Tuesday, May 05, 2009

Tags :  valvetrain valve-springsRecently, RET has covered valve spring design and measurement extensively, but it is almost impossible to discuss valve spring design without mentioning the desmodromic system which is used nowadays exclusively by Ducati.The term Desmodromic is coined from the Greek words ‘desmos’, which means ‘controlled’, or ‘linked’, and ‘dromos’, which means ‘course’, or ‘track’. In modern engineering terms desmodromic refers to a mechanism which has different controls for its actuation in different directions.In modern times, Mercedes used the system on their W196 Formula One car of 1954-55, and then in 1956 Ducati took the system up and have never looked back since.The desmodromic system is famed for the fact that it eliminates the valve return spring and indeed in racing applications it does. But what every Ducati owner and mechanic knows is that roadgoing Ducati’s do feature valve return springs, just not conventional ones.There are two major issues which a production bike faces which do not apply to racing bikes; that is, starting and emissions.Whilst racing bikes are started via motored rollers which act on the rear tyre, production bikes rely on a starter motor. The starter motor can only start the engine promptly if the engine achieves sufficient compression and to ensure that this is the case, small torsional helper springs act on the closing cams of both the inlet and exhaust valves to hold the valves closed.If these springs are omitted, the clearances in the closing mechanism, which must be built in to allow for component wear and differential thermal expansion, will mean the valve stays too far open (when it should be seated) to allow starting.The other major issue which is addressed by the torsional helper springs, is emissions. On a production bike if the valves are not held shut then unburnt fuel can pass into the exhaust system which is not just harmful to the emissions themselves but also dangerous for the catalytic converter. In the modern climate of European, and indeed worldwide emissions legislation, this is a serious consideration for a manufacturer.Smaller, but never the less important issues addressed by the torsional helper springs are idling and noise. A desmodromic equipped engine will idle more smoothly and run more quietly – probably the first time the words idle and quietly have been used in a sentence concerning Ducati’s for a while – with the helper springs installed.Torsional (or hairspring) springs can be used as valve return springs themselves, and were originally used as such on Manx Norton engines of the 1950s as a method of eliminating valve spring surge. Ducati helper springs are a different matter and are used only in addition to the closing cam, to close the valve once it is near to its seat.So it turns out that holding the valve closed on its seat is almost an un-assumed advantage of using a conventional valve return spring to shut a poppet valve; and one that must be considered even in a nominally springless system.
Leave your comment



I am still, after about 40 years of following F1 engine design, still somewhat surprised that F1 engine manufacturers have all followed the pneumatic spring path for valve closure. In conjunction with electronic engine management systems, desmo valve gear can clearly offer considerable gains over conventional camshafts – for example, the opening acceleration can be made as high as possible with a desmo system, without fear of losing control of the valve or over-opening. The opening profile can be (almost) a square of fully open during the opening period, as opposed to a curve dictated by limitations on the valve gear and (air) springs. A properly designed desmo system need not be very heavy, which imposes almost no penalty on engine weight, and the return is in reliability and higher power output.

Some desmo designs have been unnecessarily complex, but there are some designs deserving of a relook.
Hi Charles,

Friction is one major reason why they don't us Desmo systems in F1. (Ducati never run more than 4 cylinders and in that configuration the frictional losses are outweighed by the benefits of faithful valve timing adherence between theoretical and practical, and, as you infer aggressive valvtrain acceleration times)

Complexity and alternating mass is another. As you have a entirely new set of (closing) rocker arms - even if the intake valves are opened by direct camshaft operation (effectively straight) onto the lash cap or valve. Please note that it is very difficult to position all valvetrain items within a Desmo system so that no rocker arms are used. How/where would you run both sets of camshafts if no rocker arm or similar arrangement was employed?

This adds mass where you don’t want it and with the high speeds F1 motors do it presents considerations that are not always returning greater results than pneumatic systems - particularly when you have a large cylinder and valve count.

A common misconception is that valve springs or valve sprung valvetrains rob an exorbitant amount of power. While they do present some losses, they return a considerable amount of energy back to the engine. Desmo systems require energy to open and close that isn't returned. So, even the pneumatic systems that employ valve springs present significantly less losses than desmo systems.

Once the cylinder count is large enough, particularly for multicylinder engines with large outputs; desmo systems start to introduce losses that outweigh their advantages - particularly these days with the materials and sophistication that are employed in most valve springs and pneumatic systems designs.

Then of course, this additional desmo design requirement also introduces real estate considerations that (can) impose themselves on F1 motor height and width, and therefore also vehicle shape and cooling.  

All up I think F1 engineers know that valve control can be achieved more effectively and reliably with pneumatic systems. Valve opening can, in most DOHC high performance and F1 engines, be reasonably aggressive without desmo systems. It is the closing of the valve and following the camshaft (or other system dictating valve timing) profile that usually suffers with aggressive valve opening/timing profiles; as large accelerations and inertias are involved - almost defining the valve spring limitation of valve sprung valvetrains.

Camshaft driven pneumatic valve closing systems - which are what most F1 systems employ - resolve (within reason) this problem with none of the frictional losses associated with desmo systems. They pretty much remove the need to have high energy value valve springs and/or valve sprung valvetrains without introducing significant additional frictional requirements and the wide range of machining and reliability complexity associated with desmo systems.

Additionally, with desmo systems; variable duration, lift, timing and other camshaft oriented variable induction and poppet valve related philosophies become noticeably much, much harder to implement. And as a result you usually need another set of camshafts or another way to cram all operating (closing/opening) lobes safely onto the same shaft (and real estate area) without introducing more frictional and mass issues - like say rocker arms associated with transferring the camshaft’s profile to the valve’s location - which may not be near the lobe if all operating (closing/opening) lobes are on the same shaft. Obviously having multiple camshafts circumvents this, but then you trend back to the above-mentioned losses and issues - which all kind of trend back to the issues associated with valve sprung valvetrains.

Hope that helps.


Hi all,
Concerning desmodromic valvetrains, almost of people only support his theories and reasons on mostly known Desmo system from Ducati. As they exists a lot of possibilities, so different and well supported comments must appear to give a clear and neutral overview about performances and complexity, in line to compare correctly the advantages of each system.
Concerning racing applications, special requests must be taken in consideration when designing a valvetrain system. By other way, concerning current applications for daily use on our all kind of vehicles, the folio request is a little bit different. Mainly for our transportation vehicles must be taken in highly consideration the simplicity, servicing costs, performances and fuel efficiency, and highly considered too a good during life.
Is well known that a desmo engine without spark plugs into place which can be turned by hand, gives a less valvetrain losses than compared version with return springs, is a basic clear explanation how this losses can be compared.
Having a system which can be designed with different parameters for valve open, event duration and closing gives new parameters to design a most efficient engine according requirements for each application.
The Ducati racing success is a clear example how a twin cylinder engine can be optimized to battle a most “user friendly” and powerful 4 cylinders. The addition formula 1+1=2 can’t be applied correctly considering volumetric capacities and generated power. An example: in first MotoGp season with 800cc, the Ducati engines are prepared to run over 20.000 rpm’s, a real missile on track, but fuel limitation by rules has conditioned this excessive difference.
Resuming: An engine with a handicap brake inside (compressing valve springs), can’t perform the same when compared with the similar version without the same brake. Ever they exists losses. The returning power when springs release is significantly less than used to compress, due mechanical losses.
Looking away from common know idea about desmo systems, there they has a lot of possibilities for applications. The best and most advanced examples on this research field becomes from Pattakon – Greece. The desmo systems can be used by the best way to create a super-efficient multi-purpose new engine generation.