Yes I would imagine that there would be considerably more pressure in the reservior because if the valve return chambers leak, they would have to be recharged really fast to maintain proper operating pressure at high rpm.
Can anyone tell me more about this system? Is it software driven or does it have banks of pressure valves?
Desmo is a really cool system but is totally different.
The valve system for a Formula One engine versus a "conventional" DOHC layout is that instead of metal springs to close the valves, there are bellows. All the bellows are plumbed to a common plenum which maintains the pressure. In case there are any leaks, a reservoir of compressed gas can be activated, and the plenum pressure restored to optimal levels. If the leak is worse, then pit stops may be required to recharge the system. There may be weak springs on each valve, but they are there just to keep the valves closed to avoid valve-to-piston contact.
A pneumatic valve system's advantage over metal valve springs is that the pneumatic system allows higher RPM's, because in a metal spring scenario, there's an upper limit around 15,000 RPM. Pressures required, spring harmonics, valve mass inertia, all conspire to set an upper limit on how quickly you can accelerate a valve.
A desmodronic valve system is a totally different animal. It uses camshafts to open the valves, and camshafts to close the valves. Theoretically, it's fantastic, offering less power loss, high RPM. But it's very difficult to build such a mechanical monster where the opening and closing cams work in harmony with the valve. Tolerances have to tight. There is a valve spring to close the valve fully, because the closing valve has to leave some room so that in case tolerance run too tight, the valve wouldn't be mashed into the valve seat by the closing action of the camshaft.
Racing should be decided on the track, not the court room.
Pneumatic valves represent one approach to overcoming some of the disadvantages inherent in the valve train of an internal combustion engine. In a conventional four stroke engine valves open and close in defined set of sequences. For example, intake valve will open on a downstroke just before the compression stroke. If the valve opens on upstroke the compression created by the approaching piston will pump the fuel and air out of the cylinder. Which is why the exhaust stroke is always an upstroke. This is why engines run poorly, if at all, if even a tiny error in timing is present.
In order to synchronize these motions requires a number of parts. The good old American overhead valve v-8 offers a fairly simple illustration of how a valve train operates. These engines are relatively simple, with only two valves per cylinder, one intake and one exhaust. Millions of these engines have been built and they are very efficient at producing torque at low RPM.
One problem from this arrangement is that every fixed camshaft represents a compromise. As any driver knows, a car engine changes RPM constantly as road conditions vary. Yet any fixed camshaft produces maximum horsepower and torque only at a very limited RPM range. Ideally, you would like a different cam for every RPM. Practically, that is impossible. So cams are engineered for different RPM ranges. A cam concerned solely with maximum power, would be optimized for very high RPM, but would give up low RPM power, making it impractical, and possibly undriveable for the street. A camshaft designed for low RPM power can be very driveable, but gives up a lot of power for flexibility.
One answer to this problem has come with Variable valve timing, first pioneered by Honda. Most systems advance or retard the camshafts to adjust for speed. This system is mechanically simple, and offers significant improvements, but affects only one cam parameter, timing, and is thus less than optimal. The Toyota VVTi and BMW systems work like this. Others, notably the Honda V-tech shift the cam fore and aft, shifting between two different cam lobes, of different profile. This system is better in many ways, but is more complex and still only offers two cams when what you really want is a different cam for every rev. Pneumatic and electric valves represent potential approaches to overcoming this problem. In a pneumatic valve system, the valves are operated by releasing compressed air, produced by a compressor operated off the engine. Theoretically, such a system would free the engine from any of the constraints inherent to mechanical camshafts, with the valve operation defined not only by RPM, but by throttle position. Potentially, such a system can produce very efficient engines.
Such engines do not exist today, and for good reason. The valves must be operated by a compressor. That compressor requires engine power to drive. It is expensive, relative to conventional design. It might fail, which could lead to uncontrolled valve motions inside an operating engine. In some auto engines the valve and piston can may strike each other in the event of a valve train failure. Such engines are referred to as interference engines. Such contact can destroy an engine.
Another disadvantage is that the metering system is itself complex, and also prone to failure. But the biggest disadvantage is the compressor, which returns more power than it consumes only at very high RPM, usually around 13,000. Such high revving is common only in auto racing. Only in racing are pneumatic valves found.
In racing the motivations for moving to pneumatic valves were different. If you recall from our OHV V-8 engine illustration, a lot of parts are involved between cam and valve. I raced a car whose exhaust pushrods were each about a foot in length. At 6000 RPM each pushrod, rocker combo must change direction 6,000 times every minute. The long pushrods bend easily. That's a lot of reciprocating mass to overcome. For this reason OHV engines cannot be revved above 9,000 RPM and live. High performance street engines generally redline at around 6,000 RPM. Overhead Cam (OHC) design eliminates the pushrods., greatly reducing reciprocating mass and simplifying the valve train.. Overhead Cam engines can rev as high as 14-15,000 RPM before they fail. Failure occurs when the valve spring is moving so quickly that it achieves the spring metal's harmonic frequency. Yet race engine designers want to go even higher.
The reason is simple. Race engines are almost always limited in displacement, the product of cylinder swept volume times the number of cylinders. As you increase swept volume, you increase the amout of fuel and air you can get in the cylinder. Power is, after all, a product of burning fuel and air.
Renault was the first to deploy a pneumatic valve system in its RVS-9 engine. Variants of that engine powered Jacques Villeneuve and Nigel Mansell to the F1 world championship. The system operates on compressed air, and only closes the valves, replacing the valve springs. That permitted the Renault engine to rev as high as 19,000 RPM, a significant advantage. It has been adopted by all F1 engine builders out of competitive necessity. This shifted the weak point from the valve spring to the connecting rod. No one has yet thought of a way to bypass the rods, which tend to fail catastrophically. Most racing series, such as CART have banned them for cost reasons.
Pneumatic valves offer significant advantages, for a price. The price may be too high to pay for most street engines. For racing they dominate. But further development may bring pneumatic systems to the market.
note: this is not an original document but i found very interesting on the topic of valve train. i forget where i got it.
Scuderia Nuvolari wrote:From what I remember reading is the nitrogen in the return chamber is at around 200 psi and there is a mechanical steel return spring also.
All are cam driven.
Okay, knowing that the "metal bellows" have a certian amount of pressure that, when compressed the pressure is greatly increased.
When the "bellows" looses pressure, how does the valve system work to receive more compressed gas from the reservior, knowing that each "bellows" is independent of each other? Is this software controlled or is it something simple like a pressure valve?
It's frustrating to understand when these teams handle most of this as classified material.
Do you have a link to anything supporting the "bellows" you talked about? I just find it curious that "bellows" would be used.... at least "bellows"in the conventional sense (like a compressible part with pressure behind it, which compresses and expands corresponding to valve movement) (http://en.wikipedia.org/wiki/Bellows)
Again, I don't claim to know for sure, but my guess would be that the pneumatic side of the valve train would act on the valve by way of a machined high tolerance chamber. This would remove any extra moving parts in the valve train (making it far more reliable). Having a "bellows" is a step backwards IMHO...
Sealing the valve head in the machined chamber is a piece of cake and has already be proven by technology in the automotive world.
I regret not being able to post any pics with bellows in the valvetrain. I'm still searching though. They are metallic bellows, obviously of aerospace quality and tolerances. http://www.sigmanetics.com/metal_bellows_technology.htm
In place of the sliding piston inside a tube (with a total of three o-ring seals) the bellows is in place there. The obvious advantage is less parts, less moving parts, less moving mass, and less o-ring seals (just one versus 3).
Racing should be decided on the track, not the court room.
While a metal bellows would certainly eliminate the friction losses due to elastomer or teflon piston seals, they would probably have serious issues of their own. For example, to make a metal bellows capable of taking the maximum pressure forces experienced in a pneumatic valve spring system, the bellows wall would have to be fairly thick. What you'll find is that once you make the bellows wall thick enough to handle the internal pressure forces, it will be over-stressed at the convolution bends when fully compressed at the valve open position. Remember, that bellows will need to have a very low operating stress level in order to have adequate fatigue life in a high speed valve train. Even in a limited life F1 engine.
There are also dynamic issues you'll need to resolve. That metal bellows will likely have all kinds of weird axial vibratory modes that it goes thru as the engine speed goes up and down. That's one of the biggest issues with metal valve springs at high operating frequencies.
I'm afraid it's not as straightforward an issue as you might think.
"Q: How do you make a small fortune in racing?
A: Start with a large one!"
