Cam Profile Design: Follower Motion Curves and Avoiding Jerk and Vibration

Cam design is one of those areas of mechanical engineering where the math is well documented but the practical judgment calls are almost never taught properly, and I’ve spent a good chunk of my twenty years as a contract designer inside Japanese manufacturing plants cleaning up cam mechanisms that looked fine on paper but shook themselves apart or wore out prematurely in service. The core issue almost always comes back to the motion curve chosen for the follower — not just the displacement, but the velocity, acceleration, and especially the jerk (the rate of change of acceleration) that the curve produces. A cam that moves a follower from point A to point B might satisfy the displacement requirement with half a dozen different curve shapes, but only some of those shapes will avoid exciting vibration and shock loading in the rest of the machine. I want to walk through how I actually select and evaluate cam profiles on real projects, including cases where the wrong choice of motion curve caused real reliability problems.

Why Displacement Alone Never Tells the Whole Story

When I review a cam design that someone else has laid out, the first thing I ask for isn’t the displacement diagram — it’s the acceleration curve. Displacement and even velocity curves can look perfectly smooth and still hide a discontinuity in acceleration, and it’s that discontinuity, not the displacement itself, that generates the shock loading and vibration that shows up as noise, wear, and eventually fatigue failure in the follower train.

The classic example I use when training junior engineers is the simple harmonic motion curve versus the cycloidal curve. Both produce a smooth, continuous displacement and velocity profile, but simple harmonic motion has a non-zero acceleration at the very start and end of the stroke when it’s used for a rise-dwell-fall sequence, which means acceleration jumps abruptly at the dwell transition — a real discontinuity that excites vibration every single cycle. The cycloidal curve, by contrast, brings acceleration smoothly to zero at both ends of the stroke, which is exactly why it’s my default choice whenever a cam motion segment connects to a dwell period.

A Real Case: Chatter on a High-Speed Stamping Feed Cam

I was called in on a stamping line where a cam-driven feed mechanism was developing an audible chatter at the top of its stroke that got worse as the line speed increased. The mechanism had been designed using simple harmonic motion for a rise-dwell-fall cycle, which is a common enough choice that nobody had flagged it during the original design review.

When I plotted the acceleration curve, the discontinuity at the dwell transition was obvious — acceleration jumped from a substantial non-zero value straight to zero the instant the follower entered the dwell period. At the original, slower line speed, the resulting shock had been small enough that nobody noticed it, but as the plant increased throughput, the same discontinuity scaled up with the square of the speed increase and became a real vibration and noise problem, along with visibly accelerated wear on the follower roller. We re-cut the cam using a modified trapezoidal acceleration curve for the rise-dwell-fall segments, which eliminates the acceleration discontinuity at the dwell boundaries entirely. After the re-cut, the chatter disappeared and the follower roller wear rate dropped enough that the maintenance interval could be extended significantly. That project is the reason I now ask about the target line speed before I ever recommend a specific motion curve family — a curve that’s perfectly acceptable at one speed can become a serious problem at a higher one because shock loading scales with the square of speed.

Jerk: The Curve Most Engineers Forget to Check

Beyond acceleration, the derivative of acceleration — jerk — matters more than most engineers expect, particularly for high-speed cams or applications sensitive to vibration-induced noise or fatigue. A finite jump in acceleration corresponds to an infinite jerk spike at that instant, and while no real mechanism responds to a true mathematical infinity, the finite stiffness and mass of the actual hardware turn that theoretical spike into a very real, very short-duration force transient that excites the natural frequencies of the follower train.

On a cam-driven valve actuation mechanism I worked on, we were seeing occasional fatigue cracking in a follower arm that was otherwise well within its static stress allowable based on the nominal acceleration values from the motion curve. When we looked more carefully at the jerk profile, the curve we’d been using had a finite but fairly abrupt jerk discontinuity at a mid-stroke transition point, and that transient was exciting a resonance in the follower arm at the operating speed, producing dynamic stresses well above what the static calculation predicted. Switching to a motion curve family with continuous jerk throughout — one of the polynomial or “3-4-5” curve families — resolved the fatigue issue without any change to the follower arm’s static design. That case taught me to never sign off on a high-speed cam design based on peak acceleration alone; I check the jerk profile as a matter of course now, especially for anything running above a few hundred cycles per minute.

Choosing Among the Common Motion Curve Families

I keep something like this table in my head whenever a new cam project comes in, because the “right” answer depends heavily on the speed and the vibration sensitivity of the surrounding structure, not on habit or what the previous design used. Modified trapezoidal curves, for instance, actually produce lower peak acceleration than cycloidal curves for the same displacement and time, which makes them attractive for high-speed dwell transitions even though the math is more involved to lay out by hand.

Cam-Follower Interface: Where Theory Meets Real Contact Stress

A motion curve that’s perfect on paper can still produce a poor cam if the follower interface isn’t sized correctly for the resulting contact forces. I’ve seen designs where the acceleration profile was well chosen, but the follower roller diameter was too small relative to the cam’s radius of curvature at the point of maximum acceleration, producing contact (Hertzian) stresses well above what the roller bearing could sustain for the required cycle life.

On a packaging machine cam, we had premature roller bearing failures that traced back to exactly this — the motion curve itself was fine, but nobody had checked the radius of curvature of the cam profile at the highest-force point against the follower roller diameter. The local radius of curvature had gotten small enough at that point that the effective contact geometry was much sharper than the nominal roller size would suggest, spiking the Hertzian contact stress. We increased the base circle diameter of the cam, which increased the radius of curvature throughout the profile at the cost of a slightly larger overall cam package, and the bearing life came back in line with the design target. Since then, I always check minimum radius of curvature against follower geometry as a separate step from the motion curve selection, because they’re related but distinct failure modes.

Dwell Periods and the Danger of Assuming “Dwell Means No Motion Concerns”

It’s tempting to treat dwell periods as the “easy” part of a cam design, since the follower isn’t moving, but the transition into and out of dwell is exactly where most of the vibration problems I’ve traced actually originate. A dwell period with an abrupt entry — where velocity or acceleration doesn’t smoothly approach zero — transmits shock into the dwell segment even though the follower is nominally stationary there, which can be a real problem if there’s a sensitive operation, like an inspection or a precision placement, happening during that dwell.

On a vision inspection station driven by a cam-indexed table, we had intermittent false rejects that eventually traced back to residual vibration during what was supposed to be a settled dwell period, caused by an under-designed transition curve into that dwell. Switching the transition segment to a curve with continuous acceleration and jerk through the dwell boundary resolved the false reject rate without any change to the inspection system itself, which was a good reminder that “dwell” doesn’t mean “the cam design is done” — the entry and exit conditions deserve just as much scrutiny as the active motion segment.

My Practical Checklist for Any New Cam Design

Whenever I take on a new cam design, I start by confirming the required displacement, dwell locations, and target operating speed, then choose a motion curve family based on that speed rather than defaulting to whatever was used on the previous project. I always plot and review the acceleration and jerk curves, not just displacement, before committing to a profile, and I pay particular attention to any transition into or out of a dwell period, since that’s where discontinuities most often hide. Finally, I check the cam’s radius of curvature against the follower geometry separately from the motion curve analysis, because a beautifully smooth acceleration profile doesn’t protect you from a locally sharp contact geometry driving up Hertzian stress. Cam design rewards this kind of systematic checking precisely because the failure modes are subtle enough that they rarely show up in a static review — they show up months later, as chatter, wear, or fatigue cracking, once the machine has been running at speed.

Reference Standard

The exact rules behind these callouts trace back to one standard that most drawings ultimately answer to.

ASME Y14.5-2018: Dimensioning and Tolerancing

As an Amazon Associate I earn from qualifying purchases.