Angular kinematics and the link to the motion of each point, torque as a cross product and a lever arm, moments of inertia by summation, integration and the parallel-axis theorem, the second law for rotation applied to massive pulleys and spools, and static equilibrium of beams, ladders and hinged struts.
Subject: AP Physics C: Mechanics · 69 slides · diagram-first lesson
Open the interactive version of this deck
Title
AP Physics C: Mechanics — Unit 5
Angular kinematics, torque, moment of inertia, the second law for rotation, and static equilibrium
Objectives
Everything so far has treated objects as points. Real objects have size, and an object with size can turn as well as move. This deck builds the machinery for turning, and it does so by translating every linear idea you already own into a rotational one.
The organising idea: torque is to rotation what force is to translation, and moment of inertia is to rotation what mass is to translation. Hold that dictionary in your head and most of this deck is the second law written in a new alphabet.
Up to now every object in this course has been a point. A block on an incline, a cart on a track, a ball in flight — you tracked one position and ignored the fact that the object has size. That works beautifully until the object starts to turn.
Turning is not a small correction. A door, a wheel, a pulley, a ladder, a spinning skater: in each one the interesting physics is the rotation, and none of the tools so far can describe it.
The good news is that you do not have to learn a new subject. Rotation has its own versions of position, velocity, acceleration, force and mass, and they obey equations with exactly the same shape as the ones you already know. Most of this deck is a translation exercise.
Keep the dictionary from the last paragraph of the slide in front of you throughout. Torque plays the role of force. Moment of inertia plays the role of mass. Angular acceleration plays the role of acceleration. When you are stuck, ask what you would do in the linear version, and do the same thing in the rotational one.
Section
Section 1
Concept
\[ \theta = \frac{s}{r} \qquad s = r\theta \]
Figure (svg): A circle with two radii one radian apart; the arc between them is highlighted and is exactly as long as the radius.
A radian is defined as a ratio of two lengths, so it is dimensionless. That is not a technicality — it is the reason that s = r theta, v = r omega and a = r alpha have no conversion factors in them.
Use degrees and every one of those relations needs a factor of pi over 180. Use revolutions and they need 2 pi. Every rotational formula in this course assumes radians, and a calculator left in degree mode is the single most common source of wrong answers in this unit.
A radian is not an arbitrary unit like a degree. It is defined by the geometry itself: the angle at which the arc length equals the radius. That makes an angle in radians a ratio of two lengths, and a ratio of lengths has no units.
That is exactly why the relation s = r theta has no conversion factor in it. If theta were in degrees you would need s = r theta times pi over 180, and every formula derived from it — v = r omega, a = r alpha — would carry the same awkward factor.
So the rule for the whole unit is simple: angles in radians, angular velocity in radians per second, angular acceleration in radians per second squared. Convert anything given in revolutions or degrees at the very start of a problem, before it touches an equation.
And check your calculator. Sin and cos of an angle are fine in either mode as long as the mode matches the input, but the moment a radian value is fed to a calculator in degree mode, everything downstream is wrong without any visible warning.
Notation
Annotate
On: \( \omega = \frac{d\theta}{dt} \qquad \alpha = \frac{d\omega}{dt} = \frac{d^2\theta}{dt^2} \)
Compare with x, v = dx/dt and a = dv/dt. It is the same structure exactly, with angle in place of position. Everything you learned in the kinematics deck — derivatives, integrals, areas under graphs — carries over symbol for symbol.
So for constant alpha, the constant-acceleration equations reappear with the letters swapped: omega = omega0 + alpha t, theta = omega0 t + half alpha t squared, and omega squared = omega0 squared + 2 alpha theta.
Look at the two definitions side by side with their linear cousins. Velocity is the rate of change of position; angular velocity is the rate of change of angle. Acceleration is the rate of change of velocity; angular acceleration is the rate of change of angular velocity. The structure is identical.
That means every technique from the kinematics deck transfers. The slope of a theta-versus-time graph is omega. The area under an omega-versus-time graph is the change in angle. Differentiate to go down the chain, integrate to come back up, and remember the constant of integration is the initial value.
The signs work the same way too. Pick a positive sense of rotation — usually counter-clockwise — and give every angular quantity a sign relative to it. When omega and alpha have the same sign the body is speeding up; opposite signs, it is slowing down.
The practical consequence is the constant-acceleration set: omega = omega0 + alpha t, theta = omega0 t + half alpha t squared, omega squared = omega0 squared + 2 alpha theta. They are the old equations with new letters, and they come with the same restriction: alpha must be constant.
Concept
Figure (svg): A spinning disc with three marked points at increasing distance from the centre; each carries a velocity arrow whose length grows in proportion to its radius.
\[ v = r\omega \qquad a_t = r\alpha \]
A rigid body turns as one piece, so every point sweeps the same angle in the same time. That makes theta, omega and alpha properties of the whole body, not of any one point.
Linear quantities are different. A point twice as far from the axis travels twice the arc in the same time, so it moves twice as fast. The bridge between the two descriptions is always a factor of r — and only in radians.
Watch the animation and focus on the three marked points. They sweep around together, always lined up along the same radius. That is what rigid means: the shape does not change, so every point turns through the same angle in the same time.
Because of that, theta, omega and alpha describe the whole body at once. You do not ask for the angular velocity of a particular point on a wheel; there is only one, and it belongs to the wheel.
Linear speed is a different story. The outer point travels a much larger circle than the inner one in the same time, so it must move faster. Twice the radius means twice the circumference, means twice the speed — that is v = r omega.
The velocity arrows in the figure show the consequence: each is tangent to its circle and its length grows in proportion to the distance from the axis. The same factor of r relates tangential acceleration to angular acceleration, and both relations rely on the angle being in radians.
Concept
Figure (svg): A point on a circle with two acceleration arrows: one along the tangent labelled a_t, one pointing to the centre labelled a_c.
\[ a_t = r\alpha \qquad a_c = \frac{v^2}{r} = r\omega^2 \qquad a = \sqrt{a_t^2 + a_c^2} \]
The two components are perpendicular, so they add by Pythagoras. And they come from different things: the tangential part exists only if omega is changing, the centripetal part exists whenever the body is turning at all.
A wheel spinning at constant omega still has every point accelerating — toward the centre. Constant angular velocity is not zero acceleration, for the same reason uniform circular motion is not, as you saw in the forces deck.
A point on a turning body can have two quite different accelerations, and keeping them separate is essential.
The tangential part, r alpha, points along the circle. It changes how fast the point is moving. If the wheel is not speeding up or slowing down, alpha is zero and this part vanishes.
The centripetal part, r omega squared, points toward the axis. It changes the direction of the point's motion, and it exists whenever the body is turning at all — even at perfectly constant omega. This is the same centripetal acceleration you met for circular motion in the forces deck; a turning rigid body is just a collection of points all doing uniform or non-uniform circular motion together.
Since the two parts are perpendicular, the total is found by Pythagoras. In practice one often dominates enormously, as the flywheel example will show, and recognising which one matters is part of reading the problem.
Comparison
Comparison matrix
| quantity | linear | angular |
|---|---|---|
| position | x (m) | theta (rad) |
| velocity | v = dx/dt | omega = d theta/dt |
| acceleration | a = dv/dt | alpha = d omega/dt |
| constant-acceleration velocity | v = v0 + at | omega = omega0 + alpha t |
| constant-acceleration position | x = v0 t + (1/2) a t^2 | theta = omega0 t + (1/2) alpha t^2 |
| time-free equation | v^2 = v0^2 + 2a dx | omega^2 = omega0^2 + 2 alpha d theta |
Fill in the blanks from the left column. There is nothing new to memorise here — the angular equations are the linear ones with three letters changed, and they hold under exactly the same condition: constant acceleration.
When alpha is not constant, the table's bottom three rows are off-limits, and you go back to derivatives and integrals — again exactly as in one-dimensional kinematics.
This table is meant to be filled from the left column, not from memory. Every entry on the right is the entry on the left with x replaced by theta, v by omega and a by alpha.
Doing it yourself once is worth more than reading it ten times, because it convinces you there is genuinely nothing new to learn. The angular kinematic equations are not a separate set of facts; they are the same mathematics applied to a different coordinate.
The restriction travels with them. The three constant-acceleration equations need constant alpha, exactly as the linear ones needed constant a. A motor that ramps up its torque, or a torque that depends on angle, gives a changing alpha, and then you must work with derivatives and integrals directly.
When you meet a rotational kinematics problem, the first question is therefore the same one you asked in the first deck: is the acceleration constant? That single question decides which tools are legal.
Worked example
A flywheel turning at 30 rad/s slows uniformly to rest in 12 s. Find its angular acceleration, the number of revolutions it makes while stopping, and the acceleration of a point 0.40 m from the axis at t = 4.0 s.
Find alpha from the definition.
Why: Uniform slowing means constant alpha, so the change in omega over the time is all you need.
\[ \alpha = \frac{0 - 30}{12} = -2.5\,\text{rad/s}^2 \]
Find the angle turned, using the average angular velocity.
Why: With constant alpha the average omega is the mean of the start and end values, 15 rad/s.
\[ \Delta\theta = \tfrac{1}{2}(30 + 0)(12) = 180\,\text{rad} = \frac{180}{2\pi} = 28.6\,\text{rev} \]
Find omega at t = 4.0 s, then the speed of the point.
Why: Constant-alpha velocity equation, then v = r omega.
\[ \omega = 30 - 2.5(4.0) = 20\,\text{rad/s} \qquad v = (0.40)(20) = 8.0\,\text{m/s} \]
Find both acceleration components at that instant.
Why: The tangential part comes from alpha, the centripetal part from omega.
\[ a_t = r\alpha = (0.40)(-2.5) = -1.0\,\text{m/s}^2 \qquad a_c = r\omega^2 = (0.40)(20)^2 = 160\,\text{m/s}^2 \]
Combine and interpret.
Why: The magnitude is the square root of 160 squared plus 1 squared, about 160 m/s squared. The centripetal part utterly dominates — the gentle slowing is almost invisible next to the cost of going round in a circle.
The first three steps are pure translation of a linear problem: find the acceleration from a change in velocity over a time, then find the displacement from the average velocity. The average-velocity shortcut is legal only because alpha is constant.
Converting 180 radians into 28.6 revolutions is the kind of step that costs marks when skipped. Questions frequently ask for revolutions, and dividing by 2 pi is the conversion.
The last part is where the physics gets interesting. At t = 4.0 s the point 0.40 m out has a tangential acceleration of only 1.0 metre per second squared — but a centripetal acceleration of 160. The centripetal part is 160 times bigger.
That is typical of anything spinning fast. The gentle slowing that you might think of as 'the acceleration' is almost invisible next to the enormous inward acceleration needed to keep each bit of metal going round. It is also why flywheels and turbine discs can burst: the material must supply that inward force, and it grows as omega squared.
Worked example
A wheel's angle is theta(t) = 2t^3 - 6t radians. Find when it momentarily stops, its angular acceleration at that instant, and the total angle it sweeps between t = 0 and t = 2.0 s.
Differentiate to get omega and alpha.
Why: Alpha is not constant here, so none of the constant-acceleration equations apply.
\[ \omega = \frac{d\theta}{dt} = 6t^2 - 6 \qquad \alpha = \frac{d\omega}{dt} = 12t \]
Find when it stops.
Why: Set omega to zero. Only the positive root is physical.
\[ 6t^2 - 6 = 0 \;\Rightarrow\; t = 1.0\,\text{s} \qquad \alpha(1.0) = 12\,\text{rad/s}^2 \]
Evaluate theta at the start, the turning point and the end.
Why: The wheel reverses at t = 1.0 s, so the sweep must be split there.
\[ \theta(0) = 0 \qquad \theta(1) = -4\,\text{rad} \qquad \theta(2) = 16 - 12 = 4\,\text{rad} \]
Add the two legs as magnitudes.
Why: It turns 4 rad clockwise, then 8 rad counter-clockwise.
\[ \text{total swept} = 4 + 8 = 12\,\text{rad} \qquad \text{net angular displacement} = +4\,\text{rad} \]
This is the distance-versus-displacement trap from the kinematics deck, in rotational dress. Whenever omega changes sign, split the interval at the zero.
Here alpha is 12t, which is not constant, so the constant-acceleration equations are off the table from the first line. Everything comes from differentiating theta.
Finding when the wheel stops means setting omega to zero, which is the rotational version of finding a turning point. The negative root, t = -1, is before the motion started and is discarded.
The total-angle question is the classic distance-versus-displacement trap. The wheel first turns 4 radians clockwise, reaching -4 at t = 1, then turns back 8 radians to reach +4 at t = 2. Net displacement is +4 radians, but the wheel actually swept 12.
The method is the one from the kinematics deck: find every zero of the velocity inside the interval, evaluate the position at each one, and add the magnitudes of the pieces. If a question asks how many revolutions a wheel makes and the wheel reverses, it wants the total swept, not the net.
Prediction
Predict first
A merry-go-round turns at a steady rate. Rider A sits 1.5 m from the centre; rider B sits at the rim, 3.0 m out. How does B's centripetal acceleration compare with A's?
Correct: Twice as large
Why: With omega fixed, a_c = r omega squared, which is proportional to r. Doubling the radius doubles it. The v squared over r form gives the same answer once you notice v also doubles: four times as much on top, twice as much underneath, net factor two.
Figure (svg): A merry-go-round seen from above with an inner rider at half the radius and an outer rider at the rim, each with a velocity arrow.
Both tempting wrong answers come from holding the wrong thing fixed. The v-squared intuition forgets that r grew too; the 1/r intuition forgets that v grew. Pick the form whose other variable is genuinely constant — here that is r omega squared, because omega is shared.
This prediction is designed to catch the most common error with formulas that have two forms. Centripetal acceleration can be written v squared over r or r omega squared, and both are correct. The trouble is choosing which variable to hold fixed.
On a merry-go-round both riders share the same omega, because the platform is rigid. So the natural form is r omega squared, and it says the acceleration is proportional to r. Double the radius, double the acceleration.
If you reach for v squared over r, you must remember that v is not the same for the two riders. B moves twice as fast as A, so the numerator grows by four while the denominator grows by two, and the answer is again a factor of two.
The general skill is to identify what the situation keeps constant before choosing a formula. On a rigid rotating object, omega is the shared quantity. For a car taking two different curves at the same speedometer reading, v is shared. The physics decides which form is cleaner.
Trap
The rim of a spinning wheel is a blur; the hub looks almost still. So the outside of the wheel is obviously turning faster than the inside.
The rim is moving faster. It is not turning faster. Every point on a rigid body has exactly the same angular velocity.
Turning means sweeping angle. In one second the hub and the rim sweep the same angle — if they did not, the wheel would twist and tear itself apart. What differs is the arc length each point covers, and that is r times the angle.
| point on a wheel at 10 rad/s | radius | angular velocity | speed |
|---|---|---|---|
| near the hub | 0.05 m | 10 rad/s | 0.5 m/s |
| halfway out | 0.20 m | 10 rad/s | 2.0 m/s |
| the rim | 0.40 m | 10 rad/s | 4.0 m/s |
Keep the vocabulary straight and the trap disappears: angular quantities belong to the body, linear quantities belong to each point, and r converts between them.
The visual impression is genuine — the rim really is moving faster — but the word 'turning' is wrong. Turning is about angle, and every part of a rigid wheel sweeps the same angle in the same time.
If the rim really did turn faster than the hub, the spokes would have to bend, wrap round and eventually snap. A rigid body cannot do that; that is what rigid means.
The table makes the separation concrete. The angular velocity column is the same in every row, ten radians per second. The speed column grows in proportion to the radius. Same omega, different v.
Getting the vocabulary right prevents a whole class of later errors. When a problem says 'angular speed', it means a property of the body. When it says 'speed' of a point, it means r times that. Mixing them up is how students end up with different omegas for different parts of the same wheel.
Section
Section 2
Intuition
Figure (svg): A door seen from above, hinged at the left. A push near the free edge turns it easily, the same push near the hinge barely does, and a push along the door toward the hinge does nothing.
Push a door at the handle and it swings easily. Push the same door with the same force a hand's width from the hinge and you can barely move it. Push straight along the door toward the hinge and nothing happens at all.
So the turning effect of a force depends on three things: how big the force is, how far from the axis it acts, and its direction relative to the line from the axis. Torque is the single quantity that packages all three.
Every door you have ever opened has taught you the content of this slide. Push near the handle and the door swings; push near the hinge and it resists; push along the door and nothing at all happens.
The point is that force alone does not decide whether something turns. The same push produces very different results depending on where you apply it and which way it points.
Torque is the quantity that combines all three ingredients — size, distance from the axis, and direction — into one number that does decide. A large torque turns things easily; a zero torque does not turn them at all, however large the force.
Notice also what makes the door hard to open near the hinge. It is not that the force is weaker; you can push just as hard. It is that the force acts too close to the axis to have much turning effect. Distance from the axis is a multiplier on effectiveness.
Concept
Figure (svg): A pivot with a position vector r to a point where a force F acts straight up. The force's line of action is dashed; the perpendicular distance from the pivot to it is the lever arm.
\[ \tau = rF\sin\theta = r_\perp F = rF_\perp \]
All three are the same product, grouped differently. r F sin theta uses the angle between r and F directly. r-perp F shortens r to the lever arm. r F-perp keeps r and uses only the part of the force that is perpendicular to it.
Pick whichever the geometry hands you. If the angle is given, use the first. If the force's line of action is easy to draw, the lever arm is usually fastest. The units are newton metres — the same as a joule, but torque is never written in joules, because it is not an energy.
The figure shows the same torque three ways, and fluency means being able to see all three in any diagram.
The first form, r F sin theta, uses the angle between the position vector and the force. It is the most mechanical and works whenever the angle is given.
The second form uses the lever arm: extend the force into its line of action, then measure the perpendicular distance from the axis to that line. Multiply by the full force. This is often the fastest method when forces are horizontal or vertical, because the lever arm is then just a horizontal or vertical distance you can read off the diagram — the ladder and strut problems later in the deck use it constantly.
The third form keeps r and uses only the component of the force perpendicular to it. It is the natural choice when a force is resolved into components anyway.
All three are the same product grouped differently, because r sin theta is the lever arm and F sin theta is the perpendicular component. Pick whichever the geometry makes easy, and check with a second form when the numbers matter.
Notation
Annotate
On: \( \vec{\tau} = \vec{r} \times \vec{F} \qquad |\vec{\tau}| = rF\sin\theta \)
The direction comes from the right-hand rule: curl the fingers of your right hand from r toward F, and your thumb points along the torque. Out of the page is counter-clockwise; into the page is clockwise.
For the AP exam almost every torque problem is planar, so the vector reduces to a sign. But the cross product is why that sign rule exists, and it becomes essential for angular momentum in the next deck.
The cross product is the complete definition of torque, and the annotations pull out the three things to watch.
First, r runs from the axis to the point where the force acts. That makes torque depend on the axis you choose. Change the axis and r changes, so the torque changes. There is no such thing as 'the torque of a force' until you name the axis.
Second, the order matters. r cross F and F cross r point in opposite directions, so writing them the wrong way round flips every sign in your problem.
Third, the sin theta factor is why a force along r does nothing. Pull a door handle straight away from the hinge and theta is zero.
For flat problems you can mostly forget the vector and keep only its sign, using the right-hand rule: out of the page is counter-clockwise and positive. But hold on to the vector picture. In the next deck, angular momentum is also a cross product, and its direction is essential for understanding gyroscopes and precession.
Concept
Figure (svg): Two wheels: one with a counter-clockwise curved arrow labelled positive, pointing out of the page; one with a clockwise arrow labelled negative, into the page.
The convention is a choice, just like choosing which way is positive x. Counter-clockwise positive is standard because it matches the right-hand rule with the z-axis out of the page — but any consistent choice works, and choosing the direction you expect the body to turn often saves a minus sign.
What is not optional is consistency. Once you pick a sense, every torque, omega and alpha in the problem uses it. Mixed sign conventions within one equation are the commonest error in multi-torque problems.
Choosing counter-clockwise as positive is a convention, not a law. It matches the right-hand rule with the z-axis pointing out of the page, which is why textbooks use it, but any choice works if you apply it consistently.
In fact a smarter choice is often to take the direction you expect the body to turn as positive. Then the angular acceleration comes out positive if you guessed right, and negative if you guessed wrong — either way you learn something, and the algebra is cleaner.
What does not work is changing convention halfway. Each torque in the sum must be measured with the same sense, and omega and alpha must use the same sense as the torques. Most sign errors in rotation problems come from quietly switching conventions between two lines.
The dot and cross symbols at the centres of the wheels are standard notation you will meet again in magnetism: a dot is an arrow coming toward you, a cross is an arrow going away, as if you were seeing its tail feathers.
Worked example
A stepped pulley has an inner radius of 0.20 m and an outer radius of 0.50 m. A 50 N force acts tangentially on the inner rim, turning it counter-clockwise; a 30 N force acts tangentially on the outer rim, turning it clockwise; and a 40 N force acts at the outer rim at 30 degrees to the radius, turning it counter-clockwise. Find the net torque.
Figure (svg): A stepped pulley with three forces: one tangential on the inner rim, one tangential on the outer rim in the opposite sense, and one at the outer rim at thirty degrees to the radius.
Take counter-clockwise as positive and compute each torque.
Why: The two tangential forces are perpendicular to their radii, so sin theta is one. The third is at 30 degrees, so only half its magnitude counts.
\[ \tau_1 = +(0.20)(50) = +10 \qquad \tau_2 = -(0.50)(30) = -15 \qquad \tau_3 = +(0.50)(40)\sin 30^\circ = +10 \]
Add with signs.
Why: Torques about the same axis add like one-dimensional forces.
\[ \tau_{\text{net}} = 10 - 15 + 10 = +5.0\,\text{N m} \quad (\text{counter-clockwise}) \]
Check the result against intuition.
Why: The 30 N force alone beats the 50 N force, because it acts at two and a half times the radius. Torque is not force, and the biggest push does not necessarily win.
A stepped pulley is a compact way of putting several lever arms on one axis, and it makes the main lesson unmissable: torque is not force.
The 30 N force wins against the 50 N force, because it acts at 0.50 m rather than 0.20 m. Two and a half times the lever arm more than compensates for three fifths of the force.
The 40 N force is the instructive one. It acts at the rim, but at 30 degrees to the radius, so only the perpendicular component — half of 40 N — contributes. Its torque is the same 10 N m as the 50 N force on the inner rim.
The method generalises to any number of forces: compute each torque about the same axis, attach its sign, and add as numbers. Do not try to combine forces first and then find one torque; forces at different points with different lever arms cannot be added before multiplying by their lever arms.
Discrimination
Each force acts on a wheel that turns about a fixed axle through its centre. Decide whether its torque about that axle is zero or not.
Sort into buckets
The quick test: extend the force into a line. If that line passes through the axis, the torque is zero. This is why the axle force and the weight of a balanced wheel never appear in the torque equation, which is a gift — they are usually unknown.
Every zero-torque case falls into one of two families. Either the force acts at the axis, so r is zero, or its line of action passes through the axis, so the lever arm is zero.
The axle force and the weight of a uniform wheel belong to the first family: both act at the centre. The pushes directly toward or away from the axle belong to the second: they act at the rim, but along the radius.
The practical payoff is enormous. In nearly every rotation problem with a fixed axle, the axle force is unknown — and it drops out of the torque equation automatically. Likewise the weight of a balanced wheel. You can usually write the torque equation knowing only the forces at the rim.
The non-zero items are worth comparing too. The perpendicular force halfway out gives exactly half the torque it would at the rim, and the 45-degree force at the rim gives about 71 percent. Distance and angle each scale the turning effect, independently.
Trap
Torque is measured in newton metres, so it must be the force multiplied by the distance from the axis. Bigger force, farther out, more torque.
Torque is force times the PERPENDICULAR distance from the axis to the line of action — not the distance to the point of application.
Pull on a wrench along its handle and the force is 0.30 m from the bolt, yet the bolt does not turn. The lever arm is zero because the line of action passes through the bolt.
| 100 N on a 0.30 m wrench, at angle to the handle | r F (wrong) | r F sin theta (right) |
|---|---|---|
| 90 degrees | 30 N m | 30 N m |
| 60 degrees | 30 N m | 26 N m |
| 30 degrees | 30 N m | 15 N m |
| 0 degrees, along the handle | 30 N m | 0 |
The second half of the trap is the unit. A newton metre of torque is not a joule. Work is force dotted with displacement — the parallel parts multiply. Torque is r crossed with force — the perpendicular parts multiply. The units coincide; the physics does not.
This is half right, and the missing half is the word perpendicular.
The table makes the point with a wrench. The distance from the bolt to your hand never changes, so the naive formula gives 30 newton metres every time. The real torque falls from 30 to zero as the push swings round to point along the handle.
The cleanest way to avoid the trap is to train yourself to draw the line of action. Extend the force arrow into a long dashed line, then measure from the axis to that line at right angles. That distance, not the distance to your hand, is what multiplies the force.
The second half of the trap is about units. A newton metre of torque and a joule of work have the same dimensions, but they are different physical quantities. Work multiplies force by the displacement along it; torque multiplies force by the distance across it. Writing a torque in joules will be marked wrong, and it signals a misunderstanding of what torque is.
Tweak it
Parameter explorer
A 100 N force acts at the end of a 0.30 m wrench. Sweep the angle between the force and the handle from 0 to 180 degrees. Where is the torque largest, where does it vanish, and is the curve symmetric?
\[ \tau = (0.30)(100)\sin({angle}^\circ)\,\text{N m} \]
Sweeping the angle shows the sine curve directly. The torque is greatest when the push is perpendicular to the handle, falls to zero when the push points along it, and is symmetric about 90 degrees.
The symmetry has a geometric reason worth seeing. A push tilted 30 degrees toward the bolt and one tilted 30 degrees away have the same perpendicular component, so they turn the bolt equally. Only the wasted parallel component differs, and it does nothing either way.
The flatness near the top is the practical lesson. Between about 70 and 110 degrees the torque stays above 94 percent of its maximum. You do not need to be exactly perpendicular — just roughly so.
The steepness near zero is the other lesson. A push only 10 degrees from the handle direction gives just 17 percent of the possible torque. That is why a wrench slipping into line with your arm suddenly feels useless.
Section
Section 3
Concept
\[ I = \sum_i m_i r_i^2 \qquad \text{or} \qquad I = \int r^2\,dm \]
Figure (svg): Three point masses on a light rod with two candidate axes: one at the left end and one through the middle mass.
Moment of inertia is rotational inertia — how hard it is to change a body's spin. It depends on the mass, and far more strongly on where that mass sits, because each bit counts in proportion to the square of its distance from the axis.
Two consequences to fix in place now. First, I depends on the axis, so 'the moment of inertia of a rod' is meaningless until you say about which line. Second, mass on the axis contributes nothing at all.
In linear motion, mass measures how hard it is to change an object's velocity. In rotation, the corresponding resistance depends not only on how much mass there is but on where it is relative to the axis.
The formula says each bit of mass counts in proportion to the square of its distance from the axis. Twice as far out counts four times as much. That squared weighting is the defining feature, and it explains nearly every surprising result in this section.
The figure shows three masses on a rod with two possible axes. Before working out the numbers, notice that the answer must depend on the axis: about axis A the 3 kg mass is a full metre away, while about axis B it is only half a metre away. Different distances, different moments of inertia.
So always state the axis. The moment of inertia is a property of a body and an axis together, never of a body alone. The same object can have very different values, and much of the skill of this unit is choosing and naming the right one.
Worked example
Masses of 2 kg, 1 kg and 3 kg sit at 0, 0.50 m and 1.00 m along a light rod. Find I about axis A through the left end and about axis B through the middle mass, both perpendicular to the rod.
About axis A, measure each distance from the left end.
Why: The 2 kg mass sits on the axis and contributes nothing.
\[ I_A = 2(0)^2 + 1(0.50)^2 + 3(1.00)^2 = 0 + 0.25 + 3.00 = 3.25\,\text{kg m}^2 \]
About axis B, measure from the middle.
Why: Now the 1 kg mass is on the axis, and the outer two are each 0.50 m away.
\[ I_B = 2(0.50)^2 + 1(0)^2 + 3(0.50)^2 = 0.50 + 0 + 0.75 = 1.25\,\text{kg m}^2 \]
Compare.
Why: Same three masses, same rod. The moment of inertia differs by a factor of 2.6 purely because of the axis.
Notice which mass dominated I_A: the 3 kg mass at 1.00 m contributes 3.00 of the 3.25. Distance squared means the farthest mass usually controls the answer, so that is the term to compute first and check hardest.
For a handful of point masses, the sum is just arithmetic — but the arithmetic shows the logic clearly.
About axis A, the 2 kg mass sits on the axis and contributes zero. Mass on the axis has no rotational inertia at all, however heavy it is.
About axis B, the outer two masses are each half a metre away and the middle one contributes nothing. The total drops from 3.25 to 1.25, a factor of 2.6, for the same physical object.
Look at where the 3.25 came from. The 3 kg mass at one metre gave 3.00 of it. The farthest mass dominates because distance is squared, and that is a useful checking habit: compute the outermost contribution first, and if your total is not at least that big, something is wrong.
Prediction
Predict first
A uniform rod is spun about an axis through its centre, then about an axis through one end, both perpendicular to the rod. How do the moments of inertia compare?
Correct: The end axis gives four times as much
Why: About the centre, I = ML^2/12; about the end, I = ML^2/3, which is four times larger. Moving the axis to the end puts half the rod twice as far away as before, and distance counts squared. The integration on the next two slides derives both values.
Your hands know this already. Swinging a broom held at its end is far harder than twirling it held in the middle, and the factor is a full four, not a modest correction.
This is a prediction that your body has already made for you, if you have ever swung a broom or a bat. Held at the end, it is sluggish; held in the middle, it twirls easily.
The factor is four. About the centre, every bit of the rod is within L/2 of the axis. About the end, half the rod is more than L/2 away, out to a full L. The squared weighting turns that extra distance into four times the inertia.
It is worth noticing which tempting answer is most wrong. 'The same' assumes inertia depends only on mass — true for translation, false for rotation. 'Twice' assumes inertia grows linearly with distance, forgetting the square.
The next two slides derive both values from integrals, and then the parallel-axis theorem shows that the factor of four is no coincidence: it is the centre value plus M times (L/2) squared.
Worked example
Find the moment of inertia of a uniform rod of mass M and length L about a perpendicular axis through its centre, and then through one end.
Figure (svg): A rod along the x-axis centred on the rotation axis, with a thin slice of width dx a distance x from the axis highlighted.
Write dm in terms of a length element.
Why: A uniform rod has a constant linear density lambda = M/L, so a slice of width dx has mass lambda dx.
\[ dm = \lambda\,dx = \frac{M}{L}\,dx \]
Integrate x squared dm across the rod, with the axis at x = 0.
Why: The rod runs from -L/2 to +L/2.
\[ I_{\text{cm}} = \int_{-L/2}^{L/2} x^2 \frac{M}{L}\,dx = \frac{M}{L}\left[\frac{x^3}{3}\right]_{-L/2}^{L/2} = \frac{M}{L}\cdot\frac{L^3}{12} = \frac{1}{12}ML^2 \]
Repeat with the axis at one end.
Why: Only the limits change: the rod now runs from 0 to L.
\[ I_{\text{end}} = \int_0^L x^2 \frac{M}{L}\,dx = \frac{M}{L}\cdot\frac{L^3}{3} = \frac{1}{3}ML^2 \]
The recipe never changes: express dm in terms of a coordinate, write r for that element, integrate r squared dm over the body. The only skill is choosing the slice so every bit of it is at the same distance from the axis.
Continuous bodies turn the sum into an integral, and the integral is less frightening than it looks. The only real decision is how to slice the body.
For a rod, the natural slice is a short segment of length dx at position x. Every bit of that segment is the same distance x from the axis, so its contribution to the moment of inertia is simply x squared times its mass.
The mass of the slice comes from the linear density. A uniform rod has mass M spread over length L, so lambda = M/L, and the slice has mass lambda dx. That converts an integral over mass into an integral over position, which you know how to do.
Changing the axis changes only the limits of integration. With the axis at the centre, the rod runs from -L/2 to L/2 and you get one twelfth M L squared. With the axis at the end, it runs from 0 to L and you get one third. Same integrand, different limits.
Worked example
Find the moment of inertia of a uniform disc of mass M and radius R about the axis through its centre, perpendicular to its face.
Figure (svg): A disc divided into concentric rings; one thin ring of radius r and width dr is highlighted.
Choose the slice.
Why: A thin ring of radius r and width dr has all its mass at distance r, so its contribution is simply r squared dm.
\[ dm = \sigma(2\pi r\,dr) \qquad \sigma = \frac{M}{\pi R^2} \]
Integrate from the centre to the rim.
Why: Pull the constants out first.
\[ I = \int_0^R r^2 \sigma 2\pi r\,dr = 2\pi\sigma\int_0^R r^3\,dr = 2\pi\sigma\frac{R^4}{4} = 2\pi\frac{M}{\pi R^2}\frac{R^4}{4} = \frac{1}{2}MR^2 \]
Sanity-check against a hoop.
Why: A hoop has all its mass at R, so I = MR^2. A disc has most of its mass inside R, so it must come out smaller — and half is plausible.
The same ring slicing gives a solid cylinder of any length the same result, half M R squared, because stretching the disc along the axis moves no mass farther from the axis.
A disc needs a cleverer slice. A strip across the disc would have points at many different distances from the centre, which is awkward. A thin ring, by contrast, has every point at exactly the same distance r.
The mass of a ring is the area density times its area. The area of a thin ring is its circumference times its width, 2 pi r dr. So dm grows with r: outer rings hold more mass than inner ones simply because they are longer.
With dm in hand, the integral is easy, and the answer is half M R squared. The sanity check against a hoop is worth doing every time: a hoop has all its mass at R, so it gives M R squared; a disc has most of its mass inside R, so it must give less.
This slicing idea — choose slices on which r is constant — is the whole art of moment-of-inertia integrals. Rods use segments, discs use rings, and solid spheres use thin spherical shells or stacked discs. Once the slice is right, the calculus is routine.
Concept
Figure (svg): Five small shapes with their moments of inertia: a hoop, a disc, a solid sphere, a rod about its centre and a rod about its end.
| object | axis | I |
|---|---|---|
| thin hoop or thin cylindrical shell | central axis | MR^2 |
| solid disc or solid cylinder | central axis | (1/2) MR^2 |
| thin spherical shell | any diameter | (2/3) MR^2 |
| solid sphere | any diameter | (2/5) MR^2 |
| thin rod | perpendicular, through centre | (1/12) ML^2 |
| thin rod | perpendicular, through end | (1/3) ML^2 |
Read the coefficients as a ranking of how far out the mass is. A hoop puts it all at R: coefficient 1. A spherical shell is all at R but much of it close to the axis: two thirds. A solid sphere packs mass near the centre: two fifths. The AP equation sheet does not list all of these, so derive or remember them.
These six results cover almost every AP problem, and the coefficient in front tells a story about each shape.
A hoop and a thin cylindrical shell put all their mass at radius R, so the coefficient is 1 — the largest possible for a given M and R. A solid disc or cylinder spreads mass from the centre outward, giving one half.
Spheres need care because 'distance from the axis' is not the same as 'distance from the centre'. A thin spherical shell has all its mass at distance R from the centre, but much of it lies near the poles, close to the rotation axis. That brings the coefficient down to two thirds. A solid sphere concentrates mass even nearer the centre: two fifths.
The two rod values differ by a factor of four, as you predicted earlier. You do not need to memorise all six if you can derive them, but the hoop, disc and rod values come up so often that knowing them on sight saves real time in an exam.
Matching
Match the pairs
Why: The rim is a hoop, with all its mass at R. A flywheel is a solid disc. A bowling ball is very nearly a solid sphere. A metre stick is a thin rod, and the axis location decides between one twelfth (centre) and one third (end). The two metre-stick answers differ by exactly the parallel-axis term M(L/2)^2 = ML^2/4.
The real-object versions are approximations — a bicycle rim has spokes and a hub, a bowling ball has finger holes — but the idealised shape usually carries 90 percent or more of the answer. Choosing the right idealisation is a modelling skill the exam rewards.
Matching real objects to idealised shapes is a modelling skill, and the exam rewards it.
A bicycle rim is nearly a hoop: the tyre and rim are at the outside, and the spokes are light. A flywheel is deliberately made solid and uniform, so it is a disc. A bowling ball is very close to a uniform solid sphere, finger holes notwithstanding.
The two metre-stick answers are the instructive pair. Same object, two axes, coefficients of one twelfth and one third. The difference between them, one quarter M L squared, is exactly M times (L/2) squared — the parallel-axis term for shifting from the centre to the end.
When you meet an unfamiliar object, ask what simple shape it most resembles and where most of its mass sits. That usually gives you the right coefficient to within ten percent, which is enough to rank objects and check answers.
Concept
\[ I = I_{\text{cm}} + Md^2 \]
Figure (svg): A disc with its centre-of-mass axis marked and a second, parallel axis at the rim a distance d away.
If you know I about an axis through the centre of mass, you know it about every parallel axis: add the total mass times the square of the distance between them.
Two consequences. The centre-of-mass axis gives the smallest I of any axis in that direction, since M d squared is never negative. And the theorem only works from the centre-of-mass axis — a restriction important enough to get its own trap slide.
The parallel-axis theorem saves you from integrating again every time the axis moves. If you know the moment of inertia about an axis through the centre of mass, you know it about every parallel axis.
The new value is always the old one plus M d squared, where d is the distance between the axes. Physically, you can think of it as two contributions: the body spinning about its own centre, plus the whole mass M swinging round the new axis at distance d as if it were a point.
Since M d squared is never negative, the centre-of-mass axis gives the smallest moment of inertia of any axis in a given direction. That is a useful sanity check: any off-centre value must be bigger.
Two conditions apply. The axes must be parallel, and one of them must pass through the centre of mass. The second condition is the one people forget, and the trap slide two pages on shows how badly it can go wrong.
Worked example
Use the parallel-axis theorem to find I for a rod about one end, and for a disc of radius R about an axis through a point on its rim, perpendicular to its face.
Rod: start from the centre-of-mass result and shift by L/2.
Why: The end is half a length from the centre.
\[ I_{\text{end}} = \frac{1}{12}ML^2 + M\left(\frac{L}{2}\right)^2 = \frac{1}{12}ML^2 + \frac{3}{12}ML^2 = \frac{1}{3}ML^2 \]
Check against the integral.
Why: This matches the direct integration two slides back exactly, which is a good test of both methods.
Disc: shift by R from the centre to the rim.
Why: This is the axis about which a disc pivots when it hangs from a nail through its edge.
\[ I_{\text{rim}} = \frac{1}{2}MR^2 + MR^2 = \frac{3}{2}MR^2 \]
Interpret the size.
Why: Three times the central value. Pivoting at the edge puts the whole disc, on average, much farther from the axis.
The theorem turns two integrals into one. Integrate once about the centre of mass, then shift as often as you like. That is the efficient workflow whenever a problem uses an off-centre axis.
The rod calculation is a check that the theorem and the integral agree: one twelfth plus one quarter is one third, exactly the result you integrated directly. Two independent methods giving the same answer is the best evidence you have done both correctly.
The disc pivoted at its rim is a common exam setup — a disc hanging from a nail through its edge and swinging like a pendulum. Its moment of inertia about the nail is three halves M R squared, three times its central value.
That factor of three is large, and it makes sense. Pivoting at the edge puts the far side of the disc a full diameter away from the axis, and the squared weighting makes that distant mass count heavily.
The efficient workflow for any off-centre axis is therefore: find or look up the centre-of-mass value once, then shift. You will use this again in the oscillations deck, where a physical pendulum's period depends on exactly this kind of off-centre moment of inertia.
Trap
The theorem adds M d squared when you shift an axis. So to go from a rod's end axis to an axis a quarter of the way along, add M times (L/4) squared to the end value.
\[ I = \frac{1}{3}ML^2 + M\left(\frac{L}{4}\right)^2 = \frac{19}{48}ML^2 \quad \text{(wrong)} \]
One of the two axes must pass through the centre of mass. The theorem is I = I_cm + M d squared, and d is measured from the centre of mass.
The axis a quarter of the way along is L/4 from the centre. So start from the centre value and shift by L/4:
\[ I = \frac{1}{12}ML^2 + M\left(\frac{L}{4}\right)^2 = \frac{4}{48}ML^2 + \frac{3}{48}ML^2 = \frac{7}{48}ML^2 \]
The wrong answer is almost three times too large. It even fails a basic sanity check: an axis a quarter of the way along is closer to the centre than the end axis is, so its I must be smaller than one third — and 19/48 is bigger.
If you only know an off-centre value, go back to the centre first by subtracting M d squared, then forward to the new axis by adding the new M d squared.
The error on the left looks entirely reasonable: you know the end value, you want an axis a quarter of the way along, so you add M times (L/4) squared. The algebra is fine. The physics is wrong.
The theorem is not a general rule for moving axes. It says I = I_cm + M d squared, and d must be measured from the centre of mass. The end of the rod is not the centre of mass, so the end value cannot be the starting point.
The sanity check on the slide catches the error without any algebra. An axis a quarter of the way along is closer to the centre than the end is, so its moment of inertia must lie between one twelfth and one third. The wrong answer, nineteen forty-eighths, exceeds one third. Impossible.
When the only value you know is off-centre, go back to the centre first by subtracting M d squared with the old distance, then forward by adding M d squared with the new distance. Always route through the centre of mass.
Error analysis
A student finds the moment of inertia of a uniform disc about its centre by treating the radius as the integration variable.
Annotate
On: \( I = \int_0^R r^2 \left(\frac{M}{R}\,dr\right) = \frac{M}{R}\cdot\frac{R^3}{3} = \frac{1}{3}MR^2 \)
The diagnostic: before integrating, ask how much mass lies at distance r. For a rod it is the same at every r. For a disc it grows like r. For a solid sphere about its centre it grows like r squared. Getting that dependence right is the whole problem; the integral itself is easy.
The student's integral is perfectly computed. The problem is the mass element. M/R dr spreads the disc's mass evenly along a radius, as if the disc were a stick lying from the centre to the rim. That is a rod, and the integral duly gives the rod-about-its-end answer, one third M R squared.
A disc is not like that. The ring at radius r has circumference 2 pi r, so there is more material at larger r, and the correct mass element grows in proportion to r.
Because the disc's mass is weighted toward the outside, its moment of inertia is larger than the student's answer: one half rather than one third.
The diagnostic question in the final line of the slide is the habit to build. Before writing any integral, ask how the mass at distance r depends on r. Constant for a rod, proportional to r for a disc, proportional to r squared for a solid sphere about its centre. Get that right and the rest is mechanical.
Explain it to yourself
Discussion prompt
A hoop and a solid disc have the same mass and the same radius. Without quoting formulas, explain why the hoop has twice the moment of inertia, and predict which would be harder to spin up with the same torque.
Try to answer this in words before reading the model answer. Explaining a formula without quoting it is one of the best tests of whether you understand it, and the AP exam asks for exactly this kind of reasoning in its paragraph-length responses.
The core of the explanation is location. Every gram of the hoop is at the maximum possible distance R. In the disc, most of the mass is spread over smaller radii, where the squared weighting makes it count for much less.
The factor of two then comes from averaging r squared over the disc's area, which gives R squared over two. You do not need that detail to answer qualitatively, but it is satisfying to see why the number is exactly a half.
The prediction follows from torque = I alpha. With the same torque and twice the moment of inertia, the hoop gets half the angular acceleration. Mass placement, not just mass, decides how sluggish a rotating object is.
Section
Section 4
Concept
Figure (svg): A single particle on a light rod pivoted at one end, with a tangential force at the particle producing a tangential acceleration.
Write the second law along the tangent for one particle.
Why: Only the tangential force changes the speed; the radial part supplies the centripetal acceleration.
\[ F_t = ma_t = m r\alpha \]
Multiply both sides by r.
Why: The left side becomes the torque about the pivot.
\[ \tau = rF_t = (mr^2)\alpha \]
Sum over every particle of a rigid body.
Why: They all share alpha, and the internal torques cancel in pairs by the third law.
\[ \sum \tau_{\text{ext}} = \left(\sum m_i r_i^2\right)\alpha = I\alpha \]
So net torque = I alpha is not a new law — it is F = ma, applied particle by particle and summed. And just like F = ma, only external torques count, and they must all be taken about the same axis.
This slide earns the rotational second law rather than asserting it, and the derivation is short enough to reproduce on an exam.
Start with a single particle on a light rod. Only the tangential part of the force changes its speed, and Newton's second law along the tangent gives F_t = m a_t = m r alpha. Multiply both sides by r: the left becomes the torque, the right becomes m r squared alpha. For one particle, m r squared is the moment of inertia.
For a rigid body, every particle shares the same alpha, so summing over particles collects the m r squared terms into I. The internal forces between particles come in third-law pairs along the line joining them, so their torques cancel, leaving only the external torques.
So the rotational second law is not a separate postulate. It is Newton's second law, applied to every particle and summed. That is also why it comes with the same fine print: only external torques count, and every torque must be measured about the same axis as I.
Picture it
Figure (svg): Two dumbbells on vertical axes: one with its weights close to the axis labelled easy to spin, one with the same weights far out labelled hard to spin.
Hold the dictionary up against this picture. In translation, a given force produces less acceleration in a more massive object. In rotation, a given torque produces less angular acceleration in an object with a larger I.
But rotational mass has a twist that ordinary mass lacks: you can change it without adding or removing anything, just by moving mass in or out. Here the outer dumbbell's weights are 2.7 times farther out, so its I is more than seven times larger with identical mass.
That is why a figure skater spins faster by pulling their arms in, why a tightrope walker carries a long pole, and why a long hammer is harder to swing than a short one of the same weight.
The two dumbbells have identical masses. They differ only in where those masses sit, and that is enough to make one far harder to spin than the other.
This is the feature of rotational inertia that has no translational counterpart. You cannot make a car harder to accelerate without adding mass, but you can make a wheel harder to spin up just by moving its mass outward. That makes moment of inertia something you can design.
Engineers use it in both directions. Flywheels put their mass at the rim to resist changes in speed and store energy. Racing wheels are made light at the rim so they spin up quickly. Tightrope walkers carry long poles because the pole's large moment of inertia slows any tipping, giving them time to correct.
And a skater exploits it in real time. Arms out, the moment of inertia is large; arms in, it is small. The next deck shows why reducing it makes the skater spin faster — that is angular momentum conservation — but the underlying idea is already here: where the mass is matters as much as how much there is.
Worked example
A grinding wheel is a uniform disc of mass 2.0 kg and radius 0.10 m. A motor applies a 0.50 N m torque while friction in the bearings opposes with 0.10 N m. How long does the wheel take to reach 300 rad/s from rest? With the motor switched off, how long does it take to coast to a stop?
Find the moment of inertia.
Why: A uniform disc about its axle.
\[ I = \tfrac{1}{2}MR^2 = \tfrac{1}{2}(2.0)(0.10)^2 = 0.010\,\text{kg m}^2 \]
Spin-up: apply the second law with the NET torque.
Why: Friction opposes the motor, so the two torques subtract.
\[ \alpha = \frac{0.50 - 0.10}{0.010} = 40\,\text{rad/s}^2 \qquad t = \frac{300}{40} = 7.5\,\text{s} \]
Coasting: only friction acts, and it slows the wheel.
Why: Take the spin direction as positive, so the friction torque is negative.
\[ \alpha = \frac{-0.10}{0.010} = -10\,\text{rad/s}^2 \qquad t = \frac{0 - 300}{-10} = 30\,\text{s} \]
Sanity-check.
Why: The spin-up torque is four times the coasting torque, so spin-up should be four times faster: 7.5 s against 30 s. Consistent.
This problem is the rotational version of pushing a crate against friction, and it should feel familiar.
During spin-up, the motor torque and the friction torque act in opposite senses, so the net torque is their difference. Using the motor torque alone would overestimate the angular acceleration by a quarter.
During coasting, friction is the only torque, and it decelerates the wheel. Taking the spin direction as positive makes alpha negative, and the time comes out positive as it should.
The final check uses proportional reasoning rather than recomputation. The net spin-up torque is four times the coasting torque, so the wheel gains speed four times faster than it loses it, and the two times are in the ratio one to four. Checks like that catch arithmetic slips without redoing the whole problem.
Worked example
A wheel with I = 2.0 kg m^2 starts from rest. A motor applies a torque that grows as tau(t) = 4.0t newton metres. Find omega and the angle turned at t = 3.0 s.
Find alpha as a function of time.
Why: The second law holds at every instant, even though alpha is changing.
\[ \alpha(t) = \frac{\tau}{I} = \frac{4.0t}{2.0} = 2.0t\,\text{rad/s}^2 \]
Integrate alpha to get omega.
Why: It starts from rest, so the constant of integration is zero.
\[ \omega(t) = \int_0^t 2.0t'\,dt' = t^2 \qquad \omega(3.0) = 9.0\,\text{rad/s} \]
Integrate omega to get theta.
Why: Again from zero.
\[ \theta(t) = \int_0^t t'^2\,dt' = \frac{t^3}{3} \qquad \theta(3.0) = 9.0\,\text{rad} \]
Check why the constant-alpha equations would fail.
Why: Using omega squared = 2 alpha theta with alpha at t = 3 (6.0 rad/s^2) would give omega = 10.4 rad/s, which is wrong. Constant-alpha formulas need a constant alpha.
The pattern is identical to the kinematics deck: a time-dependent cause means integrate, never plug in.
When the torque changes with time, the angular acceleration changes too, and the constant-acceleration formulas no longer apply. The second law still holds at each instant, which gives alpha as a function of time.
From there it is the kinematics deck all over again: integrate alpha to get omega, then integrate omega to get theta, using the initial conditions to fix the constants.
The last step deliberately shows what goes wrong if you plug the final alpha into a constant-acceleration formula. You get an angular velocity of 10.4 rather than 9.0 radians per second — wrong, and wrong in a way that is hard to spot because it is not absurd.
So the rule is worth saying again: if anything in the problem depends on time, angle or speed, assume alpha is not constant and reach for calculus. The constant-alpha equations are a special case, not the default.
Concept
Figure (svg): A wheel with a rope wound round it and a mass hanging from the free end; the rope's acceleration equals the rim's tangential acceleration.
\[ a = R\alpha \qquad v = R\omega \]
This single line is what couples translation to rotation. The rope leaves the wheel at the rim, so the rope moves at the rim's tangential speed — and the hanging mass moves with the rope.
It is a constraint, not a law of physics: it holds because the rope does not slip on the wheel and does not stretch. Problems signal it with phrases like 'the rope does not slip' or 'wound around'. Without it, you would have one equation too few.
This constraint is the bridge between the two halves of the course. On one side of it is a rotating wheel described by alpha; on the other is a translating mass described by a.
The logic is geometric. The rope leaves the wheel at the rim. If it does not slip, the rope moves exactly as fast as the rim's surface, which is R omega. If it does not stretch, every bit of the rope, including the end tied to the mass, moves at that speed. Differentiate once and you get a = R alpha.
Without this equation, a wheel-and-mass problem has one too few equations. With it, the counting works: the mass gives one equation, the wheel gives one, and the constraint links their accelerations.
Watch for problems where the rope is wound on a radius different from the wheel's outer radius, as in a stepped spool. Then the R in the constraint is the radius where the rope leaves, not the size of the wheel.
Worked example
A 4.0 kg bucket hangs from a light rope wound around a solid cylinder of mass 8.0 kg and radius 0.20 m, free to turn on a frictionless axle. The bucket is released. Find its acceleration and the rope tension.
Figure (svg): A windlass cylinder with a rope descending to a bucket, plus separate free-body diagrams: tension down on the cylinder's rim, tension up and weight down on the bucket.
Bucket: translational second law, taking down as positive.
Why: Weight pulls down, tension pulls up.
\[ mg - T = ma \]
Cylinder: rotational second law about the axle.
Why: Only the tension has a lever arm; the axle force and the cylinder's weight act through the axis.
\[ TR = I\alpha = \left(\tfrac{1}{2}MR^2\right)\frac{a}{R} \;\Rightarrow\; T = \tfrac{1}{2}Ma \]
Combine and solve for a.
Why: Substitute T into the bucket equation.
\[ mg = \left(m + \tfrac{1}{2}M\right)a \;\Rightarrow\; a = \frac{(4.0)(9.8)}{4.0 + 4.0} = 4.9\,\text{m/s}^2 \]
Find the tension.
Why: Either equation works; use the cylinder's.
\[ T = \tfrac{1}{2}(8.0)(4.9) = 19.6\,\text{N} \]
Read the answer as an effective mass: the cylinder adds I/R^2 = 4.0 kg of inertia to the system, so the 39.2 N weight accelerates 8.0 kg instead of 4.0 kg. The general result is a = mg / (m + I/R^2), and it is worth remembering.
This is the template for every coupled translation-rotation problem in the unit, so it is worth following each step slowly.
There are two objects, so there are two free-body diagrams. The bucket gets an ordinary diagram with weight and tension. The cylinder gets its tension drawn at the rim where the rope leaves it, because the point of application matters for torque.
The cylinder's weight and the axle force both act through the axis, so neither appears in the torque equation. That is why the problem is solvable without knowing the axle force.
Combining the equations gives a = mg / (m + I/R squared). Read that as Newton's second law with an effective mass: the cylinder behaves like an extra I/R squared of mass that the falling weight must also accelerate. For a solid cylinder that extra mass is M/2, here 4.0 kg, which halves the acceleration from g to 4.9 metres per second squared.
Trap
The rope holds the bucket up, so the tension in it must be the bucket's weight, mg. That is what tension in a rope holding a mass always is.
Tension equals mg only if the bucket is not accelerating. On the previous slide the bucket weighs 39.2 N, but the tension is only 19.6 N.
If the tension equalled the weight, the net force on the bucket would be zero and it would not accelerate — yet it plainly falls. The gap between mg and T is exactly what accelerates the bucket.
| situation | tension | why |
|---|---|---|
| bucket held still | mg = 39.2 N | no acceleration, forces balance |
| bucket falls, massive cylinder | 19.6 N | less than mg so the bucket can accelerate |
| bucket falls, massless cylinder | 0 | nothing resists, so the bucket is in free fall |
| bucket hauled up, speeding up | more than mg | net force must point up |
The same error was a trap in the forces deck for the Atwood machine. It returns here because the rope now also has to turn a wheel — and the wheel is what sets how much less than mg the tension is.
The idea that tension equals weight is learned from things hanging still, and it is true there. It stops being true the moment the hanging object accelerates.
The cleanest way to see it is the bucket's own free-body diagram. If tension equalled weight, the forces on the bucket would cancel and it could not accelerate. But it does accelerate, so the tension must be less than the weight, and the difference is exactly m times a.
The table runs through the cases. Held still, T equals mg. Falling with a massive cylinder, T is less. Falling with a massless cylinder, nothing resists the rope, so T is zero and the bucket is in free fall. Being hauled up and speeding up, T exceeds mg.
The cylinder is what fixes the actual value. It needs a torque to spin up, and the tension is the only thing supplying it. A heavier cylinder needs more torque, so the tension is larger and the acceleration smaller. Tension is something you solve for, never something you assume.
Pattern
\[ \sum F = ma \qquad \sum\tau = I\alpha \qquad a = R\alpha \]
Step two is where these problems are won or lost. Consistent positive directions mean a falling bucket and a turning wheel both count as positive, so the constraint has no minus sign. Choose inconsistently and you will get a negative tension, which is the tell.
Step six is your error detector. Every answer in this unit should reduce to the massless-pulley result when I goes to zero.
This recipe works for every problem in the next section, and the order of the steps matters.
Separate diagrams for every object come first, and for the rotating body you must draw forces where they act, not at a single dot. A dot diagram hides the lever arms, and the lever arms are the whole point.
Consistent positive directions are the step where most errors are born. If you take down as positive for a falling mass, take the matching sense of rotation as positive for the wheel it unwinds from. Then the constraint is a = R alpha with no minus sign, and all your unknowns come out positive when your guess about the direction of motion is right.
Counting equations against unknowns tells you whether you are ready to solve. And the limit check at the end — set the moment of inertia to zero and see whether you recover the massless-pulley answer from the forces deck — catches almost every algebra slip.
Check
A mass m hangs from a light rope wound round a wheel of moment of inertia I and radius R, which turns freely on a fixed axle.
Check your understanding
What is the acceleration of the hanging mass?
Answer: B
Why: The mass gives mg - T = ma; the wheel gives TR = I alpha = I a / R, so T = (I/R^2) a. Adding them, mg = (m + I/R^2) a. The wheel behaves as an extra mass of I/R^2 that the weight must also accelerate, so the result is always less than g.
This is the bucket-and-cylinder result in general form, and it is worth deriving rather than recognising. Two equations — F = ma for the mass, torque = I alpha for the wheel — and the constraint a = R alpha are all it takes.
The answer has a structure you will see again and again: the driving force, mg, divided by the total effective mass. The hanging mass contributes m, and the wheel contributes I divided by R squared. That quotient has units of mass, which is a quick dimensional check that rules out choice C immediately.
The limit checks rule out the rest. As I goes to zero the answer must become g, since a massless wheel offers no resistance. As I grows the answer must shrink toward zero. Only one choice does both.
Whenever you derive a coupled-system result, run those two limits before trusting it. They take seconds and catch most algebra errors.
Section
Section 5
Worked example
Masses of 3.0 kg and 5.0 kg hang on either side of a pulley that is a uniform disc of mass 4.0 kg and radius 0.10 m. The rope does not slip. Find the acceleration and both tensions.
Figure (svg): An Atwood machine with a disc pulley: a lighter mass on the left rising and a heavier mass on the right falling, with different tensions T1 and T2 on the two sides.
Choose consistent directions.
Why: The 5.0 kg mass falls, so take down positive for it, up positive for the 3.0 kg mass, and clockwise positive for the pulley.
\[ T_1 - m_1g = m_1a \qquad m_2g - T_2 = m_2a \qquad (T_2 - T_1)R = I\frac{a}{R} \]
Add all three, after dividing the torque equation by R.
Why: The tensions cancel, leaving the driving weight difference against the total effective mass.
\[ a = \frac{(m_2 - m_1)g}{m_1 + m_2 + I/R^2} = \frac{(2.0)(9.8)}{3.0 + 5.0 + 2.0} = 1.96\,\text{m/s}^2 \]
Find each tension from its own mass.
Why: Note I/R^2 = M/2 = 2.0 kg for a uniform disc.
\[ T_1 = m_1(g + a) = 3.0(11.76) = 35.3\,\text{N} \qquad T_2 = m_2(g - a) = 5.0(7.84) = 39.2\,\text{N} \]
Verify with the pulley's equation.
Why: The torque is (39.2 - 35.3)(0.10) = 0.39 N m; I alpha = (0.020)(19.6) = 0.39 N m. They agree.
This is the Atwood machine from the forces deck with one change: the pulley now has mass, and therefore a moment of inertia. That single change makes the two tensions different and lowers the acceleration.
The sign choices are the key to clean algebra. The 5.0 kg mass falls, so down is positive for it; the 3.0 kg mass rises, so up is positive for it; and the pulley turns clockwise, so clockwise is positive for it. With those choices every acceleration is the same positive a, and the constraint has no minus signs.
Adding the three equations eliminates both tensions at once and gives the acceleration as the driving weight difference over the total effective mass. The pulley contributes I/R squared, which for a disc is half its mass.
The final check is worth copying as a habit. Recomputing the pulley's torque from the two tensions and comparing with I alpha confirms all three equations at once. If those numbers did not agree, one of the tensions would be wrong.
Socratic
Discussion prompt
In the forces deck the Atwood tension was the same on both sides of the pulley. Here it is not. What changed, and what would happen if the tensions were equal with a massive pulley?
This question is about understanding where a familiar rule comes from, so that you know when it stops applying.
In the forces deck, the tension was the same on both sides of the pulley. That was not a law; it was a consequence of the pulley being massless. A massless pulley has zero moment of inertia, so any net torque would give it an infinite angular acceleration. The only way to avoid that is zero net torque, which means equal tensions.
Give the pulley mass and the argument reverses. Now it needs a net torque to spin up, and the only source of torque is the difference between the two tensions. Equal tensions would leave the pulley unable to accelerate while the rope around it accelerates — a contradiction.
So the general rule is: every straight segment of rope between two objects with mass gets its own tension. Equal tensions are a special case for light, frictionless pulleys, and you should never assume them otherwise.
Edge cases
Parameter explorer
For the 3.0 kg and 5.0 kg Atwood machine, sweep the mass of the disc pulley from zero to 40 kg. What does the acceleration do, what is it at zero, and does it ever reach zero?
\[ a = \frac{(5.0 - 3.0)(9.8)}{8.0 + {M}/2}\,\text{m/s}^2 \]
Discussion prompt
Suppose the pulley were a hoop of the same mass instead of a disc. Would the acceleration be larger or smaller, and by how much does the denominator change?
Start the slider at zero. The acceleration is 2.45 metres per second squared, the massless-pulley result from the forces deck. That is the limit check in action: the general formula reduces to the special case.
As the pulley gets heavier, the acceleration falls, because the same driving weight difference must now also spin up more rotational inertia. The effective mass added is half the pulley's mass, since it is a disc.
The acceleration never reaches zero. However heavy the pulley, the weight difference still supplies a net torque, and any net torque produces some angular acceleration. A very heavy pulley makes the system sluggish, not stationary.
The follow-up about a hoop is a good test of the effective-mass idea. A hoop of the same mass has twice the moment of inertia of a disc, so it adds its full mass rather than half of it, and the acceleration drops further.
Worked example
A spool with I = 0.050 kg m^2 turns on a fixed frictionless axle. A 4.0 kg mass hangs from a rope wound on its inner radius, 0.10 m, on the left; a 1.0 kg mass hangs from a rope wound on its outer radius, 0.20 m, on the right. Find the angular acceleration, each mass's acceleration, and both tensions.
Figure (svg): A two-radius spool: a heavy mass hangs from the inner radius on the left and a light mass from the outer radius on the right; the heavy mass falls and the light mass rises.
Decide which way it turns by comparing torques at rest.
Why: Left: (4.0)(9.8)(0.10) = 3.92 N m. Right: (1.0)(9.8)(0.20) = 1.96 N m. The left side wins, so the 4.0 kg mass falls and the spool turns counter-clockwise. Take that as positive throughout.
Write each equation, with each rope's own constraint.
Why: The ropes leave at different radii, so a1 = r alpha and a2 = R alpha.
\[ m_1g - T_1 = m_1 r\alpha \qquad T_2 - m_2g = m_2 R\alpha \qquad T_1 r - T_2 R = I\alpha \]
Eliminate the tensions and solve for alpha.
Why: Substitute both tensions into the torque equation.
\[ \alpha = \frac{(m_1 r - m_2 R)g}{I + m_1r^2 + m_2R^2} = \frac{(0.40 - 0.20)(9.8)}{0.050 + 0.040 + 0.040} = 15.1\,\text{rad/s}^2 \]
Convert to linear accelerations and tensions.
Why: Note the light mass accelerates twice as fast, because its rope leaves at twice the radius.
\[ a_1 = 1.51\,\text{m/s}^2 \;\text{down} \qquad a_2 = 3.02\,\text{m/s}^2 \;\text{up} \qquad T_1 = 33.2\,\text{N} \qquad T_2 = 12.8\,\text{N} \]
Verify the spool equation.
Why: (33.2)(0.10) - (12.8)(0.20) = 0.76 N m, and I alpha = (0.050)(15.1) = 0.76 N m. They agree.
The stepped spool is the most demanding coupled problem in the unit, and it rewards a careful first step: deciding which way the system turns before writing any equations.
Compare the torques with the system held at rest. The heavier mass hangs at the smaller radius and the lighter mass at the larger one, so neither the masses nor the radii alone decide it. The heavier side wins here, 3.92 newton metres against 1.96, so take its direction as positive.
The new feature is that the two ropes leave the spool at different radii, so they have different accelerations. The spool has a single alpha, but the inner rope moves at r alpha and the outer rope at R alpha. The light mass ends up accelerating twice as fast as the heavy one.
After that it is the recipe: three equations, three unknowns once the constraints are substituted. The closing check — recomputing the net torque from the two tensions and comparing with I alpha — confirms the whole solution in one line.
Check
A spool turns on a fixed frictionless axle. A 2.0 kg mass hangs from its inner radius, 0.10 m, on one side; a 0.80 kg mass hangs from its outer radius, 0.30 m, on the other. The system is released from rest.
Check your understanding
What happens?
Answer: B
Why: Compare torques, not weights. The heavy side gives (2.0)(9.8)(0.10) = 1.96 N m; the light side gives (0.80)(9.8)(0.30) = 2.35 N m. The light mass wins because it hangs at three times the radius, so it falls and lifts the heavier mass.
This question tests whether you have fully absorbed that rotation responds to torque, not to force. The heavier mass is the obvious pick, and it is wrong.
The comparison takes two multiplications. The heavy mass gives 1.96 newton metres; the light mass, hanging three times farther out, gives 2.35. The light mass wins and falls, lifting the heavier one.
The distractor about the moment of inertia is subtle and worth understanding. The spool's I decides how quickly the system accelerates, because it appears in the denominator of alpha. It cannot decide the direction, because direction is set by the sign of the net torque, and the moment of inertia is always positive.
This is the principle behind every lever and pulley system that lets a small force lift a large load: trade distance from the axis for force.
Anomaly
Predict first
A uniform rod is hinged at one end and held horizontal, then released. At the instant of release, how does the downward acceleration of its free tip compare with g?
Correct: Greater than g — one and a half times g
Why: About the hinge, torque = Mg(L/2) and I = ML^2/3, so alpha = 3g/(2L). The tip's tangential acceleration is L alpha = 3g/2. The tip outruns free fall: a coin placed on the tip is left behind.
Figure (svg): A horizontal rod hinged at the left end; acceleration arrows at the centre and at the tip, the tip's longer than a free-falling ball's g arrow beside it.
This feels impossible — nothing is pushing the tip down except gravity, which only ever gives g. But the tip is not a free particle. It is part of a rigid body, and the rod's rotation drags the tip down faster than gravity alone could, using the hinge as the anchor.
The rod's centre of mass, by contrast, accelerates at only 3g/4 — less than g, as it must be, because the hinge pushes up with Mg/4. The worked example on the next slide computes all three numbers.
This result surprises almost everyone, which is exactly why it is worth studying. A coin placed on the end of a horizontal hinged rod is left behind when the rod is released: the rod's tip drops away from it faster than the coin can fall.
Nothing violates gravity. The coin is a free particle, and gravity gives it g. The tip is not free — it is attached to the rest of the rod, which is pinned at the hinge. The rod rotates as a whole, and a point far from the hinge must move a long way for each bit of rotation.
The numbers come from torque = I alpha about the hinge. The weight acts at the centre, half a length out, while the moment of inertia is one third M L squared. Their ratio gives alpha = 3g over 2L, and the tip, a full L from the hinge, gets 3g over 2.
The centre of mass, by contrast, falls at only 3g over 4, slower than free fall. That has to be true, because the hinge is pushing up on the rod. The anomaly is confined to points beyond two thirds of the length.
Worked example
A uniform rod of mass M and length L is hinged at one end and released from rest in a horizontal position. At the moment of release find its angular acceleration, the acceleration of its centre and tip, and the force from the hinge.
Torque about the hinge.
Why: Only gravity has a lever arm about the hinge, and it acts at the centre, a distance L/2 away, perpendicular to the rod.
\[ \tau = Mg\frac{L}{2} \]
Apply the second law with the end-axis moment of inertia.
Why: Taking torques about the hinge means the hinge force drops out, so it is the natural axis.
\[ \alpha = \frac{\tau}{I} = \frac{MgL/2}{ML^2/3} = \frac{3g}{2L} \]
Find the linear accelerations.
Why: At release omega is zero, so there is no centripetal part: the accelerations are purely tangential, which here means straight down.
\[ a_{\text{cm}} = \frac{L}{2}\alpha = \frac{3}{4}g \qquad a_{\text{tip}} = L\alpha = \frac{3}{2}g \]
Find the hinge force from the translational second law.
Why: The centre of mass obeys F = ma with the external forces: gravity down, hinge force up.
\[ Mg - F_h = M\left(\tfrac{3}{4}g\right) \;\Rightarrow\; F_h = \tfrac{1}{4}Mg \;\text{upward} \]
This is the full power of the two second laws used together: torque = I alpha gives the rotation, F_net = M a_cm gives the force you could not see. Using only one of them would leave the hinge force unknown.
This problem uses both second laws, and each one supplies something the other cannot.
The rotational law, taken about the hinge, gives the angular acceleration. The hinge force has zero lever arm about the hinge, so it drops out — which is convenient, but it also means this equation can never tell you what the hinge force is.
The translational law, applied to the centre of mass, supplies the missing piece. The centre accelerates at 3g over 4, and the net external force must equal M times that. Gravity provides Mg downward, so the hinge must provide a quarter of Mg upward.
Note the role of the instant of release. Because omega is zero then, there is no centripetal acceleration, and every point accelerates straight down. A moment later the rod is moving and the centripetal parts appear, and the hinge force gains a horizontal component. Questions that say 'just after release' are choosing the instant where the algebra is cleanest.
Section
Section 6
Concept
\[ \sum \vec{F} = 0 \qquad \text{and} \qquad \sum \tau = 0 \;\text{about any point} \]
A body at rest that stays at rest must have no linear acceleration and no angular acceleration. The first condition stops it sliding; the second stops it turning. Neither implies the other.
The phrase 'about any point' is the powerful part. If a body is in equilibrium, the net torque is zero about every point — not just its hinge, not just its centre. You are free to choose whichever point makes the algebra easiest, and the next slide shows how to choose.
In a planar problem this gives three equations: forces in x, forces in y, and torques. So you can solve for at most three unknowns. A problem with more is statically indeterminate, and the AP exam does not set those.
Equilibrium means no acceleration of any kind. Zero net force guarantees the centre of mass does not accelerate. Zero net torque guarantees the body does not start to spin. You need both, and the counterexample slide later shows a case with one and not the other.
The torque condition is unusually generous. In equilibrium, the net torque is zero about every point, not just about a hinge or the centre. That is not true in general for accelerating bodies, but when the net force is zero, the choice of reference point stops mattering for torque.
That freedom is a tool, and the next slide shows how to use it. You get to pick the pivot that makes the problem easiest.
For flat problems, the conditions give three independent equations: horizontal forces, vertical forces, and torque. Three equations solve for three unknowns. If you find yourself with four unknowns, look for a geometric condition you have missed, or recheck whether the problem has told you something like 'the wall is frictionless'.
Concept
Figure (svg): A beam with two unknown support forces; torques taken about the left support make that force vanish from the equation.
Put the pivot where an unknown force acts. That force passes through the pivot, so its lever arm is zero and it vanishes from the torque equation. If two unknowns act at the same point — like the two components of a hinge force — one pivot choice removes both.
Then the torque equation has a single unknown, which you solve immediately, and the force equations clean up the rest. This is the single most important technique in equilibrium problems, and it is why the 'any point' freedom matters.
This is the most important single technique in equilibrium problems, and once it becomes a habit these problems shrink dramatically.
Any force that acts at the pivot has zero lever arm about it and vanishes from the torque equation. So if you place the pivot where an unknown force acts, that unknown disappears, and the torque equation contains fewer unknowns — ideally just one.
Hinges are the best place of all, because a hinge force has two unknown components and both act at the same point. One pivot choice removes two unknowns at once. The same is true at the foot of a ladder, where the floor's normal force and friction both act.
After the torque equation gives you one unknown directly, the two force equations mop up the rest. The order matters: torque first, about the clever point; forces second.
Worked example
A uniform 4.0 m plank of mass 30 kg rests on supports at its left end and 3.0 m from the left end. A 50 kg person stands 3.5 m from the left end. Find both support forces, and how far right the person can walk before the plank tips.
Figure (svg): A plank on two supports, one at the left end and one at three metres, with a person standing beyond the second support near the right end.
Torques about the left support, counter-clockwise positive.
Why: N1 drops out. The plank's weight acts at its centre, 2.0 m out.
\[ 3.0N_2 - (294)(2.0) - (490)(3.5) = 0 \;\Rightarrow\; N_2 = \frac{588 + 1715}{3.0} = 768\,\text{N} \]
Vertical forces give N1.
Why: The two supports together carry the total weight of 784 N.
\[ N_1 = 784 - 768 = 16\,\text{N} \]
Find the tipping point.
Why: Tipping begins when N1 reaches zero: the plank pivots on the right support. Take torques about it with N1 = 0.
\[ (294)(1.0) = (490)d \;\Rightarrow\; d = 0.60\,\text{m beyond the support, i.e. at } 3.6\,\text{m} \]
N1 at only 16 N tells you the person is already close to the edge of safety. A support force can only push, so the instant your algebra calls for a negative N1, the plank has started to tip.
The torque equation is written about the left support so that N1 drops out immediately, leaving N2 as the only unknown. The vertical force balance then gives N1 in one line.
The small value of N1, only 16 newtons, is a warning sign. The left support is barely being pressed on. The person is standing so far out that they are nearly lifting the left end off its support.
The tipping question formalises that. A support can push but not pull, so the plank tips when the algebra would require N1 to become negative. At the threshold N1 is zero, the plank is on the verge of pivoting about the right support, and taking torques about that support finds the tipping position.
The result, 3.6 metres, is just beyond where the person currently stands. Physically, the person's torque about the right support has to beat the plank's weight acting on the other side of that support, and the plank's weight only has a one-metre lever arm to work with.
Worked example
A uniform 5.0 m ladder of mass 12 kg leans against a frictionless wall with its foot 3.0 m from the wall. Find the force from the wall, the friction at the floor, and the minimum coefficient of static friction that keeps it from slipping.
Figure (svg): A ladder leaning on a smooth wall: the wall pushes horizontally at the top, the floor pushes up and with friction toward the wall at the base, and weight acts at the midpoint.
Torques about the base.
Why: Both the floor normal and the friction act at the base, so one pivot choice eliminates two unknowns.
Find each lever arm from the geometry.
Why: The wall force is horizontal, so its lever arm is the vertical height, 4.0 m. The weight is vertical, so its lever arm is the horizontal distance to the midpoint, 1.5 m.
\[ N_w(4.0) = (12)(9.8)(1.5) \;\Rightarrow\; N_w = \frac{176.4}{4.0} = 44.1\,\text{N} \]
Forces in both directions.
Why: Horizontally, friction balances the wall; vertically, the floor carries the full weight.
\[ f = N_w = 44.1\,\text{N} \qquad N_f = mg = 117.6\,\text{N} \]
Find the minimum coefficient.
Why: Static friction can supply at most mu N_f, so it must be at least the friction required.
\[ \mu_{\min} = \frac{f}{N_f} = \frac{44.1}{117.6} = 0.375 \]
Move the foot farther out and the weight's lever arm grows while the wall's shrinks, so more friction is needed. That is why a ladder at a shallow angle slides out from under you.
The ladder is the classic equilibrium problem, and it shows why choosing the pivot well matters so much.
There are four forces: the wall's normal force, the weight, the floor's normal force and floor friction. Two of them, the floor forces, act at the base. Taking torques about the base removes both, leaving a single equation for the wall force.
The lever arms come straight from the geometry once you remember to use perpendicular distances. The wall force is horizontal, so its lever arm about the base is vertical — the height of the top. The weight is vertical, so its lever arm is horizontal — half the base distance.
Once the wall force is known, horizontal balance says friction must equal it, and vertical balance says the floor carries the full weight. The minimum coefficient follows. It rises as the ladder leans more, because the weight's lever arm grows while the wall's lever arm shrinks — the formal reason a shallow ladder slides out from under you.
Worked example
A uniform 2.0 m strut of mass 10 kg is hinged to a wall and held horizontal by a cable from its far end, making 30 degrees with the strut. A 20 kg sign hangs from the far end. Find the cable tension and both components of the hinge force.
Figure (svg): A horizontal strut hinged to a wall, supported by a cable from its outer end up to the wall at thirty degrees, with a sign hanging at the end.
Torques about the hinge, counter-clockwise positive.
Why: This removes both hinge components. Only the vertical part of the tension, T sin 30, has a lever arm along the horizontal strut.
\[ (T\sin 30^\circ)(2.0) - (98)(1.0) - (196)(2.0) = 0 \;\Rightarrow\; T = \frac{490}{1.0} = 490\,\text{N} \]
Horizontal forces.
Why: The cable pulls the strut toward the wall, so the hinge pushes it away.
\[ H_x = T\cos 30^\circ = 424\,\text{N} \;\text{(away from the wall)} \]
Vertical forces.
Why: The total weight is 294 N; the cable supplies T sin 30 = 245 N of it.
\[ H_y = 294 - 245 = 49\,\text{N} \;\text{(upward)} \]
The cable tension is five times the sign's weight. A shallow cable has a small vertical component, so it must pull very hard to supply enough of it — and that large horizontal pull is what the hinge must resist.
A hinged strut has three unknowns — the cable tension and the two components of the hinge force — and the hinge is the obvious pivot because it holds two of them.
About the hinge, only the vertical component of the tension has a lever arm along the horizontal strut. Its horizontal component points straight at the hinge and contributes nothing. That is why T sin 30 appears in the torque equation and not T.
The tension comes out at 490 newtons, five times the sign's weight. A cable at a shallow angle is inefficient at holding things up: most of its pull is horizontal, and only half of it, at 30 degrees, is vertical.
The horizontal component of that large tension pushes the strut into the wall, so the hinge must push back with 424 newtons. The vertical balance then needs only a small upward hinge force, 49 newtons. A useful check: the hinge components should point in physically sensible directions, and here they do.
Trap
In equilibrium problems you take torques about the pivot — the hinge, the support, the place the object would turn about. Any other point would give a different, wrong answer.
In equilibrium, net torque is zero about EVERY point. The hinge is just usually the most convenient one.
Check it on the ladder. Take torques about the TOP of the ladder instead. Now the wall force drops out, and the floor forces appear: the normal force acts 3.0 m horizontally from the top, friction acts 4.0 m vertically below it, and the weight acts 1.5 m horizontally.
\[ N_f(3.0) - f(4.0) - mg(1.5) = (117.6)(3.0) - (44.1)(4.0) - (117.6)(1.5) = 352.8 - 176.4 - 176.4 = 0 \]
Zero, exactly as it must be. The choice of point changes which unknowns appear, never the physics.
A useful consequence: when you have solved an equilibrium problem, take torques about a different point as a check. If it does not come out zero, there is an error somewhere.
This is the misconception that makes students afraid to choose a clever pivot. The worry is that some points are 'correct' and others give wrong answers.
In equilibrium they all give correct answers. The net torque about the top of the ladder is zero, just as it is about the base. The slide verifies this with the numbers already found: the three torques about the top cancel exactly.
What changes with the pivot is which forces appear in the equation. About the base, the wall force appears and the floor forces vanish. About the top, the reverse. The underlying physics is the same; the algebra is more or less convenient.
That gives you a free check on any equilibrium solution. After solving with one pivot, take torques about a different point using your answers. If the result is zero, your solution is consistent. If not, there is an error, and you have found it before the examiner does.
Counterexample
A student claims: 'If the forces on a body add to zero, the body is in equilibrium.'
Figure (svg): A steering wheel with two equal and opposite forces applied on opposite sides of the rim, forming a couple that turns it.
Discussion prompt
Find a counterexample to the student's claim, and say what extra condition is needed.
Look for a counterexample before reading the answer. A good one needs forces that cancel as vectors but still turn something.
The steering wheel is the standard example. Equal and opposite forces on opposite sides of the rim add to zero, so the wheel's centre does not accelerate. But both forces turn it the same way, so their torques add rather than cancel. The wheel spins.
A pair of equal, opposite, non-collinear forces like this is called a couple, and it has a remarkable property: its torque is the same about every point. That is the quantitative reason the 'any point' rule for equilibrium works.
So the student's claim misses half the story. Zero net force controls translation. Zero net torque controls rotation. A body in equilibrium needs both, and neither condition implies the other.
Real world
A tower crane lifts loads at the end of a long jib, sometimes 60 m from the tower. The tower itself is slender.
Discussion prompt
Use torque about the top of the tower to explain why a crane carries a large counterweight on its short back arm, and why the counterweight alone does not make the crane safe at every load.
This is torque balance on a very large scale, and the numbers are what make it convincing.
A load far out on the jib has a huge lever arm about the top of the tower. Even a modest load produces millions of newton metres of torque, and a slender tower would have to resist all of it by bending.
The counterweight supplies an opposing torque on the short back arm. Because the back arm is short, the counterweight has to be heavy — four times the load in the example, at a quarter of the distance. With the torques balanced, the tower is loaded straight down, which is what a tall column handles well.
The interesting part of the answer is the limitation. A fixed counterweight balances one particular load-times-distance. Any other load leaves an unbalanced torque, so cranes are rated by load moment rather than by load alone, and the operator's display monitors that product continuously.
Check
A uniform 4.0 m plank of mass 20 kg rests on two supports, 1.0 m and 3.0 m from its left end. A 60 kg painter walks from the middle toward the right end.
Check your understanding
How far from the left end can the painter stand before the plank begins to tip?
Answer: B
Why: At the tipping point the left support carries nothing, so take torques about the right support at 3.0 m. The plank's weight acts at 2.0 m, 1.0 m to the left: (20)(g)(1.0) = (60)(g)d gives d = 0.33 m beyond the support, so 3.33 m from the left end. The value of g cancels.
This is the tipping calculation from the plank example, with new numbers. The key move is the same: at the threshold of tipping, the far support carries nothing, so the plank is balanced on the near support alone.
Taking torques about the right support removes its unknown force and leaves two torques: the plank's weight acting one metre to the left, and the painter acting some distance d to the right. Setting them equal gives d = one third of a metre.
Notice that g cancels, as it always does when every force in a torque balance is a weight. You can work in kilogram metres throughout.
The common wrong answer, 3.67 metres, comes from giving the plank's weight a two-metre lever arm. That is its distance from the left end, not from the support you are pivoting about. Always measure lever arms from the pivot you chose.
Check
A uniform disc of mass M and radius R hangs from a nail through a point on its rim and swings about it, in the plane of the disc.
Check your understanding
What is its moment of inertia about the nail?
Answer: C
Why: The parallel-axis theorem from the centre: I = (1/2)MR^2 + MR^2 = (3/2)MR^2. The nail is a distance R from the centre of mass, and the shift always adds M d^2 to the centre-of-mass value.
A direct parallel-axis question, and the most common way it appears on the exam: a disc or a rod hung from a point other than its centre.
Start from the centre-of-mass value, half M R squared, and add M times the distance squared. The nail is R from the centre, so the extra term is M R squared, giving three halves in total.
The distractors are the three ways the theorem gets misapplied. Forgetting to shift at all gives the central value. Forgetting the body's own I_cm gives just M R squared. Using the wrong shape's I_cm, the hoop's, gives 2 M R squared.
You will meet this exact moment of inertia again in the oscillations deck, where a disc swinging from its rim is a physical pendulum whose period depends on it.
Section
Section 7
Comparison
Comparison matrix
| idea | translation | rotation |
|---|---|---|
| what causes change | force F | torque tau |
| what resists change | mass m | moment of inertia I |
| second law | F = ma | tau = I alpha |
| equilibrium condition | net force = 0 | net torque = 0 |
| depends on axis choice? | no | yes |
| link between them | a = R alpha | v = R omega |
The row that breaks the symmetry is the axis row. Mass is a fixed property of a body; moment of inertia is not. Every rotational quantity — torque, I, and next deck angular momentum — is defined relative to an axis, and must be computed about the same one.
Fill in the blanks from the dictionary you have been using all deck. Torque causes rotational change, moment of inertia resists it, and the second law and equilibrium condition have identical shapes.
The row that matters most is the one where the analogy breaks. Mass is a single number attached to a body. Moment of inertia depends on the axis, as does every torque. That is the one genuinely new idea in rotational dynamics, and it is the source of most errors: computing I about one axis and torques about another.
The last row is the bridge you used for every coupled problem: v = R omega and a = R alpha link a rotating object to anything moving with its rim.
Next deck extends the table with energy and momentum. Rotational kinetic energy will be half I omega squared, and angular momentum will be I omega — once again, the linear formulas with the dictionary applied.
Pattern
Point four is the one students forget. Torque = I alpha about the hinge tells you nothing about the hinge force, because that force has no lever arm there. The translational second law for the centre of mass is what finds it, exactly as on the falling-rod slide.
Next deck adds the energy and momentum versions of all this — rotational kinetic energy, rolling, and angular momentum — which often bypass the torque equation entirely.
This is the decision procedure for the whole unit, and the first question sorts problems into two very different families.
If nothing is accelerating, it is an equilibrium problem: zero net force, zero net torque, and a well-chosen pivot. If something is accelerating, you need the second laws, both translational and rotational.
The fourth point deserves attention because it is so easily forgotten. A torque equation about an axis can never tell you the force at that axis, since that force has no lever arm. To find an axle or hinge force on an accelerating body, use F_net = M a_cm for the body's centre of mass. The falling rod showed exactly this.
And the last point is the kinematics warning one more time. The angular constant-acceleration equations are only for constant alpha. A torque that varies with time or angle means calculus.
Estimation
Predict first
A 70 kg cyclist stands with their full weight on one pedal when the crank arm is horizontal. Bicycle cranks are about 17 cm long. Roughly what torque does this apply about the crank axle?
Correct: About 120 N m
Why: The weight is about 690 N and acts straight down; with the crank horizontal the lever arm is the full 0.17 m. 690 times 0.17 is about 117 N m. When the crank points straight down the lever arm is zero, which is why standing on a pedal at the bottom of the stroke does nothing.
For scale, tightening a car's wheel nuts takes about 100 to 140 N m — so a cyclist standing on a pedal is applying a car-mechanic's torque, twice per revolution.
The lever-arm variation through each revolution is why pedalling feels uneven, and why experienced riders learn to push forward over the top and pull back across the bottom: they are trying to keep the force perpendicular to the crank.
Estimation questions reward simple models with the right numbers. Here the model is a single force — the cyclist's weight — acting at the end of a horizontal crank arm.
A 70 kilogram person weighs about 690 newtons. With the crank horizontal, the weight is perpendicular to it, so the lever arm is the full 17 centimetres. The product is about 120 newton metres.
The comparison with tightening wheel nuts makes the number meaningful. Mechanics use a torque wrench set to roughly that value, so the cyclist is applying a serious torque — and doing it twice per revolution, once on each pedal.
The variation through the stroke is the deeper point. With the crank pointing straight down, the weight acts along the crank, the lever arm is zero, and standing on the pedal does nothing. Good cyclists learn to sweep their feet so the force stays as close to perpendicular as possible throughout the circle.
Exit ticket
Discussion prompt
From memory: (1) Two points on a spinning wheel are at r and 2r. Compare their omega, v and centripetal acceleration. (2) When is a force's torque about an axis zero? (3) A rod's I about its centre is ML^2/12 — what is it about an axis L/4 from the centre? (4) In an Atwood machine with a massive pulley, are the two tensions equal?
Answer all four from memory before looking at the model answers. These are the four ideas most likely to be tested directly, and each one has a common wrong answer.
The first checks the difference between angular and linear quantities. The second checks the line-of-action idea behind zero torque. The third checks that you route the parallel-axis theorem through the centre of mass, rather than through the end. The fourth checks that you no longer assume equal tensions over a massive pulley.
If any of them took you more than a few seconds, go back to the slide on that idea before the next deck, because the energy and angular momentum material is built directly on top of all four.
Recap
| the idea | the one-line version |
|---|---|
| v = r omega, a_t = r alpha | one omega for the body, a different v for each point |
| tau = r F sin theta | only the perpendicular part, at a distance, turns things |
| I = sum m r^2 | mass far from the axis counts quadratically more |
| I = I_cm + M d^2 | shift only from the centre-of-mass axis |
| net tau = I alpha | F = ma, summed particle by particle |
| equilibrium | zero net force AND zero net torque, about any point |
Next deck: energy and momentum of rotating systems. Rotational kinetic energy, rolling without slipping, and angular momentum — whose conservation explains the spinning skater properly, and which follows from net torque = I alpha exactly as momentum conservation followed from the third law.
This deck turned every object from a point into an extended body, and gave you the tools to describe how it turns.
The kinematics carried over directly, with theta, omega and alpha replacing x, v and a. Torque gave you the rotational version of force, with the lever arm as the essential new idea. Moment of inertia gave you the rotational version of mass — with the twist that it depends on the axis and on how the mass is arranged.
Net torque = I alpha then let you solve dynamics problems, and the constraint a = R alpha let you couple rotating wheels to translating masses. Static equilibrium added the torque condition to the force condition, and the freedom to choose your pivot.
The next deck asks the energy and momentum questions about rotating bodies. Many problems in it can be solved without ever computing a torque, but every idea there rests on what you built here.
Want this taught 1-on-1? Alexander tutors AP Physics C: Mechanics — $55/session, free consultation.