling. While tens of thousands of Americans learned to enjoy small-displacement Japanese step-throughs and light motorbikes in the early 1960s, British twins continued to define what the sporting motorcyclist wanted.

How and why did Turner make the choices he did? Magic insight is not an explanation.

Turner arrived in motorcycling by buying a small dealership, Chepstow Motors in SE London, which happened to have a Velocette franchise just as that company was switching from two-stroke to overhead-cam four-stroke. Turner had previously hand-prototyped his own motorcycle design, also an OHC single, and had shown it to the press and offered a later build for sale. Anyone who has built a complex machine piece by piece knows this is a slow, hard game, though one with much to teach the player. Velocette’s price for its production 350 OHC single was far below what Turner’s design cost to build. This told him his designs were useless without a proper factory to produce them.

He talked his way into the actual motorcycle industry. At Ariel, he evidently impressed wheeler-dealer Jack Sangster enough to be hired as a designer, working with the very experienced Valentine Page, lately come from J.A.P., and draftsman Bert Hopwood, later to be an influential engineer in his own right. This was 1929, with the Great Depression about to hit. Hardly an ideal time to launch something as complex as what Turner had sketched: a twin-crankshaft four-stroke OHC square four.

Driven by hot ambition and personal certainty, Turner worked long hours to push his square four and other projects into being. Unwilling to sharpen pencils for other men, Turner achieved and talked his way into command of Sangster’s next bargain acquisition: the motorcycle side of Triumph. The deal was made in mid-January 1936.

It was here that Turner’s creativity would bring success. Shifting his thinking away from innovative complexity to production practicality, in 1936 he drew a parallel twin that fit into the chassis of a single, had a completely different sound, and propelled a motorcycle much more smoothly than any single. It was shown to the press at the end of July 1937 and entered production in 1938. Turner’s design introduced little that was new, but broke away from decades of single-cylinder sameness in ways buyers did not resist.

Turner had given his original square four complexity that required multiple redesigns. He gave this new twin a simplicity achieved by combining proven existing elements in the most basic way. It would reward him by selling in great numbers for over three decades.

Firing Order

Why, in trying to replace the sameness of British singles with a twin, would Turner give his new design a 360-degree firing order, with both pistons moving up and down together? Didn’t that vibrate like a single?

Bert Hopwood, Page’s draftsman at Ariel, remembered, “While Page was still with us (he left for Triumph in 1932) we had modified a square-four engine by discarding two pistons and had developed a parallel twin in this way.”

That entailed assembling the two rear crank halves onto their central flywheel at a 360-degree crankpin spacing, rather than the original 180, and provision of a suitable camshaft. All present were surprised at how smooth this 250 twin was. When Page arrived at Triumph in 1932, part of his work was design of an engine that is acknowledged as the first British parallel twin, the 6/1, which he gave 360-degree crankpin spacing. Why? This gave an even firing order that allowed operation on a single carburetor. It also did not rock from side to side as 180-degree parallel twins do.

Turner, arriving as general manager at Triumph during 1936, was annoyed by what he regarded as the unnecessary complexity of Page’s 6/1 parallel twin, so he did away with it and began his own design.

Turner Influenced by the Riley ‘Nine’ Auto Engine?

In the spring 2002 issue of The Antique Motorcycle, English journalist David G. Styles suggested that Edward Turner adopted features of the engine in his Riley “Nine” auto (a model originally launched in 1926) in his “Speed Twin.” The Riley Nine sought the reduced valve train weight of OHC by mounting its cams high but not overhead, one on each side of the cylinder block. This allowed use of lighter short pushrods. (In this same period, Vincent featured “high cams” on its new 500 single.) The Riley’s engine had deep hemispherical combustion chambers, each with two valves at an included angle of 90 degrees. All of these did become features of the Triumph Speed Twin, but the hemi chambers and 90-degree valve angle were already present in Page’s Triumph 6/1.

Forged Aluminum Connecting Rods

An unusual feature of most British twins, originating with Turner, was use of forged aluminum con-rods, some running directly against the crankpins as their own bearing material (although replaceable insert bearings were later adopted). Anyone who, like Turner, had spent time at Brooklands Speedway, knew of W.C. Devereux and his company, High Duty Alloys (formed 1927). Brooklands brought together people from automotive, aviation, and motorcycle backgrounds, making the racetrack a practical university of modern transportation. Frank Halford, who would later design Napier’s 3,500 hp H-24 Sabre World War II aircraft engine, had been an associate and friend of Harry Ricardo, racing his Triumph-Ricardo 500 in the Isle of Man Senior TT of 1922. Halford had initiated the use of forged aluminum pistons and con-rods in his 1,500cc AM Halford Special racing car of 1925, and then put RR56 aluminum rods in his mid-to-late 1920s Cirrus engine for light aircraft. RR56 was one of a new family of aluminum alloys whose unusual strength came from precipitated intermetallic compounds of copper, iron, and nickel within them.

Why would Turner choose aluminum? He is said to have remarked that aluminum’s lighter weight allowed such rods to be light while giving greater bending stiffness. But there may be another aspect. The two pistons and rods of a 500cc twin do not together weigh more than the single piston and rod of a 500 single, and in fact may very well be lighter. The result could be reduced vibration. Knowing this, and likely being familiar with Halford’s use of aluminum rods in his projects, made such rods attractive to Turner. Therefore, during his second redesign of the Ariel Square Four in 1935, he chose rods forged in RR56 aluminum alloy. The following year, drawing the Speed Twin, he again specified such rods.

Why Put the Flywheel at the Center of the Crank?

Turner’s Speed Twin has a two-bearing 360-degree crankshaft with its flywheel at the center. Why not locate the flywheel at one end of the crank, as on Guzzi’s 120-degree twin or BMW’s flat twins? This was implicit in Turner’s vision of making the new engine fit existing chassis designed for a British single, an engine whose large-diameter crankshaft is its flywheel. He was just putting the flywheel of his new twin where room for it already existed: at the center.

He had already done just this when hired at Ariel in 1929 to build his original 498cc square four. To make it narrow, Turner gave this engine’s two crankshafts central flywheels, doubling as crank phasing gears by having helical teeth cut into their outside diameters (this was later dropped). As on Scott two-stroke twins, the early Square Four’s roller crankpins were “overhung”; they were not supported by outer crank discs and outer main bearings, narrowing the engine by the width of the two deleted crank discs and main bearings.

Turner’s Square Four achieved both primary balance, meaning as much reciprocating mass was, symmetrically, always going up as was going down; and the elimination of any side-to-side rocking couple by phasing diagonal pairs of pistons to rise and fall together. This same self-balancing arrangement has appeared in more recent engines, such as Suzuki’s 1970s RG500 two-stroke.

Only Two Main Bearings?

Speed Twin was given only two main bearings, a ball on one side and a roller on the other. Why no center support from a third bearing? The obvious answer is that, at the Speed Twin’s max of 6,000 rpm, there was no need. And in the mid-late 1960s, when the Meriden Triumph organization got scientific about winning Daytona and even entered GP races with a twin, it was able to rev the bike to nearly 10,000 rpm. So why worry?

Turner Capable of Error

In design, Turner could on occasion have more personal conviction than understanding. Bert Hopwood, who worked with him for years, observed that Turner dismissed analytical calculations as “schoolboy work” irrelevant to practical design. That personal shortcoming doomed his Square Four to limited power in relation to its growing weight, because its rear pair of cylinders were “cooled” by air that had already been heated by passing through the fins of the front cylinders and heads.

Apologists for the square four have blamed that design’s chronic overheating on anything but the obvious: You can’t cool second-row cylinders with hot air. Successful radial aircraft engines have been built with as many as four rows of cylinders, but always with careful provision of 100 percent fresh cold air to every cylinder. The square four’s design could cool adequately in the early days of low power, but when worked harder, trouble appeared. In the end, Turner’s Ariel Square Four was no Vincent-like road-burner and accepted the role of hefty and sedate sidecar hauler.

An Earlier Precedent for the Central Flywheel?

In 1947, when Bert Hopwood, by then a respected engineer in his own right, was designing Norton’s “export twin,” he too adopted the central flywheel, taking care to alter details to avoid patent infringement action from Triumph. As a second line of defense, he found a design using a similar bolted central flywheel that considerably predated both his own and Turner’s. That was in Harry Ricardo’s 1922 design for the “Vauxhall TT,” an inline auto racing four. Just like Turner’s later Triumph Speed Twin, it joined its two crank halves to a central flywheel by means of flanges and bolts.

Val Page, who left Ariel for Triumph in 1932, updated that company’s singles and designed a 650cc parallel twin, the 6/1, making 25 hp at 4,500 rpm. Many of its features would appear on Turner’s Triumph Speed Twin four years later: deep hemi combustion chambers with two valves in each, set at a 90-degree included angle; OHV valve operation by pushrods and rockers from a gear-driven camshaft; 360-degree crankpin phasing; dry-sump oiling via a plunger oil pump, Val Page’s usual choice.

Which example, if either, did Turner follow? Or was the two-valve deep hemi chamber with a 90-degree or greater included angle just “in the air” among designers since being pioneered in Fiat’s early 1920s GP car engines?

By the time Edward Turner arrived at Triumph in 1936, he’d lost his appetite for complexity (those square-four redesigns!) and for needless expense (the herringbone primary gear drive in Val Page’s Triumph 6/1!). Once at Triumph, he had a clear idea of what he wanted in a new twin.

Design Details?

As valve operation by overhead cam was in vogue, and because he had adopted OHC for the 498cc and 601cc versions of the Square Four, his first drawings for a new twin included it. He soon reverted to pushrods and rockers. Speed Twin’s crankshaft “was built up from three parts, a central flywheel and two single [forgings], each of which formed a crank, bobweight, and mainshaft” (page 21, Whatever Happened to the British Motorcycle Industry, Hopwood).

Connecting rods were of forged RR56 aluminum. The double-piston plunger oil pump that Page had given all models at Ariel in 1929 was chosen to operate Speed Twin’s dry-sump oiling system, driven by an eccentric pin on the end of the intake camshaft. The smaller plunger supplied oil pressure to the right end of the crank, thence by drillings to the two crankpins, while the larger plunger scavenged the oil that drained to the bottom of the two-piece crankcase, returning it to an external oil tank.

Choosing an Oil Pump

How was the plunger oil pump chosen? In the 1920s motorcycle engines gave up total-loss oiling in favor of pumped recirculating systems, but there were alternatives. BSA chose double-gear oil pumps, and Turner put one into his early square fours. Val Page adopted plunger pumps, possibly for their reputation of being better able to lift oil after long use. Because cavitation could develop at high pump speed, it was usual to drive oil pumps from a camshaft, rotating at half crank speed.

Although Turner had originally given his Square Four a radical (for the time) unit three-speed gearbox (“unit” meaning located within the main engine castings), his learned distaste for novelty and the unproven put a conventional separate four-speed Triumph gearbox behind the Speed Twin’s engine.

A young Edward Turner. (meriden.triumph/)

Dry Sump or Wet?

Why dry sump rather than wet? Turner had given the original Square Four a wet sump, with roller connecting-rod big-ends lubricated by “dippers” and a gear-type oil pump. If a wet sump is designed with a flat bottom, oil can move away from the oil pump pick-up during acceleration or braking, interrupting oil flow to the pump. A dry sump was a solution to this problem, as its oil flowed by gravity to the oil pump from a remote oil tank.

The Speed Twin’s valve tappets were paired in round tappet guides, one before and one behind the cylinder; the guides looked just like those adopted by Turner for the 1935 Ariel 4G 997cc Square Four. Cams were supported by bronze bushings, not by the ball bearings of the Square Four.

Although Val Page had designed his 6/1 parallel twin with exposed valve springs, Turner saw that full valve train enclosure was the future.

Cam Drive: Gear or Chain?

Bert Hopwood, in his later British twin designs for Norton and BSA, drove their cams as Harley-Davidson does today on the Milwaukee-Eight: by quiet chains, rather than by gears whose clicking backlash increases as the aluminum crankcase warms up and expands.

Turner drove his cams by gears. Hopwood noted in his book that “…the characteristic ticking and clicking noises seemed to be accepted and forgiven by Triumph enthusiasts” (Hopwood, page 30).

Imitation the Sincerest Form of Flattery

Because Turner’s Speed Twin was the first British Twin to achieve market success, its design inspired competitors to copy—and improve. An example of this is the more compact shallow combustion chamber with the narrowed valve included angle of 57 degrees, given to Norton’s line of OHV twins by Hopwood in 1947.

Beyond scholarly argument and historical documents lies the fact that Turner had real insight into what motorcyclists wanted and would buy. He may have been as disagreeable as some contemporaries remember him, but it was his version of the parallel-twin concept that became the enduring sales leader.

Who may have copied whom is unimportant. The bottom line is that having educated himself through hard experience, Edward Turner chose for his Triumph Speed Twin the simplest features that added up to attractive functionality, produced at an affordable price.

View the full article