Mistake Master
Rotational kinetic energy is $K_\text{rot} = \tfrac{1}{2}I\omega^2$, the rotational twin of $K = \tfrac{1}{2}mv^2$. A spinning wheel on a fixed axle has no center-of-mass motion, so the 3.1 formula misses it; $K_\text{rot}$ catches the motion of every particle in the body, packaged through $I$ and $\omega^2$. Same quadratic trap as 3.1: doubling $\omega$ quadruples $K_\text{rot}$, not doubles it.
3.1's formula $K = \tfrac{1}{2}mv^2$ handles anything that translates: a block, a ball, a cart. It doesn't handle a flywheel spinning on a fixed axle. The flywheel's center of mass is at rest, so 3.1 would say $K = 0$, yet the wheel is clearly carrying energy.
The fix is a second kinetic-energy formula for the rotational piece: $K_\text{rot} = \tfrac{1}{2}I\omega^2$. It plays the same role for a rotating body that $\tfrac{1}{2}mv^2$ plays for a translating one. The two are additive: for a rigid body that's doing both at once, like a rolling ball, $K_\text{tot} = K_\text{trans} + K_\text{rot}$.
Two warnings up front. (1) $K_\text{rot}$ scales with $\omega^2$, not $\omega$. Doubling the angular velocity quadruples the rotational KE; this is the headline trap of the topic. (2) $K_\text{rot}$ is a scalar. The direction of rotation matters for angular momentum (Topic 6.3), but for kinetic energy the $\omega^2$ scrubs the sign just like $v^2$ does for translational KE.
For a rigid body rotating with rotational inertia $I$ about a fixed axis at angular speed $\omega$:
$$K_\text{rot} = \dfrac{1}{2}I\omega^2.$$Units: kg, m, rad/s, joules. The $\tfrac{1}{2}$ falls out the same way it did in 3.1: every particle in the body has translational $\tfrac{1}{2}m_iv_i^2$, with $v_i = r_i\omega$ for a rigid rotation. Adding those up gives $\tfrac{1}{2}\omega^2 \sum m_i r_i^2 = \tfrac{1}{2}I\omega^2$. So $K_\text{rot}$ isn't a new kind of energy; it's the sum of the translational KEs of all the pieces, repackaged into one formula that uses $I$ instead of tracking each piece by hand.
The structure mirrors $\tfrac{1}{2}mv^2$ exactly: an inertia term times the speed squared, with a half out front. Replace $m$ with $I$ and $v$ with $\omega$ and you've moved from translation to rotation.
A symmetric dumbbell: two $0.5$ kg point masses on a light rod, each $r = 0.6$ m from the axis through the rod's center. The dumbbell spins at $\omega = 4$ rad/s. Find $K_\text{rot}$.
Two point masses at distance $r$ from the axis give $I = 2mr^2 = 2(0.5)(0.6)^2 = 0.36 \text{ kg}\cdot\text{m}^2.$
$$K_\text{rot} = \tfrac{1}{2}I\omega^2 = \tfrac{1}{2}(0.36)(4)^2 = \tfrac{1}{2}(0.36)(16) = 2.88 \text{ J}.$$
Scalar, positive, no direction. Squaring matters: using $\tfrac{1}{2}I\omega$ would give $0.72$ J, dimensionally wrong and off by a factor of 4. The $\omega^2$ is the heart of the formula.
A flywheel on a fixed axle has only $K_\text{rot}$. A sliding block has only $K_\text{trans}$. A rolling ball has both at once: its center of mass is moving (translation) and the body is spinning about that center (rotation). The total kinetic energy is the sum:
$$K_\text{tot} = K_\text{trans} + K_\text{rot} = \tfrac{1}{2}Mv_\text{cm}^2 + \tfrac{1}{2}I\omega^2.$$Both terms, every time, for any rigid body that's doing both at once.
For rolling without slipping, the linear and angular speeds are tied together by $v_\text{cm} = R\omega$, so $\omega = v_\text{cm}/R$. That lets you write $K_\text{rot}$ in terms of $v_\text{cm}$ alone, which is what makes the ramp energy-conservation problems work out cleanly.
A uniform solid cylinder of mass $M = 2$ kg and radius $R = 0.1$ m rolls without slipping from rest down a ramp of height $h = 0.6$ m. Find its center-of-mass speed at the bottom. Use $g = 10$ N/kg and $I_\text{cyl} = \tfrac{1}{2}MR^2$.
At the top, at rest: $E_i = Mgh = (2)(10)(0.6) = 12 \text{ J}.$
At the bottom, rolling: $K_\text{tot} = \tfrac{1}{2}Mv^2 + \tfrac{1}{2}I\omega^2.$ Use $\omega = v/R$ and $I = \tfrac{1}{2}MR^2$:
$$K_\text{rot} = \tfrac{1}{2}\cdot\tfrac{1}{2}MR^2\cdot\dfrac{v^2}{R^2} = \tfrac{1}{4}Mv^2,$$so $K_\text{tot} = \tfrac{1}{2}Mv^2 + \tfrac{1}{4}Mv^2 = \tfrac{3}{4}Mv^2.$ The $R$ cancels.
Set $Mgh = \tfrac{3}{4}Mv^2$: $v^2 = \dfrac{4gh}{3} = \dfrac{4(10)(0.6)}{3} = 8,$ so $v = \sqrt{8} \approx 2.83$ m/s.
A frictionless sliding block on the same ramp would reach $v = \sqrt{2gh} = \sqrt{12} \approx 3.46$ m/s. The cylinder is slower because a third of its kinetic-energy budget is tied up in rotation instead of translation. The "missing" energy isn't missing; it's spinning.
Limiting cases. If $I = 0$ (a point particle, or a massless cylinder), $K_\text{rot}$ disappears and you recover $v = \sqrt{2gh}$, the standard frictionless-slide answer. At the other extreme, a hoop has $I = MR^2$, so $K_\text{rot} = \tfrac{1}{2}Mv^2$, $K_\text{tot} = Mv^2$, and $v = \sqrt{gh}$, slower still. The hoop puts the most mass at the rim, so most of the $mgh$ budget goes into rotation and the least into translation.
Each error below mishandles rotational kinetic energy in a different way — how it scales with spin, what motion it counts, and whether it has a direction at all. You'll meet all three in your diagnostic.
$K_\text{rot}$ scales with $\omega^2$, not $\omega$. Most physics quantities you've seen scale linearly (twice the force, twice the acceleration), so the instinct is to scale $K_\text{rot}$ linearly with $\omega$ too. It doesn't.
Fix. When $\omega$ changes by a factor $r$, $K_\text{rot}$ changes by $r^2$: double $\omega$ and you quadruple $K_\text{rot}$; triple $\omega$ and you multiply by 9; halve $\omega$ and you quarter it. The same trap as 3.1 with $v$, dressed up for rotation.
3.1 said $K = \tfrac{1}{2}mv^2$ for a moving object, full stop, so this looks complete. But a rolling cylinder is doing two things at once: its center of mass is moving (translation) and it's spinning about that center (rotation). Both carry kinetic energy.
Fix. For any rigid body that translates and rotates, $K_\text{tot} = K_\text{trans} + K_\text{rot}$. Both terms, every time. Reporting only $\tfrac{1}{2}Mv^2$ for a rolling object misses the $\tfrac{1}{2}I\omega^2$ piece; for a solid cylinder, that's a third of the actual total.
Angular velocity has a direction (CCW or CW), and $K_\text{rot}$ contains $\omega$, so a direction "ought to" survive. But $\omega$ enters squared, and squaring a signed quantity erases the sign.
Fix. $K_\text{rot}$ is a scalar, like $K_\text{trans}$. Two flywheels spinning at $\omega = +5$ rad/s and $\omega = -5$ rad/s carry the same $K_\text{rot}$. The direction of rotation matters for angular momentum (Topic 6.3); it does not matter for rotational kinetic energy.