Camshaft Basics
Here are a couple numbers for you. A Honda camshaft lobe has three wear signs, and a vintage Honda camshaft follower (“rocker arm”) is made one of two different ways. Let’s jump right in.

The cam’s wear patterns are: normal wear, pockmarking, and galling/tearing. First, normal wear. With rare exception (late hypersports models’ chrome-moly cams), Honda cams are made of soft, easily-drilled (even with a Harbor Freight drill bit!) cast iron or steel and wear at a surprisingly fast rate. By 10,000 miles these cams are considerably worn, and valve timing thus appreciably altered. But this is not a defect. So commonly is this encountered that I have perked up many engines simply by ordering aftermarket replacement cams made to stock specs which happily restored proper valve timing and full stock engine performance. The second cam wear presentation is a pockmarked surface on the cam lobe, cratering in other words. This is due to casting “inclusions” (air pockets) that are below the surface and appear only as the cam wears. Though very common on many Honda cams, this is definitely a cam defect. Despite that, it is not always a bad thing; it does not always require cam replacement. Cam lobe pockmarking is mild on some engines, worse in others. The issue is, does it aggravate the cam follower? Here is where the follower type comes in. Honda made cam followers (“rocker arms”) in two ways. On many 60s and 70s Hondas they are simply chrome plated. The chrome is a less expensive way to get the required hardness. But the chrome eventually buckles and fails, most often due to the aforementioned cam lobe pockmarking. The plating starts peeling, and this rocker arm then becomes a cheese grater and takes out the cam. Fortunately, many of Honda’s twins—the DOHC 450 and the very popular early 350--got properly heat-treated followers instead of chrome-plated ones (you can tell the hardened ones by the copper plating), and these followers do not self-destruct, though they can go bad for other reasons connected with lubrication. Unfortunately, you may have observed that Honda’s currently-available replacement followers for the DOHC 450 are chrome-plated instead of the original induction hardening, making them just like those on a CB750 and thus much more vulnerable than the production parts. The third cam wear observation is just plain destruction, that is, galling and tearing. The most common reason in vintage Honda engines is bad followers, but lubrication faults obviously can enter into it also.

Many seem to stress over cam timing, assuming that Honda made the marks to line up perfectly. They didn't and they don't. Add to this cam chain wear and some models’ very interpretive cam timing marks, and newcomers to the task can be challenged. Mechanics learn to judge correct timing by noting whether one sprocket tooth's variation is closer or farther relative to the mark.

The official Honda manual sometimes has you measure the cam lobe, heel (base circle) to nose. While having a purpose, the actual outcome of this measurement is not to gauge the potential of your cam or compare it to another cam. You can't do those things by this measurement; it's too gross, it spans too many of the cam lobe's important regions all at once, areas of the cam that need to be separated to get meaningful specifications. This measurement is not intended to tell you anything about the cam's specifications. It is simply a crude and easy way to track cam wear, and you'll discover an astonishing rate of wear doing this—Honda cams can wear as much as 0.003" every thousand miles. Some people feel that measuring heel to nose and then measuring again ninety degrees to that and subtracting, is a way to derive the cam's valve lift specification. As with many things, this sounds good but ignores reality. Three problems make this idea unworkable. First, the cam's heel or base circle is not necessarily continuous for 180 degrees. So you're not really measuring the base circle with your micrometer at ninety to the nose. In fact you're not measuring anything of any consequence. Second, when the cam is from a rocker arm equipped engine, the cam's movement profile does not match the valve's; the engine sees different movement due to the rocker arm's leverage ratio. Therefore valve movement—the thing you really want to know—is most accurately measured at the valve itself. Third, even on a given model engine, the cam's base circle diameter changes slightly from year to year, making the heel to nose measurement of absolutely no use in comparing cams. You can tell almost nothing significant from measuring the size of a cam lobe. The manual's reason for doing this is to track cam wear and nothing more.

In fact it's actually a lot more work than this to properly measure a cam. The correct way is what is called "profiling" the camshaft. The cam is left in the engine, a degree wheel is affixed to the crankshaft, and a dial indicator is placed onto the valve. As the crankshaft is rotated, the indicator's readings are marked on a piece of graph paper every five degrees, and the result is a right-angle, X and Y, duration versus lift, depiction of the valve movement caused by that cam.

A related note: Many inexperienced people have objected to the common career mechanic’s advocacy of running slightly looser valve clearances than stock. Some argue that it isn’t factory and thus can’t be right. Others are perplexed as the to value of such an adustment. And a few even object that a looser clearance will appreciably reduce the amount the valve is opened. This last argument demonstrates an unawareness of crankshaft-to-camshaft geometry. On a right angle graph that zooms in on a portion of a cam profile, you can easily visualize the effects of increased valve clearance on valve duration and valve lift. Guess what? Valve clearance has many times the effect on duration as it does on lift. That’s after all the reason for increasing it. Cam profiling shows this and a lot more, and is a tremendously useful tool for getting a meaningful "snapshot" of a cam. And far from being esoteric, once you begin to understand engines it is very educational and even indispensable when modifying an engine.

Camshaft Anatomy
The purpose of valve clearance is to compensate for heat expansion, right? Wrong. This assumption may be intuitive, but it is incorrect. Actually, as the valve in an overhead cam engine heats up, its head expands more than the castings around it, resulting in the valve being pulled into the combustion chamber, increasing valve clearance, not decreasing it. Why then, clearance? Valve clearance is wholly a product of cam lobe geometry.

A cam lobe is not just an ellipse machined onto a shaft. It's a complex shape with several distinct radii, each with its own job to do, and one of those dimensions is where the need for valve clearance comes from. Let’s start by looking at the cam’s "base circle", also called the heel. The base circle is that part of the cam lobe that looks mostly perfectly round, a constant radius. This is home base, the point of departure from which everything else on the cam is measured. It is also the valve's fully-closed point, giving the valve its cooling time. Opposite the cam's base circle is the "nose", the point at which the valve is fully open. The nose affects valve open spring pressure, an important consideration in performance engine building. On either side of the nose are very steep areas called "flanks". They open and close the valve.

But the most interesting of the cam lobe's parts is the "ramp". Detectable only with a dial indicator and degree wheel, the ramp is that tiny transition zone at which the radius of the base circle becomes the very broad surface of the flank. Though mysterious and complex, the ramp's job is simple: to cushion the valve. For a handful of crankshaft degrees, the ramp eases the valve into its sudden climb up the lobe's opening flank. Then after the valve has fully opened and is skiing down the closing flank, an identical ramp on the closing side of the lobe gently decelerates the valve so it doesn't crash onto its seat. However, there isn't actually room enough on any cam lobe for as much transition as the valve really needs, and that’s where valve clearance comes in. Valve clearance supplements the ramp’s cushioning. Different engines have different valve clearance specifications because they have different cams with different amounts of ramp, thus valve clearances that to differing amounts supplement those ramps. Modern engines run three to five times the valve clearance of engines of fifty years ago because their larger lobes necessarily take real estate away from the ramps (there aren’t more than 360 degree in a circle, after all), making the ramps smaller. Less ramp means less opening and closing transition, resulting in the need for more clearance. Small ramps go with big valve clearance, large ramps with small clearance. It's that simple. And heat has nothing to do with it. And neither does noise.

And here’s the really interesting thing about ramps. The fact is, these ramps accelerate and decelerate the valve so slowly that interpreting exactly when the valve has opened or closed during a cam profiling session is challenging. Because of this, performance engine builders have developed a method of ignoring the ramps when profiling a cam. The ramp isn't there to move the valve anyway, its sole purpose is to graduallly take up clearance—to "babysit" the valve. Performance engine builders know to disregard the valve's movement until the cam has rotated past the ramp and the valve has lifted a certain amount, then to start their graph paper markings, thus ensuring that the valve is well clear of the ramp. That delayed amount is called the cam's "checking height." A checking height is merely a predetermined starting and stopping point in cam measurement.

This is great. Checking height qualifies the process of measuring the cam, makes it more accurate. But now we have another problem: ramps aren't all the same size. The ramps on a pushrod engine's cam for example are the largest, because that engine's spindly valve train demands very gentle opening and closing curves. The ramps on overhead cam engines by contrast are much smaller, because there are fewer valve-related parts flailing about needing cushioning. Each engine design has its own unique ramp spec and correspondingly its own relevant checking height; some 0.020", others 0.040", and still others 0.050. Comparison of these different cams therefore becomes meaningless, as two different checking heights give two very different timing measurements. What's needed is a system of measuring cam timing that is independent of the checking height used. Is there such a thing? Yes! That's the “lobe center” procedure. That is why it was developed. Though it still requires the use of a checking height, the lobe center approach indexes the cam based on an average of its opening and closing points—the cam's mathematical "center"—not one of those points. For the purposes of data recording, we then count the number of crankshaft degrees from there to top dead center overlap. This means different checking heights can be used on a given cam and the lobe centers (thus timings) will still come out the same. No matter what the starting and stopping points are, the center of a thing is always the center. Checking heights, though necessary for profiling, are removed from comparisons of one cam to another, because not all cams require the same checking heights. This makes the whole thing a scientific process. Make sense? Lobe center divorces checking height, once an insurmountable hurdle in ecumenticizing cams, from cam measurement, making legitimate cam comparisons possible. That's why it was invented.

You might argue that no one compares the cam from a Harley Big Twin with that from a Hayabusa and of course you're right. That isn't the kind of comparison we're talking about. Then why do we need the lobe center method to compare two differently-speced cams for the same engine? Their checking heights will be the same, won't they? That's just it, they won't necessarily! Cam makers, each one having their own history and influences (for example, those favoring 0.050” checking height have automotive backgrounds), continue to use their favorite (and therefore different from the next guy’s) checking heights even on cams made for the exact same model motorcycle! Well, why isn't the cam lobe's center measured directly? Because it isn't the lobe's physical center we're looking for, it's its mathematical center--the average of its opening and closing points.

The Hard Stuff
Everyone thinks they know more about cams than they actually do. Aftermarket cams aren't something you can just bolt up and go. Dozens of painstaking measurements are used by engine builders and have been for nearly a hundred years to ensure a non-original camshaft’s compatibility with the engine. Such a camshaft absolutely must be installed carefully and correctly, and part of this involves making several valve-related checks on the engine. This is the real work, the mandatory stuff that needs to be done when installing an aftermarket camshaft. Because of this many of the cams in a catalog have become purposely dumbed-down to protect the cam manufacturer from inexperienced installers.

The first check is valve spring free-length. This is the length of the valve spring uninstalled and is usually the only spring specification given in the Honda service manual. Honda’s purpose is merely to enable you to easily spot fatigued springs, which interestingly enough is not that common on Hondas. However, when building a high performance engine we go way beyond this simple goal, to using spring free-length as the jumping off point for calculating several other important spring specifications.

One of these is installed spring height. This is the height of the valve spring installed, that is, slightly compressed already, even before the valve has moved. The purpose of knowing installed height is primarily to gauge valve seat pressure. Valve seat pressure is important in any engine, stock or modified, but the modified engine's increased cylinder pressure requires no-nonsense sealing. Also, at high rpm there is less time for valve cooling, so good seating becomes more crucial on that basis as well.

Then there’s valve spring full-open length. This is the spring's shortest working length, when the valve is fully open, the hardest the spring can be expected to work. Among other things, knowing spring full-open length helps the engine builder determine if there is sufficient spring pressure to control valve float. All performance camshafts increase how quickly the valve opens, adding to valve acceleration and making controlling valve "float" more of a concern. The spring's full-open pressure is your first line of defense against valve float, and it will often need to be more than stock in a higher-performing engine.

And then valve spring coilbind. This is when the spring's coils touch one another, metal to metal. The spring is compressed past the valve's full-open point. Coilbind must never occur during engine operation, and there's something wrong with a spring that comes near to doing so. In fact, the purpose of checking for coilbind clearance is to validate your choice of aftermarket valve springs, to point out when they are not the correct rate.

And don’t forget valve free travel. Also called retainer-to-guide, this is the amount that the valve can travel from closed all the way to where the spring retainer hits the valve guide (or valve guide seal). One of the most important checks, many engine builders have failed to respect this one and have paid dearly.

And of course, valve-to-piston clearance. Valve to piston clearance necessarily changes in an engine modified with a different piston or camshaft, or when having remachined castings almost anywhere in the top end, or different thickness gaskets. Even when the cam is stock. To do this check, carburetor springs are put in place of the valve springs. A degree wheel on the crankshaft and a dial indicator on the valve tell us how much the valve can move (when pushed with a finger) before touching the piston, at points from 0 to 15 degrees either side of TDC overlap, the danger zone, in 5 degree steps. This is unarguably the most important of all hot-rod engine building top end clearance checks. Don't believe what high performance parts retailers tell you. It's your engine, find out for yourself.

You can see from this that most of the work centers around the valve and valve spring. Honda designed both for the exact forces and geometry inherent in the stock engine, and changes to that engine can run afoul of this careful design. A dealership I worked at once wanted to give away high performance camshafts as part of a promotion. What a disaster that could have been! Few average Joes practice or even know about these measurement techniques. Not only are these tests required, but there are several other "gotchas" to look out for, such as clearance of a larger cam's lobe with engine castings, rocker arm interference, valve protrusion, and more. Hopefully it is clear that a professional performance engine builder never "throws" an aftermarket cam at an engine.

Speaking of valve protrusion, in most vintage Hondas it is not a real concern. The adjusting screws on most rocker arm valve trains provide considerable compensation when a valve sticks up farther in the head due to valve seat machining, certainly enough to accommodate the typical 0.005”-0.010” extra valve height. But those Hondas having limited adjustabiltiy—the eccentrically-adjusted 350 and DOHC 450, plus all of Honda’s later shim type engines—will present difficulty. In these cases valve protrusion becomes an issue to be dealt with. This is one more reason to prefer stainless steel replacement valves. Unlike the plated stock valves and even their aftermarket substitutes, stainless valves can safely be shortened, thereby reducing excess valve protrusion and eliminating problems associated with limited-range eccentric adjusters or with a fixed range of available valve adjustment shims.

Duration and Other Secrets
No part of the four-stroke high performance engine embodies more mystery to the average enthusiast than the camshaft. But a camshaft is not a silver bullet. In fact, if anything is really that special, it’s the cylinder head, which when removed from the engine most race mechanics keep wrapped in a towel to ward off prying eyes. That’s where the power is, and not all of it centers around the ports.

But the cam does play an important role. Better cylinder filling is of course the goal of opening the valves for a longer period via an aftermarket cam’s extended valve duration. But there is a drawback. Exposing the cylinder to the outside air for longer periods slows the intake airstream. The result is less optimum mixture distribution in the combustion chamber at lower engine speeds—the fuel separates from the charge. In addition, extended valve duration cuts into the time the cylinder has to compress its mixture—the engine has less compression. So while added valve duration often yields a high-rpm power increase, a side-effect is often lost low-rpm power. The engine is “cammy”. Extended valve duration also increases valve overlap, that period during which both the intake and exhaust valves are open. As valve overlap increases, the wave (lengths and angles based) tuning of the intake and exhaust systems becomes demonstrably more effective. These latent effects are “wakened up” and utilized. However, again there is a tradeoff because the engine thus modified depends less on its mechanical parts and more on its wave pulses, which are somewhat tenuous, to run well at all. Rather like how a highly-developed two-stroke engine behaves. Therefore, though power-adding, wave tuning is effective for only a very narrow rpm range—the engine runs better than stock, but only above a certain rpm, below which it runs noticeably worse than stock.

A lot can be said about tuned exhausts: diameters, angles, lengths, the speed of the pulsed waves, and all the rest. It makes for interesting reading; there really is some physics going on there that excites the mathematically-minded. But one important thing should be considered at the outset: it does no good to put on a tuned exhaust if the camshaft isn’t there to support it. The camshafts in Honda street bikes made until well into the 1990s are all designed with almost no valve overlap. Thus exhaust tuning can’t work—the valves aren’t open to receive the waves resulting from exhaust tuning. A tuned exhaust on a vintage Honda cannot improve engine performance unless the camshaft is also changed, although the weight reduction and reduced back pressure (in very loud exhausts having large baffles) can have a small performance effect. For that matter, not all “performance” exhausts are tuned. Exhaust pervayers Dennis Manning and Terry Vance went on record decades ago as saying their products owe more to serendipity and style than to engineering. And despite Internet conversations to the contrary, there is no appreciable tuning in stock exhausts.

There is another, less often considered, way the camshaft can be used to increase cylinder filling, by opening the valve farther instead of holding it open longer. More valve lift, in other words. Like the valve that is held open longer, the valve that is opened farther allows more air into the cylinder, potentially increasing power. However, unlike longer-duration cams, since cylinder exposure to atmosphere is not increased, the gases moving in the tract are not slowed, and the charge enters the cylinder still well mixed and combusts nicely. And the engine isn't made more dependent on wave tuning by extra valve overlap. And there is no loss of cylinder compression. There is plenty of low-rpm power. That sounds like a win/win. Why aren't all camshafts made this way then, with an emphasis on lift instead of on duration, especially performance cams? There are three reasons. First, there is valve acceleration. The valve that is opened farther in the same amount of time is opened more quickly, i.e. accelerated harder. The valve is then closer to the point of "float," and valve float is a serious threat. Thus high-lift cams are potentially problematic. It takes some careful reengineering of the rest of the engine to make them work, making such cams a greater liability for the manufacturer in the hands of home-brew engine builders. Second, because of the motoring world's historic fixation on valve duration, a certain politic of never opening a valve farther than 25 percent of its diameter has prevailed. Until the mid 1980s, few if any production Honda valves did. This 25 percent figure came from an interesting geometrical fact: when the valve is opened to a distance equal to one-fourth its diameter, the flow curtain around the valve's periphery is as effective as it will ever be. At least that’s what the math says. However, consider for how long that valve is fully open. An almost infentessimal period. Openings beyond 0.25d effectively extend the amount of time the valve is fully open, without actually increasing valve open duration and thus without adversly affecting tract speed and ultimately combustion. More open time without more duration. In other words, when the valve is opened to say 0.35d-0.40d—as most are today—the valve, instead of “seeing” 0.25d for a thousandth of a second, actually dwells a bit at 0.25d, increasing flow potential without the drawback of extended duration. Engine builders discovered this principle and engineers have confirmed it—it has almost become the new tradition. It works. But the third and the real reason not many aftermarket camshafts are designed for lift is that such cams are very difficult to market. Their numbers aren't as impressive.

Speaking of unimpressive, 60s and 70s Honda cams are pretty tame, typically producing only 5-6mm lift and 10-40/40-10 timing yielding a scant 230 degrees duration of the valve. Efforts at special degreeing of them, in the absence of other engine changes, has not proven very effective. But there is one exception: worn cams and cam chains. A worn cam opens its valves later than normal, as does also a worn chain. Thus a timing that when new was 10 degrees BTDC can only a few thousand miles later profile at 5 degrees or even 0. I have observed this firsthand and it is very common in Hondas, whose cams wear quickly. A stock engine therefore can regain some of its lost power by either replacing the cam and cam chain—which of course is best—or by re-degreeing the worn cam to the original 10 degrees through the expedient of slotting its sprocket. Camshaft degreeing of a stock engine is in this instance fruitful, and it is something I have done many times when the customer has not wanted to replace the cam.