HOW DUCATI BEAT THE WORLD
World Superbike’s rules were intended to put twins and fours on a level playing field – but nobody counted on Ducati’s fabulous desmodromic valve system
How Ducati’s clever desmodromic valves gave them the WSB edge
World Superbike was conceived as a road-based formula in which four-cylinder bikes were limited to 750cc and twins could be 1000cc. Many said this gave Ducati’s 851cc V-twin a massive and unfair advantage over bikes like Honda’s RC30 748cc V4, and was handed out to ensure a competitive Italian contender. Is it true? Let’s grab a metaphorical back of an envelope (the engineer’s greatest tool) and do some sums…
Limits to engine performance are many and varied, but firstoff the reality is engines are air consuming devices. How much power they make is proportional to how much air they can flow. Both Ducati’s 851 and Honda’s RC30 were four-valve engines, and the 851’s bore was 92mm while the RC30 had 70mm. Assuming valve throats the same proportion of their bore diameters, then on a per cylinder basis the Ducati could flow about 73% more air (remember, we are talking areas here). However, the Honda had twice as many cylinders, and so might be expected to flow about 15% more air than the 851. Pretty much in-line with the difference in swept volumes.
However, smaller cylinders can be expected to produce more power than larger ones for other reasons. Among these is the ability to rev harder due to lower mechanical stresses, and the fact the flame does not have to travel as far to complete combustion (so knock is less likely). Larger cylinders waste less heat, but the total amount of heat lost to the piston, putting it under stress, is larger. To get an idea for the mechanical stress a piston is under (and, let’s face it, the piston has the hardest time in an engine) we can use the ‘mean piston speed’, which is essentially a product of twice the stroke and the engine speed. For a high-performance four-stroke of the late ’80s this was in the region of 20m/s, and at maximum power the production Ducati’s mean piston speed was 20.5m/s. However, for the Honda it was only 17.8 m/s, so the V4 could rev harder before hitting a mechanical limit and be tuned to give more within it.
So far, one would probably back the engine with more cylinders to have the advantage, so there are sound engineering reasons for letting twins be larger. Though the discrepancy becomes smaller as the ratio of the number of cylinders reduces, it is clear that the twin’s engineers will have to use whatever tools they can to make a difference. In the case of Ducati, the ace up their sleeve was desmodromic valves.
Rewind slightly. Mean piston speed is actually a fairly blunt tool to compare engines with. Given one can make the piston and crank train survive, and that the amount of air an engine can pump through itself essentially governs performance, we need to look at other factors affecting air flow. Engines need to have some mechanism valving the combustion chambers to let mixture in and exhaust gases out – a function which for most four-strokes is provided by a poppet valve driven by a cam. So just considering throat areas won’t really do, at least until the poppet valve is sufficiently clear of the seat that the ‘curtain area’ (the cylindrical surface formed by the edge of the valve head and the valve seat) exceeds throat area. Do the maths and this is at a valve lift of just over about 25% of throat diameter.
Then it gets a bit complicated. Heavier valves need heavier springs, and the maths ensuring the spring is kept in contact with the cam mean that higher engine speeds require longer cam profiles, which can start to break into compression and expansion phases (for intake and exhaust cams respectively) – cutting effective compression and expansion. Valve springs are more of a real limit to performance than you’d think.
How do you circumvent the problem of the valve spring, given the need to sell road bikes to homologate racers? One approach would be to have many more smaller valves, which weigh less and so have to lift less far until curtain area equals throat area. This is what drove Honda to five-cylinder 125cc and six-cylinder 250cc engines in the 1960s, when valve spring design and materials were much less advanced. Fabio Taglioni at Ducati was aware of another approach, however – positive valve control, or ‘desmodromic’ valve gear (coming from the Greek meaning ‘captive running’).
Taglioni wasn’t the first to successfully use desmodromics – for example, Mercedes had been pretty dominant in Grand Prix racing in the 1950s with their system – but he is the only one to productionise it. To the casual observer his method, where the spring is deleted and one cam-and-rocker system is used to open the valve before another set closes it, might seem an extravagant indulgence when the crude old valve spring could do the job instead. But eliminating the spring brings several advantages. First is that heavier valves can be controlled more accurately up to higher speeds, and any margin can be used to facilitate a more aggressive lift profile. This gives the required ‘time area’ of the system – effectively its breathing capacity – in a shorter time period. Because the
‘TAGLIONI WASN’T FIRST WITH DESMODROMICS, BUT HE IS THE ONLY ONE TO PRODUCTIONISE IT’
profile is now governed by material stresses in the rocker systems, and not a spring, faster opening and closing can be provided, increasing air flow and reducing the profile length – to the benefit of real compression and expansion.
Secondly, the energy for closing the valves is not provided while opening them, as has to be the case with a spring. The result is a softer input (or ‘stab’) torque profile into the cam drive system, which can lead to lower friction. For proof, look how Ducati can use timing belts, and are the only company currently offering hydraulic cam phasing (variable timing).
Finally, because there is no spring seat, the head height can be lower for a given port shape. With a V-twin this is very important. And Massimo Bordi, in forcing the Ducati Quattrovalvole engine into existence, hadn’t wasted any opportunity here – he’d consulted with Cosworth, at the time one of probably only two engine consultancies who really understood anything about four-valve engine port design. He can’t be accused of squandering the opportunity.
Put this together and you can see the 851 and RC30 should have been evenly matched. Ducati’s marketing genius was not to go straight to full litre capacity, thrash everyone and invoke the wrath of the rule makers. Ducati won the first WSB round in 1988, and though Honda won the title in ’88 and ’89, the Italians were just getting started. Gradual capacity increases kept them on par with fours which had to be tuned harder and harder to compete. The 851 begat 888, 916 and 996, riders including Roche, Polen, Bayliss and Fogarty finished the job off, and Ducati became the dominant WSB force. In the end Honda went twin-cylinder, but does anyone remember that? Aprilia never won the championship with a twin. And Ducati built a legend for performance while selling more bikes as they increased the capacity incrementally. Frankly, this was genius.
Ducati even consulted to Ferrari to apply desmo to the 1990 3.5-litre V12 F1 motor. Put ‘Testa motore desmodromico F1 Ferrari’ into Youtube and try not to be massively impressed by the fact that the engine shown had five valves per cylinder and thus logically must have had 120 rockers…
It was desmodromics that gave Ducati their WSB advantage, rather than additional capacity. And, after 62 years using the system, they now understand better than most what it can be used for. In today’s Motogp they are believed to have the most powerful engine – yet it is the only one not using air valve springs, an expensive and frankly non-road-viable technology. Ducati use the same technology that’s in their road bikes to win races and championships; if you want to buy an engine embodying Motogp, buy a Ducati with desmo valves.
As a mechanical engineer, my guilty pleasure is that I really like desmodromic valve control. And it is World Superbike that really gave it a platform on which to shine.
‘DUCATI EVEN CONSULTED TO FERRARI TO APPLY DESMO TO THEIR 1990 V12 F1 MOTOR’
Fabio Taglioni (left) and Massimo Bordi with the prototype (in 748cc Endurance guise), August 1986
Ducati’s prototype eight-valve ‘desmoquattro’ V-twin was based on a Pantah bottom end
Who is Jamie Turner? Classic Bike’s expert is Professor of Engines and Energy Systems for the Department of Mechanical Engineering at the University of Bath. A firm two-stroke fan, Jamie developed rotary engines for Norton, worked at Cosworth and...
Desmodromic camshafts have a unique appearance Four-valve head – this is a prototype, from a time when factories weren’t so guarded
Honda’s RC30 engine had gear-driven double overhead cams, titanium conrods (a road bike first) and a slipper clutch. Exotic? Yes. But note those regular valve springs...