falloutboy's Fundgrube und Texte

Rocker Geometry 01

[falloutboy]

Der wohl beste Artikel den ich jemals über dieses Thema gelesen habe. Teil01

Rocker Geometry

by JIM MILLER

rocker01

ROCKER ARM GEOMETRY seems to raise its head every now and then, and when it does, I rarely ever see it stated accurately. Too often a sound bite of only a small piece of information is taken out of context and then used as the Gospel, totally ignoring the other dynamics that revolve around it. In some cases, something totally erroneous is stated that is not only wrong, but makes no sense for anyone who just stops and thinks about it objectively.

When lecturing at trade shows, schools, engine shops or just getting pinned down on the phone by a knowledgeable engine builder going deeper than most on a technical issue, I have found that I spend about half my time trying to undo various misconceptions about rocker geometry

before I ever begin explaining the facts. There has been so much info put out there by reputable companies (and by my reckoning, incorrect), that people are reluctant, by nature, to see something different from the prejudice of what they already know or think they know. If people are used to doing something a certain way, they see everything from that perspective. Usually, my getting through to them involves discrediting what I think is wrong with what they were doing and then begin to explain what they needed to change. At that point I could break down the simple rules for what geometry really is, and why.

Background of Rocker Geometry

Rocker geometry (or the lack of it), goes way back to many fathers on both sides of the ocean, to when the Wright Brothers were still studying the theories of lift in an airfoil. But for our purposes here, and to avoid boring the curious who’ve managed to get this far on this story, I’ll come to the point about rocker arms and explain as needed how the “mistakes” got to where they did.

In the old days, rocker arms were all pretty much what we term a “shoe” design; meaning the contact pad with the valve had a large radius scruff surface that depressed upon the valve tip as the rocker moved through its rotation. The term of course comes from the appearance to a shoe’s sole, but also to the mechanical motion much like a foot and boot would do, as it pushes off. This pushing off motion, as many will already know, has the effect of the rocker arm stretching itself as it moves through the depressing (lift) cycle. It is actually lengthening itself as it moves across the valve tip, and you see this by the wide foot print (we call a “witness mark”) atop the valve tip.

The use of rocker arms goes back to many things predating engines, but the principles were never required to be so specific on axis point heights and their consequences, as it is for helicopter bell cranks, and racing engines! There was no rocket science to designing these parts a hundred years ago, which ended up on our prehistoric cars and early airplanes. Engineers simply made designs that tried to minimize the degree of how much scuffing was imposed on the valve tip; got it close, and moved on to more important questions. Somewhere along the line, there became a principle to get this in a general ballpark, that someone later coined as the 1/3-2/3 theory (or either of the two). This placed the pivot point of the rocker arm so that it was 2/3 of the way below the valve tip, or the valve tip was 1/3 of the way above the

rocker shaft, depending on your point of view. But the answer was the same. This thinking was originally derived from the intention that a near 90 degree arc could be realized when the valve reached its intended full open position. Bear in mind that valve lifts back then were usually

in the quarter inch or so range, on little two and four cylinder engines. So being off a little really had no measurable difference in performance of the engine, and wear and tear was the real yardstick engineering back at the turn of the century was aimed at. Also, the ability to accurately measure wear and tear, horsepower, thermal loss and many other cool things we take for granted on today’s computers, wasn’t even a possibility back then.

The advent of more valve lift, and thus pushing a budding internal combustion engine technology higher to produce more power was really inspired for leaps and bounds by the advent of aviation, not Henry Ford. Not to take anything away from the automotive crowd’s contributions, but only aviation imposed the second requirement that defined “efficiency” – and that was light weight. Making Goliath engines that had more power was a lot easier than making more power from light weight engines that would be flying over somebody’s head, somewhere. So the

whole “thinking process” for efficiency in engine technology really found its impetus in aviation, because racing back in the early 1900s was still done on the back (or behind) of one-horsepower whose exhaust was more easily stepped in than emitted from a pipe. And as far as my 30 year old memory serves me on the research, aviation was also the first use of a “roller tip” rocker arm, on radial engines as far back as the 1930s, and perhaps before. In fact, to this day, I never cease to be amazed at the foresight and creativity of both aviation and automotive engineers in the 1920s, and ‘30s, and ‘40s. Four valve Pent roof combustion chambers, roller cams, fuel injection, nitrous oxide, water injection, two stage superchargers, turbo chargers, and many other cool things we assume were concepts of the last 20 or 30 years, were actually done and done quite well, seventy and eighty years ago.

Aside from the roller tip rockers of aviation long ago, the fundamental rules of rocker arm design were based around the shoe tip, contact pad design still used today. Many of you may know that you can’t use a roller tappet on a flat tappet cam, and of course vice verse not only because of hardness of material difference, but because of “geometry.” The principles of trailing motion and dynamics between something that is making direct LINEAR contact upon another object that is imposing or receiving a RADIAL (circular) is entirely different than if that contact is occurring with a roller tip making or following the contact. This isn’t rocket science either, and you can see how this happens by drawing a roller tappet in various stages of lift as the cam lobe goes around to push upon it, and see a straight line from its axis to the cam lobe is constantly

shifting around as the tappet goes from the close (base circle) position, up along the acceleration ramps, then over the nose. When finally, as it crosses dead center at full LOBE lift, this straight line between its axis and the cam centerline is also in alignment with the tappet bore itself. At all other times, the tappet is actually receiving some level of side thrust in its bore (engine block) from the pushing out forces that the lobe imposes as it chases it up.

Even when engineers were chasing efficiency with aircraft engine and power development, the threshold for seeing measurable loss of engine life, like valve stem or valve guide wear, was not easy unless things were really out of whack. They weren’t worried about loss of cam

events through small changes in rocker geometry, even though they knew the variables existed. So if they kept to this 1/3-2/3 rule, everything looked good on the valve tip, and the leverage of the rocker arm upon the spring – or more accurately stated – the leverage of the

spring on the rocker arm was at near perpendicular relationship with the valve when rates were at their highest. Even though by today’s standards for valve springs, spring rates back in the 1920s and ‘30s were negligible. This “attitude” continued on throughout the decades

afterwards. More importantly, and unfortunately, it bled over into the soon to come roller rocker arm market, that got its impetus in the 1950s. The first person who made a working aluminum “roller tip” rocker arm for automotive application belongs to my dear old and

departed friend, the late Harland. Some other garage efforts might have been getting tinkered with out in California about the same time, but it is pretty well undisputed that 1958 is the beginning of what we know today as “the aluminum roller rocker.” Keep in mind that aviation

roller rockers existed twenty or more years before, but they were steel, they were radial engines, and they didn’t comport to the automotive needle bearing aluminum body that Harland introduced.

Just like the flat tappet cam and the roller tappet having entirely different geometry because of where the measurement for motion is made, the same rules apply to the shoe tip rocker versus the roller tip rocker arm. But when Harland made his silhouette, he didn’t allow for this, and inadvertently moved the axis of the roller roughly .300” of an inch higher than what it should have been. The axis of the roller should have

been in the same place as the contact pad. So when his rocker was placed on the engine, and the roller was positioned for a good “eye-ball” track on the valve, now the push-rod cup was too high. The result, was that it went way up and in toward the stud. In actuality though,

most engine builders in the sixties continued to keep using standard length pushrods, and the excessive motion from this mistake was occurring on the valve tip, which was deemed “normal” because the roller rolled! Believe it or not, even to this day, people think that the roller

tip is for rolling on the valve. It is NOT. The roller tip is for one reason only, and that is to convert the shifting length of the rocker’s arc (that moves across the valve on a shoe design), to a fixed length that moves far less in its effect, because it is always point down in line with the

valve’s motion, just as a roller tappet of a cam is always aiming its contact tangent line with the axis of the camshaft.

This error stayed, and was copied by many manufacturers and eventually by everyone in some measure or another. It would take several decades before enough trial and error, and even a patent would be studied to make manufacturers rethink this, slowly improving their

designs. Ironically, some of the most well known names continue to promote designs they never changed, and even promote the less accurate means of using little tools, that tell an engine builder what pushrod they need, while never even taking into account the valve lift that will be used. Make no mistake; you cannot set rocker geometry without knowing exactly how much the rocker arm is going to move.

It seems logical that since a roller cam can provide all the acceleration any of today’s heads need, for any rocker geometry scenario, then why not set the rocker geometry to ONE STANDARD that has the least amount of wasted motion, and will always duplicate the same percentage of cam information, regardless of what cam you use? For understanding this, understanding a little history is always best. This ends a lot of rhetoric.

Whenever pushrods leave their inline paths to now have their end follow around the rocker shaft by any amount, this is LOST CAM INFORMATION. The cam literally has to turn more degrees to affect the same LIFT at a later point on the crank. Velocity, too, is lost. So you’ve lost duration, throughout the entire lift cycle (not just overall), and you’ve lost RATE of acceleration, by slowing the rocker down.

To put this in perspective, let’s take a simple even value of cam lift, like .400” to make a point. Rocker geometry is usually thought about as only what is happening at the valve. In our .400” cam lift example, and a 1.50:1 ratio rocker arm, this would theoretically yield .600”. To have the optimum use of in-line motion being converted into circular motion, you need to divide these values into two equal parts. Engine builders do this at the valve, but the real deal is happening with the CAM. The cam is the source of the information. So for the CAM side of this value, we’re only ending up with .200”, which is one half of our .400” cam lift. If you fail to place the axis of the tip of the pushrod at the proper length as to divide that cam lift accurately, so the rocker arm is at a 90 degree angle, then you have a pushrod that is going to move in and out more than it needs. The result will be wasted cam information that can require the crank to move several degrees more to effect the same lift of the valve. Those lost degrees were absorbed in the excessive motion the rocker arm had. You will spend hours and hundreds of

dollars to get a camshaft that is ground to fourth decimal accuracy, and chosen to give you a specific degrees of duration at .050” tappet lift, and you will change a cam to gain as little as four or five degrees if you think the engine needs it, but you just threw away more than that

because of the PIVOT POINT on your rocker arm didn’t establish the correct angles with the pushrod and valve.

Why did this continue?

Back in the early 1960s, because there was so much inefficiency and experimenting with cylinder heads, cams, induction systems and so on, this valve train flaw went under the radar. Now I was still a punk, barley sixteen in 1969, but if memory serves me right, it wasn’t until a real student of engineering and racing stuck his nose into the situation, and started shifting the pivot point around for his own purposes and seeing

distinct changes. He had some odd name car in a new class of drag racing, called Pro Stock; I think it was “Grumpy’s Toy.” Bill Jenkins was one of the real pioneers for many things in not just fixes to problems, but also a more scientific Yunick-like approach to analyzing. He didn’t follow other people; other people went out of their way to follow him. It was a short list of real pioneers to both cylinder head and valve train

development back then, and Bill was on the short-short list. But prior to Bill and a few others like him of that era, rocker geometry was totally ignored beyond the vague generalities of the 1/3 rule. But technology in the cams and heads was soon catching up. Right about this time,

in 1969/70, Chrysler approached Crane Cams for a new camshaft for the factory backed Hemi teams of Sox & Martin, Herb McCandless and the “Motown Missile” (later Mopar Missile). That development was the beginning of the .700”-plus valve lift boundary being broken.

The late sixties and early seventies were really exciting times for factory muscle cars, and the stepping stones of technology that has brought us to where we are today. It all began back in this limited, golden era. And the fundamentals established then, cut in stone, have not changed to this day either. They’ve only gotten repackaged, renamed and resold, even though other boundaries in valve lift, cylinder heads and so forth have been elevated. The principles for cam technology and specifically rocker arm geometry that would soon come along in 1980, but spawned in 1973, have not changed to this day.

Definition

What is rocker geometry? Rocker geometry is “angles of motion.” It is not some linear reference point on the tip of the valve, that trying to adjust the wear pattern will guarantee it being correct. What is correct? Correct, is “efficiency.” It is having the least amount of wasted motion being used to do the greatest amount of work (that is designed to be done by the cam). This last point is important, because the rocker arm

can be used to add to the cam, besides what it usually does by error, which is take away from the cam. But I will get into each of these below. I just wanted a simple “mission statement” that defines what geometry is and is not, so that the following hopefully makes sense.

Importance

WHY is rocker geometry so important? When you change the pivot point of where the rocker arm is, in relation to the valve tip, it CHANGES THE CAM. It doesn’t matter whose rockers you use, it doesn’t matter what style rocker you have, it doesn’t even matter what application your engine is; whenever you change location of the rocker pivot point in relation to the tip of the valve, you are changing cams. You are changing all three parameters simultaneously: LIFT, DURATION and VELOCITY (rate of acceleration).

The degree you change these depends on how much you move the pivot point. And one or two of these three parameters may be affected more than the other. But if you don’t LOCK your geometry in to the SAME THING all the time, which has the least amount of wasted motion, then you are aiming at a moving target with every cam change. Whatever results you get from one cam to another is tainted by the diluted effects of wasted motion in the rocker arm.

Rocker arms are a “radial” device being ordered to do a “linear” thing. They rotate on an axis, moving in a circle. But what they have to impart is a straight line command. They get their order from the camshaft, in the form of IN-LINE information that they then have to ROTATE around an axis and then MULTIPLY it by some ratio, and finally TRANSFER this result back to another IN-LINE component of greater movement. This movement has THREE values: LIFT, DURATION (of lift), and VELOCITY (acceleration of lift). If the rocker arm does ANYTHING ELSE

besides this, then it is NOT efficient, and SOME of this information is being lost.

Let me make a point about something on this. Your camshaft is ground to ten-thousandths of an inch precision. It is computer designed to millionths of an inch, you (or your cam manufacturer) selected it for a division of duration values where you considered two or three degrees important; any more was too big, and any less was too small. Hopefully by now, you realize that moving this rocker pivot point will change this at the valve.

You just don’t know how much. Well, it is MORE than the two or three degrees you think is important. In some cases it can exceed TEN degrees, and is often five or six degrees. As if this isn’t enough reason, consider this: It is that value of loss throughout most of the lift cycle, not just total – where you’re only inclined to measure it from, and where your cam card is limited to. That is what’s missing. Engine builders only check at FULL LIFT. They check rocker ratio and total valve lift, and that’s that. But when your rocker geometry is off, you’ve lost those

degrees of duration throughout most of the entire cam profile. Which means rate of acceleration is lost, but you may only see a small change in lost valve lift, thinking the difference is just flex in the rocker ratio.

Two Geometries

Rocker Geometry is the correct DESIGN and INSTALLATION of the rocker arm so that its relationship to both sides of up-and-down motion is fully realized by BOTH. This is of course, the pushrod, and the valve (respectively).

The rocker arm is a RADIAL tool, asked to do a LINEAR job. It pivots around an axis in a reciprocating radial (circular) motion, and has all the dynamics that anything revolving around an axis will have. But on either end of the rocker arm’s connecting points are two other instruments each live and breathe by the laws of linear (in-line) motion.

Now all this may sound like “rocker arms 101” and we may all know this, but few people I have found over 35 years, seem to understand how sensitive, precise and important this observation is. I think this true because most treat both the design of the rocker arm and its installation with casual regard.

To have the “most efficient” design and use of the rocker arm requires TWO things: (1) The rocker arm must be designed to mirror the inherent angles of each engine’s pushrod to valve geometry. (2) Rocker geometry must have an accurate location of its rotational axis with the valve tip’s height. The first of these is called DESIGN geometry; and the second is called INSTALLED geometry.

Every engine has an inherent acute angle (I call the “attack angle”) which the pushrod leans either into or away from the valve centerline. The small block Chevy for instance is 19 degrees positive (leans into). This comes from the engine block having a 4 degree angle of its tappet bores with the piston cylinder centerline. The head is already a 23 degree valve angle (actually it is a 67 degree, because it references from

the deck), so you simply subtract the 4 degrees that the tappet and pushrod “aim” toward an already inclined 23 degree valve, and you end up with simple math: 19 degrees. Every engine is unique. The SB Ford is 20 degrees for this same value. This is what rocker geometry needs to be designed to, otherwise your efforts of installing the rocker arm accurately will be limited to just one side of the rocker or the other.

What Rocker Geometry IS NOT

Rocker geometry is NOT where the roller (or contact point) is at on the top of the valve. Forget that. Rocker geometry, especially is NOT the idea that you want to place the roller or wear pattern (shoe tip rocker) in the “middle of the valve.” Forget that, too.

The valve tip is everything. It is ground zero. This is where all leverage, and the full stroke of valve lift begins, this is our reference point. There are many ways to measure rocker geometry, but there is ONLY ONE way to SET IT. Now, you can set rocker geometry in the closed position, or you can set it in the MID-LIFT position (half open), or you can do like most people have done, and simply roll the engine over a couple dozen times watching the valve open and close to see your witness mark (foot print) atop the valve tip and play hit and miss with chasing a “minimal wear pattern.”

The problem with this latter point is that this is a symptom, it is not geometry. Granted, when you have the rocker geometry set properly, you will have the least amount of wear pattern, but to try to set geometry through moving the rocker and trying to see how small you can get it, is better than nothing – not quite good enough. You can easily be off by .005” to .010” (or a great deal more) on even seeing this actual width, let alone measuring it. And being off .010” on the horizontal plane of what you are really trying to measure, which is the vertical

plane (valve lift), will multiply out quite easily to .030” or .050” or .080” or more in your error of where the trunnion is to the valve tip! Those kinds of errors will cost you several degrees of crank rotation to open the valve a like amount.

What about using a tool, or dial indicator designed to measure this in and out motion, resting on the spring retainer? Well, this is a better way of the same thing, but it is still measuring a horizontal plane for a vertical plane result. Error can be off several times more than measuring directly in the vertical plane, or parallel to the valve motion itself. When using these tools, just like a dial indicator on the top of a piston, as soon as the piston reaches perfect TDC, you will have two or three degrees of crank movement before you see the dial indicator move. There is a float time there, and so too is the effect by using a tool on the roller tip of a rocker to measure in and out motion. It floats enough to allow the valve lift to be off by .005” to .010” or more. But if you can set it dead nuts within .002” to .003” without having to buy such a tool, why wouldn’t you do it in a more precise way?

Alternative Geometries

Before getting too deep into philosophies, history and facts, let me restate the key point of what rocker geometry is, then I will mention the comparative arguments people (and companies) have made against this. MID-LIFT geometry is rocker geometry that has the ultimate “efficiency,” in that it is doing the greatest amount of work with the least amount of effort. It has the least amount of wasted motion in the pushrod and valve, commonly referred to as the “inand-out” motion. It affects the maximum response through linkage of whatever the

cam’s instructions are. If you don’t have geometry set precisely, your consequences range from simply losing a little cam information at the valve, to excessive side loads in various directions, on various parts that will at the very least rob you of power. In more extreme cases, wearing out parts or outright catastrophic part failure may occur.

There are arguments to using different geometry than mid-lift. I simply don’t agree with them because they violate the principles of efficiency. One of these theories is to adjust the rocker arm’s height so it reaches a 90 degree relationship when the valve is about 2/3 open, not half open. The logic being, that the spring loads on the rocker body are less. Another reason I’ve heard is that it accelerates the valve in the “mid-range” better, thus making more power. Other variations of this approach shift the rocker arm’s pivot point higher on the valve tip to create this 90 degree effect sooner in accelerating the opening of the valve at a lower point of lift, thus increasing what is termed “area-underthe-

curve.”

Both of these are a way to add different cam information to the valve by using the rocker arm’s geometry. The only reason you would use the rocker arm for creating a “second dynamic” of valve acceleration, is if the cam was unable to give you the acceleration you needed. Now, in some cases this limitation exists. They would be flat tappet cams, of mechanical or hydraulic operation, and an engine where cylinder heads had the flow potential, and/or cubic inches had the demand which required a crazy acceleration to midlift flow values off the seat. In other words, the engine was so big, and the heads were so big but for rules or some other illogical reason, the cam they HAD TO USE was a flat tappet profile that had limited “rate of acceleration” by its limited tappet diameter and base circle constraints. I won’t get into cam technology and limits, but that is the first reason I can think of for using the rocker arm as its own cam tuner. In some Stock classes where the original cam must be used, a creative (and well funded) engine builder can play games with rocker geometry to change valve lift rates, but

these are very limited differences, usually not worth the trouble, and most of all, in ALL these examples, there is going to be detriments that outweigh benefits.

In the first place, for those who have an engine of large flowing heads, and big cubic inches, or heads for very high rpm’s, they will have the benefit of using a roller cam. So the issue of how fast you can open the valve is not even a consideration, because by nature of roller tappet geometry, any value of acceleration up to and through suicidal parts destruction can be implemented on the cam profile. And for those classes

where a flat tappet cam is required, the cubic inches and head limitations of most rules I’ve seen over the years, fall within airflow and rpm limits that a flat tappet cam fit just fine. Too many times, cam companies talk customers into roller profiles that are not needed, and in fact don’t make as much power as a well chosen flat tappet would, because it takes more power to operate the roller. Using rocker geometry as a second cam shaft is not a good idea. The velocity of the rocker arm increases where its motion line reaches a 90 degree angle, and trying

to pick a particular segment of the valve lift that you want to impose that thinking over what the cam manufacturer has done, is bad news. But there’s another point to consider on this issue.

The rocker arm is a symmetrical device to whatever geometry it is set at. In other words, whatever acceleration it exhibits on the opening side of the valve gets reversed on the closing side. Simply, if you set geometry with a HIGH pivot point, so it increases its velocity quickly off the valve seat, then slows to full lift... Guess what? It’s going to accelerate back to the close position when it leaves full lift. Because it will always mirror whatever its settings are.

To see the difference all you have to do is take an old fashioned needle pointer torque wrench and turn the engine over without spark plugs and measure the drag. Then set the geometry to MID-LIFT, and see the difference. I’ve had people do this and fall out of their seats. They see as much as 45 foot pounds or more and LESS torque to turn the engine over. Usually it is 15 or 20 foot pounds but it depends on how much spring pressure you have, how complex the rocker geometry is (Hemi versus SB Chevy, for instance), and how wrong their geometry really was. Either way, it quickly shows you the definition of efficiency. If they MAP their valve acceleration, throughout the entire lift curve, then I need say no more. The best way is to DIVIDE the arcs EQUALLY and standardize this for all cam and cylinder head testing. Change your cam

as you need.