Turn an unpowered permanent-magnet motor slowly by hand and you can feel it: a series of gentle notches, as if the shaft wanted to settle into preferred positions. That is cogging torque, a purely magnetic effect that exists whether or not the motor is energized. It comes from the rotor magnets being pulled toward the iron teeth of the stator, and it produces a rippling, uneven torque that is most troublesome at low speed. This guide defines cogging torque, distinguishes it from load-driven torque ripple, and describes how motor designers and drives work to smooth it out.
Cogging Torque in one line: Cogging torque is a position-dependent torque in a permanent-magnet motor caused by the attraction between the rotor magnets and the slotted iron teeth of the stator, which pulls the rotor toward preferred alignments. Because it varies with rotor angle and exists even when the motor is unpowered, it produces a periodic velocity ripple that is most noticeable at low speed, and it is reduced through motor design and, where needed, active compensation in the drive.
Cogging torque arises from the basic construction of a slotted permanent-magnet motor. The stator windings sit in slots between teeth of iron, and the rotor carries strong permanent magnets. Magnets are attracted to iron, so as the rotor turns, its magnets are drawn toward the nearest stator teeth, preferring positions where the magnetic path through the iron is easiest. At those preferred alignments the rotor sits in a shallow magnetic detent, and moving away from them takes a little extra torque, while approaching them the magnets pull the rotor in. The result is a torque that rises and falls with rotor position, entirely from magnetism and geometry, with no current involved.
Because it depends only on the relative position of magnets and teeth, cogging torque is present even in a completely de-energized motor, which is why you can feel it when turning the shaft by hand. It repeats many times per revolution, at a spatial frequency set by the number of magnet poles and stator slots, so the rotor passes through a series of these detents as it turns. The detents are what give a cogging motor that characteristic notchy feel and what the term detent torque refers to when discussing the same effect in the unpowered state.
The strength of cogging depends heavily on how the motor is built. The size and shape of the slots, the strength and placement of the magnets, and the geometry of the teeth all influence how strong the pull toward each alignment is. A motor with large open slots and strong magnets tends to cog noticeably, while careful design choices can make the detents much shallower. This is why cogging is fundamentally a motor-design matter first, with the drive doing what it can afterward for what remains.
It is important to separate cogging torque from other sources of uneven torque, because they have different causes and cures. Cogging is intrinsic to the motor's magnetic structure and exists with no current flowing. Torque ripple more broadly can also come from the way the drive commutates the motor, from imperfections in the phase currents, and from the load itself, for example a pump, gear, or mechanism that presents a torque that varies with position. Load-driven ripple originates outside the motor's magnetics, whereas cogging originates inside them, so telling them apart is the first step to smoothing an axis.
Whatever its source, this kind of position-dependent torque variation is felt most acutely at low speed. When a motor turns quickly, its own inertia and the inertia of the load act as a flywheel that smooths over the rapid ripples, so the speed barely wavers between detents. At low speed there is little such smoothing, and each detent produces a visible surge and lull in velocity as the rotor is pulled into and out of alignment. The motor may even move in tiny jerks rather than gliding, which shows up as velocity ripple and, in precise applications, as a defect in whatever the axis is producing.
This low-speed roughness is exactly what matters in the applications where smoothness is prized, such as slow, steady scanning motions, direct-drive stages, and any process where uneven speed leaves a mark. In those cases even a small cogging torque that is invisible at running speed becomes the limiting factor on quality. The demand for smooth low-speed motion is therefore the main reason so much effort goes into reducing cogging, both in how the motor is designed and in how the drive is set up to counter it.
Motor designers have well-established ways to reduce cogging at its source. Skewing, in which the slots or magnets are twisted slightly along the length of the motor rather than running straight, spreads the detents out so they no longer all line up and reinforce, greatly softening the effect. Careful choice of the pole and slot combination, shaping of the magnets and teeth, and specialized constructions such as slotless designs all reduce cogging, though often with trade-offs in cost, torque density, or manufacturability. A motor intended for smooth low-speed work is usually designed from the start to cog as little as possible.
The drive can then compensate for the cogging that remains. Because cogging torque is a repeatable function of rotor position, a drive can learn its shape, essentially mapping how much extra torque is needed at each rotor angle to cancel the detent, and then inject that corrective torque as the motor turns. This active cogging compensation, combined with smooth field-oriented commutation and clean phase currents, can markedly improve low-speed smoothness beyond what the motor delivers on its own. Higher-resolution feedback helps here too, because cancelling a position-dependent effect requires knowing rotor position finely.
The cogging behaviour of a motor is fixed deep in the drive and motor, but its effects on process quality are the kind of thing a plant monitors through cloud SCADA. Velocity ripple, low-speed roughness, and the resulting variation in whatever a machine produces can be trended by a platform such as Merobix, and a machine that has become rougher at low speed, or whose product quality has drifted on slow moves, points either to a motor or drive problem or to compensation that is no longer matching the motor. For operations running many similar machines across sites, comparing low-speed smoothness and quality metrics centrally helps a team spot the outlier axis and decide whether it needs re-tuning, re-compensation, or replacement, without evaluating each machine by hand and feel.
They describe the same underlying effect, the pull of the rotor magnets toward the stator teeth, from slightly different angles. Detent torque usually refers to the torque you feel when turning an unpowered motor by hand, the resistance holding it in its preferred positions. Cogging torque is the term used when that same position-dependent torque disturbs a running motor's smoothness. Both come from magnet-to-slot attraction and exist with no current flowing.
At higher speeds the combined inertia of the motor and load acts like a flywheel that smooths over the rapid position-dependent torque ripples, so the speed barely wavers. At low speed there is little inertial smoothing, so each cogging detent produces a visible surge and lull in velocity as the rotor is pulled into and out of alignment. This is why cogging matters most in slow, steady motions where velocity ripple would show up in the product.
It is tackled first in the motor design, through techniques such as skewing the slots or magnets, choosing a favorable pole and slot combination, shaping the magnets and teeth, or using slotless constructions, all of which soften the detents. The drive can then compensate for what remains by learning the cogging as a function of rotor position and injecting corrective torque to cancel it, aided by smooth commutation and high-resolution feedback. Design and drive compensation together give the smoothest result.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.