Desmodromic valvegear

Mechanical valve closure has many benefits – and it helped long-time proponent Ducati win the MotoGP championship
By Keith Howard

In most internal combustion engines the valves are opened by direct action of the camshaft lobe but closed by means of a spring. This works fine at low revs, but as revs increase, resonant ‘surge’ effects within a metal spring can play havoc with valve operation and reliability. This was why Renault began the trend to today’s ubiquitous use of pneumatic valve springs in Formula 1 – to break through the circa 14,000rpm rev barrier imposed even by titanium alternatives.

But there is another way to obviate valve spring deficiencies: do away with springs altogether and use a desmodromic system instead, in which the valve is both opened and closed by the camshaft. One conventionally shaped cam lobe pushes on the valve stem, via a follower, to open the valve, while a second D-shaped cam alongside pushes a second follower which pulls on the stem head to close it, usually via a forked end.

Desmodromic valve systems were briefly employed in early Grand Prix car engines (1914 Delage and Schneider, 1922 Rolland-Pilain), and enjoyed a short resurgence in the 1950s in the Mercedes W196 GP car – a development that encouraged others to investigate the system. But in only one place did that interest take root and carry through for the past 50 years – in Ducati road and racing motorcycles, pioneered by Ing Fabio Taglioni.

Shortly before Ducati Corse’s historic day at this year’s Japanese MotoGP race in late September – where Casey Stoner secured the Italian squad’s first MotoGP world championship and team-mate Loris Capirossi beat the might of the Japanese motorcycle industry on home territory for the third year in a row – I spoke to Ducati Corse director Filippo Preziosi about the desmodromic system, its development and why it still enjoys technical advantages.

“In the beginning the desmodromic system was developed because the spring materials of the time were not so good, so it was impossible to reach high revs. Now we are in a different situation in which you can achieve extremely good performance in terms of flow dynamics, and hence power, with pneumatic valve springs. But a pneumatic valve system is complex – it needs a reservoir, which is especially a problem for a motorcycle where the space is very tight. And a pneumatic valve system is also very expensive in terms of power losses compared with the desmo system, especially when you are running at low and medium revs.

“When you design a valve spring system, whether it uses pneumatic or metal springs, you have to put in enough spring force to allow it to work well at peak revs. Friction is related to this spring force, so when you are riding through a corner at 9000rpm rather than 19,000rpm there are big friction losses. The desmodromic system is better because it only generates the inertial force necessary to move the valve at that engine speed, so its friction is less.

“This is a great advantage in a championship where fuel consumption is so important because of the 21-litre fuel capacity limit. Also, if you use racing to develop your street bikes, it is no good using pneumatic valves, because they are not suitable for production machines. We can develop the desmo system in racing and our customers reap the advantage in their street bikes.

“Maximum engine revs are not limited by the desmodromic system but by the need to achieve a good balance between performance and fuel consumption. This season we are achieving something over 200bhp [from 800cc] and 19,000rpm, so we achieve similar engine speeds to F1 units – without using pneumatic valves.

“In the past we have had to detune engines to make the bike more rideable, or because the rider asked for it. Traction control helps a little but whatever we are able to do technically, the rider makes the biggest difference. In F1 you might get the two McLarens on the front row, the two Ferraris on the second row. So it’s clear the car plays a huge part in who wins the title. In MotoGP this year Casey Stoner had good tyres and a good bike, of course, but it was his excellent riding which won him the title.

“We suffered many problems of reliability with the desmo system during the early 1990s because we had designed it without having considered its dynamics. The behaviour of the desmodromic system seems very straightforward when you move it slowly, but it becomes extremely complex when it is running at very high revs. So we began two projects: experimental measurement of the system to determine the position, speed and the acceleration of each part – that’s easy to say but is difficult to do – and the mathematical modelling of the system to know better what happens in the valve mechanism when it is running at high speed. We have developed both of these approaches down the years, the experimental side and the calculations, so that now we feel that we understand the system very well and it is easy to manage.

“We found there were various things we could improve: materials, surface treatments, things like that. We could also improve the design to increase each part’s strength and stiffness. And we could properly design the cam lift.

“But there are small improvements still to be made in every part of the engine, including the desmodromic system. We have to reduce the power losses as much as we can, even though the desmo has an advantage in this respect.”

Fifty-one years on – still more to come.