Cams, or motion designed on paper

Drawing a motion profile and turning it into a physical shape.

A cam is motion designed on paper first and manufactured second, because its entire job, however complicated the sequence it produces, is drawn as a simple graph of position against rotation before a single curve of its actual outline is ever cut, the physical shape existing purely to reproduce whatever motion that graph specified.

Sketching out a cam's motion profile on paper before cutting anything, a rise here, a pause there, a fast drop right at the end, and then watching a follower trace that exact sequence back out once the shape existed in metal, made something click that a description of the mechanism alone never quite had, that a cam is really a shape that has been asked to remember a sequence and simply plays it back, once per rotation, for as long as it keeps turning.

Introduction and overview

A cam converts a simple rotation into whatever more complicated motion its designer actually wants, by giving a follower, a roller, a pin, a flat-faced pusher, a surface to ride against whose distance from the cam's own centre changes as the cam turns. Where that surface sits far from the centre, the follower is pushed out; where it sits close, the follower is allowed to drop back in; and the entire sequence of pushes, pauses and drops the follower experiences across one full rotation is encoded directly into the cam's outline, point by point, the way a recipe's steps are encoded in the order its instructions are written down.

The music-box comparison

A music box's rotating cylinder is studded with small pins arranged around its surface in a specific pattern, and as the cylinder turns, each pin in turn lifts and releases a tuned metal tooth, producing one note at exactly the right moment relative to all the others, the melody itself existing nowhere except in the precise arrangement of pins around that one cylinder. Nothing about the cylinder's own rotation is musical on its own, a plain, pinless cylinder turning at the same steady speed produces nothing but silence, and the entire tune is a direct, physical translation of a sequence someone worked out on paper first, note by note, before it was ever pressed into the pins' actual positions. A cam works by exactly the same principle applied to motion instead of sound, its outline standing in for the pins, a follower's rise and fall standing in for the notes, and the whole sequence existing nowhere but in the shape someone deliberately gave the cam before it ever turned a single revolution.

Why a cam can do what a linkage cannot

A linkage, however cleverly proportioned, is limited to whatever motion the geometry of its connected rigid bars is capable of tracing, a family of curves determined entirely by a handful of link lengths chosen in advance. A cam carries no such restriction, since its outline can be drawn to specify literally any motion profile a designer can put on paper, a follower that dwells motionless for most of a rotation and then leaps suddenly, a motion no simple linkage could ever reproduce cleanly, becomes entirely routine for a cam, since the cam's shape simply is that motion, transcribed directly from the graph that defined it. This is the real reason cams show up wherever a precise, possibly unusual timing matters more than mechanical simplicity, an engine's valves opening and closing at exactly the right instant in a cycle, a sewing machine's needle pausing at the top of its stroke long enough for the thread to be caught.

One figure worth keeping in mind

A cam's outline can be specified at as many points around its rotation as the motion profile actually needs, sometimes hundreds of individually calculated positions feeding into a single smooth curve, giving a designer effectively unlimited control over the follower's motion at the cost of needing a correspondingly precise manufacturing process to actually cut that curve accurately. A simple four-bar linkage, by contrast, is fully described by only four numbers, its link lengths, which is far easier to specify and build but offers nowhere near the same freedom over the exact shape of the resulting motion.

What follows from this

Designing with a cam means starting from the motion actually wanted, not from a mechanism assumed in advance, drawing out how the follower should move across one full cycle and only then working backwards to the physical shape that produces exactly that. This inverted design process, motion first, shape second, is the opposite of how a linkage typically gets designed, where the shape of the mechanism is chosen first and the resulting motion is whatever falls out of that choice, and recognising which direction a given design problem actually runs in saves a great deal of wasted effort trying to force the wrong tool to answer the wrong kind of question. A designer who starts sketching cam outlines before ever writing down the motion they actually want is, in a real sense, working backwards, guessing at a shape and hoping the motion that falls out of it happens to be useful, rather than starting from the one thing that was actually meant to be designed.

What this does not explain

A cam's flexibility comes at the cost of a permanently fixed motion profile, since changing what the cam does at all means cutting a new cam, unlike a linkage or an electronically controlled actuator, either of which can sometimes be adjusted without remaking a physical part. A cam is an excellent answer to a motion that is known precisely and will not need to change, and a poor one for a motion that is still being worked out or that genuinely needs to vary between one cycle and the next.

This trade-off is itself the reason cams and linkages tend to specialise in different parts of the same machine rather than competing directly, a linkage handling the coarse, repeatable transport motion a design is unlikely to ever need changed, and a cam handling the fine, precisely timed detail work where getting the exact profile right matters more than being able to adjust it easily afterwards.