A case for the Pushrod Engine:

[dwightlooi]

The Push Rod engine has been much maligned for being ancient, inefficient, low tech etc. It has been called “clunkers”, “American Pig Iron” and a whole host of different derogatory nicknames. But is it really the piece of obsolete trash that it is made out to be? I must say I am a DOHC fan and I applaud GM’s move to put the 3.6 DOHC V6 in everything imaginable. I think that the engine from GM’s lineup I’ll most like to have in my car is the LNF 2.0 liter DI turbo. I love the turbo whistle and I don’t really have a taste for the V8 rumble. But, let’s be honest about the whole Push Rod thing and clear up all the myths shall we?

Why Push Rods?

There were overhead cam engines l before the push rod small blocks came along. Heck, if you want a really old example of a really “high-tech” OHC 4-valve per cylinder, twin spark, direct injection (yes DI), gasoline engine I suggest that you look at a WWII vintage Daimler-Benz DB601/603/605 inverted V-12 engine. These engines even have a continuously variable supercharger drive by means of a hydraulic coupling that automatically adjusts the drive ratio (based on barometric altitude) to provide the right absolute boost pressure. You’ll find these in the Messerschmitt Bf109, Bf110 fighters. The Rolls-Royce Merlin is also a 4-vavle overhead cam design BTW (albeit carbureted) , and the Junkers-Jumo 213 in the Fw190D9 is a SOHC 3-valve, DI, twin spark design very similar in valve layout to contemporary 3.2 liter MB V6es! This is WAYYY before the American muscle car era; way before Impalas, GTOs, Camaros and Corvettes.

The pushrod design was invented to do primarily two things – make the V-type engine more compact and make it lighter. By throwing out the tall and bulky SOHC or (especially) DOHC heads, and actuating the valves from a single in-block cam you save three camshafts, lots of head height, lots of head width, lots of parts and yes lots of weight. You also make the cam drive much simpler, more compact and lighter.

If we look at the above, you can see that by removing the four overhead camshafts and half of the valves, we can have just eliminated a whole bunch of metal. The heads become roughly 1/3 as wide, about ½ as tall and less than half as heavy. Of course, the pushrod designs also have less parts and lower costs. I think we can see the contrasting difference between that DOHC engine drawing and the naked LS2 below it. See how compact the engine is and how much dieting it accomplished? FYI, the 4.6 liter DOHC 32v Northstar V8 for instance, is actually a larger and heavier engine than the 6.0 liter LS2 V8. Its output is also only roughly comparable to the 5.3 liter LS4 OHV power plant.

Why Overhead Cam?

Well, one word – Airflow. In contrast to what many people believe, neither displacement nor rpms make any power by themselves. To make a given amount of power you have to burn a given amount of fuel. To burn a given amount of fuel you need roughly 14 to 15 times as much air as the fuel you intend to burn. The overhead cam designs allow for intake and exhaust valves to be set at opposed angles to each other each with straighter shot intake pots. This is good for airflow. You can also actuate four (or rarely five) valves per cylinder to maximize valve area if you so desire. You can also reap auxiliary benefits such as a dead center spark plug location, light actuated mass for each valve to allow for high speed operation without valve float (or excessively heavy springs) and of course in recent years the ability to implement a CVVT design which advances/retards the intake and exhaust cams separately. It has also been argued that high tumble designs (most OHC engines) promote better fuel-air mixing that high swirl designs (most OHV engines) although this has not been demonstrated conclusively.

The combination of high rev breathing capacity and light actuated valve train components allow the OHC engines to also have a much more refined character and sound. Today, this is arguably more important than theoretical airflow limits (more on that later).

The problem with DOHC 4-vavle designs is that you incur a lot of weight, a lot of bulk in those cylinder heads. Also, none of the airflow benefits actually show up until pretty high up in the rev range. Airflow capacity, you see, is not a free lunch. And more is not always better. At low engine speeds, storm drain sized runners and ports along with big fat valves and/or a lot of valve lift create very low intake velocities which result in poor intake charge mixing, low power output and high emissions. High rpm breathing enhancement techniques like high exhaust/intake valve duration overlap and advancing the valve timings can downright lead to an engine that won’t idle or run right at low engine speeds.

In fact, when DOHC engines first saw mainstream use in the early 1980s, they all performed badly at low engine speeds because of intake velocity issues. Various techniques like blocking off half the intake tract with butterfly valves (eg, the TVIS system on the Toyota 1.6 4AGe and 2.0 3SGE engines) is commonly used to get them to not stumble off idle and at low engine speeds.

What happened in the last 20 years?

Well, to put it simply, DOHC engines have mellowed and OHV (Push Rod) engines have really shaped up.

You see, the typical street car, even those with sporting pretensions are typically not going to rev to 8000 or 10000 rpms. In fact, most DOHC engines today are optimized to reach their maximum output at between 5500 and 6500 rpm. Over the years manufacturers have also shied away from monster sized intake ports and runners. Narrower runners and smaller ports are intentionally adopted to so the engine operates well in from idle and up. In short, the original reason for which the DOHC 4-vavle designs where coined – airflow maximization – is not longer pursued! At up to about 6500 rpm, the airflow required can be achieved with pushrod OHV designs which are smaller, lighter and less complex! How small are the differences? Well, a good 3.5 liter push rod V6 like the LZ4 is good for about 224 hp on 87 octane. A good 3.5 liter DOHC engine like the Nissan VQ35 is good for 245hp on 87 octane. That’s about a 10% difference. The LZ4 is an iron block engine that is physically much smaller than the VQ35, and had it been an aluminum block powerplant it would be significantly lighter as well! A So the question today should really be why anyone should want to employ DOHC valvetrains in a typical car?

The answer comes down to the consumer’s perception. A car equipped with a Push Rod OHV engine is deemed outdated by the consumer. The perceived refinement in terms of vibrations and acoustic signature also heavily favors the DOHC engines. They don’t exhibit much valve clatter at higher RPMs because their reciprocating valve train pieces are lighter, they tend to also not appear to run out of breath as easily in most cases as revs climb past 5000 rpm or so. In most cases, they are also slightly more economical because on the average DOHC designs mix fuel and air better, and have slightly lower pumping losses at cruise, than push rod designs. Part of the reason is that they tend to have better technological content such as VVT, coil-on-park ignitions, etc. The ability to vary intake and exhaust cam separately is also good for emissions control – good enough to eliminate the need for an EGR subsystem. Even though some of these are not necessarily tied to their DOHC design, they are nonetheless associated with DOHC engines because it is true that when you compare the typical DOHC engine to the typical Pushrod engine, you are more likely to find these “high-tech” features on DOHC ones.

However, TODAY, if you are after the highest power density with respect to engine dimensions or weight, and you don’t care about perception or refinement or anything else, you’ll find it in a well engineered Push Rod engine like the LS2 or LS7. I’ll tell you this right now. A 400hp 6.0 liter LS2 is probably smaller or at worst the same size as the BMW 4.0 liter DOHC V8 that is going into the next M3. It is probably no heavier as well. It has better low end tractability and it is cheaper than any V8 engine built for 8,000+ rpm duty will ever be. Sure, it’ll sound and feel less refined. But, as far as performance goes it is a better solution.

Where do we go from here?

Let’s put it this way… a lot of the DOHC 4-valve engines’ advantages come from their technological content, not the design layout. Things like direct injection, VVT, high energy individualized ignition system, advanced intake design, advanced engine control electronics have nothing to do with Push Rods or Overhead cams. No, you typically find them more on DOHC engines, but that is because market forces and manufacturers put them there not because of the choice of valvetrain layout permits or forbids them!

Nonetheless, I see the continued dominance of DOHC valvetrains in family sedans and typical cars. It sounds and feels more refined – HONESTLY IT’S TRUE! The consumer expects it. And they will get it. However, I think that if the GM folks are smart about it they will be very successful in pushing HIGH TECH pushrod designs in sporty cars. As far as HIGH TECH goes, I expect the first direct injection pushrod small block within the next five years. I also expect synchronous VVT and AFM (DoD) -- which has already happened. I expect to see at least an attempt to incorporate either variable length or variable resonance volume intakes. If they really want to push it is actually possible to do a VTEC/VVTL-i like cam lobe switching system on a push rod V8 too. All you have to do is chuck the traditional lifter and replace each with a pair of roller rockers in the block. One follows a lower-lift/low-duration cam lobe. The other follows an aggressive high-lift high-duration lobe. The pushrod is connected directly to the low speed cam. The rocker following the high speed lobes bounces on a spring actuating nothing but itself. At a certain RPM, a steel pin, slipper or collar locks the two together. And the valve is then forced to follow the high speed cam. You can still have hydraulic valve lash adjustment – you just have to do it via lash adjusters on the pivoting end of the rocker. If all these features are used, a 500 hp 6.0 liter push rod V8 with a velvety idle is possible. The engine will weigh less and be smaller than any 500hp DOHC engine using the same block and head materials. We they want to go even more, may I suggest a centrifugal blower driven directly from the crank via a Van Doorne type continuously vectoring transmission? The CVT should be able to vary the total supercharger drive ratio from about 24:1 at lower engine speeds to about 6:1 at high engine speeds linearizing the power adding characteristic of the efficient but peaky centrifugal impeller.

Edited February 16, 2007 by dwightlooi

[dwightlooi]

Here is my predictions concerning how various "high-tech" features can progressively improve the Push Rod V8.

The starting point (red line) is the actual torque curve of the 6.0 liter LS2. The subsequent pictures illustrate how various technology can progressively improve performance until a 500hp engine is achieved. Not that I am not trying for ultimate horsepower here but rather an extremely linear and tractable engine which also happens to achieve ~83.3 hp/liter and ~75.8 lb-ft/liter. I am sure even power is possible with the mentioned technologies if emissions, tractability and linearity is sacrificed. However, I for one, believe that these things can be more important than maximum power output.

That'll be a pretty darn nice Small Block "Advanced" won't it?

[Sixty8panther]

Excellent Article.

Althought I learned a few things about both DOHC and Pushrods my theory is

stands very much true. DOHC is a smoke and mirrors marketing catchphraze

and not muhc else. I sure as heck feel no need to ever own a DOHC motor.

They have never impressed me much. I look at Acuras, Toyotas & Mazdas at

the junkyard withthe velvecover off and it always just makes me think of how

much added weight and complexoty there is. It's not at all "sophisticated" or

more advanced in my eyes.

DOHC was advanced on Deusenbergs and it was sort of neat on the Northstar,

which is the only V8 that it belongs on in the GM lineup. I hope pushrods never

die, simplicity, durrability & low weight are GREAT attributes in a motor.

If it aint broke, don't fix it!!!

[91z4me]

You can also actuate four (or rarely five) valves per cylinder to maximize valve area if you so desire.

You do know that there are 4 valve OHV engines in produciton, right? The Duramax is probably the best example.

You can also reap auxiliary benefits such as a dead center spark plug location,

Didn't the original 426 Hemi engines have a central spark plug?

light actuated mass for each valve to allow for high speed operation without valve float (or excessively heavy springs)

No instead of heavy springs you get 2 much heavier cams on top of each cylinder for a total of 4 for a V engine, instead of 1 cam in the valley. Also with the weight being higher on the engine the vehicle would have a higher tendancy to rotate because the weight is closer to the center of gravity, or slightly above it.

and of course in recent years the ability to implement a CVVT design which advances/retards the intake and exhaust cams separately.

GM holds patents on a twin cam CVVT system for OHV engines. Not only that by the VVT Gen IV V8s are already in production.

It has also been argued that high tumble designs (most OHC engines) promote better fuel-air mixing that high swirl designs (most OHV engines) although this has not been demonstrated conclusively.

And has the opposite been demostrated conclusively? If not the point is moot.

The combination of high rev breathing capacity and light actuated valve train components allow the OHC engines to also have a much more refined character and sound.

I don't think the actuation of the valves plays a huge role in the sound of the engine. In fact I would wager that the shape of the valves, engine firing sequence, and shape of the exhaust port plays a much larger role in the exhaust note.

Please note I am not trying to argue with you. Just pointing out some, apparent to me, flaws in judgement.

[dwightlooi]

You do know that there are 4 valve OHV engines in produciton, right? The Duramax is probably the best example.

The Duramax head is not particularly good for airflow (the intake and exhaust ports run into one valve first then the other since the valves are tandem not side by side). Also, it is also possible only in a diesel because the Diesel engine uses a a completely flat combustion chamber roof and all four valves are parallel. The approach -- using one rocker to push down on two parallel valves connected by a bridge is not feasible for a gasoline engine.

Didn't the original 426 Hemi engines have a central spark plug?

Kinda central, but not quite. It is very much like the BMW E30 era 2-valve SOHC heads. The valve is between the two opposed valves but displaced slightly to one side.

No instead of heavy springs you get 2 much heavier cams on top of each cylinder for a total of 4 for a V engine, instead of 1 cam in the valley. Also with the weight being higher on the engine the vehicle would have a higher tendancy to rotate because the weight is closer to the center of gravity, or slightly above it.

GM holds patents on a twin cam CVVT system for OHV engines. Not only that by the VVT Gen IV V8s are already in production.

It is not that heavier springs make the engine heavier. I meant the spring force is significantly heavier because it has to actuate the much heavier total inertial of the lifter, push rod, rocker and relatively large valve. In a simplest OHC design (tappets) the spring only as to actuate the valve and the shim (or shimless) buckets. Even for the popular roller follower designs the total mass of the roller followers and the valve is significantly lighter (~1/2 to 1/3 the mass) of the driven mass of typical push rod linkage set. Low spring force is directly related to low startup wear.

The main problem with 3-valve, twin in-block cam designs such as those proposed by GM a few years ago is that the rods themselves compete for space with the intake runners and ports. It is impossible to have straight, unobstructed runners and intake ports. Perhaps the reason these designs did not make it to production is because the intake port issues limits flow improvements to a point where improvements over 2-valve designs become intangible.

And has the opposite been demostrated conclusively? If not the point is moot.

I don't think the actuation of the valves plays a huge role in the sound of the engine. In fact I would wager that the shape of the valves, engine firing sequence, and shape of the exhaust port plays a much larger role in the exhaust note.

Actually, the push rod setup has A LOT to do with the noisy racket push rod motors tend to make at higher revs. Let me explain... Basically, valve train racket comes from valve float. In general, all valve trains "float" to a tiny degree at higher revs. Just like with suspension systems, when rebounding from a compressed state it the spring is unable to keep everything fully loaded and in contact. Higher spring rates help, but spring rates high enough to eliminate all slop will also cause excessive wear due to friction and hurt economy. I tiny amount of "float" during those small periods of time where the engine is actually reving at the upper 1/4 of its rev range does not hurt the engine. The difference between push rod engines and OHV designs is that the actuated mass of each valve element is very high. There is the lifter, the rod, the rocker and finally the valve itself. The heavier the actuated mass the higher the impact momentum when the valve train elements floating a tiny bit at higher revs remakes contact. This creates louder noise and more notable vibrations. On top of that, the OHV design also has more interface layers than the OHC designs. In a typical DOHC design there is one to two interface layers. With bucket tappets there is basically just one interface -- between the cam lobe and the tappet. In roller follower designs there is two -- cam to follower and follower to valve. In push rod designs you have four -- cam to lifter, lifter to rod, rod to rocker and finally rocker to valve. Hence, there are four locations where you'll have slop during high rev operations and a lot more noise making locations. Its like an anchor chain with five links makes more noise when you rattle it than one with three links.

Things like firing order (whether the V8 is flat or cross plane for instance; whether it is an even fire design), 1st and 2nd order vibrations due to engine configuration (I4s and 90 degree V8s have 2nd order vibrations at 2 x crank speed; I3s, 60 and 90 degree V6es have 1st order vibrations at crank speed; I6s, 60degree V12s and horizontally oppose designs are perfectly balanced, etc), and vibrations due to imperfect crank balancing and variance in rod and piston weights across each cylinder are completely independent of the valve train design.

Please note I am not trying to argue with you. Just pointing out some, apparent to me, flaws in judgement.

NP. I am happy to discuss these things.

Edited March 3, 2007 by dwightlooi

[balthazar]

>>"...chuck the traditional lifter and replace each with a pair of roller rockers in the block. One follows a lower-lift/low-duration cam lobe. The other follows an aggressive high-lift high-duration lobe. The pushrod is connected directly to the low speed cam. The rocker following the high speed lobes bounces on a spring actuating nothing but itself. At a certain RPM, a steel pin, slipper or collar locks the two together. And the valve is then forced to follow the high speed cam."<<

Interesting. Do any DOHC motors use a 'split' cam lobe (as far as the cam grind goes)?

Not positive of the workings of your propsal from the sectional view; is the "pair of roller rockers" going to actuate twin pushrods, or singles? This pair of rollers is always in contact with the cam, right? but unless each pushrod's lobes are synchronized on the cam centerline (optimum for performance/emissions??)- this obviously won't work. It the lobes are synchronized and you attempt to connect the 2 at some mid-point in the RPM scale, there would be tremendous valvetrain noise, vibration & wear, not to mention long-term durability issues- unless computer controls can pinpoint the nanosecond the 'holes' would be aligned and make a near-seamless connection.

OR.... what if the cam lobe featured a 'ramped' lobe allowing a 'CVT-ish' transition between low- and high-profiles?

>>"tiny amount of "float" during those small periods of time where the engine is actually reving at the upper 1/4 of its rev range does not hurt the engine."<<

To my knowledge, valve float does not occur until the last 5-10% of the RPM range, not the last 25%, and this small margin tends to be beyond peak power anyway (near redline). But the average joe will attribute 'pushrod clatter' to ANY RPM, even idle. There shouldn't be much --if any-- valvetrain sound difference between bluepinted IBC & OHC engines until valve float; a IBC engine has higher spring pressures but a DOHC had double the valves. Many people obviously make this judgement -valid or not- standing outside the car at idle or 'goosing' RPMs - where valve float is not occuring.

I would have to believe "interferance layers" is the cause of pushrod 'clatter' far more than valve float- at redline inside a car, with wind roar, tire roar, exhaust roar heavily involved, valve float would be very difficult to pinpoint as a factor in dubbing pushrod motors 'clattery'. BTW-- your illustrated proposal has added another 'interferance layer' to the pushrod design.

Great, informative article, DL.

[dwightlooi]

Interesting. Do any DOHC motors use a 'split' cam lobe (as far as the cam grind goes)?

Not positive of the workings of your propsal from the sectional view; is the "pair of roller rockers" going to actuate twin pushrods, or singles? This pair of rollers is always in contact with the cam, right? but unless each pushrod's lobes are synchronized on the cam centerline (optimum for performance/emissions??)- this obviously won't work. It the lobes are synchronized and you attempt to connect the 2 at some mid-point in the RPM scale, there would be tremendous valvetrain noise, vibration & wear, not to mention long-term durability issues- unless computer controls can pinpoint the nanosecond the 'holes' would be aligned and make a near-seamless connection.

OR.... what if the cam lobe featured a 'ramped' lobe allowing a 'CVT-ish' transition between low- and high-profiles?

All Honda "VTEC", Toyota VVTL-i and previous generation Mitsubishi MIVEC engines use lobe switching. Even though the actual implementation of the follower mechanism differ -- Honda uses steel pins for locking, whereas Toyota uses a sliding slipper for instance -- the concept is the same.

The answer to your question regarding synchronization is that the lobes and the followers that rides on them are NOT synchronized during their lift duration. That is one lobe may open the valve earlier or later than the other, it may open it with a greater or lesser lift, and it may have a completely different opening and closure ramp. However, both lobes have the same base circle. The switching occurs when the valves are closed and both followers are riding their base circle. When the followers are actually being lifted by the lobes the pin and holes do not align and the steel pins do not slide in place even when hydraulic pressure is being applied to them. They'll slide in place when the followers all drop back to the base circle and the locking mechanism align. The next time the valves open they'll follow the high lift profile.

If you look at VTEC in detail, it actually has three lobes per pair of valves. The followers on the sides actuate the pair of valves directly. These provide the velvety idle and smooth, low emissions, drivability at low and mid engine speeds. The one in the middle has a completely different timing, lift, duration and ramp profile. This lobe is practically a racing lobe which will not idle even at 2000 rpm and definitely won't pass emissions at low speeds. At about 5200 rpm, a pair of steel pins lock the three followers together and they all follow the center aggressive lobe. In some VTEC engines the system is actually a 3-stage one. This is because the two side lobes are not identical. One practically does not even open the valve! At low engine speeds, one valve opens regularly, the other is barely cracked open. This causes high swirl and provides high intake velocity at low speeds. At a certain RPM, the barely opening valve is locked to the center "racing cam" while the normally opening one continues to follow its lobe. At an even higher RPM, the normally opening one also follows the high lift cam lobe. Hence there are three stages of cam aggressiveness. On top of that the system adjusts cam timing continuously with the traditional CVVT mechanism on the drive sprocket. Honda engines using both CVVT and cam profile switching are given the moniker i-VTEC.

Here are some pictures of the real thing...

To my knowledge, valve float does not occur until the last 5-10% of the RPM range, not the last 25%, and this small margin tends to be beyond peak power anyway (near redline). But the average joe will attribute 'pushrod clatter' to ANY RPM, even idle. There shouldn't be much --if any-- valvetrain sound difference between bluepinted IBC & OHC engines until valve float; a IBC engine has higher spring pressures but a DOHC had double the valves. Many people obviously make this judgement -valid or not- standing outside the car at idle or 'goosing' RPMs - where valve float is not occuring.

I would have to believe "interferance layers" is the cause of pushrod 'clatter' far more than valve float- at redline inside a car, with wind roar, tire roar, exhaust roar heavily involved, valve float would be very difficult to pinpoint as a factor in dubbing pushrod motors 'clattery'. BTW-- your illustrated proposal has added another 'interferance layer' to the pushrod design.

Well, there shouldn't be any float related clatter at idle. In fact, if you listen to a modern pushrod V-6 at idle it is pretty darn quite. However, because the valve train of a push rod engine comprises of twice as many elements that are not rigidly coupled together, there tend to be considerably more slop in the whole setup.

The whole float thing is harder to quantify. It is not as if float suddenly happens at a certain speed. It doesn't. For a given spring rate and actuated mass, there is progressively more slop. In general people define float as when the valve actually closes after the cam lobes have past over it extending the cam duration. However, even before that happens the valve lifter will tend to "lift off" the cam minutely between its peak and the base circle, but catching up before arriving at the base circle and hence close on time. The 3.5 LZ4 for instance -- in my experience -- exhibit some notable amount of what appears tto be valve clatter at about 4500~5000 rpm onwards.

[balthazar]

Thanks for the vtec detail- I've never investigated the mechanics of it before. What is the central, encased spring on the left-side valve set?

Looks like 'slider' cam followers rather than rollers- rollers would eliminate a lot of friction- esp since each lobe in a 3-lobe set-up is much narrower- providing far less surface/wear area. I assume there is no room for rollers, then?

However, because the valve train of a push rod engine comprises of twice as many elements that are not rigidly coupled together, there tend to be considerably more slop in the whole setup.

"slop" is a rather extreme term. There will not be any momentary 'gaps' in any valvetrain with properly engineered spring pressures until near-redline speeds.

It is not as if float suddenly happens at a certain speed. It doesn't.

Well, not univerally, but valve float is valvetrain speed fast enough to overcome spring pressure, so this should be consistant within one engine/spec. Increasing spring pressure will eliminate valve float (or at least further 'bump up' it's RPM occurance), but sometimes with the downsides you mention above. Roller valvetrains also reduce valve float, as the flat end of the hydraulic/solid lifter is eliminated.

...the valve lifter will tend to "lift off" the cam minutely between its peak and the base circle...

This should never happen in a properly-engineered engine. A proper-spec engine doesn't allow the pushrods to rotate by hand under low pressure- there is not going to be any clatter unless something is wrong. Spring pressure dictates valvetrain 'snugness'- grossly insufficient pressure could result in microscopic lift, tho this incorrect pressure obviously could occur in both IBC and OHC engines- it is not a result of valvetrain design as it is of valvetrain spec.

Valve float should unload the entire valvetrain and allow multiple points of lift/clatter/noise- not just at the lifter, but again- this should not be a discernable issue in street cars since we're talking about redline engine speeds. What, exactly, provides 'pushrod noise' at low speeds, I'm not sure of, beyond the multiple contact areas in a IBC engine.

[gm4life]

I see a good and a bad for both. Having never owned a OHC vehicle I can say OHV have a broader torque/power band than a non-VVT-VVTi/VVTi-L-VTEC whatever you call it, but OHV struggle higher up in most cases past 4-5K. The LS7 is proof that it is not all, that puppy screams past 5-6-7K! I still like the low weight of an OHV motor, and the powerband. I hate how people on the Car and Driver Forums bash OHV. I am not a OHC hater, gee I might own a N* powered DTS (slighty used of course someday), or a 3.6L DOHC Torrent GXP FWD/Enclave FWD or even a G8 but it prolly would be a GT V8 hey never know I might fall in love with the 3.6L DOHC V6 engine in the base G8.

[76ChevyTrucker]

Overhead Cam Engines have their place, if you like high-revving engines. If you like to not have to get so deep into the throttle to get your power, Overhead Valve Engines work more to your liking.

[dwightlooi]

Thanks for the vtec detail- I've never investigated the mechanics of it before. What is the central, encased spring on the left-side valve set?

That's the idler spring for the center follower (which follows the "aggressive" lobe). When the steel pins are not engaged, the center follower does not open any valve, it simply rides its idler spring while the followers on its side open the their respective valves under as dictated by their own mild and gentle lobes.

The primary reason rollers are not employed is because the space under the swiping "pads" are where the steel locking pins are housed.

[balthazar]

>>"OHV struggle higher up in most cases past 4-5K. The LS7 is proof that it is not all, that puppy screams past 5-6-7K! "<<

It's all in how it's spec'd, not in valvetrain style! Guy I know builds motors for a living. He built what he called a "junkyard" Pontiac 462 (uhh, that's 7.4L for you newbs) that turned 8900 RPM and propelled a '70 Firebird drag car to consistant very low 9s in the quarter for a number of years of regular racing. Most bench racers would tell you that's 'impossible' 'because it's a pushrod, and pushrods don't rev, you know.'

[gm4life]

I understand that I and I said not all OHV motors are like that the LS7 is one.

[dwightlooi]

>>"OHV struggle higher up in most cases past 4-5K. The LS7 is proof that it is not all, that puppy screams past 5-6-7K! "<<

It's all in how it's spec'd, not in valvetrain style! Guy I know builds motors for a living. He built what he called a "junkyard" Pontiac 462 (uhh, that's 7.4L for you newbs) that turned 8900 RPM and propelled a '70 Firebird drag car to consistant very low 9s in the quarter for a number of years of regular racing. Most bench racers would tell you that's 'impossible' 'because it's a pushrod, and pushrods don't rev, you know.'

Actually, it is easier to build a high reving DOHC 4-valve motor that lasts than a Push rod one. If you want a 7000 rpm motor and you are using push rods, you'll be using very heavy valve springs. You may also be using exotic materials or construction techniques to lighten the lifter, rod, rocker, valve retainer and all of that stuff. Heavy springs actuating a heavy valve train linkage is bad for durability. Skeletonized and aggressively lightened parrts may also narrow your margin to material failure.

In DOHC designs the reciprocating components are much lighter and the spring tension is usually much lower for a given maximum rated engine speed. This is good for durability and reliability. 7000 rpm is no big deal; something an economy car designed for low manufacturing costs can easily pull off. With the same kind of effort and wear acceptance as you may experience with an LS7 you can easily push a DOHC design to rev at 10000 rpm and still cover it with manufacturers warranties. In fact, with DOHC designs today piston speed is a bigger concern than the limits of the valve train. The spring actuated DOHC valve train is good for about 18000 rpm is small engines (~0.15 liters per cylinder; ~600cc I4) and about ~13000 rpm medium sized engines (0.5 liters per cylinder; ~2.0 I4 or ~4.0 V8). It is when you want to go higher than these stratospheric numbers that you need to think of something other than metallic coil springs. This is why innovations like Ducati's Desmodromic valve train are really novelties which addresses what is today a non-existent problem. The more limiting factor is piston speed. The Honda F20C engine (2.0 liter S2000s) reaches a piston speed of 1512 m/min (4961 ft/min) at the rev limit -- thats roughly Formula 1 territory.

[David]

Thank you Everyone, THis is one of the best threads I have read. Very informative and educational.

[balthazar]

>>"If you want a 7000 rpm motor"<<

Why is this "wanted"? The only reason I can see to intentionally increase reliable RPM ceilings is to extract more power or to sound 'racy'; to raise the ceiling merely because high RPMs are sometimes associated with 'high tech' is absurd.

>>"and you are using push rods, you'll be using very heavy valve springs. You may also be using exotic materials or construction techniques to lighten the lifter, rod, rocker, valve retainer and all of that stuff. Heavy springs actuating a heavy valve train linkage is bad for durability. Skeletonized and aggressively lightened parrts may also narrow your margin to material failure."<<

No exotic construction techniques, but yes, 'exotic' materials, but that's 'high-tech', which = more gooder. And those same exotic materials (titanium, etc) do not make for a heavier IBC valvetrain, but a lighter & stronger one, which increases longevity/durability (more 'more gooder'). Yes; OHC configurations should always be lighter than IBC set-ups.

Not to mention, higher RPM in and of itself reduces longevity over a more moderate average operating RPM thru increased wear & tear.

IMO, technology should push for slower-revving engines, as long as the same power levels could be obtained.

[rkmdogs]

This is a fantastic thread, and I really do appreciate all of the detail and specifics.

However there is one point that has been overlooked, or left out of the discussion. That is, the

desmodromic valve consideration.

For those of you who were not around in the 50's, desmodromic valves do not use springs for valve

closure, they use mechanical actuation for both opening and closing.

Two of the best examples of this concept were the Mercedes W196 engine, and the Pegaso 504.

I wrote a theme paper about the Pegaso when in college, explaining this feature.

The Pegaso was a very expensive car to build. It was financed by the Spanish government and

the design was by Spanish college students, so expense was not even considered as a factor.

I would love to hear some comments from you contemporary guys on this, if you have any research data.

The other factor not mentioned in this discussion other than casually, is internal parasitic frictional losses.

Most college-level internal combustion engine design textbooks raise the issue of internal efficiency, and the

ways to measure it for a given design. Does anyone have any input on this factor, in the valve system design

discussion?

Again, love this thread & would like to see a lot more like it. This is the kind of stuff that C & G should promote,

not the "feel" of the upholstery!

[balthazar]

Friction amounts from the 2 valvetrains should, theoretically, be minimally different. Roller tech reduces the lion's share of it. In contrast, a DOHC design (not to mention a 2- or 3-stage cam lobe) should, again: theoretically, add that much more friction to that valveltrain (double the number of valve stems, valve springs, cam lobes, etc). How much for which is a guess at this point.

[dwightlooi]

(1) The ONLY Desmodromic valvetrain still in production is used by Ducati Motocycle engines. The concept is relatively simple -- each valve requires two cam lobes and two followers, one to open the valve and one to close it. The purpose of the design is the elimination of valve springs and hence valve float at higher RPMs. Back when it was originally invented, metallurgy and lubricant quality limited the amount of pressure a valve spring can apply on the contact surfaces before wear becomes unacceptable. Metallurgy back then also had problems with producing valve springs that are strong, light, compact and which does not lose its memory over the expected lifetime of the engine. Desmodromic valvetrains eliminated all of these problems. However, Desmodromic valve trains are noisier -- because it'll require zero tolerance for there to be no zero slop at all points between the opening and closing fingers and the valve, and zero tolerance does not exist. Also, when the system gets out of adjustment, it can tension the valve against the valve seat or fail to close it completely, both of which can have very serious consequences. Its demise however was that today -- or since the 80s for that matter -- metallic valve springs, contact surface material, valvetrain lightening and lubricants have already conquered valve float issues beyond any sane RPM. And for insane RPMs (say 20,000 rpms in an F1 car) pneumatic valve closure is the superior solution compared to Desmodromic. Hence, it became irrelevant. Ducati uses it because it is a brand identity thing.

(2) Parasitic frictional drag is higher in OHC designs than in IBC designs. It is also higher the more moving parts you have in a valve train -- more valves, more lobes, more followers, etc. However, the total parasitic loss of an engine is not dependent on frictional drag alone. It is a combination of frictional drag and pumping losses. In fact, more of it is due to pumping losses than frictional losses. An engine is like a syringe. The total surface area of the pistons is a lot larger than the surface area or the intake valves or the cross section of the intake plumbing. In general DOHC engines have less pumping losses but higher frictional losses, but because frictional losses is a smaller fraction of the total parasitic loss they are generally slightly more efficient.

Edited May 9, 2007 by dwightlooi

[balthazar]

>>"the total parasitic loss of an engine is not dependent on frictional drag alone. It is a combination of frictional drag and pumping losses. In fact, more of it is due to pumping losses than frictional losses."<<

True- but pumping losses by themselves (the rotating assembly: piston/ rod/ crank) are not directly effected by valvetrain friction quantities. That is: a given rotating assembly is not going to "know" if it is working with a IBC or OHC valvetrain as far as friction goes... altho the crank may see some slight degree of load difference from a IBC valvetrain vs. a OHC valvetrain, but I don't know which way that would go; have any comparison studies of this ever been published?

But, in fact, OHC rotating assembly friction quantities would be greater due to higher piston speed and operational/ redline RPM capabilites. Add increased frictional quantities with OHC valvetrains & belt-drives and I strongly question: >>"In general DOHC engines have less pumping losses "<<. Can you cite evidence to support this?

Only thing that springs to mind as a possible contributor is valve timing/overlap, which may increase combustion compression/expansion efficiency, which may in turn work better with the rotational pumping pressures to aid rather than oppose those forces.... but again- in order to legitimately analyse these two engine designs, we need to eliminate variables. Valve timing --for the purposes of this duscussion-- needs to remain equal between the 2, so that as a contributor has to be striken from consideration.

Likewise, lubrication & metallurgy need to remain a constant in this theoretical analysis of different engine designs, as these factors are not a mandate of valvetrain design.

>>"...since the 80s for that matter -- metallic valve springs, contact surface material, valvetrain lightening and lubricants have already conquered valve float issues beyond any sane RPM"<<

Valve float is a mechanical byproduct of engine speed & valve spring pressure alone; lubrication and seat hardness do not enter into it. I would question whether spring material has changed dramatically in the last 40 years as far as a demonstratable advantage goes in production engines. Weight would have some (miniscule) contribution, but pressure is the primary factor.

Edited March 16, 2007 by balthazar

[dwightlooi]

But, in fact, OHC rotating assembly friction quantities would be greater due to higher piston speed and operational/ redline RPM capabilites. Add increased frictional quantities with OHC valvetrains & belt-drives and I strongly question: >>"In general DOHC engines have less pumping losses "<<. Can you cite evidence to support this?

Evidence is not the right word. However, indications are abound. The most obvious, is that in general DOHC-4 valve designs have not just higher power but also higher torque yield than IBC-OHV designs of the same displacement. Remember output yield is really gross - total loss. If the frictional qualities are inferior or even equal, and accessory tap off is roughly the same, then the only reason this can be the case is higher pumping loss. From a more elemental standpoint, 4-valve DOHC designs almost always have larger valve area to piston area ratio. Basically, it is very simple... if you have higher output yield per unit displacement, and you have equal or higher frictional loss, you also necessarily have lower pumping losses.

Valve float is a mechanical byproduct of engine speed & valve spring pressure alone; lubrication and seat hardness do not enter into it. I would question whether spring material has changed dramatically in the last 40 years as far as a demonstratable advantage goes in production engines. Weight would have some (miniscule) contribution, but pressure is the primary factor.

But they do because the wear resistance of the contact surface is very much affected by the material and lubrication. This governs how much spring pressure you can apply before wear becomes unacceptable. Higher valve spring tensions increases frictional wear -- the classic (idealized) equation for frictional force is normal force x friction co-efficient. Let's put it this way.... if we make cam lobes and lifters out of say soft iron, we can apply less pressure between the two than if we use titanium nitride coated high speed steel and it last the required number of cycles. Modern lubricants which leaves a decent film even after the oil drains away also helps. In short, if you have advanced, very wear resistant materials and very good lubricants, you can use higher spring rates and still have the parts last the same amount of time.

[Drew Dowdell]

There hasn't been a suitable explanation as to why a pushrod engine can't have more than two valves....

If you're going to compare pumping loses, compare a SOHC 2 valve engine to a pushrod 2 valve engine. Pumping loses are likely to be similar.

Taking that into account, the location of the cam in the engine becomes moot. It's the number and location of valves that is important.

[balthazar]

>>...in general DOHC-4 valve designs have not just higher power but also higher torque yield than IBC-OHV designs of the same displacement.

...If the frictional qualities are inferior or even equal, and accessory tap off is roughly the same, then the only reason this can be the case is higher pumping loss."<<

What about increased airflow, possible increased combustion efficiency and increased RPM of a DOHC head?

A straight displacement/output comparison is not 'apples-to-apples' and is not valid proof of anything here other than a broad generality. We're attempting to discern the pros & cons of 2 different valvetrain designs- factors such as metallurgy & lubricant quality are not design parameters nor should they be factors.

>>"But they do because the wear resistance of the contact surface is very much affected by the material and lubrication. This governs how much spring pressure you can apply before wear becomes unacceptable."<<

Wear is achieved over great spans of time- even a brand-new, blueprinted engine with no wear can achieve valve float if improperly spec'd; these are 2 different issues.

The only way to objectively & scientifically determine the facts here is obtain 2 identical engines (displacement, materials, # of cylinders, piston speed, clearances, redline, etc), but have one a 2-valve IBC design and 1 a 2-valve OHC design. Only then could friction be measured & compared.

>>"In short, if you have advanced, very wear resistant materials and very good lubricants, you can use higher spring rates and still have the parts last the same amount of time. "<<

Truw, if not longer... have you ever tried to manually cut tool or spring steel?

Edited March 16, 2007 by balthazar

[Carguy]

There hasn't been a suitable explanation as to why a pushrod engine can't have more than two valves....

If you're going to compare pumping loses, compare a SOHC 2 valve engine to a pushrod 2 valve engine. Pumping loses are likely to be similar.

Taking that into account, the location of the cam in the engine becomes moot. It's the number and location of valves that is important.

Good point Oldsmoboi!!!---I would like to know what went wrong with the OHV 3VPC heads GM was working on? We could have had a HAPPY MEDIUM with that design both a more compact design and more then 2VPC for better flow at higher RPM's. This move toward DOHC's and 4VPC V6's by GM will dictate lower axal ratios and inturn less fuel efficency to get higher RPM power so they can be more compairable to the Asian models. But it seems to me a DI AFM VVT OHV 3VPC 4.0L V6 could put out the same HP as a 3.5L DOHC V6 at a lower RPM and more torque and in turn have a higher Axel Ratio for equal if not BETTER fuel economy in a midsize sedan! Edited March 16, 2007 by Carguy

[dwightlooi]

There hasn't been a suitable explanation as to why a pushrod engine can't have more than two valves....

If you're going to compare pumping loses, compare a SOHC 2 valve engine to a pushrod 2 valve engine. Pumping loses are likely to be similar.

Taking that into account, the location of the cam in the engine becomes moot. It's the number and location of valves that is important.

The purpose of a DOHC 4/5-valve design is mainly airflow maximization and optimization. In a pushrod OHV design you have a big problem with the pushrods and the intake ports fighting for the same space on the side of the heads. Basically, on the intake side you need a minimum of two pushrods and their tunnels. Typically these are smack in the middle with the intake port snaking to one side. If you have two intake valves there is not enough room to put in the intake tracts to support them. In addition, if you put two small intake ports on each side to feed two intake valves, the airflow collides in the middle and eliminates any swirl that you typically get. In addition, you'll need additional hardware on top of the heads to bridge both valves so the can be operated by one push rod. This is a problem because any hardware that bridges two valves on the intake side also blocks access to the exhaust valve by the other pushrod and its rocker.

The only single in block cam, push rod actuated 4-valve per cylinder design is the Duramax 6600. In this engine, if you look at a cylinder from the intake side, the two valve on the left are intake valves and the two on the right are exhaust valves. In otherwords, the valves are tandem in arrangement. A single pushrod tips a rocker arm which pushes down on the middle of a bridge connecting both valves. This is possible because the diesel engine has a totally flat roofed combustion chamber. That is all four valves are parallel to each other. The tandem arrangement does not permit or support the intake port dimensions any where near as large as typical DOHC design, this is bad for airflow, but it is good enough for a diesel which revs in the 3250 rpm redline. In a gasoline engine it would have been completely impractical.

What about increased airflow, possible increased combustion efficiency and increased RPM of a DOHC head?

Increased airflow = reduced pumping losses. Period. Pumping losses is precisely the power it takes to overcome airflow restrictions. The the cylinder head off and spin the crank pushing the cylinders into open air and you'll have zero pumping losses just frictional losses.

[balthazar]

I think the DuraMax 'pulls off' a IBC/4V system by being a diesel (high CR reduces the need for a particular-shape, combustion-friendly chamber) and by being a diesel (no spark plug... tho where is the glow plug?)

>>"That is all four valves are parallel to each other."<< = valve faces are all in the same plane.

>>"this is bad for airflow, but it is good enough for a diesel which revs in the 3250 rpm redline."<< Hence: turbochargin', baby.

>>"In general DOHC engines have less pumping losses "<<

>>"Increased airflow = reduced pumping losses. Period."<<

So did we not just agree?; a DOHC will appear to have a lower quantity of friction due to increased airflow, not because of actual lesser frictional quanitities from the valvetrain. This does not prove but does lends creedence to the theory that a 2V IBC valvetrain may have less friction than (esp) a DOHC design. Score one more for the IBC.

[aldw]

Wouldn't Magnaride tech be applicable for electromagnetic valve control systems?

[dwightlooi]

Wouldn't Magnaride tech be applicable for electromagnetic valve control systems?

In short, no.

Magnaride uses a ferros magnetic fluid in the shock absorbers. By passing an electric field through the liquid, the vicousity of the liquid can be adjusted. This changes how easily (or hard) the fluid is to pass through the valving in the shock tubes, hence making the shock absorbers stiffer or less stiff. The system does not actually move anything at all -- it merely changes the properties of the oil if you will. I don't see how this can be applicable in the valve train.

[dwightlooi]

The Duramax is not a good example of a multi-valve (meaning greater than 1-intake and 1 exhaust valve per cylinder) pushrod engine. The design is workable only because its a diesel with a flat roofed combustion chamber and no sparking plug.

Here is a better one:-

This is what GM toyed with for a while when they were thinking about how to improve on the LS1.

Problems with this design may include the rather asymmetrical offset rockers with interconnecting rods having poor high rpm stability and projected asymmetrical wear. The significantly increased valve train mass driven by a single lifter also practically double the spring pressure applied on the lifter and hence increases frictional wear on the intake side. In the end, airflow improvements over an optimized 2-valve design may not have been demonstrated.

The design was not adopted. Instead the LS2 got LS6 style optimized 2-valve heads with bigger, straighter ports.

[ZL-1]

A question from a non-engineering inclined person: what about the camless engine Chrysler was supposed to be working on a few years ago? IIRC the idea was to have no pushrods or overhead cams, but electronically controlled individual valve actuators integrated with the rest of the engine's electronics. Has the idea been abandoned altogether?

[FAPTurbo]

A question from a non-engineering inclined person: what about the camless engine Chrysler was supposed to be working on a few years ago? IIRC the idea was to have no pushrods or overhead cams, but electronically controlled individual valve actuators integrated with the rest of the engine's electronics. Has the idea been abandoned altogether?

It's likely been abandoned because Chrysler and 'high-tech' powertrains don't mix.

[dwightlooi]

A question from a non-engineering inclined person: what about the camless engine Chrysler was supposed to be working on a few years ago? IIRC the idea was to have no pushrods or overhead cams, but electronically controlled individual valve actuators integrated with the rest of the engine's electronics. Has the idea been abandoned altogether?

Electro-magnetic valve actuation hasn't panned out.

It hasn't panned out because:-

(1) The coils needed to drive valves at the requisite speeds to operate an engine at the typical redline of say 5000~6500 rpm is pretty darn big -- about the size of a can of coke.

(2) The voltage needed to operate these coils effectively (100~300 volts) typically are not available on typical cars which only has 12~14v power supplies.

(3) Because electro-magnetic actuation is not a positive engagement drive it causes a whole magnitude of control problems. When a cam lobe pushes down on a valve it is not going anywhere but down in exactly the amount that the lobe depresses it. When an electro-magnetic force is applied on the valve it is like you are pushing it down against the valve spring with a heavier spring. There is no positive engagement there, and how much movement actually happens depends on various things like the rpm (hence valve speed), the residual pressure in the chamber and the pressure or vaccuum stacked on the valve in the intake tract. In otherwords, it keeps changing with driving conditions and the computer must smartly, accurately and promptly map the right amount of current to apply on the coil each and every time.

(4) Valve float at higher rpms is actually a lot worse than with cam and spring drives because the valve itself has to have a spring, a stem and a magnetic puc. This makes the driven mass heavier.

(5) Assuming that none of the above matters, the system of at least 8 coils, a high voltage power supply and control system is notably heavier and a pair of camshafts and a chain or belt.

The ONLY advantage of EM valve actuation is that it is by default capable of continuously and infinitely variable timing, lift and actuation profile. Both timing and lift can already be varied with mechanical valvetrains -- CVVT is everywhere, Valvetronic does CVVL. We can also switch between two or three completely distinct camlobes with systems like VTEC, MIVEC or VVTL-i.

[thegriffon]

Pneumatic/hydraulic valve actuation is coming. The first road cars should have the systems (there are more than one supplier) by 2009.

Edited April 28, 2007 by thegriffon

[ZL-1]

Thanks for the other guy's replies (even the Chrysler&'high-tech' comment - I admit I LOLed) but this got me interested:

Pneumatic/hydraulic valve actuation is coming. The first road cars should have the systems (there are more than one supplier) by 2009.

Can you share how that works for the layman like me?

[Flybrian]

I have to say this is one of the most enlightening threads ever on C&G. Thanks for all who have contributed so far and I hope the discussion continues on and on...

[thegriffon]

Ahh, electro-hydraulic actuators move the valves down, usually with a built-in spring return. They can either be used to move the intake valves, with a normal cam for the exhaust valves, or both intake and exhaust valves. Lift and timing are infinitely variable throughout the rev range, and cylinder or port deactivation simply requires the right engine control program.

[balthazar]

Tucker was working on a hydraulic valve system in '47. There was not enough time in his timeframe to work thru all the engineering- ended up going with a ready-built motor with conventional valvetrain. Had the business end of the venture not gotten in the way, this discussion could be going down an entirely different route.

[ZL-1]

Ahh, electro-hydraulic actuators move the valves down, usually with a built-in spring return. They can either be used to move the intake valves, with a normal cam for the exhaust valves, or both intake and exhaust valves. Lift and timing are infinitely variable throughout the rev range, and cylinder or port deactivation simply requires the right engine control program.

Thanks!

What about interaction with the other softwares: are there benefits (power/torque, fuel efficiency, even smoothness of operation) in integrating this feature with engine and transmission softwares?

[dwightlooi]

Pneumatic/hydraulic valve actuation is coming. The first road cars should have the systems (there are more than one supplier) by 2009.

Actually, I think we need to clear things up a little. To the best of my knowledge, nobody is attempting pneumatic actuation. So far it doesn't exist and there is no good reason for it to exist. Hydraulic actuation is something else altogether and there is nothing pneumatic about it.

I know that Sturman Industry has been working on a Hydraulic valvetrain since the mid 90s. I haven't heard of a major manufacturer signing on for it though. Their design uses a central hydraulic pump to supply a reservoir of pressurized hydraulic fluid. These are fed to each valve but blocked by a electrically controlled solenoid valve. The valve can either supplies this pressurized liquid to the driving piston which depresses the valve or it bleeds it off to the return system leading to the hydraulic pump (and ultimately back to the reservoir, or do neither (blocks access both ways). The speed at which the valve opens cannot be controlled. But when it opens (feed connect) and when it closes (return connect) can be controlled precisely. To control how far the valve opens, the system cuts off the connection to both the feed and return system and causes the valve to "stay" at its current position which maybe partway open. The valve is closed by a mechanical spring. The system offers infinitely variable timing, lift and duration. The potential problem with the system lies in the longevity of the solenoid valves and the limits in "ramp" rates (opening speed) which a pumping a liquid through small passages can support. They have a running engine. But I don't know if they have convinced anyone to adopt their system.

Pneumatic Actuation on the other hand does not really exist. What is commonly called Pneumatic Actuation and used in F1 since the 80s is not really pneumatic actuation but rather pneumatic closure. The F1 engines use camshafts and cam lobes to open the valves. The "pneumatic" part is simply that instead of metal springs they uses gas pressure to close the valves. Each valve has a cylinder where the spring resides. This is kept pressurized with an inert gas (usually nitrogen). Each cylinder has a one-way valve for the most parts which means the gas can get into it but not out of it. The gas supply is regulated (usually 100~150 psi). When the valve opens it compresses the gas in the cylinder and this pressure helps push the valve back. The system has two advantages. The first is that metallic springs react too slowly when operating at 15000 to 20000 rpms. Actually, this is also the reason why valved 2-stroke engines couldn't be made feasible -- because the valving speeds are too slow for a closure system that relies on a solid bending and contouring itself from one shape to another. The gas closure system is fast enough as there is no "plasticity lag" inherent of solid metal springs. The other advantage is that gas pressure is at its highest when the valve is fully opened and is a fraction of that when it is closed and the cylinder is not compressed. With a compression ratio of 10:1 the pressure is roughly 10 times higher when valve starts to close than when it is closed. This highly progressive nature of pneumatic closure puts applies maximum "spring rates" when it is most needed, but very little pressure when it is not. This not only reduces overall wear, but also reduces mechanical frictional drag.

[Camino LS6]

Fascinating thread!

This takes things up a notch at C&G, thanks to all involved!

OH, and keep it up.

[balthazar]

Hydraulic sounds appealing IMO- properly engineered it's a very repeatable and controllable science, but the complexity & weight involved is a tremendous penalty.

I could see the theoretical need for Pneumatic activation in the realm of ultra high RPMs if mechanical means are insufficient (they would be in this case) and it sounds like it should entail a much smaller weight penalty, but I still question the demonstratable advantages in a typical 'street' engine that merely sees 6000-7000 RPM.

[Guest ArrrBeee]

Great article. A couple of thoughts. The bench mark for engine designs has come to 100 hp per liter, normally aspirated.. It's been that way for motorcycle engines for a long time. HP from a push rod engine is done by means of the "Big Hammer" method. I find it an embarrassment that a 6.2L push rod engine struggles to put out 400+ hp.

DOHC engines may be wider, but they're not necessairly heavier, because of the use of aluminum heads and blocks. Yes, the same metal can be used in a OHV engine.

The Ford Coyote engine is a fine example of a modern U.S. engine design. They're real close to achieving the 100hp per liter.

But really, the reason for the persistent OHV engine is the sound, rumble and the Good O'l Boy percepton of what performance should sound like. You can thank NASCAR NHRA for this

[dwightlooi]

The key qu

Great article. A couple of thoughts. The bench mark for engine designs has come to 100 hp per liter, normally aspirated.. It's been that way for motorcycle engines for a long time. HP from a push rod engine is done by means of the "Big Hammer" method. I find it an embarrassment that a 6.2L push rod engine struggles to put out 400+ hp.

DOHC engines may be wider, but they're not necessairly heavier, because of the use of aluminum heads and blocks. Yes, the same metal can be used in a OHV engine.

The Ford Coyote engine is a fine example of a modern U.S. engine design. They're real close to achieving the 100hp per liter.

But really, the reason for the persistent OHV engine is the sound, rumble and the Good O'l Boy percepton of what performance should sound like. You can thank NASCAR NHRA for this

Why is hp per liter of any importance?

If both engine makes (let's say) 400 hp. If the pushrod engine is externally smaller, lighter and have better fuel economy. Why is it of any importance if it makes that 400hp out of a static displacement of 6 liters instead of 4 liters?

The most common argument for the relevance of hp/liter being important is that a 4 liter engine is smaller and lighter. But this isn't true. In fact, the wide DOHC heads, 3 addtional camshafts, 3 times as many cam phasers and double the number of valves often causes the 4 liter DOHC engine capable of 450hp to be actually bulkier on the outside and heavier on the scales. Yes, it is possible for a DOHC engine to be lighter, but that only happens when there is a drastic inequality in the amount of lightening technology and advanced materials applied on the two engines being compared. It's just like saying that it is possible for a mid-sized car to be lighter than a compact. When construction techniques and materials are equal, the pushrod engine is always lighter.

The other argument is that if you make that 400 hp out of 4 liters instead of 6 liters, you get better fuel economy. That too isn't true as displacement is pretty low on the list of features the most affect fuel economy despite popular misconceptions. The problem with DOHC designs is that when you have three times as many camshafts and twice as many valves, these components along with their bearings and chain drives incur extra parasitic frictional drag that otherwise isn't there. In addition, when you rely on small displacements to make the same power output as a larger, lower revving engine, you end up with less torque. The lower torque output often requires a lower gearing for equal performance and/or drivability. Higher frictional losses + lower gearing often result in inferior fuel economy despite a lower static displacement. As a matter of fact, if you want to maximize fuel economy the most effective thing you can do is not to reduce displacement, but to increase it while adopting an Atkinson Cycle cam. An Atkinson cycle engine of the same output, but larger displacement is always more economical on fuel.

A good comparison of two engines can be found with the BMW S65 DOHC 4.0 V8 and the GM LS3 Pushrod 6.2 V8 in the Camaro SS.

• The S65 with 414 hp / 295 lb-ft in a 3704 lbs M3 posts an EPA rating of 14 mpg (City) / 20 mpg (Hwy) . The S65 engine itself weighs 202 kg.
  
• The LS3 with 426 hp / 420 lb-ft in a 3849 lbs Camaro SS posts an EPA rating of 16 mpg (City) / 24 mpg (Hwy). The LS3 engine itself weighs 183 kg.
  

The Pushrod, low hp/liter engine of a larger displacement is lighter than the DOHC, high hp/liter engine of a smaller displacement. It produces slightly more power, a lot more torque and better mileage from a heavier car. How is that embarassing?

The entire notion that DOHC is "modern" and Pushrods are "ancient" also begs a history lesson. The DOHC design was invented BEFORE the Pushrod design was. The DOHC design was pioneered by Bentley in 1919. The Pushrod OHV design was invented in 1930 by Arthur Chevrolet. So, which is more modern?

Edited August 14, 2012 by dwightlooi

[Drew Dowdell]

Great article. A couple of thoughts. The bench mark for engine designs has come to 100 hp per liter, normally aspirated.. It's been that way for motorcycle engines for a long time. HP from a push rod engine is done by means of the "Big Hammer" method. I find it an embarrassment that a 6.2L push rod engine struggles to put out 400+ hp.

I don't compare by HP alone... any engine can produce 100 hp/liter if you spin it fast enough. A 6.2 liter pushrod will produce a much more satisfying torque curve than than a 4.0 liter DOHC assuming both use the same type of aspiration.

One of the things I like most about the Chrysler Hemi V8 is the effortless woosh of power that comes on at very low RPM.

[A Horse With No Name]

Having a good friend who has an SRT-8 300...I can agree!

[dwightlooi]

In case you are still wondering why it is possible for a 6.2 liter pushrod engine to be smaller and lighter than a 4.0 liter DOHC V8, here are the two engines side-by-side. Notice the amount of room and extra material needed by the DOHC heads, quad cams, their sprockets and gears?

BMW S65 4.0 liter DOHC V8 (Left), GM LS3 6.2 liter Pushrod V8 (Right)

Edited August 14, 2012 by dwightlooi

[Drew Dowdell]

What does the extra chain at the bottom of the BMW V8 do? Oil pump?

[dwightlooi]

What does the extra chain at the bottom of the BMW V8 do? Oil pump?

Yes... BMW S65 Oil Pump Assembly (Below)

[A Horse With No Name]

Although the Ford OHC motors have proven to be really good V8's....

[balthazar]

If you throw your preconceived notions out the window, the DOHC arrangement sure looks like a decades-old precursor to the IBC motor; all spread out with long bicycle-style chains before they found a way to make it more compact.