Normal Force

The normal force is having a renaissance in robotics. A walking robot is essentially a machine for managing $N$: every footstep must land with enough normal force for friction to hold ($f \le \mu N$), but not so abruptly that the impact spike destabilizes the gait. The balance controllers in today's humanoid robots estimate the normal force at each foot hundreds of times per second — from force-torque sensors in the ankles or from motor currents — and reposition the body so the zero-moment point, the effective location of the resultant normal force, stays inside the foot's support polygon. When you watch a 2026-generation humanoid recover from a shove, you are watching real-time normal-force bookkeeping. Grasping is the same story at the fingertips: warehouse manipulators squeeze each object with just enough normal force that friction beats gravity, but not enough to crush — the two-inequality problem ($\mu N \ge mg$, $N \le N_{\mathrm{damage}}$) solved over and over, package by package. Vehicle engineering plays the opposite game: manufacturing normal force beyond $mg$. A Formula 1 car's wings and floor generate aerodynamic downforce exceeding the car's own weight above roughly $180\ \mathrm{km/h}$ — so $N > 2mg$ — which multiplies the friction budget and lets the car corner at accelerations no street tire could dream of. Same physics as the sandbag in a pickup bed in winter, executed at 300 km/h.

Press your palm on a table, gently, then harder. The table answers every push with an exactly matching push back — soft when you are soft, firm when you are firm. Nothing in the wood is "deciding" anything. Under your hand, the top layers of atoms are shoved a few nanometres closer to the layers below, and the interatomic bonds — stiff electromagnetic springs — compress and push back. The harder you press, the more they compress, the harder they resist. The table is a mattress of springs so stiff that the sag is invisible, and the normal force is the springs' collective reply.

Three habits of the normal force are worth internalizing before any algebra. First, it is self-adjusting. Put a heavier book on the table and $N$ grows; press down on the book and $N$ grows again; tilt the table and $N$ shrinks. The atomic springs compress by exactly the amount that stops further penetration, so $N$ always takes the value the situation demands. Second, it is unilateral: compression can be resisted, but a table cannot grab your hand and pull it down. The normal force can only push away from the surface, so $N \ge 0$ — and the moment the physics would need $N < 0$, contact simply ends. Third, it is perpendicular by definition ("normal" is the geometer's word for perpendicular): whatever direction the surface faces — floor, ramp, or wall — $\mathbf{N}$ points along the outward perpendicular. A wall's normal force is horizontal; an incline's is tilted. The tangential part of the contact force has its own name, friction, and its own article. Because $N$ answers to the situation rather than to a formula, the question you ask to find it is always the same: what perpendicular force must the surface supply so that the object's acceleration perpendicular to the surface matches what the surface itself is doing? For a block on a stationary floor that perpendicular acceleration is zero, and $N$ balances gravity exactly. For the elevator passenger it is not zero, and $N$ must differ from $mg$ — which is precisely what your stomach and the scale both register.

Everything about the normal force fits in two statements — a direction and an inequality. They look almost too slight to be called a law, and that is the point: they are a constraint, not a force law.

where: $\mathbf{N}$ is the force the surface exerts on the object and $N$ its magnitude. The perpendicularity defines which part of the contact force we are naming (the tangential part is friction); the inequality encodes "surfaces push, never pull." Everything else — the actual number — comes from Newton's second law applied to the geometry at hand. Start with the simplest geometry: a block of mass $m$ resting on a horizontal floor, nothing else touching it. Vertically there are exactly two forces, gravity $mg$ down and $N$ up, and the block's vertical acceleration is zero. Newton's second law along the vertical reads $N - mg = 0$, so

where: $g \approx 9.81\ \mathrm{m/s^2}$ near Earth's surface. Read this as a result, not a definition: $N$ equals $mg$ here only because zero vertical acceleration and no other vertical forces happened to demand it. A $2\ \mathrm{kg}$ book gets $N = 19.6\ \mathrm{N}$; push down on the book with an extra $10\ \mathrm{N}$ and the floor instantly answers with $N = 29.6\ \mathrm{N}$. The equality with $mg$ is a coincidence of the simplest case — treating it as a law of nature is the single most common error in introductory mechanics. Tilt the surface and the geometry does the talking. On a frictionless ramp at angle $\theta$, gravity still points straight down, but "perpendicular to the surface" now means tilted by $\theta$ from the vertical. Only the perpendicular component of gravity, $mg\cos\theta$, must be balanced for the block to stay on the ramp:

where: $\theta$ is the incline angle above horizontal. The parallel component $mg\sin\theta$ never enters $N$ — it drives the block down the slope (or is held by friction). Check the dial at its ends: $\theta = 0$ returns the flat floor $N = mg$; $\theta = 90°$ gives $N = 0$, a block beside a vertical wall that the wall does not need to support at all.

The normal force also holds the purse strings of friction. Both static and kinetic friction scale with how hard the surfaces are pressed together, and $N$ is that pressing:

where: $\mu_s$ and $\mu_k$ are the static and kinetic friction coefficients. Every trick for changing grip — race-car aerodynamic downforce, sandbags in a pickup truck in winter, leaning into a sprint start — is really a scheme for changing $N$. And when a car crests a hill and $N$ momentarily drops (as we derive below), the friction budget drops with it: crests are where tires lose grip first. When the support itself accelerates, the constraint changes from "zero perpendicular acceleration" to "match the surface's acceleration," and $N$ moves away from $mg$. For a person of mass $m$ in an elevator accelerating upward at $a_y$:

where: $a_y > 0$ means the car accelerates upward (speeding up on the way up, or braking on the way down — only the acceleration's direction matters, not the velocity's). This $N$ is what a scale reads and what your body interprets as weight, which is why it is called the apparent weight. Set $a_y = -g$ and $N = 0$: free fall, the floor no longer pressing, the sensation astronauts live with continuously. Finally, curved paths. At the crest of a hill of curvature radius $R$, the center of the circle is below the car, so the net downward force must supply the centripetal $mv^2/R$:

where: $v$ is the speed at the crest. As $v$ grows, $N$ shrinks — the road needs to push less and less because gravity is increasingly "used up" steering the car around the curve. At $v = \sqrt{gR}$ the road pushes with nothing at all, and beyond that the car is airborne: the unilateral constraint $N \ge 0$ has run out of room.

Which equation answers which question

Here is a small heresy with large consequences: nothing in your body senses gravity. Gravity pulls every atom of you equally, producing no internal compression, no signal. What you feel as weight is the normal force — the floor pressing your soles, your chair pressing your back — and the internal stresses it sets up as it holds you against gravity's pull. A bathroom scale is just a calibrated spring reporting how hard its surface pushes you: it reads $N$, never $mg$. The two agree only when you are unaccelerated. Let the floor accelerate and they part company — that is equation (5), and it deserves a careful derivation because it explains everything from elevator stomach-drops to orbital weightlessness.

The inequality in equation (1) is not decoration; it is a prediction machine. Because a surface can only push, any situation whose algebra demands a negative $N$ is telling you the contact assumption itself has failed: the object leaves the surface, and from that instant it is a projectile with $N = 0$. The boundary case $N = 0$ marks the exact speed, angle, or acceleration at which contact is lost — and finding it is one of the most common exam maneuvers in mechanics.

1 · The elevator commute (confidence builder)

A $75\ \mathrm{kg}$ person stands on a scale in an elevator. Find the scale reading (a) at rest, (b) accelerating upward at $1.2\ \mathrm{m/s^2}$, (c) accelerating downward at $1.2\ \mathrm{m/s^2}$, and (d) cruising upward at a steady $3\ \mathrm{m/s}$. Take $g = 9.81\ \mathrm{m/s^2}$.

Step 1. At rest, $a_y = 0$: $N = mg = 75 \times 9.81 = \boxed{736\ \mathrm{N}}$. This is the person's ordinary weight — the calibration point.

Step 2. Accelerating up, $a_y = +1.2\ \mathrm{m/s^2}$: $N = m(g + a_y) = 75 \times (9.81 + 1.2) = 75 \times 11.01 = \boxed{826\ \mathrm{N}}$ — about $12\%$ heavier. This happens both when starting upward and when braking at the bottom of a descent: same acceleration direction, same reading.

Step 3. Accelerating down, $a_y = -1.2\ \mathrm{m/s^2}$: $N = 75 \times (9.81 - 1.2) = 75 \times 8.61 = \boxed{646\ \mathrm{N}}$ — the floaty feeling as the car starts to descend.

Step 4. Cruising at constant $3\ \mathrm{m/s}$: $a_y = 0$, so $N = \boxed{736\ \mathrm{N}}$, identical to standing still. The scale senses acceleration, never velocity — mid-ride, with your eyes closed, a smooth elevator is indistinguishable from a stationary room.

2 · The crest of Baldwin Street (exam level)

A $1200\ \mathrm{kg}$ car drives over a hill crest whose vertical curvature radius is $R = 40\ \mathrm{m}$. (a) Find the normal force on the car at $v = 15\ \mathrm{m/s}$. (b) Find the speed at which the car leaves the road. (c) At $15\ \mathrm{m/s}$, what fraction of its usual grip does the car have for braking at the crest, if grip is proportional to $N$?

Step 1. Radial equation at the crest (center of curvature below): $mg - N = mv^2/R$, so $N = m(g - v^2/R)$.

Step 2. At $v = 15\ \mathrm{m/s}$: $v^2/R = 225/40 = 5.625\ \mathrm{m/s^2}$. Then $N = 1200 \times (9.81 - 5.625) = 1200 \times 4.185 = \boxed{5.0\ \mathrm{kN}}$, compared with the flat-road value $mg = 11.8\ \mathrm{kN}$.

Step 3. Loss of contact at $N = 0$: $v_{\max} = \sqrt{gR} = \sqrt{9.81 \times 40} = \sqrt{392.4} = \boxed{19.8\ \mathrm{m/s} \approx 71\ \mathrm{km/h}}$. A modest highway speed — sharp crests launch cars at speeds that feel entirely ordinary.

Step 4. Grip fraction: friction available $\propto N$, and $N/mg = 4.185/9.81 = \boxed{0.43}$. At the crest this car has $43\%$ of its normal braking grip — precisely where a driver spotting an obstacle beyond the crest needs it most. This single calculation is why road-design codes set minimum crest curvature radii for each speed limit.

Texts. Kleppner & Kolenkow, An Introduction to Mechanics, ch. 2 (Newton's laws and contact forces) · Morin, Introduction to Classical Mechanics, ch. 3 (Forces and free-body diagrams) · Taylor, Classical Mechanics, ch. 1 · Goldstein, Poole & Safko, Classical Mechanics, 3rd ed., ch. 1 (constraints and their classification). The key results developed in this article are: (1) $N = mg\cos\theta$ and $a = g\sin\theta$ on a frictionless incline, with the $\theta \to 0, 90°$ limits; (2) the apparent-weight law $N = m(g + a_y)$ and its free-fall corollary $N = 0$; (3) the crest/loop threshold $v = \sqrt{gR}$ from the condition $N = 0$. The worked-example numbers ($N = 736/826/646\ \mathrm{N}$ for the $75\ \mathrm{kg}$ elevator passenger; $N = 5.0\ \mathrm{kN}$, $v_{\max} = 19.8\ \mathrm{m/s}$, grip fraction $0.43$ for the $1200\ \mathrm{kg}$ car at $R = 40\ \mathrm{m}$) are arithmetic computed from the boxed formulas with $g = 9.81\ \mathrm{m/s^2}$.

Builds on: Newton's Second Law · Newton's Third Law · Vectors, Coordinate Systems and Kinematics · Gravitational Force Near Earth's Surface. Leads to: Friction Forces · Circular Motion and Centripetal Acceleration · Newton's First Law and Inertial Frames.