Why no 5 Valve Engines?

[abisec]

what you are goin to leave out from a f1car....is not goin to add on to its weight and is not going to give you problems...besides the universal definition of a racecar would be......

[riff_raff]

with 3 intake valves instead of 2 each valve is lighter and the valve lift is lower so you have lower mass and lower accelerations

This means that you have less stress on the valve springs, and, as you said in another thread

Reca,

With regards to valve spring stress (helical compression springs are loaded primarily in torsion), there are lots of factors that determine stress limits: deflection, wire diameter, modulus of elasticity, coil diameter, etc. Fitting 3 coil springs in the same space as 2 coil springs naturally requires that the 3-coil-spring configuration requires a smaller spring OD. Helical coil spring torsional stresses increase at the square of the spring diameter, but only linearly with wire diameter.

Of course, this whole argument is moot, because F1 valve systems are pneumatic.

[Reca]

Of course, this whole argument is moot, because F1 valve systems are pneumatic.

That’s the crucial point indeed and a point I already made almost 1 year and half ago in this same thread.
Ferrari used 5 valves per cylinder in all the V12 they made from 1989 to mid 1993, then they adopted 4 valves per cylinder. Do you know what happened during the 1992 season ? They introduced, for the first time on their engines, pneumatic springs instead of metal springs. Is it a coincidence that the first modification they worked on after the introduction of pneumatic springs was the 4 valve per cylinder head ? Some years ago a bird, a red bird, told me it wasn’t.

[riff_raff]

Reca,

I believe the first manufacturer to use pneumatic valve return springs in F1 was Renault (in 1986?). Their system was the "distribution pneumatique".

The major benefits of the pneumatic system were that it was somewhat self-dampening and it also provided a highly progressive spring rate as RPM's increased. Thus allowing much higher engine speeds.

Sorry if I'm rehashing an old topic. I wasn't on this thread 1-1/2 years ago. Regardless, I appreciate your feedback, as you seem very knowledgable on the subject!

[Reca]

riff_raff :
first of all I’m sorry for the delay of this answer... I did visit f1technical few times in the last few weeks and I completely missed your post until today...

Obviously it was Renault that introduced in F1 the pneumatic valve springs (in 1986 in the turbo engine and then in 1989 in the NA V10), what I wanted to say is that 1992 was the first time for Ferrari, that’s what I meant with “in their engines”, sorry if I wasn’t clear enough.
Pneumatic is a very clever and simple solution; the main problem of the valvetrain in term of rpm is the mass of the spring, if you reduce, almost eliminate actually, the mass of the spring then the natural frequency of the assembly valve + spring becomes lot higher and the resonance isn’t anymore the limit for rpm increment.
Before to adopt the pneumatic springs Ferrari engineers, in an attempt to solve the problem, also evaluated the desmodromic. At Galleria Ferrari in Maranello should be visible the desmo head of an experimental V12. It’s an impressive piece of mechanics, but clearly the pneumatic valve springs are lot more simple.

Regardless, I appreciate your feedback, as you seem very knowledgable on the subject!

Thanks, I appreciate your feedback too, it’s an enjoyable discussion.

[riff_raff]

Reca,

I would make the argument that reduced valvetrain mass is secondary to system frequency, when considering benefits of the pneumatic valve train. A pneumatic valve return "spring" system has no critical (or resonant) frequency per se. The system stiffness, and thus natural frequency, can be changed by raising/lowering the ambient pressure in the feed manifold. The pneumatic spring system is also not an adiabatic process, so there is a fair amount of energy loss (or dampening) taking place each cycle.

Having the ability to increase valve "spring rate" as engine speed increases also ensures that friction losses in the valve train are kept to an absolute minimum. On average, valve train friction accounts for about 20% of mechanical losses in a typical engine. On a 900hp F1 engine, that equates to about 50hp. Although peak hp won't benefit, fuel consumption at part throttle conditions will be improved substantially with a progressive valve spring rate.

Regards,
Terry

[bmwpower8]

they do not use five valve cylinders because they need to balance the intake flow with the exhaust flow. and if there is two valves on both side this can be done. however if they did use five valve cylinders they would want to use them on the exhaust side of the cylinder, so the gases could exit faster. but the problem with that is the the valve overlap, the period of time when the intake valves and the exhaust valves are both open. the off balance of two intake valves and three exhaust valves would disturb the valve overlaps pressure that keeps the exhaust going out of the exhaust valve, causing the exhaust gases to come out of the intake valves, causing an over heat of an engine.

[Reca]

Bmwpower8 :

First of all the exhaust valve area has to be smaller than the intake valve area. Main reason is that the pressure jump between cylinder and exhaust duct during the exhaust phase is a lot higher than the pressure jump between intake duct and cylinder in the intake phase so you need a larger effective area at the intake than at the exhaust. Then for other reasons related with fluid dynamics to pass thru the valve going from the cylinder to the duct (as in the exhaust), is more efficient than in the opposite direction (from duct to cylinder) hence the same geometrical area gives an higher effective area (that is what really matters) at the exhaust than at the intake. Experience shows that the geometrical area of the exhaust valves should be in the order of 0.6-0.7 the geometrical area of the intake valves. So, if you wanted to increase the number of valves that would surely be at the intake, not at the exhaust. And in fact the 3 valves engines have 2 intake valves and 1 exhaust valve and the 5 valves engines have 3 at the intake and 2 at the exhaust.

As for the gasses going in the intake runners during overlap, that’s a problem that doesn’t depend by how many valves you have, it’s related with the relative pressure between intake duct and cylinder/exhaust duct during the overlap, relative pressure that depends by the “game” of the pressure waves. In the intake (but also in the exhaust) duct there are high pressure and low pressure waves moving between the two ends of the duct. In a correctly tuned engine you have an high pressure wave arriving at the valve exactly during the overlap so it “pushes” the products of the combustion out of the cylinder and down in the exhaust “cleaning” the cylinder. In a badly tuned engine the opposite happens so you have a rarefaction (low pressure) wave arriving at the valve during the overlap => disaster, what should go in the exhaust is sucked back in the intake instead and will get back again in the cylinder in the following intake phase reducing the amount of “fresh” charge.
The tuning of the waves is related with the geometry of the ducts, mainly with the length, each rpm should have a given optimal geometry, so the same geometry that is optimal at a given rpm could be the worst at a completely different rpm. Hence the two stages intakes or variable length intakes, and also the variable valve timing.

[riff_raff]

bmwpower8,

Take a look at SAE paper 860032. Yamaha examined 4, 5, 6 and 7 valve cylinder head configurations. Their conclusion was that the 5-valve configuration, with a centrally located spark plug, was optimum. They analysed steady-state flows up to 13,000 rpm.

[Reca]

At Galleria Ferrari in Maranello should be visible the desmo head of an experimental V12. It’s an impressive piece of mechanics

And here, finally, a couple of pics of it, it was made in 1991.

A global view :


And a close up. As it’s evident from the number of cams, I was shooting from the opposite side of the display case compared to the previous pic so the intake now is on the right :

[DaveKillens]

Sweet piece of engineering, I wonder how many hours it took to set up all the proper clearances with all those valves, cams, and followers.

[Guest]

bmwpower8,

Take a look at SAE paper 860032. Yamaha examined 4, 5, 6 and 7 valve cylinder head configurations...

Honda NR 750, 8 valves per cylinder...
http://www.bikez.com/motorcycles/honda_nr_750_1992.php

[manchild]

That was me... sorry

[Reca]

That one had oval pistons (and 2 con rods per piston) in an attempt to have more piston area with just 4 cylinders (the concept was derived from racing were Honda was trying to limit the 4 stroke disadvantage compared with the 2 stroke). To fit 8 valves per cylinder in this case was simple, more or less like fitting 4 valves on a circular piston.

http://www.billzilla.org/nrpiston.jpg
http://www.survivalskills.clara.net/nrpiston.jpg

[manchild]

That one had oval pistons....

I know that it had oval pistons just thought that this topic was about number of valves regardless on shape of cylinder