Engineering Mechanics

Newton's second law says that the acceleration of an object is proportional to the net force on it and inversely proportional to its mass. Push twice as hard and it speeds up twice as fast; make it twice as heavy and it speeds up half as fast. The equation also defines the unit: one newton is exactly the force that accelerates one kilogram at one metre per second squared, so $1\ \text{N} = 1\ \text{kg·m/s}^2$. Newton himself did not write $F = ma$; he wrote that force is the rate of change of momentum, $F = \mathrm{d}p/\mathrm{d}t$, which is the more general statement and reduces to $ma$ whenever the mass is not changing.

A 1400 kg car reaching 100 km/h — that is 27.8 m/s — in 8.5 s has an average acceleration of $a = 27.8/8.5 = 3.27$ m/s², so the net force driving it forward is $F = 1400 \times 3.27 \approx 4580$ N. The engine has to supply more than that, because drag and rolling resistance are pulling the other way and the 4580 N is what is left over after they have taken their share.

Almost every other force relation on this site is this one wearing a hat. Weight is $F = ma$ with $a = g$. Centripetal force is $F = ma$ with $a = v^2/r$. Impulse is this law integrated over time, and the work–energy theorem is it integrated over distance. Its rotational twin, $\tau = I\alpha$, swaps torque for force and moment of inertia for mass and behaves identically.

The word doing the most work in the law is "net", and it is the word most often dropped. A crate being pushed with 200 N across a floor that resists with 150 N of friction accelerates as though 50 N were acting, not 200. If an object moves at constant speed the net force on it is zero, however many forces are actually pushing on it. The second mistake is arithmetically worse: kilograms are not a force. A "70 kg load" is a mass, and the force it exerts on its hanger is $70 \times 9.81 \approx 687$ N. Feeding 70 into this calculator as a force when you meant a mass is wrong by a factor of 9.81. The imperial world hides the same trap behind identical words — the pound-mass and the pound-force are different quantities related by 32.174 ft/s², which is where the notorious $g_c$ conversion factor comes from. Use the unit selectors and let them handle it.

Mass and weight are different quantities, and this formula is the exchange rate between them. Mass is how much matter an object contains, measured in kilograms, and it is the same everywhere in the universe. Weight is the gravitational force acting on that mass, measured in newtons, and it changes with where you are standing. The relation $W = mg$ is just Newton's second law with the acceleration set to whatever gravity supplies locally.

A 70 kg person weighs $70 \times 9.80665 \approx 687$ N on Earth. Take the same person to the Moon, where $g = 1.62$ m/s², and they weigh about 113 N — roughly a sixth — while still being made of exactly 70 kg of person. On Mars, $g = 3.72$, the answer is 260 N. Nothing about the body changed; only the field it sits in did.

The $g$ used here, 9.80665 m/s², is standard gravity, a defined constant fixed by the third General Conference on Weights and Measures in 1901 so that engineering calculations would have one agreed number. Real gravity varies: about 9.780 m/s² at the equator and 9.832 at the poles, a 0.5% spread caused by the Earth's rotation and its slightly flattened shape, with smaller local variations from altitude and the density of the rock underfoot. This whole formula is a near-surface shortcut for universal gravitation — put Earth's mass and radius into $g = GM/R^2$ and 9.8 is what comes out.

The everyday confusion is baked into the instruments. A bathroom scale measures force, using a spring or a load cell, then divides by a value of $g$ programmed at the factory and displays the result as kilograms. Take that scale to the Moon and it will confidently report that you have lost five-sixths of your mass. A balance, which compares your weight against known masses, would read correctly anywhere — which is the older and more honest instrument. The same confusion lives in the kilogram-force, a legacy unit equal to 9.80665 N, and in the imperial pound, which does double duty as a mass and a force. One more, because it comes up constantly: astronauts on the space station are not weightless because gravity has run out. At 400 km altitude $g$ is still about 8.7 m/s², nearly 89% of its surface value. They float because they and the station are both in free fall around the Earth together — falling, and missing.

Pressure is force divided by the area it is spread over, $P = F/A$. The same force can be gentle or destructive depending entirely on how much surface it acts through, and that is the whole content of the idea. One pascal is one newton per square metre, which is a very small pressure — a sheet of paper lying on a table exerts about 1 Pa — so real numbers are almost always in kilopascals or higher. Standard atmospheric pressure is 101.325 kPa.

An 80 kg person weighs about 785 N. Standing in winter boots with maybe 350 cm² of sole in contact — 0.035 m² — the pressure under them is $785/0.035 \approx 22$ kPa, enough to break through crusted snow. Strap on snowshoes with 0.30 m² of bearing surface and the same 785 N becomes 2.6 kPa, and the snow holds. Run it the other way: lean on a thumbtack with 20 N through a point of about 0.01 mm², which is $1\times10^{-8}$ m², and the pressure at the tip is around 2 GPa — well past what wood fibre can resist.

The same relation is behind hydraulics. Pascal's principle says pressure applied to a confined fluid is transmitted undiminished throughout it, so a small force on a small piston becomes a large force on a large one: same $P$, bigger $A$, bigger $F$. A hydraulic jack with a 200:1 area ratio multiplies force 200-fold, at the price of moving 200 times as far. It also underlies the hydrostatic pressure page, where the force is simply the weight of the fluid standing above.

The area to use is the area actually in contact, and it is usually smaller than it looks. A tyre bears on its contact patch, not on the outline of the tread; a bolted flange bears on the annulus under the washer, not on the whole plate. Getting that wrong understates the real pressure, sometimes badly. The unit conversion is the other reliable trap: a square metre is 10 000 square centimetres and a million square millimetres, so an area entered in cm² against a force in newtons is out by four orders of magnitude. And $F$ must be the component perpendicular to the surface — a force at an angle contributes only its normal component to pressure, with the rest showing up as shear. Finally, be clear whether a quoted pressure is gauge or absolute. Gauge pressure is measured against the surrounding atmosphere, so a tyre reading "zero" is not empty; it holds about 101 kPa absolute. The suffixes psig and psia exist precisely because this goes wrong so often.

Torque is the rotational cousin of force: how hard something is twisted about a pivot. Push a 0.3 m wrench with 100 N perpendicular to the handle and you apply 0.3 × 100 × sin 90° = 30 N·m. Only the force component perpendicular to the lever arm turns the bolt — pushing along the handle (θ = 0°) accomplishes nothing, which is exactly what sin θ encodes. Archimedes understood the leverage half of this over two millennia ago: "give me a place to stand and I will move the Earth".

The practical lesson is in the trade-off — a longer wrench delivers the same torque with less force, which is why breaker bars exist and why door handles sit far from the hinges. When solving for θ, the arcsin returns only the principal branch, angles up to 90°; the supplementary angle 180° − θ produces the identical torque, so check which geometry matches your setup.

Split the weight along the slope and you get mg sin θ — the force a rope, a brake or friction must hold to keep an object from sliding down. A 20 kg crate on a 30° ramp pulls down-slope with 20 × 9.80665 × sin 30° ≈ 98 N, exactly half its weight, because sin 30° = ½. Galileo built his entire kinematics programme on this: an incline "dilutes" gravity by sin θ, slowing a falling body enough to time it with a water clock in 1604, centuries before anything could time a free fall directly.

Highway grades use the same maths in disguise — a 6% grade means a rise of 6 m per 100 m, θ ≈ 3.43°, so a 40-tonne truck feels about 40,000 × 9.80665 × sin 3.43° ≈ 23 kN pushing it downhill, which is precisely why runaway-truck ramps exist. Pair this with N = mg cos θ and you have the complete free-body diagram for any slope; mixing up which one takes sine and which takes cosine is the standard exam trap, so check the limits: on level ground the down-slope force must vanish, and sin 0° = 0 does exactly that.

On a slope, gravity still pulls straight down, but the surface can only push back perpendicular to itself — so it carries just the cos θ share of the weight: N = mg cos θ, with g = 9.80665 m/s². A 10 kg block on a 60° ramp presses in with only 10 × 9.80665 × cos 60° ≈ 49 N, half its 98 N weight. Level ground (θ = 0°) recovers N = mg, and a vertical wall (θ = 90°) gives N = 0, which is why nothing rests on a wall.

This is the quiet half of every incline problem, because friction is proportional to N: as the slope steepens, the down-slope pull grows while the friction budget shrinks, and the object eventually lets go. The classic error is using the full weight for N on a ramp, which overestimates friction and predicts that boxes stay put when they will actually slide.

Static friction is the only force in introductory mechanics that is written with an inequality: f ≤ μₛN. It supplies exactly whatever is needed to prevent sliding, up to a ceiling of μₛN, and this formula computes that ceiling. Press a 500 N normal load onto a surface with μₛ = 0.6 and you can push with anything up to 300 N and nothing moves; at 301 N the object breaks free and the weaker kinetic friction takes over.

That inequality is the single biggest trap: plugging μₛN in as "the friction force" on a stationary object is wrong unless the object is on the verge of slipping. The distinction pays real dividends — a car's tyres grip through static friction as long as they roll, which is why ABS pumps the brakes to keep them from locking into a lower-μₖ skid, and why a driven wheel spinning on ice suddenly has far less traction than one that is merely rolling.

Once a surface is already sliding, friction settles to a nearly constant value proportional to how hard the surfaces are pressed together: f = μₖN. Push a 200 N-loaded crate across a floor with μₖ = 0.3 and it resists with 60 N no matter how fast you shove it. The startling part — that friction depends on load but essentially not on contact area or speed — was established by Guillaume Amontons in 1699 and confirmed in Charles-Augustin de Coulomb's prize-winning 1785 study of rope, axles and rigging for the French navy, work so thorough that dry friction is still called Coulomb friction.

Typical coefficients: rubber on dry asphalt around 0.7, steel on steel about 0.6 lubricated down to 0.05, waxed ski on snow near 0.05, and PTFE on steel about 0.04. The trap is assuming N equals the weight — true only on level ground with no extra push or pull. On a slope N shrinks to mg cos θ, and a downward-angled push or a car's aerodynamic downforce raises it. Note also that kinetic friction always runs slightly below the static maximum, which is why a heavy box lurches forward the instant it breaks free.

Tilt a plank until the block on it just begins to slide: at that angle the down-slope pull mg sin θ exactly equals the friction ceiling μₛ mg cos θ, the mass cancels, and μₛ = tan θ. It is the cheapest friction experiment in existence — no force gauge, no scale, just a protractor. A block that lets go at 31° reports μₛ = tan 31° ≈ 0.60.

The same angle governs bulk solids: pour sand, grain, gravel or cement and the cone stabilises at its angle of repose, roughly 34° for dry sand and 40° for crushed stone. Silo designers, mining engineers and highway embankment crews all size their structures around it, and geologists read it back out of scree slopes. The trap is dimensional intuition — since the mass cancels, a heavy block does not slide at a gentler angle than a light one of the same material, which surprises nearly everyone the first time they see it.

Torque is the turning effect of a force about a pivot, and $\tau = rF\sin\theta$ says it depends on three things: how hard you push, how far from the pivot you push, and the direction you push in. Only the component of force perpendicular to the lever actually turns anything, which is the whole job of the $\sin\theta$. Push straight along a wrench handle, $\theta = 0$, and the bolt does not care at all. Push at right angles, $\theta = 90°$, and every newton counts.

A 200 N push applied perpendicular at the end of a 300 mm wrench delivers $\tau = 0.30 \times 200 \times \sin 90° = 60$ N·m. Slip a 600 mm breaker bar on instead and the same 200 N gives 120 N·m — the reason breaker bars exist, and the reason a wheel nut torqued to 120 N·m can be undone by a person who could not possibly generate that force directly.

The idea is old. Archimedes set out the law of the lever in the third century BC and is supposed to have said that with a place to stand he could move the Earth, which is this formula pushed to its limit: any torque is available if $r$ is large enough. The rotational world is built on it — the moment of a force in statics, the bending moment in a beam, and $\tau = I\alpha$ in dynamics are all the same quantity in different contexts.

The measurement error is measuring $r$ to the wrong point. The lever arm runs from the axis of rotation to the point where the force is applied, and both ends get mistaken. On a torque wrench with an extension or a crow's-foot adapter fitted, the effective length is no longer the wrench's marked length, and the reading on the scale no longer equals the torque at the fastener — that correction catches experienced people. On a bolted joint, the pivot is the bolt axis, not the edge of the bracket. The second mistake is a symbol swap: work is $W = Fd\cos\theta$ and torque is $\tau = rF\sin\theta$, and the two angles are measured the same way but enter through different trigonometric functions, because work wants the component along the displacement and torque wants the component across the lever. Getting them the wrong way round turns a maximum into a zero. One more, for when you solve for $\theta$: arcsine returns only the principal branch up to 90°, and the supplementary angle $180° - \theta$ produces exactly the same torque, so check which geometry your setup actually has.

Balance the torques about a pivot, $F_e \cdot d_e = F_l \cdot d_l$ — and the force multiplication falls out as the ratio of the arms. A 1.2 m crowbar with its fulcrum 0.3 m from the load gives MA = 4, so 200 N of effort lifts 800 N. Archimedes summed it up around 250 BC with "give me a place to stand and I will move the Earth", and the three lever classes still classify tools today: a seesaw (class 1), a wheelbarrow (class 2), and tweezers or a human forearm (class 3, where MA is deliberately less than 1 to trade force for speed and range).

Nothing is free — the effort end must travel the same factor further, so a MA of 4 means moving your hand 4 cm to raise the load 1 cm, and the work in equals the work out. Real levers fall short of the ideal because friction at the pivot and flex in the bar eat a few percent, which is why the actual mechanical advantage measured from forces is always a little below this ideal value computed from lengths.

In a block and tackle the load hangs from several sections of the same rope, and since the tension is identical everywhere along an ideal rope, each section carries an equal share: F = W ⁄ n, where n counts only the sections actually pulling up on the moving block. An 800 N load on a four-part tackle needs just 200 N of pull. Plutarch tells of Archimedes using a compound pulley to draw a fully laden ship out of Syracuse harbour single-handed, to King Hiero's astonishment.

The price, again, is rope: to raise the load 1 m you must haul n metres through your hands, so the work is unchanged. Count carefully — the classic error is including the rope section you are pulling on when it runs off a fixed pulley (a fixed pulley only changes direction, adding nothing to n) and forgetting to include it when it leaves the moving block. Real tackle loses roughly 5–10% per sheave to friction, so a nominal 4:1 rigging typically pulls closer to 3.3:1.

No machine returns everything you put in: friction, elastic hysteresis, windage and noise take a cut, so $\eta = W_{\text{out}} / W_{\text{in}}$ always lands below 1. Feed a hoist 1000 J and get 750 J of lifting done and it is 75% efficient, with 250 J warming the bearings. Enter η as a plain ratio or switch the unit to % — the calculator handles both.

The historical spread is enormous. Thomas Newcomen's 1712 atmospheric engine converted well under 1% of its coal's energy into work; Watt's separate condenser roughly tripled that; a modern combined-cycle gas turbine reaches about 60%, a large electric motor 95%, and a bicycle drivetrain around 97%. Efficiency chains multiply, which is why a 90%-efficient gearbox behind a 90%-efficient motor delivers 81% overall — and why the honest way to quote a system is end to end, not stage by stage.

Reactions come straight from statics, and the lever-arm logic is worth internalising rather than memorising: each support takes the fraction of the load proportional to its distance from the opposite end. A 40 kN load 2 m along a 6 m span puts $40 \times 4/6 = 26.7$ kN on the near support and $40 \times 2/6 = 13.3$ kN on the far one. The two sum to 40 kN, as they must, and the near support — the one the load is closer to — takes the larger share.

The "opposite end" part is where people go wrong, and it is worth a sanity check every time: slide the load right up against a support and that support should take essentially all of it. This crossed relationship is why an off-centre load punishes one support badly. Park a crane outrigger at the quarter point of a beam and that support carries 75% of the load, not half, and the beam-to-column connection, the bearing plate and the column below all have to be checked for that number rather than for the average.

Two practical follow-ons. Reactions are what the supporting structure has to carry, so this is the number that migrates down into the column, the wall, the footing and eventually the soil — and it is also what a scaffold, shoring tower or crane mat gets sized on. And for a run of loads, superposition works: compute each load's contribution to each reaction and add. It is exact, unlike most rules of thumb, because reactions on a determinate beam come from equilibrium alone and never depend on the beam's stiffness or material at all.

Put a single load anywhere on a simply supported span and the moment diagram is two straight lines meeting under the load, peaking at $M = Pab/L$ where $a$ and $b$ are the distances to each support. A 40 kN load at 2 m of a 6 m span gives $40{,}000 \times 2 \times 4/6 = 53.3$ kN·m. Check it independently: the near reaction is $Pb/L = 26.7$ kN, and walking out to the load gives $26.7 \times 2 = 53.3$ kN·m. Same number.

The useful shape of this result is that it is a parabola in $a$, maximised at midspan where it becomes the familiar $PL/4$. Our 53.3 kN·m is only 89% of the 60 kN·m a midspan load would produce, and moving the load from the third point out to the quarter point drops it further to 75%. So off-centre loads are always kinder than centred ones, which cuts two ways: it means a lifting beam checked at midspan is conservative wherever the hook actually is, and it means a moment measured in the field cannot tell you where the load was without more information — two positions symmetric about midspan give exactly the same peak.

Two traps. First, the maximum moment is under the load, but the maximum deflection is not — it sits nearer midspan, at $\sqrt{(L^2-b^2)/3}$ from the far support, and for a badly off-centre load the two locations are far apart. Second, this is one load in isolation. Add a second point load and neither result transfers; you have to build the shear diagram and find where it crosses zero, because that crossing, not the load position, is what locates the peak moment.

Under constant acceleration, velocity changes at a steady rate, so the final speed is simply the starting speed plus the acceleration multiplied by the elapsed time. Picture a car merging onto a highway: entering the ramp at 15 m/s and holding a steady 2 m/s² for 5 seconds, it reaches v = 15 + (2)(5) = 25 m/s — right at highway pace. Deceleration works the same way with a negative a, which is how stopping times are estimated from braking data.

This is the first of the SUVAT equations, the toolkit of uniformly accelerated motion that traces back to Galileo's inclined-plane experiments in the early 1600s, where he showed that falling bodies gain equal speed in equal times. Because the relationship is linear in every variable, each of the four rearrangements has exactly one answer — no square roots, no ambiguity — making it the friendliest member of the kinematics family.

When acceleration is constant, displacement has two parts: the distance you would cover at your initial velocity alone, plus the extra distance contributed by speeding up — and that extra grows with the square of time. A jet starting its takeoff roll from rest and holding 2 m/s² covers d = 0 + ½(2)(30²) = 900 m in 30 seconds, which is why runways are measured in kilometres.

The ½ appears because the acceleration term is built from the average of a speed that grows linearly from zero. Galileo uncovered the underlying pattern — distances in successive equal time intervals follow the odd numbers 1, 3, 5, 7 — by rolling bronze balls down inclined planes. Note that solving for t would mean solving a quadratic with potentially two positive roots, so this calculator rearranges only for d, v₀, and a, where the answer is always single-valued.

This is the SUVAT equation with the clock taken out of it. Every other member of the family needs a time; this one relates the starting speed, the finishing speed, the acceleration and the distance directly, which makes it the right tool whenever you know where something ended up but not how long it took getting there. You obtain it by solving $v = v_0 + at$ for $t$ and substituting into $d = v_0 t + \tfrac{1}{2}at^2$; the time cancels and $v^2 = v_0^2 + 2ad$ is what survives.

A car braking from 25 m/s — 90 km/h — at a firm $-8$ m/s² comes to rest in $d = (0 - 25^2)/(2 \times -8) \approx 39$ m. That is the distance the tyres need after the brakes are on, and adding the driver's reaction time at 25 m/s puts another 25 m or so in front of it. Braking-distance tables in road-safety pamphlets are this equation, run once per speed.

Multiply the whole thing by $\tfrac{1}{2}m$ and it turns into something familiar: $\tfrac{1}{2}mv^2 = \tfrac{1}{2}mv_0^2 + (ma)d$, which is the work–energy theorem, final kinetic energy equals initial kinetic energy plus the work done by the net force. That is not a coincidence — it is the same statement written twice, once in the language of kinematics and once in the language of energy. It also explains the squares: energy has always gone as $v^2$, so a relation involving distance and force had to.

Signs are where this equation punishes carelessness. Pick a positive direction, then stay in it: a car braking while travelling in the positive direction has a negative $a$, and entering $+8$ instead of $-8$ returns a stopping distance of $-39$ m, which is the calculator telling you the car would have had to be reversing. The physical consequence of the squares is the one every driving instructor tries to convey and this formula proves: stopping distance goes as the square of speed. The 39 m from 90 km/h becomes about 156 m from 180 km/h. A 20% increase in speed is a 44% increase in the distance you need. Two smaller points: solving for $v$ or $v_0$ takes a square root, and this page returns the principal non-negative branch — if the object actually reversed direction, the negative root is the physical one and you should supply the sign yourself. And $d$ is displacement along the direction of motion, not path length, so it is not the right tool for a curved route.

Strip friction away and the down-slope force mg sin θ divided by the mass m leaves a = g sin θ — the mass cancels completely, so a marble and a bowling ball slide identically. A 30° ramp yields 9.80665 × 0.5 ≈ 4.90 m/s², exactly half of free fall. This is Galileo's "diluted gravity": by 1604 he had shown that distances on an incline still grow as t², and he extrapolated to θ = 90°, where a = g and the ramp becomes free fall.

Watch the assumption — sliding, not rolling. A ball that rolls without slipping must also spin up its own moment of inertia, so a solid sphere manages only (5/7)g sin θ and a hoop just (1/2)g sin θ. That difference is the whole point of the classic race down a ramp, where a solid cylinder always beats a hollow one of identical mass and radius.

Gravity pulls the block down-slope with mg sin θ while friction drags back with μₖ mg cos θ; divide the difference by m and the mass drops out, leaving a = g(sin θ − μₖ cos θ). A 30° ramp with μₖ = 0.2 gives 9.80665 × (0.5 − 0.2 × 0.866) ≈ 3.20 m/s², a third slower than the frictionless 4.90 m/s². If the bracket comes out negative, the block was never sliding in the first place — the slope sits below the angle of repose, and the honest answer is a = 0.

Solving for θ uses the amplitude-phase identity sin θ − μ cos θ = √(1+μ²)·sin(θ − arctan μ), which is why the rearranged angle carries an arctan and an arcsin. The equation is the working model behind ski-slope grooming, luge run design and conveyor chute angles, all of which are chosen to land the acceleration in a narrow, controllable band.

Newton's second law on a hoisted load reads T − mg = ma, so the rope carries T = m(g + a). Lift a 50 kg crate while accelerating upward at 2 m/s² and the rope feels 50 × (9.80665 + 2) ≈ 590 N, about 20% more than the 490 N it holds at rest or at constant speed. Decelerate on the way up — or accelerate downward — and a goes negative, easing the tension; at a = −g the rope goes completely slack and the load is in free fall.

This is exactly why you feel heavy as an elevator starts up and light as it starts down: the floor is the "rope", and the scale under your feet reads m(g + a). Rigging engineers turn the same relation into a dynamic load factor, sizing slings and hooks for the accelerating case rather than the static weight — snatching a load, or an emergency stop, can spike the tension well past the crane's nameplate figure.

Hang two masses over a light, frictionless pulley and only the difference in weight drives the system, while the total mass has to be accelerated — giving a = (m₁ − m₂)g ⁄ (m₁ + m₂). With 3 kg against 2 kg the acceleration is (1 ⁄ 5) × 9.80665 ≈ 1.96 m/s², a fifth of free fall. Make the masses nearly equal and the acceleration becomes as gentle as you like, which is precisely the point: George Atwood built his machine in 1784 to slow gravity down enough to verify Newton's laws with the crude clocks of the day.

Two sanity checks fall out immediately. Equal masses give a = 0, the system balances. Let m₂ → 0 and a → g, plain free fall. The idealisation to watch is the pulley: a real one has rotational inertia and bearing friction, so measured accelerations run a few percent low, and the rope's own mass matters once the masses are small. Elevator counterweights are the industrial version of the same trick — balancing most of the car's weight so the motor only has to handle the difference.

Above walking pace, resistance through a fluid follows the quadratic law $F = \tfrac{1}{2} C_d \rho A v^2$, where $C_d$ is a shape factor measured in a wind tunnel: about 1.1 for a flat plate, 0.47 for a sphere, 0.25–0.35 for a modern car, and roughly 0.04 for a sailplane fuselage. A car with $C_d$ = 0.3 and 2.2 m² of frontal area meets 0.5 × 0.3 × 1.225 × 2.2 × 30² ≈ 364 N of drag at 30 m/s. Gustave Eiffel, having finished his tower, spent his later years dropping instrumented shapes down its side and then building France's first serious wind tunnel, producing the earliest reliable drag coefficients.

The v² is the entire story of highway fuel economy: drag force quadruples when you double speed, and since power is force times velocity, the power needed to overcome it grows with the cube — 8× the power from 50 to 100 km/h. That is why the last few km/h of top speed cost so much engine, why cyclists draft, and why the formula multiplies $C_d$ by A rather than treating them separately; a slippery shape on a huge frontal area still pushes a lot of air.

A falling body speeds up until drag, which grows as v², matches its weight; after that the net force is zero and the speed locks in at $v_t = \sqrt{2mg / (\rho A C_d)}$. For a belly-to-earth skydiver — roughly 80 kg, 0.7 m² of frontal area, $C_d$ ≈ 1.0, air at 1.225 kg/m³ — that works out to about 43 m/s, near the familiar 190 km/h. Pull into a head-down dive and A collapses, pushing terminal velocity past 90 m/s; deploy a parachute and A jumps by two orders of magnitude, dropping it to a survivable 5 m/s.

Density matters as much as shape, which is why Felix Baumgartner exceeded the speed of sound in 2012 at 39 km altitude: with ρ perhaps 1% of sea-level air, $v_t$ rises roughly tenfold. The formula also explains why small animals survive falls that kill large ones — mass grows with the cube of size while area grows with the square, so a mouse's terminal velocity is a fraction of a horse's. Note that $v_t$ is an asymptote, not a speed reached at a definite moment; a skydiver is within a few percent of it after about 12 seconds.

In physics, work is force applied through a distance — and only the component of force along the motion counts, which is where the cos θ comes from. Pull a sled with 100 N on a rope angled 30° above the snow for 20 m, and you do W = 100 × 20 × cos 30° ≈ 1732 J, not the full 2000 J. The term itself was coined by the French engineer Gaspard-Gustave de Coriolis in 1826, precisely to compare what steam engines and horses could deliver.

Two useful edge cases: a force perpendicular to the motion (θ = 90°) does no work at all — the Moon's orbit costs gravity nothing — and a force opposing the motion, like friction, does negative work, showing up as cos θ below zero. Solving for θ uses the arccos principal branch, 0° to 180°, which conveniently covers the entire physical range of angles between two directions.

Kinetic energy is the energy an object carries because it is moving, $E_k = \tfrac{1}{2}mv^2$, and equivalently it is the work you would have to do to bring it from rest up to that speed — or the work it can do on something else in coming back to rest. The two odd-looking features, the half and the square, both fall out of that second statement. Push with force $F = ma$ through a distance and integrate: $\int ma\,\mathrm{d}x = \int mv\,\mathrm{d}v = \tfrac{1}{2}mv^2$. The square is not a modelling choice, it is what the integral hands back.

A 1500 kg car at 50 km/h — 13.9 m/s — carries $\tfrac{1}{2} \times 1500 \times 13.9^2 \approx 145$ kJ. The same car at 100 km/h, 27.8 m/s, carries about 580 kJ. Twice the speed, four times the energy, and since the brakes can only dissipate energy at roughly a fixed rate per metre of road, roughly four times the distance to stop.

Which form of energy is the "real" one was a genuine dispute. Descartes and his followers backed $mv$; Leibniz argued for what he called vis viva, $mv^2$. Willem 's Gravesande settled the experimental half of it in the 1720s by dropping brass balls into soft clay and finding that a ball arriving twice as fast sank about four times as deep, and Émilie du Châtelet made the theoretical case in the 1740s alongside her translation and commentary on the Principia. Both quantities turned out to matter — momentum $mv$ is conserved in every collision, kinetic energy only in elastic ones — which is why this site has pages for each.

Everything that goes wrong here goes wrong at the square. Doubling the speed does not double the energy, and the intuition that it does is what makes highway speeds feel deceptively similar to city ones. The unit trap follows directly: enter a speed in km/h where the formula wants m/s and you are wrong by $3.6^2 = 12.96$, not by 3.6 — the error is an order of magnitude and it looks plausible. Two smaller ones. Kinetic energy is a scalar with no direction, so two cars closing head-on do not have "negative" energy relative to each other, and you cannot cancel them the way you cancel momenta. And it is frame-dependent: a coffee cup on a train table has zero kinetic energy in your frame and a great deal in the frame of the platform. That is not a flaw; it is why the work–energy theorem only ever deals in changes.

Lifting a mass banks energy in the gravitational field, and near Earth's surface the deposit is simply mgh, with g = 9.80665 m/s². A roller coaster earns its entire ride on the first climb: a 500 kg car hauled 30 m up stores 500 × 9.80665 × 30 ≈ 147 kJ, which the drops and loops then spend as speed. Only differences in height matter — you are free to call the ground floor, the table top, or sea level "zero", as long as you stay consistent.

The same idea runs entire power grids: pumped-storage hydro plants push water uphill when electricity is cheap and let it fall through turbines at peak demand, storing gigawatt-hours as nothing more than elevated water. The formula is a near-surface approximation — it treats g as constant, excellent for heights small compared with Earth's radius.

Whatever the forces, the net work done on an object shows up entirely as a change in its kinetic energy: W = ½m(v² − v₀²). Accelerate a 2 kg mass from rest to 10 m/s and exactly 100 J went in, no matter whether it took 1 m of huge force or 100 m of gentle push. Gaspard-Gustave de Coriolis formalised both "work" and the ½mv² form of kinetic energy in 1829, precisely so factory owners could compare what different machines actually delivered.

The theorem's power is that it skips time entirely, making it the fastest route to braking distances: a 1360 kg car slowing from 26.8 m/s to 13.4 m/s sheds about 367 kJ, and dividing that by the braking force gives the stopping distance directly. Sign discipline is the trap — friction and braking do negative work, so W comes out negative whenever the object slows, and this calculator will happily return a negative number to tell you so.

Robert Hooke published this law in 1676 as a Latin anagram — ceiiinosssttuv — unscrambled two years later to "ut tensio, sic vis": as the stretch, so the force. An ideal spring pushes or pulls back in proportion to how far you displace it from rest. The spring constant k is the stiffness: a 200 N/m spring stretched 0.1 m pulls back with 20 N, while a car's suspension spring might run tens of thousands of N/m.

This calculator uses the magnitude form; strictly the restoring force points opposite the displacement, which is written F = −kx and is what makes released springs oscillate. The law holds only up to the elastic limit — stretch a spring too far and it deforms permanently. Within that limit it underpins spring scales, force gauges, vehicle suspensions, and even the atomic bonds that make solids springy.

Compressing or stretching a spring puts energy into it, and $U = \tfrac{1}{2}kx^2$ says how much. The half is not a fudge factor. By Hooke's law the force you must apply grows steadily from zero at the start of the stretch to $kx$ at the end, so the average force over the whole displacement is $\tfrac{1}{2}kx$, and work is average force times distance. Draw the Hooke's law line on a force-versus-displacement graph and the stored energy is the area beneath it — a triangle, and the area of a triangle carries a half.

A 400 N/m spring compressed 50 mm stores $U = 0.5 \times 400 \times 0.05^2 = 0.5$ J. Compress the same spring 100 mm and it stores 2 J, not 1 — the square again. That is the arithmetic behind a mousetrap, a valve spring, and the reason a bow drawn to full draw stores so much more than one drawn halfway.

The shape of this expression is worth recognising because it recurs everywhere. Kinetic energy is $\tfrac{1}{2}mv^2$; rotational kinetic energy is $\tfrac{1}{2}I\omega^2$; the energy in a capacitor is $\tfrac{1}{2}CV^2$ and in an inductor $\tfrac{1}{2}LI^2$. Every one of them is the integral of a quantity that grows linearly, and every one of them therefore comes out as a half times a coefficient times a square. Spot the pattern once and four formulas stop needing to be memorised separately.

The mistake that matters in real machinery is measuring $x$ from the wrong place. The $x$ in this formula is displacement from the spring's free length — its length when nothing is touching it — not from its installed length and not its total length. A spring installed with 20 mm of preload and then compressed a further 10 mm has not stored $\tfrac{1}{2}k(0.010)^2$. It has gone from 20 mm to 30 mm of deflection, so the energy added is $\tfrac{1}{2}k(0.030^2 - 0.020^2)$ — five times as much. Preloaded springs are everywhere in mechanisms, and this catches people every time. Two lesser traps: spring rates are quoted in N/m and in N/mm, and mixing them is a factor of a thousand in $k$; and the formula holds only inside the elastic limit, so a spring stretched until it takes a permanent set has absorbed energy this equation cannot account for, because some of it went into deforming the metal rather than into recoverable storage.

Power is the rate at which work is done, $P = W/t$. It answers a different question from work: not "how much energy did this take?" but "how fast was it delivered?" Two apprentices carrying identical toolboxes up the same stairs do exactly the same work against gravity, and the one who takes the stairs two at a time develops more power. One watt is one joule per second, which makes the watt a small unit — a person working steadily manages perhaps 75 W, and a fit cyclist holds around 250 W for an hour.

Take 20 kg of tools hauled 12 m up a ladder. The work is $mgh = 20 \times 9.80665 \times 12 \approx 2354$ J regardless of how it is done. Take 40 s over it and $P = 2354/40 \approx 59$ W. Rush it in 15 s and the same job demands 157 W. The energy bill is identical; only the rate has changed, and it is the rate that decides whether a motor is big enough.

James Watt coined horsepower in the 1780s as a sales tool. He needed to tell mill owners how many horses one of his engines would replace, measured a horse turning a mill wheel, and settled on 33 000 foot-pounds per minute — about 745.7 W. It was a marketing unit before it was an engineering one, and it has outlived the argument it was built to win. The same rate appears in two other forms on this site: $P = Fv$ when the work is a force moving something along, and $P = \tau\omega$ when it is a torque turning a shaft.

The commonest error is treating power as though it were energy. A kilowatt-hour is not a unit of power — it is a power multiplied by a time, so it is an energy, equal to 3.6 MJ. A 100 W bulb does not consume "100 watts per hour"; it consumes 100 watts, which over an hour amounts to 0.1 kWh. The phrase "watts per hour" is almost always a symptom that the two ideas have been mixed. Second, watch which horsepower a figure is quoted in: mechanical horsepower is 745.7 W, but metric horsepower — PS, cv, ch — is 735.5 W, and European engine ratings are usually the latter, a 1.4% difference that quietly walks into converted specifications. Third, $t$ must be the time over which the work was actually done, not the length of the shift; a hoist that lifts for 20 s and then sits idle for 10 minutes has a duty cycle, and averaging over the whole ten minutes describes the energy consumption honestly but badly understates the motor the job needs.

Divide both sides of $W = Fd$ by time and the $d/t$ turns into velocity, leaving $P = Fv$. It is the same statement as $P = W/t$, rewritten for the common case where a steady force is pushing something along at a steady speed — a car holding a cruise, a conveyor dragging material, a tug pulling a barge. The virtue of this form is that it needs no clock and no distance, only what is happening right now.

A car on the highway is fighting drag and rolling resistance. If those total 600 N at 30 m/s — about 108 km/h — the engine must deliver $P = 600 \times 30 = 18\ 000$ W, or 18 kW, roughly 24 hp, purely to keep the needle where it is. Nothing is accelerating and no height is being gained; that power is going straight into stirring air and warming tyres.

The relation has an unpleasant surprise buried in it for anyone chasing top speed. Aerodynamic drag rises with the square of speed, so the power needed to overcome it rises with the cube. Doubling highway speed takes roughly eight times the power, which is why an engine of twice the output buys only about a 26% higher top speed, and why fuel consumption climbs so steeply above about 90 km/h. The rotational version, $P = \tau\omega$, is the same equation on a shaft and is what a dyno chart is plotting.

The conceptual trap is expecting power to feel like force. At a fixed power the two trade off exactly: a truck in low gear applies enormous force at a crawl, and the same engine in top gear applies a small force at speed, with the identical power in both cases. That is the whole job of a gearbox. It also means the equation misbehaves at the ends — at $v$ near zero it would demand infinite force for any finite power, and what actually limits you there is traction and the clutch, not the engine. Two mechanical cautions as well. $F$ must be the component of force along the motion; for a force at an angle, take $F\cos\theta$ first, exactly as in the work formula. And this is instantaneous power unless both $F$ and $v$ hold steady — during acceleration both are changing, and the average power over the run is not $F_{\text{avg}}v_{\text{avg}}$.

Momentum is mass in motion: multiply how much material is moving by how fast it moves. A 40,000 kg truck creeping along at 1 m/s carries 40,000 kg·m/s of momentum, while a 100 kg bicycle-and-rider at 10 m/s carries only 1,000 — which is why the slow truck is far harder to stop. Newton called momentum the "quantity of motion" and framed his second law around its rate of change, not around ma.

Momentum's real power is conservation: in any collision or explosion with no outside force, the total momentum before equals the total after. That single principle lets crash investigators reconstruct vehicle speeds from wreckage and lets rocket engineers plan burns — the exhaust thrown backward and the craft pushed forward always balance. In this one-dimensional form, direction simply rides along as the sign of v.

Impulse is force multiplied by how long that force acts, $J = F\,\Delta t$, and its whole significance is that it equals the change in momentum produced. That is not a separate law — it is Newton's second law in its original form. Newton wrote that force is the rate of change of momentum, $F = \Delta p/\Delta t$, and multiplying both sides by $\Delta t$ gives this. The units confirm it: a newton-second and a kilogram-metre per second are the same thing.

Here is the calculation that motivates every piece of passive safety equipment in a car. A 70 kg occupant travelling at 15 m/s must lose $p = 70 \times 15 = 1050$ kg·m/s of momentum in a frontal impact. That number is fixed by the crash; nothing can reduce it. If the body stops against a rigid dashboard in 0.1 s, the average force is $F = 1050/0.1 = 10\,500$ N. Stretch the same stop to 0.5 s using a crumple zone, a seatbelt that pays out under load, and an airbag, and the force falls to 2100 N. The impulse is identical in both cases. Only the time changed, and the time is the only variable an engineer gets to design.

The same trade-off is why you bend your knees on landing, why a boxer rolls with a punch, why gymnasts land on foam, and why a fall onto concrete is dangerous and the identical fall onto a mattress is not. In the other direction it is how rockets are specified: a model rocket motor's class is its total impulse in newton-seconds, because that is what determines the velocity change it can give a vehicle, regardless of whether it burns fiercely for a moment or gently for several seconds.

Impulse is a vector, and the sign flip on a bounce is the classic error. A 0.15 kg ball thrown at a wall at 20 m/s and rebounding at 20 m/s has not undergone a momentum change of zero, and not one of 3 kg·m/s either. Its momentum went from $+3$ to $-3$, a change of 6 kg·m/s — twice what stopping it dead would have required. A bouncing collision always demands more impulse than a catching one, which is why a bouncy object hits harder than a soft one of the same mass and speed. The second caution is that $F$ here is the average force over the contact. A real impact pulse is peaked, often two to three times the average at its maximum, so a component designed only against the average will be under-rated for the moment that actually breaks it.

With no external force acting, the total momentum of two colliding bodies before the impact equals the total after — always, whether they bounce, stick, or shatter. Kinetic energy is not so obliging; it survives only in a perfectly elastic collision. The principle was nailed down in 1668 when the Royal Society set the collision problem as a challenge and John Wallis, Christopher Wren and Christiaan Huygens independently sent in the answer.

Signs are everything in one dimension: pick a positive direction and stick with it, so a body moving the other way carries a negative velocity. A 2 kg cart at 5 m/s striking a stationary 3 kg cart and slowing to 1 m/s leaves the second cart at (10 + 0 − 2) ⁄ 3 ≈ 2.67 m/s. Because momentum is conserved in every collision, crash investigators use exactly this equation with skid-mark evidence to back out pre-impact speeds, and it works equally well for recoil: the rifle and the bullet start with zero total momentum and must end with zero.

When two bodies lock together on impact they share one final velocity, and momentum conservation hands it to you directly: the combined momentum divided by the combined mass. A 1000 kg car at 20 m/s rear-ending a stationary 1500 kg van and tangling with it leaves the wreck moving at 20000 ⁄ 2500 = 8 m/s. Kinetic energy, by contrast, is not conserved — here 200 kJ goes in and only 80 kJ comes out, the missing 120 kJ spent deforming metal, which is exactly what crumple zones are designed to do.

The classic application is the ballistic pendulum, devised by Benjamin Robins in 1742: fire a bullet into a hanging block, measure how high the block swings, and work backwards through this equation to get the muzzle velocity — the first accurate method of measuring how fast a bullet flies. Perfectly inelastic collisions dissipate the maximum energy any collision can while still conserving momentum, which is why "sticking together" is the worst case for occupant survival and the best case for a crash-test energy budget.

Newton's experimental law of restitution says the relative speed after a collision is a fixed fraction of the relative speed before: e = separation ⁄ approach. e = 1 is perfectly elastic, e = 0 is perfectly inelastic (the bodies move off together), and everything real lands in between. Newton reported measurements in the Principia itself, swinging pendulum balls together and recording roughly 5/9 for glass and 15/16 for tightly wound wool.

Sports bodies now legislate the number. A regulation basketball must return 1.2–1.4 m when dropped from 1.8 m onto hardwood, an e of about 0.85; a tennis ball tested per ITF rules comes in near 0.75; golf drivers are capped by a related "COR" limit of 0.83 to keep drives in the stadium. Two traps: e is not a property of one object but of the pair of surfaces, and it drops measurably at high impact speeds, which is why the test conditions are specified so precisely.

Angular velocity is the rotational twin of ordinary speed. Where speed counts metres covered per second, angular velocity counts angle swept per second: $\omega = \theta/t$. The reason it deserves its own quantity is that every point on a rigid spinning body shares the same $\omega$ while having a completely different linear speed — the hub of a wheel and the tread of the tyre turn through the same angle in the same second, which is exactly what makes the body rigid.

A record turntable at 33⅓ rpm turns through 33.33 revolutions each minute. Convert properly: $33.33 \times 2\pi = 209.4$ rad per minute, divided by 60 s gives $\omega \approx 3.49$ rad/s. A four-pole induction motor on 60 Hz mains runs near 1750 rpm, which is about 183 rad/s. A minute hand on a clock manages $2\pi/3600 \approx 0.00175$ rad/s.

Radians are the natural currency here, and it is worth knowing why rather than just accepting it. A radian is defined as the angle that subtends an arc equal to the radius, so arc length is $s = r\theta$ with no constant attached — but only when $\theta$ is in radians. Everything downstream inherits that cleanliness: $v = \omega r$, $a_c = \omega^2 r$, $\tau = I\alpha$, $P = \tau\omega$ all come out with no conversion factor at all. Use degrees and every one of them needs a $\pi/180$ glued on, which is precisely the kind of thing that gets forgotten.

Feeding rev/min into an equation that expects rad/s is the single most common error in rotational mechanics, and it is worth memorising the conversion once: multiply rpm by $2\pi/60$, which is 0.10472. Do it the other way and multiply by 9.5493. The error is a factor of 9.55, and because it is not a round number it does not announce itself the way a factor of ten would — the answer just comes out wrong and plausible. It gets worse downstream: anything with $\omega^2$ in it, like rotational kinetic energy or centripetal acceleration, is then wrong by a factor of 91. A related slip is treating one revolution as 360 in this formula; in radians one revolution is $2\pi \approx 6.283$, and mixing the two is a factor of 57.3. This page's unit selectors will do the conversion for you, but the habit of checking which unit a supplier's data sheet used is worth more than any calculator.

Angular acceleration is how quickly a rotation rate is changing, $\alpha = (\omega - \omega_0)/t$. It is the exact rotational mirror of $a = (v - v_0)/t$, and that mirroring is not a coincidence or a mnemonic — the entire structure of linear kinematics carries across intact, with $\theta$ in place of $x$, $\omega$ in place of $v$, and $\alpha$ in place of $a$. Every uniform-acceleration technique you already know works unchanged in the rotational world.

A hard-disk platter spinning up from rest to 7200 rpm in 4 s: convert first, $\omega = 7200 \times 0.10472 \approx 754$ rad/s, so $\alpha = (754 - 0)/4 \approx 188$ rad/s². A large industrial fan coasting down from 1200 rpm (126 rad/s) to rest over 90 s has $\alpha = -1.4$ rad/s², and the negative sign is the whole of what distinguishes a spin-down from a spin-up.

What makes $\alpha$ worth computing is that it is the bridge to torque. Once you know the angular acceleration a duty cycle demands, $\tau = I\alpha$ tells you the torque the drive must supply, and that is how motors get sized for anything that has to start and stop repeatedly — a centrifuge, a machine-tool spindle, a conveyor, a robot joint. The starting torque is very often larger than the running torque, and this equation is where that requirement first appears.

The unit mistake here has an extra layer to it. As always, $\omega$ must be in rad/s before it goes anywhere near the formula, but the output also has to be read correctly: $\alpha$ comes out in rad/s², and a figure quoted as "rpm per second" is not the same quantity — it needs the same 0.10472 factor applied. A drive rated to accelerate at "500 rpm/s" is delivering 52.4 rad/s². The second issue is signs. Pick a positive direction of rotation and keep it: braking a shaft turning positively gives a negative $\alpha$, and the calculator's own answer will tell you if you have them mixed, because a spin-down entered with both velocities positive and a positive $\alpha$ is describing something that cannot happen. Finally, this is the average angular acceleration over the interval. A real motor's torque varies enormously with speed, so the instantaneous $\alpha$ during a start is nothing like constant; the average is the right number for a duty-cycle estimate and the wrong one for a stress calculation at the moment of highest load.

Every point on a rigid spinning body shares the same angular velocity, but the farther a point sits from the axis, the faster it actually travels through space. The relationship is exactly linear in the radius: $v = \omega r$. It comes straight from the definition of the radian — arc length is $s = r\theta$, so dividing both sides by time gives $v = r\omega$, with no constant to remember. This is why the outer horses on a carousel feel so much livelier than the inner ones, and why the difference between the two is precisely the ratio of their radii.

A 300 mm circular saw blade at 3600 rpm: convert first, $\omega = 3600 \times 0.10472 = 377$ rad/s, and the radius is 0.15 m. The rim speed is $v = 377 \times 0.15 \approx 57$ m/s, over 200 km/h. That number is not academic — it is the figure a blade's maximum-rpm rating is derived from, and it is why fitting a larger blade to a faster saw is dangerous even when it physically fits.

The relation is the whole basis of gearing and belt drives. Two pulleys joined by a belt share a common belt speed $v$, so $\omega_1 r_1 = \omega_2 r_2$ and the speed ratio is the inverse of the radius ratio. Meshing gears share a common pitch-line velocity and behave identically. In machining it appears as surface speed, the quantity that actually governs tool wear, which is why a lathe's constant-surface-speed mode continuously changes the spindle rpm as the tool moves across a face.

The measurement that goes wrong is which radius. Component data is quoted in diameters far more often than radii — a "300 mm blade", a "600 mm sheave", a "50 mm bar" — and entering the diameter doubles the answer. Worse, on a V-belt drive the correct figure is the pitch diameter where the belt actually sits in the groove, not the outside diameter of the sheave, and the two can differ by 10 mm or more on a small pulley; on a gear it is the pitch diameter, not the tip diameter. A speed ratio computed from outside diameters is close enough to look right and wrong enough to matter. The other error is the familiar one of feeding rpm straight in where the equation wants rad/s, which multiplies the answer by 9.55 — a 3600 rpm blade would appear to have a rim speed of 540 m/s, which is faster than sound and ought to prompt a second look.

Anything that spins or orbits sweeps exactly $2\pi$ radians in one complete cycle, so the angular velocity and the period are locked together: $\omega = 2\pi/T$. This is the $\omega = \theta/t$ relation with one full turn substituted in, and it is the form to reach for whenever what you actually know is how long one revolution takes — which is most of the time for orbits, pendulums, alternating current and anything described by its cycle time.

The Earth turns once on its axis in one sidereal day, 23 h 56 min 4 s, which is 86 164 s. So $\omega = 2\pi/86\,164 \approx 7.292\times10^{-5}$ rad/s. That number is the input to every Coriolis calculation, every geostationary orbit computation, and the design of every inertial navigation system. At the other end of the scale, a crankshaft at 3000 rpm has a period of 0.02 s and $\omega \approx 314$ rad/s.

Because $T$ and frequency $f$ are reciprocals, this can equally be written $\omega = 2\pi f$, which is the form electrical engineering uses constantly: the 60 Hz mains supply corresponds to $\omega = 377$ rad/s, and that 377 turns up in every reactance calculation. The same relation ties the period of a pendulum, a mass on a spring, or an LC circuit to the angular frequency that appears inside the sine function describing its motion.

The trap here is that frequency and angular frequency are both often called "frequency" and they differ by a factor of $2\pi$. Mains power is 60 Hz and 377 rad/s; those are the same physical thing with different units, and substituting 60 where 377 belongs is wrong by 6.283. Watch the symbols, $f$ in hertz, $\omega$ in radians per second — and be suspicious of any source that writes "frequency" without saying which. The second point is subtler and specific to the Earth example. The day you live by is the solar day, 86 400 s, the time for the Sun to return to the same place in the sky. The Earth's actual rotation period is the sidereal day, about four minutes shorter, because the planet has moved along its orbit and has to turn a little extra to face the Sun again. Using 86 400 gives $7.272\times10^{-5}$ rad/s, a 0.3% error — negligible for a rough estimate and quite unacceptable for satellite work.

This is nothing more than speed equals distance over time, applied to a circle. One complete trip round covers a circumference of $2\pi r$, and it takes one period $T$, so the speed is $v = 2\pi r/T$. Everything difficult about circular motion lives in the direction of the velocity, which is changing constantly; the magnitude is this piece of grade-school arithmetic and nothing more.

The International Space Station orbits about 410 km above the surface, and Earth's mean radius is 6371 km, so its orbital radius is $r \approx 6781$ km. It completes one orbit in roughly 93 minutes, which is 5580 s. That gives $v = 2\pi \times 6.781\times10^6 / 5580 \approx 7630$ m/s, or 7.6 km/s — about 27 500 km/h, and the reason its crew see sixteen sunrises a day.

Combine it with the two neighbouring pages and a lot falls out. Since $\omega = 2\pi/T$, this equation is exactly $v = \omega r$ with the period substituted in. Put it into the centripetal acceleration $a_c = v^2/r$ and you get $a_c = 4\pi^2 r/T^2$; set that equal to the gravitational acceleration $GM/r^2$ and rearrange, and $T^2 \propto r^3$ drops out — Kepler's third law, derived in three lines from a circumference and Newton's law of gravitation.

The radius is measured from the centre of rotation, and for orbits that means from the centre of the Earth, not from the ground. Using the ISS's 410 km altitude as $r$ instead of its 6781 km orbital radius understates the speed by a factor of about four and is the most common way this calculation goes wrong. The same principle applies on a smaller scale: for a point on a flywheel, $r$ runs from the shaft axis, and for a car on a banked track it runs to the centre of the curve, not to the inside edge of the road. Two more. $T$ is the time for one full revolution — a rotation rate given in rev/min has to be inverted first, $T = 60/N$ seconds, so 1800 rpm is a period of 0.0333 s, not 1800 of anything. And this describes uniform circular motion. A real planetary orbit is an ellipse on which the speed varies continuously, fastest at perihelion and slowest at aphelion; $2\pi r/T$ with a mean radius gives an average, not the speed at any particular moment.

This is $F = ma$ for things that spin. Torque takes the role of force, angular acceleration takes the role of linear acceleration, and moment of inertia takes the role of mass: $\tau = I\alpha$. The correspondence is exact, and it is the reason rotational dynamics feels familiar once you accept that $I$ is doing a job that mass alone cannot.

What makes $I$ different from mass is that it depends not only on how much material a body contains but on where that material sits relative to the axis. Mass far from the axis resists being spun up far more than the same mass close in, because the far-out mass has to be accelerated to a much higher linear speed to achieve the same angular one. Apply 10 N·m to a flywheel with $I = 2$ kg·m² and it accelerates at $\alpha = 10/2 = 5$ rad/s². Build a flywheel of the same mass but twice the radius and $I$ becomes four times as large, so the same torque manages only 1.25 rad/s².

Run backwards, this is how drives get sized. Decide the angular acceleration a duty cycle requires, look up or compute the moment of inertia of everything the motor has to turn, and the product is the torque the motor must produce over and above whatever the load already demands. It is also the equation behind the reflected-inertia calculation in a geared drive, where a gearbox of ratio $n$ reduces the load inertia seen at the motor by $n^2$.

The mistake that invalidates the whole calculation is using a moment of inertia quoted about the wrong axis. Unlike mass, $I$ is not a property of a body — it is a property of a body and a chosen axis. A rod's moment of inertia about its centre is $\tfrac{1}{12}mL^2$ and about one end it is $\tfrac{1}{3}mL^2$, four times larger, and the parallel-axis theorem $I = I_{cm} + md^2$ is what converts between them. Taking a handbook value without checking which axis it refers to is the standard way to get a plausible answer that is wrong by a factor of several. The second point is the same one that $F = ma$ has: $\tau$ is the net torque. Bearing friction, windage, and whatever load is attached all subtract from the driving torque, and it is only what remains that produces $\alpha$. A motor rated at 10 N·m driving through a gearbox that absorbs 2 N·m accelerates the load as though 8 N·m were applied — which is why a rig that models beautifully can still fail to reach speed in the time the specification allows.

A solid disk has its mass spread evenly from the axis out to the rim, and integrating $r^2$ over that distribution gives exactly half of $mr^2$: $I = \tfrac{1}{2}mr^2$. The coefficient is a pure statement about geometry. Compare the two extremes and it makes sense: a thin hoop with all its mass at the rim has $I = mr^2$, and a mass concentrated at the very centre would have $I = 0$. A uniform disk lands halfway between, though not for the reason the word "halfway" suggests — most of a disk's area is in its outer half, and the two effects happen to cancel to a clean $\tfrac{1}{2}$.

A steel grinding wheel 300 mm across and weighing 8 kg has $r = 0.15$ m, so $I = 0.5 \times 8 \times 0.15^2 = 0.09$ kg·m². At 3600 rpm, $\omega = 377$ rad/s, it stores $\tfrac{1}{2} \times 0.09 \times 377^2 \approx 6.4$ kJ — enough that it will keep turning for a long time after the power is cut, and enough to do serious harm if it lets go.

Nothing in the formula mentions thickness, and that is not an omission: a solid cylinder of any length has the same $\tfrac{1}{2}mr^2$ about its central axis, because stacking disks along the axis does not move any mass closer to or further from it. So this one expression covers a coin, a flywheel, a roller and a shaft alike. It is the value behind the stored energy in flywheels, the spin-up time of hard-disk platters, and the rolling behaviour of any wheel treated as a uniform disk.

This is the moment of inertia about the central axis — the one the disk naturally spins about — and only that one. Flip the disk so it turns about a diameter instead, like a coin rolled on edge and spun about a horizontal line through its centre, and the correct value is $\tfrac{1}{4}mr^2$, half as much. The two get confused because the same disk is involved. The second error is applying $\tfrac{1}{2}mr^2$ to something that is not solid. A tube, a pipe, or a rim-heavy flywheel with a light web has more of its mass out near the radius, and its moment of inertia is $\tfrac{1}{2}m(r_1^2 + r_2^2)$ using both the inner and outer radii — for a thin-walled tube that approaches $mr^2$, twice the solid-disk figure. Treating a fabricated flywheel as a solid disk understates its inertia badly, and since flywheels are deliberately built rim-heavy, it understates exactly the case where you were relying on the number. And, as ever, $r$ is a radius: a "300 mm wheel" gives 0.15, and using 0.3 overstates $I$ fourfold.

This is the building block from which every other moment of inertia is assembled: a compact mass $m$ at distance $r$ from the axis contributes $I = mr^2$. It comes directly from the energy. That mass moves at $v = \omega r$, so its kinetic energy is $\tfrac{1}{2}m(\omega r)^2 = \tfrac{1}{2}(mr^2)\omega^2$, and the bracketed quantity is what has to be called $I$ if $\tfrac{1}{2}I\omega^2$ is going to work. The $r^2$ is not chosen, it is inherited from the square in kinetic energy.

A 2 kg mass whirling on a 1.5 m tether has $I = 2 \times 1.5^2 = 4.5$ kg·m². Move it to 3 m and $I$ becomes 18 kg·m² — the mass has not changed, the distance has doubled, and the resistance to being spun up has quadrupled. That square is the most important fact in rotational mechanics, and it is why a diver tucks, why a tightrope walker's pole is long rather than heavy, and why putting material at the rim of a flywheel is worth far more than putting it near the hub.

Every extended body's moment of inertia is this formula summed over all its mass, $I = \sum m_i r_i^2$, or integrated for a continuous body. That integration is exactly where the coefficients on the other pages come from: the $\tfrac{1}{2}$ of a solid disk, the $\tfrac{2}{5}$ of a solid sphere, the $\tfrac{1}{12}$ of a rod about its centre. Each is a statement about how the mass of that shape is distributed relative to the axis, and nothing more.

The $r$ is the perpendicular distance to the axis — a line — not the distance to a point. For a mass sitting 3 m along the axis and 0.5 m out from it, the moment of inertia about that axis is $m \times 0.5^2$, not $m \times 3.04^2$; the along-axis position is irrelevant, because that part of the displacement is not being swung round. This trips people up as soon as the geometry stops being flat. The related habit worth building is always naming the axis before quoting an $I$: the same point mass has a completely different moment of inertia about a different line, and unlike mass, which is one number for the object, $I$ is one number per axis. When several masses share an axis, their moments of inertia simply add, which makes building up a real assembly a matter of arithmetic rather than calculus — and the parallel-axis theorem, $I = I_{cm} + md^2$, is just this formula treating a whole body as though it were a point at its own centre of mass.

An object going round a circle at perfectly constant speed is nevertheless accelerating, and that sentence is the first thing to make peace with. Acceleration is the rate of change of velocity, and velocity is a vector with a direction as well as a size. Something moving in a circle is having its direction changed continuously, so its velocity is changing continuously, so it is accelerating — even though a speedometer strapped to it would never move. The acceleration points at the centre of the circle, and its size is $a_c = v^2/r$.

A 1200 kg car rounding a 50 m curve at 20 m/s — 72 km/h — experiences $a_c = 20^2/50 = 8$ m/s², about 0.82 g. That is close to the limit of what a good tyre on dry pavement can supply, which is why that corner at that speed feels like it is asking a real question. Take the same corner at 30 m/s and the demand rises to 18 m/s², about 1.8 g, and no ordinary road tyre will hold it.

The derivation is short enough to be worth carrying. Over a small time $\Delta t$, the position vector sweeps through an angle $\Delta\theta = v\Delta t/r$. The velocity vector, always at right angles to the position vector, must rotate through exactly the same angle, and rotating a vector of length $v$ through a small angle changes it by $|\Delta v| = v\Delta\theta$. Put the two together: $|\Delta v| = v^2\Delta t/r$, so $a = v^2/r$. Christiaan Huygens published this result in 1673, and it is what let Newton check the inverse-square law against the Moon's orbit.

There is no outward force, and this is the single most persistent misconception in mechanics. In the ground frame nothing pushes you outward in a turning car. What happens is that your body would continue in a straight line, the car turns underneath you, and the door pushes you inward. The sensation of being flung out is your inertia, not a force. "Centrifugal force" is a bookkeeping term that appears only when you insist on doing the physics in the rotating frame, where it is added artificially so Newton's laws balance. It is a real effect and a useful device; it is not a force in an inertial frame, and there is no third-law partner to it. A more mundane error costs just as much: the $r$ is a radius, not a diameter. A component described as "600 mm diameter" has $r = 0.3$ m, and entering 0.6 halves the answer. Last, this is only the component of acceleration perpendicular to the motion. If the object is also speeding up or slowing down, there is a tangential component too, and the total acceleration is the vector sum of the two.