Free-Body Diagrams & Problem-Solving Strategy

Every bridge you have ever crossed was certified by an act of imagination: an engineer mentally sawed the structure apart — one girder, one cable at a time — lifted each piece onto a clean page, drew every force the rest of the world exerts on it, and demanded that the piece obey Newton's laws all by itself. That ritual has a name, the free-body diagram, and it is the most valuable picture you will learn to draw in physics. Examiners know it: in most mechanics problems the marks are won or lost before any algebra happens, in the two centimetres of page where you drew (or failed to draw) the forces. Every board runs on this skill. CBSE and ISC award explicit method marks for a correct diagram; IGCSE and A-Level ask you to draw and label the forces as a question in its own right; AP Physics 1 has a free-response style where a wrong diagram forfeits the page. JEE and NEET never say "draw a free-body diagram" — they pose two- and three-block systems unsolvable without one, then plant options matching every classic diagram error: the phantom "force of motion", the mama arrow, the N=mgN = mg reflex. This module turns the skill into a five-step algorithm you run identically on every problem, plus the strategic choice experts make and beginners miss: where to cut — the whole system, or one body — and why you usually do both. It is the operating manual for everything U3 throws at you next, from the Atwood Machine to the block-on-incline family.

Key takeaways
  • A free-body diagram (FBD) isolates one body and shows every external force on it — weight, normal, tension, friction, applied pushes — and nothing else.
  • The admission test for every arrow: name the agent. "Table pushes book" qualifies. "Motion carries it forward" does not — motion is not a force, and ma⃗m\vec a is the result of forces, never an arrow.
  • Forces the body exerts on other things go on their diagrams — third-law partners never share a diagram.
  • The algorithm: isolate → inventory (touch test + gravity) → label with symbols → axes along the acceleration → ΣF=ma\Sigma F = ma per axis.
  • System vs component: cut around several bodies for the acceleration (internal forces cancel); cut around one body to expose the internal force the question asks for. Most exam problems need both passes.
  • Key idea: cut one body free, draw every force acting on it and nothing else, then read ΣF=ma\Sigma F = ma along well-chosen axes — the diagram is the complete interface between the physical situation and the mathematics.
The core insight

Cut one body free, draw every external force acting on it and nothing else, then project ΣF⃗=ma⃗\Sigma \vec F = m\vec a along axes matched to the acceleration — the free-body diagram is the complete interface between a physical situation and Newton's laws, and where you cut (whole system or single body) decides how hard the algebra will be.

Where this lives today

The free-body diagram never retired; it multiplied. Every finite-element simulation — the software that certifies aircraft wings and crash-tests cars before they exist — is millions of free-body diagrams solved at once: the structure is meshed into tiny elements, and each node's equation is ΣF=ma\Sigma F = ma for the forces its neighbours exert on it. The engineer's mental saw became an algorithm, but the physics per element is this module's recipe unchanged. Robotics runs the method in real time: a legged robot deciding where to plant a foot solves a contact problem — which normal and friction forces keep the body on its intended motion — hundreds of times per second, and game-physics engines resolve thousands of per-body ledgers every frame. Engineering licensure exams still open the way your next problem set does: draw the free-body diagram first.

The physical picture

Think of a free-body diagram as an interrogation. Grab one object, lift it out of its surroundings, and ask: who is pushing or pulling on you, and which way? Each truthful answer becomes an arrow. Earth pulls the block down — arrow. The table presses up — arrow. The rope pulls — arrow. That is the complete testimony: anything the block does to the table is the table's problem, on the table's diagram, not here. The inventory is mercifully short. Contact forces act only where something touches the body — run a finger around its boundary and you find them all: a normal force from every surface, friction along it, tension from every string, the push of a hand. Add the one long-range force of this course, gravity, from the centre. Trace the boundary, add weight, and you are provably done.

Figure
From scene to free-body diagram — the block pulled at an angle, interrogated
the scene m rope, angle θ three agents act: Earth, table, rope the free-body diagram N W = mg T at θ every arrow has a nameable agent · symbols, not numbers the tilted T is why N ≠ mg here — the vertical equation decides

Two illegal arrows account for most wrong diagrams ever drawn. The first is the "force of motion" — the feeling that a moving object must carry a forward push. It died in 1687: uniform motion needs no force, only changes do (the whole of Force, Inertia & Newton's First Law). A coasting puck has exactly two arrows, weight and normal, however fast it slides. The second is drawing ma⃗m\vec a as a force: mama is the verdict the forces produce, the right-hand side of the law — smuggling it onto the diagram double-counts the physics. Slogan: the left side of ΣF=ma\Sigma F = ma lives on the diagram; the right side is what the diagram predicts.

Figure
The two illegal arrows — a puck sliding right on smooth ice, wrong ledger vs right ledger
WRONG RIGHT v (a state, not a force) v mg N "force of motion" ✗ ma ✗ mg N …and nothing else no agent pushes a coasting puck forward — it keeps moving because nothing needs to (first law)

The five-step algorithm

Run these five moves, in order, on every problem — they are the marking scheme's checklist.

Step 1 — Isolate. Choose the body and say it aloud: "the block", not "the block and rope, sort of". Treat it as a point at its centre (fine while nothing rotates).

Step 2 — Inventory. The touch test: trace the boundary — every touch contributes a normal force, possibly friction, or a tension. Add gravity. Every candidate arrow must pass the agent test: who is exerting you?

Step 3 — Draw and label. Arrows start on the body, point in the force's direction, and carry symbols (WW, NN, TT, ff, FF) — never numbers yet. Algebra first, arithmetic last: early substitution hides errors and destroys part marks.

Step 4 — Choose axes. Any perpendicular pair works, but the algebra collapses when one axis lies along the acceleration. On an incline, tilt the axes with the slope — the constraint direction then reads a=0a = 0.

Step 5 — Project and solve. Each axis gives one scalar equation. Count unknowns; if bodies are connected, add the constraint (same aa along an inextensible string) and repeat for another body. The law the diagram feeds is Newton's second, written with the subscript that carries the whole philosophy — only forces exerted by other objects on this body enter:

ΣF⃗ext=ma⃗(1)\Sigma \vec F_{\text{ext}} = m\vec a \qquad (1)

where: F⃗ext\vec F_{\text{ext}} — each force drawn on the diagram; mm — the body's mass; a⃗\vec a — its acceleration. It is a vector statement, which is exactly why the diagram matters: the bookkeeping that guarantees no force forgotten, none invented, each with the right direction. Axis by axis it becomes

ΣFx=max,ΣFy=may(2)\Sigma F_x = ma_x, \qquad \Sigma F_y = ma_y \qquad (2)

where: the components are signed — a force at angle θ\theta to an axis contributes its cos⁡θ\cos\theta piece there and its sin⁡θ\sin\theta piece to the other axis. Each body supplies two scalar equations in 2D — that count tells you in advance how many bodies you must isolate.

The force inventory

forcedirection ruleknown or unknown?
weight W=mgW = mgvertically down, from the centreknown
normal NNperpendicular to the surface, pushing awayunknown — solve the ⊥ equation; N≥0N \ge 0
tension TTalong the string, always pullingunknown — same TT throughout an ideal string
friction ffalong the surface, opposing sliding or its tendencystatic: unknown up to a ceiling · kinetic: μN\mu N
applied push/pullas givenknown
spring forcealong the spring, toward natural lengthkxkx (Hooke's Law module)
Derivation ARunning the algorithm — block pulled at an angle

We solve the hero-figure problem in full: a block of mass mm on a smooth horizontal table, pulled by a rope with tension FF at angle θ\theta above horizontal; find the acceleration and the normal force. It is the cleanest end-to-end run of the five steps, and it carries the exam's favourite subtlety — the tilted pull changes NN, so anyone who wrote N=mgN = mg by reflex has already lost the thread. Watch the vertical equation quietly derive the normal force.

Step 1. Isolate the block. Touch test: the table touches it (normal NN; smooth, so no friction), the rope touches it (tension FF at θ\theta), gravity acts (mgmg down). Three arrows, each with an agent.

Step 2. Axes: horizontal xx (the block stays on the table, so aa is horizontal) and vertical yy. Only the pull needs decomposing: Fcos⁡θF\cos\theta along xx, Fsin⁡θF\sin\theta along yy.

Step 3. Project equation (2) on yy, where ay=0a_y = 0:

N+Fsin⁡θ−mg=0⟹N=mg−Fsin⁡θN + F\sin\theta - mg = 0 \quad\Longrightarrow\quad N = mg - F\sin\theta

Why NN dropped below mgmg: the rope's vertical slice carries part of the weight, so the table supplies less. At Fsin⁡θ=mgF\sin\theta = mg, N=0N = 0 — the verge of lift-off.

Step 4. Project on xx:

Fcos⁡θ=ma⟹a=Fcos⁡θmF\cos\theta = ma \quad\Longrightarrow\quad a = \frac{F\cos\theta}{m}

Why only the cosine drives: the vertical slice fights gravity, not inertia — only the horizontal slice accelerates the block.

Step 5. Limits: θ=0\theta = 0 gives a=F/ma = F/m, N=mgN = mg (flat drag, the reflex case); θ=90°\theta = 90° gives a=0a = 0, N=mg−FN = mg - F (a pure lift attempt). Both endpoints are right, so the decomposition is trustworthy. Verified: m=4.0 kgm = 4.0\ \mathrm{kg}, F=20 NF = 20\ \mathrm{N}, θ=30°\theta = 30°, g=9.8g = 9.8: N=29.2 NN = 29.2\ \mathrm{N}, a=4.33 m/s2a = 4.33\ \mathrm{m/s^2} — recomputed in Worked Example 1; NCERT Ch. 5 pattern.

Where to cut — system vs component

Newton's third law (its own module) makes cutting legal. When two bodies interact, a cut through their connection exposes an equal-and-opposite pair — one arrow plunging into each body. Separate diagrams show both arrows, one each. Fuse the bodies into a single system and the pair hides inside, cancelling exactly. Choosing where to cut is choosing which forces you can see:

system cut: internal forces invisiblecomponent cut: internal forces exposed(3)\text{system cut: internal forces invisible} \qquad \text{component cut: internal forces exposed} \qquad (3)

where: "internal" means both ends of the interaction live inside the cut. The system cut is the fast lane to the acceleration — one equation, no internal unknowns. The component cut is the only way to find a tension or contact force, because a force must cross the cut boundary to appear on a diagram. Skilled solvers run both, in that order.

Figure
One problem, two cuts — the system finds a, the component finds the contact force
cut 1 · the system — internal pair cancels m₁ m₂ F a = F/(m₁+m₂) only external forces survive — one line, a is out cut 2 · m₂ alone — the contact force appears m₂ N₁₂ the only horizontal force on m₂ the ledger system: F = (m₁+m₂)a component: N₁₂ = m₂a ⇒ N₁₂ = F·m₂/(m₁+m₂) the contact force feeds the mass downstream of the cut rules of the cut: · system cut → fast a, blind to internal forces · component cut → exposes the forces crossing it
Go deeper · Derivation B · 1 minTwo blocks in contact — the two-pass method
Derivation BTwo blocks in contact — the two-pass method

A hand pushes with force FF on block 1 (mass m1m_1), which pushes block 2 (m2m_2) ahead of it on a smooth floor. We want the contact force between them — invisible on any diagram of the pair, visible only when we cut between them. The strategy is the two-pass method every connected-body problem reuses: system first for the acceleration, then one strategic component cut. The moral is worth memorising — the contact force is exactly the fraction of FF needed to accelerate the mass downstream of the cut.

Step 1. System cut around both blocks (m1+m2m_1 + m_2). The pushes between them are now an internal third-law pair and cancel; the only horizontal external force is FF:

F=(m1+m2) a⟹a=Fm1+m2F = (m_1 + m_2)\,a \quad\Longrightarrow\quad a = \frac{F}{m_1 + m_2}

Why the pair cancels: the two contact forces are equal and opposite and both act inside the cut, so they vanish from the external sum — that cancellation is the entire advantage of the system view.

Step 2. Component cut around block 2 alone. The hand touches only block 1, so the single horizontal force on block 2 is the push from block 1, N12N_{12}:

N12=m2 a=m2m1+m2 FN_{12} = m_2\,a = \frac{m_2}{m_1 + m_2}\,F

Why block 2 is the smarter cut: isolating block 1 gives F−N21=m1aF - N_{21} = m_1 a — two forces instead of one. Both agree; the cleanest diagram is the body with the fewest arrows.

Step 3. Test by reversing the push: apply FF to block 2's side and the downstream mass becomes m1m_1, so the contact force is F m1/(m1+m2)F\,m_1/(m_1+m_2). Same blocks, same push, different squeeze — the cut feeds whatever mass lies beyond it. Verified: F=9 NF = 9\ \mathrm{N}, m1=2.0m_1 = 2.0, m2=1.0 kgm_2 = 1.0\ \mathrm{kg}: a=3.0 m/s2a = 3.0\ \mathrm{m/s^2}, N12=3.0 NN_{12} = 3.0\ \mathrm{N}; reversed, 6.0 N6.0\ \mathrm{N}. Ledger: 3+6=9=F3 + 6 = 9 = F ✓. Limits: m2→0m_2 \to 0 zeroes the contact force; m1→0m_1 \to 0 passes the full FF. HC Verma Ch. 5 pattern.

The same two-pass logic, plus the string constraint (one ∣a∣|a| shared along an inextensible string, one TT throughout an ideal one — Tension & Ideal Strings), is the entire method behind pulleys. The Atwood Machine & Variants archetype runs it as the worked setup, so we leave pulleys to it.

Limiting cases

caseresultmeaning
a⃗=0\vec a = 0ΣFx=0\Sigma F_x = 0, ΣFy=0\Sigma F_y = 0statics — Equilibrium of Concurrent Forces is this row grown into a module
system cutinternal pairs cancelfastest route to aa; internal forces invisible
component cutconnection forces exposedthe only way to see a tension or contact force
coasting bodyonly WW and NN on the pageuniform motion needs no force — first law
massless stringone TT along the stringa massless body obeys ΣF=0\Sigma F = 0 even accelerating
circular motionnet inward part =mv2/r= mv^2/rcentripetal is a role, never an extra arrow

Worked examples

1 · Board level — the angled pull (CBSE/ISC pattern, 3 marks)

A 4.0 kg4.0\ \mathrm{kg} block rests on a smooth horizontal table. A rope pulls it with 20 N20\ \mathrm{N} at 30°30° above the horizontal (g=9.8 m/s2g = 9.8\ \mathrm{m/s^2}). Draw the free-body diagram and find the block's acceleration and the normal force.

Step 1. FBD: weight 39.2 N39.2\ \mathrm{N} down, normal NN up, tension 20 N20\ \mathrm{N} at 30°30°. Decompose the pull: horizontal 20cos⁡30°=17.32 N20\cos 30° = 17.32\ \mathrm{N}, vertical 20sin⁡30°=10 N20\sin 30° = 10\ \mathrm{N}.

Step 2. Vertical equation (ay=0a_y = 0): N+10−39.2=0⇒N=29.2 NN + 10 - 39.2 = 0 \Rightarrow N = \boxed{29.2\ \mathrm{N}} — the rope carries a quarter of the weight.

Step 3. Horizontal equation: 17.32=4.0 a⇒a=4.33 m/s217.32 = 4.0\,a \Rightarrow a = \boxed{4.33\ \mathrm{m/s^2}}.

2 · Board level — two blocks in contact (IGCSE-to-CBSE ladder)

A 30 N30\ \mathrm{N} force pushes a 4.0 kg4.0\ \mathrm{kg} block, which pushes a 2.0 kg2.0\ \mathrm{kg} block ahead of it on a frictionless floor. Find aa and the force between the blocks.

Step 1. System cut (6.0 kg6.0\ \mathrm{kg}): a=30/6.0=5.0 m/s2a = 30/6.0 = \boxed{5.0\ \mathrm{m/s^2}}.

Step 2. Component cut around the 2.0 kg2.0\ \mathrm{kg} block — one horizontal force, the contact push: N12=2.0×5.0=10 NN_{12} = 2.0 \times 5.0 = \boxed{10\ \mathrm{N}}.

Step 3. Audit with the other block: 30−10=20=4.0×5.030 - 10 = 20 = 4.0 \times 5.0 ✓ — checking the unused equation is five-second insurance.

Step 4. Push the 2.0 kg2.0\ \mathrm{kg} block first instead and the contact force feeds the downstream 4.0 kg4.0\ \mathrm{kg}: 20 N20\ \mathrm{N} — double, for the same applied force.

3 · Competitive level — the three-block train (JEE-Main pattern, ~3 minutes)

Blocks of 1.01.0, 2.02.0 and 3.0 kg3.0\ \mathrm{kg} sit touching in a row on a smooth floor. A 12 N12\ \mathrm{N} force pushes the 1.0 kg1.0\ \mathrm{kg} block from the left. Find both contact forces.

Step 1. System cut (all three, 6.0 kg6.0\ \mathrm{kg}): a=12/6.0=2.0 m/s2a = 12/6.0 = 2.0\ \mathrm{m/s^2}.

Step 2. Contact between blocks 1 and 2: cut so the downstream mass is 2.0+3.0=5.0 kg2.0 + 3.0 = 5.0\ \mathrm{kg}. N12=5.0×2.0=10 NN_{12} = 5.0 \times 2.0 = \boxed{10\ \mathrm{N}}.

Step 3. Contact between blocks 2 and 3: downstream mass 3.0 kg3.0\ \mathrm{kg}. N23=3.0×2.0=6.0 NN_{23} = 3.0 \times 2.0 = \boxed{6.0\ \mathrm{N}}.

Step 4. Audit the middle block (two contacts): 10−6=4=2.0×2.010 - 6 = 4 = 2.0 \times 2.0 ✓. Timing note: ~3 min the first time, ~90 s once "downstream mass × a" is a reflex. The trap options are 12, 10, 6 N shuffled — a contact force matched to the wrong cut. Say it aloud: the force feeds "everything beyond this face".

4 · Competitive level — horizontal force on an incline (JEE-Main pattern, ~2.5 minutes)

A 2.0 kg2.0\ \mathrm{kg} block is held at rest on a smooth 30°30° incline by a horizontal force PP (g=10 m/s2g = 10\ \mathrm{m/s^2}). Find PP and the normal force.

Step 1. FBD: weight 20 N20\ \mathrm{N} down, normal NN perpendicular to the incline, PP horizontal. Tilt the axes; two forces now decompose — weight (1010 along, 17.3217.32 into the slope) and PP (Pcos⁡30°P\cos 30° up-slope, Psin⁡30°P\sin 30° into it).

Step 2. Along-slope equilibrium: Pcos⁡30°=20sin⁡30°P\cos 30° = 20\sin 30°, so P=20tan⁡30°=11.5 NP = 20\tan 30° = \boxed{11.5\ \mathrm{N}}.

Step 3. Perpendicular equation: N=20cos⁡30°+Psin⁡30°=17.32+5.77=23.1 NN = 20\cos 30° + P\sin 30° = 17.32 + 5.77 = \boxed{23.1\ \mathrm{N}}.

Step 4. Note what happened: N=23.1>mgcos⁡θ=17.32 NN = 23.1 > mg\cos\theta = 17.32\ \mathrm{N}, because the horizontal push presses into the slope. Anyone who wrote N=mgcos⁡θN = mg\cos\theta from memory lost this question — the diagram, not the formula sheet, decides. (Check: N=mg/cos⁡θ=23.1N = mg/\cos\theta = 23.1 ✓.) Timing note: ~2.5 min. Budget the first 30 s for the diagram and axes; the two projections are then mechanical. The planted wrong answer is always mgcos⁡θmg\cos\theta for NN.

Exam traps

The phantom arrows. Force of motion, mama, "centripetal force" as an extra agent — all fail the name-the-agent test. When asked to draw an FBD (AP, A-Level, IGCSE), examiners deduct for extra arrows, not just missing ones.

N=mgN = mg written from memory. The vertical (or perpendicular) equation derives NN; any angled pull, push or extra contact changes it — Examples 1 and 4 both punish the reflex, and a wrong NN poisons friction via μN\mu N next module.

Third-law partners on one diagram. The pair of "table pushes block up" is "block pushes table down" — a force on the table, filed on the table's diagram. Weight and normal on one block are not partners; they merely balance when a=0a = 0.

Axes fighting the acceleration. On an incline, horizontal-vertical axes put both unknowns in both equations. Tilt the axes so one equation reads a=0a = 0 and the algebra halves. Heavy simultaneous equations usually mean wrong axes.

Numbers substituted too early. Label arrows symbolically and solve; substitute in the last line. Boards award method marks for the symbolic line even when arithmetic slips — a numbers-first page loses both.

Sign chaos in multi-body problems. Give each body its own positive direction along its own motion, so the shared aa stays positive in every equation. Mixing conventions mid-problem is the classic Atwood killer.

Which gg? CBSE/ISC 9.8 m/s29.8\ \mathrm{m/s^2}; JEE/NEET as stated (usually 1010); A-Level 9.819.81; AP-1 accepts 1010. State it once at the top.

Origins · 1 minLeonhard Euler and how this was discovered
LE
Leonhard Euler
1707–1783 · Basel, St Petersburg, Berlin

Newton stated the laws; Euler made them an algorithm. His Mechanica (1736) recast the Principia's geometry as equations for a single isolated body, and in 1750 he published a "new principle of mechanics" — F=maF = ma per component, valid for any body you choose to cut free from its surroundings, rigid or fluid, whole or part. That cut principle is the license behind every free-body diagram: pick any piece of the world, replace everything you removed with forces, and the second law holds for the piece. The idea that "any body, cut free, obeys ΣF=ma\Sigma F = ma" is arguably the most-used single idea in applied mechanics.

History
1687
— Newton's Principia states the three laws — argued in Euclidean geometry, without a single force diagram.
1736
— Euler's Mechanica turns the laws into differential equations for one isolated particle.
1750
— Euler's "new principle of mechanics": F = ma per component for any body cut free from its surroundings — the cut principle behind every FBD.
1937
— Timoshenko and Young's Engineering Mechanics makes the FBD the compulsory first move of statics education worldwide.
2009
— Physics-education research (Rosengrant, Van Heuvelen, Etkina) confirms students who draw correct FBDs dramatically outperform those who skip them.

Going further

Texts. NCERT Physics Class 11, Ch. 5 · HC Verma, Concepts of Physics Vol. 1, Ch. 5 · Halliday, Resnick & Walker, Ch. 5–6.

Provenance. Standard Newtonian method. A verifier should recompute: the angled pull (4.0 kg4.0\ \mathrm{kg}, 20 N20\ \mathrm{N} at 30°30°, g=9.8g = 9.8: N=29.2 NN = 29.2\ \mathrm{N}, a=4.33 m/s2a = 4.33\ \mathrm{m/s^2}); the contact-pair law N12=Fm2/(m1+m2)N_{12} = F m_2/(m_1+m_2) (9 N9\ \mathrm{N}, 2.0/1.0 kg2.0/1.0\ \mathrm{kg}: 3.0 N3.0\ \mathrm{N}, reversed 6.06.0); the two-block example (30 N30\ \mathrm{N}, 4+2 kg4+2\ \mathrm{kg}: a=5.0a = 5.0, contact 10 N10\ \mathrm{N}, reversed 2020); the train (12 N12\ \mathrm{N}, 1+2+3 kg1+2+3\ \mathrm{kg}: a=2.0a = 2.0, contacts 1010 and 6 N6\ \mathrm{N}, audit 4 N4\ \mathrm{N}); the incline hold (2.0 kg2.0\ \mathrm{kg}, 30°30°, g=10g = 10: P=11.5 NP = 11.5\ \mathrm{N}, N=23.1 N=mg/cos⁡θN = 23.1\ \mathrm{N} = mg/\cos\theta).

Prerequisites. Newton's Second Law: F = ma · Newton's Third Law & Interaction Pairs · Weight, Normal Force & Apparent Weight. Leads to. Equilibrium of Concurrent Forces · The Atwood Machine & Variants · The Block-on-Incline Family — and, for the deep dive, the university-tier Free-Body Diagrams (Advanced Mechanics).

Why doesn't the force I exert on the wall appear on my own free-body diagram?

Because your diagram lists only forces exerted on you, and that force acts on the wall — it belongs on the wall's diagram. What appears on yours is its third-law partner: the wall's push back on you. Equal and opposite, but on different bodies — which is why they can never cancel. This one filing rule — every force filed under the body it acts on — is the discipline of the method. Push a wall while standing on ice and your diagram predicts the outcome: the only horizontal force on you is the wall's push, so backward you slide.

Is centripetal force a separate arrow on the diagram?

No — adding one is a classic double-count. "Centripetal" names a role, not an agent: it is whatever net inward component the already-drawn forces provide. A ball whirling on a string has two arrows, tension and weight; the tension is (most of) the centripetal force. Correct procedure: draw the ordinary forces, point one axis at the circle's centre, set the net inward component equal to mv2/rmv^2/r. An arrow labelled "centripetal" with no nameable agent fails the admission test — delete it.

Where does ma go, if not on the diagram?

On the other side of the equals sign. The diagram supplies the left of ΣF=ma\Sigma F = ma — the inputs; mama is the output those forces produce. Drawing it as a force double-counts the physics and wrecks the signs. Clean habit: write the force sum straight off the diagram, then the equals sign, then mama — the geography of the equation mirrors the logic. (Non-inertial frames, where a pseudo-force −maframe-ma_{\text{frame}} legally joins the diagram, have their own module — note it uses the frame's acceleration, not the body's.)

Weight and normal force on a resting book are equal and opposite — aren't they a third-law pair?

No — the most-tested subtlety in the chapter. Partners are one interaction seen from both ends: "Earth pulls book" pairs with "book pulls Earth". Weight (from the Earth) and normal (from the table) are two different interactions that balance only because the book has a=0a = 0. In a lift accelerating downward, N<mgN < mg instantly, while every true pair stays exactly equal and opposite. Equal-because-of-equilibrium is the second law's doing, not the third's.

How do I decide which way to draw friction?

Ask what the contact surfaces are doing — or trying to do — relative to each other. Kinetic friction opposes actual slipping; static friction opposes the tendency to slip (if friction vanished, which way would the surfaces slide? Static friction points the other way). This produces surprises: friction points forward on a car's driven wheels, and forward on a crate in an accelerating truck — there it is the only force accelerating the crate. The full machinery lives in Static Friction and Kinetic & Rolling Friction; the diagram needs only the direction.

When can I lump bodies into one system, and when must I separate them?

Cut wherever you like, provided everything inside the cut shares one acceleration. Lumping is the fast lane to aa: internal forces cancel by the third law. The price is blindness — internal forces are invisible from outside the cut. The moment a question asks for a tension or contact force, cut through that connection so it crosses the boundary and appears. Professionals do both in sequence: system for aa, then one component cut. More unknowns than equations means too few cuts.

Before this

Where this leads