ARCHETYPE: Block on an Accelerating Wedge

Every incline problem you have solved so far kept one silent promise: the incline itself stood still. The wedge archetype breaks that promise, and JEE-Advanced is built on the wreckage. Push the wedge horizontally and the "incline" becomes an accelerating platform; set it free on a smooth floor and the block's descent shoves the wedge out from under itself — two bodies, coupled by a single normal force, each rewriting the other's motion. The comfortable results of the fixed incline ( $N = mg\cos\theta$ , $a = g\sin\theta$ ) all fail, in ways examiners can price to the mark. What makes the wedge the chapter's boss fight is not new physics — it is the choice of weapon. Route one: stay in the ground frame, write Newton's law for each body, and encode "the block stays on the face" as a constraint between their accelerations. Always correct, occasionally miserable. Route two: ride the wedge, pay the pseudo-force toll $ma_0$ from the non-inertial frames module, and watch a two-body dynamics problem collapse into a one-body incline problem. The driven wedge (find the push that freezes the block on the face) falls in three lines to route two; the free wedge (block slides, wedge recoils) is route one's showcase, with a momentum-conservation shortcut hiding inside. This article drills both weapons on both wedges: the master results $a = g\tan\theta$ and the recoiling-wedge formulas, the demoted normal force ( $N < mg\cos\theta$ — the free wedge presses lighter!), the friction windows, and the route-choice instinct that separates a five-minute solution from a fifteen-minute one. Boards: this is JEE-Main and JEE-Advanced territory, and it is where your pseudo-force training pays its dividend.

Key takeaways

A wedge that accelerates is an incline whose rules change: $N \ne mg\cos\theta$ , and the block's fate depends on the wedge's motion.
Driven wedge (pushed at $a_0$ ): the block rides without sliding when $a_0 = g\tan\theta$ (smooth face) — derived in three lines in the wedge frame, where the pseudo-force $ma_0$ balances gravity along the face. The face then presses harder: $N = mg/\cos\theta$ .
Free wedge (smooth floor): the descending block drives the wedge backward with $A = \dfrac{mg\sin\theta\cos\theta}{M + m\sin^2\theta}$ , and the normal force drops to $N = \dfrac{Mmg\cos\theta}{M + m\sin^2\theta} < mg\cos\theta$ .
Route choice: "block stationary relative to wedge" → wedge frame + pseudo-force (statics). "Both bodies move" → ground frame + constraint (the block's acceleration relative to the wedge lies along the face), with horizontal momentum conservation as the free-wedge shortcut.
The coupling is one third-law pair: the block presses the face with $N$ , the face pushes back with $N$ — the horizontal component of that pair is what recoils the wedge.
Key idea: an accelerating wedge is solved by either paying the pseudo-force toll in the wedge frame or writing the on-the-face constraint in the ground frame — and the normal force is a live unknown, never a reflex $mg\cos\theta$ .

The core insight

Wedge problems couple a block and a movable incline through one normal-force pair: ride the wedge frame with pseudo-force $ma_0$ when the block is at rest relative to the wedge, or work in the ground frame with the along-the-face constraint when both bodies accelerate — in both routes the normal force is solved for, not assumed.

Where this lives today

The driven wedge is the physics of every tilted surface that accelerates with cargo aboard. Airport tilt-tray sorters throw parcels off exactly when their tray's $a$ exits the friction window you computed in Example 3 — sorter firmware enforces acceleration limits per tray angle. Aircraft-carrier catapult officers face the inverse problem: everything on a sloped deck surface must stay put through a $3g$ launch, which is why loose-object ("FOD") discipline is a launch-critical checklist item. The free wedge, meanwhile, is the entry ticket to analytical mechanics: "block on a movable wedge" is the canonical first problem of Lagrangian dynamics at university, where the two coordinates ( $s$ along the face, $X$ of the wedge) and the conserved horizontal momentum you met here fall out of one function with no normal force ever drawn. Master the wedge with forces now, and you will recognize it — solved in four lines — on the first page of your UG mechanics course.

The setup

Geometry first, because in an archetype the geometry is the physics. A wedge of mass $M$ , face angle $\theta$ , sits on a floor; a block of mass $m$ rests on the face. Two canonical stories:

Figure

The two wedge stories — driven wedge freezes the block, free wedge recoils under it

The driven wedge. An external agent pushes the wedge at a controlled $a_0$ ; the question is the $a_0$ that keeps the block from sliding (or, with friction, the range of survivable $a_0$ ). The whole system then shares one acceleration — the problem is secretly statics in the wedge frame.

The free wedge. Nothing external pushes horizontally (smooth floor). The block slides down; because the block presses the tilted face, the face's reaction has a horizontal component, and the wedge recoils. Both bodies accelerate; the problem is genuinely two-body. One force couples them in both stories: the normal pair $N$ , perpendicular to the face. The block pushes the face with $N$ (its horizontal component $N\sin\theta$ shoves the wedge); the face pushes the block with $N$ (its vertical component $N\cos\theta$ fights the block's weight). And the crucial demotion from fixed-incline life: $N$ is an unknown. It is whatever the coupled motion demands — larger than $mg\cos\theta$ for a wedge driven into its block, smaller for a wedge escaping from under it.

The math: two routes, two wedges

Route selection, the exam instinct:

situation	route	why
block at rest relative to the wedge	wedge frame + pseudo-force	one body, statics, three lines
block slides while wedge moves	ground frame + constraint	genuine two-body dynamics
"wedge displacement when block descends $h$ "	horizontal momentum / centre of mass	no forces needed at all
rough face, "range of $a_0$ "	wedge frame, friction at ceiling both ways	the incline window, imported

Derivation AThe driven wedge — pseudo-force route versus ground-frame route

We find the horizontal acceleration $a_0$ that keeps a block motionless on a smooth wedge face, both ways, to calibrate the two weapons on a problem where both work cleanly. Target results: $a_0 = g\tan\theta$ and $N = mg/\cos\theta$ . Watch the pseudo-force's direction — opposite the wedge's acceleration, i.e. into the slope when the wedge is pushed toward the block's downhill side.

Step 1. Wedge-frame route. Ride the wedge: the block is at rest — equilibrium. Forces: $mg$ down, $N$ perpendicular to the face, pseudo-force $ma_0$ horizontal (opposite the wedge's acceleration).

Step 2. Resolve along the face (down-slope positive): gravity contributes $mg\sin\theta$ ; the pseudo-force contributes $ma_0\cos\theta$ up-slope. Equilibrium:

mg\sin\theta = ma_0\cos\theta \;\;\Rightarrow\;\; a_0 = g\tan\theta

Step 3. Resolve perpendicular to the face:

N = mg\cos\theta + ma_0\sin\theta = mg\cos\theta + mg\frac{\sin^2\theta}{\cos\theta} = \frac{mg}{\cos\theta}

Why: the driven wedge presses HARDER than a fixed incline ( $mg/\cos\theta > mg\cos\theta$ ) — it must both hold the block up and shove it sideways. Note the mass independence of $a_0$ : every block freezes at the same $g\tan\theta$ , the equivalence-principle signature of pseudo-force problems.

Step 4. Ground-frame route, for calibration. The block moves horizontally with acceleration $a_0$ (it rides the wedge). Newton in ground axes: horizontal $N\sin\theta = ma_0$ ; vertical $N\cos\theta - mg = 0$ . Divide: $\tan\theta = a_0/g$ — same answer; and $N = mg/\cos\theta$ from the vertical line directly. Verified: limits — $\theta \to 0$ needs no push ( $a_0 \to 0$ ); $\theta \to 90°$ demands $a_0 \to \infty$ (you cannot pin a block to a vertical wall by finite horizontal acceleration of the wall alone, without friction). Both routes agree line by line; the wedge frame wins on speed, the ground frame on transparency.

Go deeper · Derivation B · 1 minThe free wedge — recoil, reduced normal force, and the constraint that runs it

Derivation BThe free wedge — recoil, reduced normal force, and the constraint that runs it

We solve the showcase: block $m$ released on the smooth face of wedge $M$ , smooth floor. Targets: the wedge's recoil acceleration $A$ , the normal force $N$ , and the block's acceleration relative to the wedge, $a_{\text{rel}}$ , along the face. Strategy: ground frame; unknowns $A$ , $a_{\text{rel}}$ , $N$ ; three equations — wedge horizontal, block horizontal, block vertical — with the constraint built in by writing the block's acceleration as (wedge's $A$ ) + ( $a_{\text{rel}}$ along the face). Watch signs: take the wedge recoiling in the $-x$ direction while the block's relative slide has components ( $+a_{\text{rel}}\cos\theta$ , $-a_{\text{rel}}\sin\theta$ ).

Step 1. Wedge, horizontal. The block presses the face with $N$ ; the horizontal component $N\sin\theta$ drives the recoil:

N\sin\theta = M A

Step 2. Block, horizontal. Ground-frame horizontal acceleration of the block $= a_{\text{rel}}\cos\theta - A$ (slide component forward, wedge carrying it backward). The only horizontal force on the block is the face's push, $-N\sin\theta$ ... choosing forward (down-slope horizontal) as positive:

-N\sin\theta = -m\left(a_{\text{rel}}\cos\theta - A\right)\;\;\Rightarrow\;\; N\sin\theta = m(a_{\text{rel}}\cos\theta - A)

Why: block and wedge push each other apart horizontally with the same $N\sin\theta$ — the third-law pair doing double duty. Adding Steps 1 and 2 kills $N$ : $m\,a_{\text{rel}}\cos\theta = (M+m)A$ — which is horizontal momentum conservation in disguise (no external horizontal force on the pair).

Step 3. Block, vertical. Vertical acceleration $= a_{\text{rel}}\sin\theta$ downward; forces $mg$ down, $N\cos\theta$ up:

mg - N\cos\theta = m\,a_{\text{rel}}\sin\theta

Step 4. Solve the three. From Steps 1+2: $a_{\text{rel}} = \dfrac{(M+m)A}{m\cos\theta}$ . Substitute into Step 3 with $N = MA/\sin\theta$ :

A = \frac{m g \sin\theta\cos\theta}{M + m\sin^2\theta}, \qquad N = \frac{M m g\cos\theta}{M + m\sin^2\theta}, \qquad a_{\text{rel}} = \frac{(M+m) g\sin\theta}{M + m\sin^2\theta}

Why: read the structures. $N$ is $mg\cos\theta$ times the factor $M/(M + m\sin^2\theta) < 1$ — the escaping wedge presses lighter, because the face is partly "falling away" from the block. And $a_{\text{rel}} > g\sin\theta$ : the block slides down the face FASTER than on a fixed incline, since the face retreats beneath it. *Verified: limits — $M \to \infty$ : $A \to 0$ , $N \to mg\cos\theta$ , $a_{\text{rel}} \to g\sin\theta$ , the fixed incline recovered exactly; $\theta \to 90°$ : $N \to 0$ , $a_{\text{rel}} \to g$ , free fall beside a slab; $\theta \to 0$ : everything sleeps. Numbers ( $m = 1$ , $M = 3$ kg, $\theta = 30°$ , $g = 10$ ): denominator $3.25$ ; $A = 4.33/3.25 = 1.33\ \mathrm{m/s^2}$ , $N = 26.0/3.25 = 8.0$ N (fixed incline would say $8.66$ N), $a_{\text{rel}} = 20/3.25 = 6.15 > 5 = g\sin 30°$ ✓.*

Figure

The free wedge anatomy — one N-pair couples descent to recoil, and the constraint ties their accelerations

The momentum shortcut

The free wedge carries a hidden conservation law: with a smooth floor, no external horizontal force acts on the block+wedge pair, so horizontal momentum stays zero (from rest). Two consequences examiners bill separately: (i) the velocity ratio at any instant, $Mv_{\text{wedge}} = m v_{\text{block},x}$ ; (ii) the displacement ratio — while the block's horizontal position advances $x_b$ , the wedge retreats $x_w$ with $m x_b = M x_w$ . A block descending a height $h$ travels $h/\tan\theta$ horizontally relative to the wedge, so the wedge's ground displacement is

x_w = \frac{m}{M+m}\cdot\frac{h}{\tan\theta} \qquad (1)

where: the factor $m/(M+m)$ is the centre-of-mass bookkeeping: the pair's COM cannot move horizontally, so the two slide in opposite directions in inverse proportion to their masses. No forces, no accelerations, no time — the fastest three marks on the paper when asked.

Limiting cases

case	result	meaning
$M \to \infty$	fixed-incline formulas return	the whole incline family is the wedge's special case
$\theta \to 0$	$a_0 \to 0$ , $A \to 0$	a floor needs no driving and gives no recoil
$\theta \to 90°$	$N \to 0$ (free), $a_0 \to \infty$ (driven)	a vertical face can't hold a block without friction
$m \gg M$ (free)	$A \to g\cot\theta$ , $N \to \frac{Mg\cos\theta}{\sin^2\theta}$	light wedge squirts out; block nearly free-falls
$a_0 < g\tan\theta$ (driven, smooth)	block slides down the face	under-driving fails downhill
$a_0 > g\tan\theta$ (driven, smooth)	block slides up the face	over-driving fails uphill — friction widens both edges into a window

Common trap"N = mg cos θ — it's an incline, after all"

The fixed-incline reflex is the wedge's number-one casualty. The normal force is set by the perpendicular dynamics of the block, and once the wedge itself accelerates, that dynamics changes: a driven wedge presses its block harder — up to $N = mg/\cos\theta$ at the frozen condition (the face must supply the block's sideways $ma_0$ as well as fight gravity) — while a free wedge presses lighter, $N = Mmg\cos\theta/(M + m\sin^2\theta)$ , because the retreating face partially drops away from the block. Both errors propagate straight into friction budgets, recoil rates, and every MCQ option built from them. The discipline is the same one the incline family taught, upgraded: the perpendicular equation is your first line of algebra, and on a wedge it contains acceleration terms, not just weights. If your wedge solution never solved for $N$ — if $N$ entered as a memory — the answer is wrong in exactly the way the examiner planned.

Worked examples

1 · Board level — the driven wedge, both deliverables (JEE-Main opener, full method)

A $2.0$ kg block rests on the smooth $30°$ face of a wedge. With what horizontal acceleration must the wedge be driven so the block does not slide? What normal force does the face then exert? ( $g = 10\ \mathrm{m/s^2}$ .)

Step 1. Wedge frame; declare the toll: pseudo-force $ma_0$ horizontal on the block, opposite the drive. Block at rest in this frame ⇒ equilibrium.

Step 2. Along the face: $mg\sin 30° = ma_0\cos 30° \Rightarrow a_0 = g\tan 30° = \boxed{5.8\ \mathrm{m/s^2}}$ .

Step 3. Perpendicular: $N = mg\cos 30° + ma_0\sin 30° = 2(10)(0.866) + 2(5.77)(0.5) = 17.3 + 5.8 = \boxed{23.1\ \mathrm{N}}$ .

Step 4. Sanity: $N = mg/\cos 30° = 20/0.866 = 23.1$ N ✓, and $N > mg = 20$ N — the face is doing two jobs. Ground-frame cross-check: $N\sin 30° = 11.5 = ma_0$ ✓.

2 · Board level — how far does the wedge slide back? (momentum route)

A $1.0$ kg block starts at the top of the smooth face of a $4.0$ kg wedge ( $\theta = 45°$ ) resting on a smooth floor, and descends a height $h = 0.80$ m. How far does the wedge move backward?

Step 1. Smooth floor ⇒ zero net horizontal force on the pair ⇒ horizontal momentum stays zero ⇒ the pair's centre of mass never moves horizontally.

Step 2. Relative to the wedge, the block moves $h/\tan 45° = 0.80$ m horizontally along the descent.

Step 3. Split the relative displacement in inverse mass ratio (COM fixed): wedge's share $x_w = \dfrac{m}{M+m} \times 0.80 = \dfrac{1.0}{5.0} \times 0.80 = \boxed{0.16\ \mathrm{m}}$ backward.

Step 4. Check: block's ground displacement $= 0.80 - 0.16 = 0.64$ m forward, and $1.0 \times 0.64 = 4.0 \times 0.16$ ✓ — momenta of displacement balanced. No forces, no accelerations: three marks in ninety seconds.

3 · Competitive level — the rough driven wedge window (JEE-Main pattern, ~4 minutes)

A block sits on the $37°$ face of a wedge, $\mu_s = 0.5$ between block and face; $\sin 37° = 0.6$ , $\cos 37° = 0.8$ , $g = 10\ \mathrm{m/s^2}$ . Find the minimum and maximum horizontal acceleration of the wedge for the block not to slide.

Step 1. Wedge frame: forces on the block are $mg$ , $N$ , pseudo-force $ma_0$ (into the slope side), and friction $f \le \mu_s N$ along the face. Two verge cases, friction at ceiling, opposite directions.

Step 2. Perpendicular line (both cases): $N = mg\cos\theta + ma_0\sin\theta = m(8 + 0.6a_0)$ .

Step 3. Minimum $a_0$ (block verges on sliding down; friction up-slope at ceiling). Along the face, the down-slope pull must be met by the pseudo-force's up-slope share plus friction: $mg\sin\theta - ma_0\cos\theta = \mu_s N$ . The masses cancel; in numbers: $10(0.6) - 0.8\,a_0 = 0.5(8 + 0.6\,a_0)$ ⇒ $6 - 0.8a_0 = 4 + 0.3a_0$ ⇒ $a_0 = \dfrac{2}{1.1} = \boxed{1.8\ \mathrm{m/s^2}}$ .

Step 4. Maximum $a_0$ (verge of sliding up; friction down-slope): $a_0\cos\theta \cdot m = mg\sin\theta + \mu_s N$ ⇒ $0.8a_0 = 6 + 0.5(8 + 0.6a_0)$ ⇒ $0.8a_0 - 0.3a_0 = 10$ ⇒ $a_0 = \boxed{20\ \mathrm{m/s^2}}$ . Window: $1.8 \le a_0 \le 20\ \mathrm{m/s^2}$ . Timing note: this is the incline's equilibrium-window problem wearing a pseudo-force — the same two-verge ritual, with $ma_0\cos\theta$ as the "applied force along the slope" and $ma_0\sin\theta$ feeding $N$ . Set both verges up as a pair before solving either; each is then three lines of arithmetic. The wide window (×11) is the $N$ -feeding effect again. Budget 4 min; masses cancel throughout — if your window contains $m$ , a term went astray.

4 · Competitive level — JEE-Advanced stretch: the free wedge in numbers (~5 minutes)

A $1.0$ kg block is released on the smooth $30°$ face of a $3.0$ kg wedge that rests on a smooth floor; $g = 10\ \mathrm{m/s^2}$ . Find (a) the wedge's acceleration, (b) the normal force between block and face, (c) the block's acceleration relative to the wedge, and (d) the time for the block to slide $1.3$ m along the face.

Step 1. Denominator once for everything: $M + m\sin^2\theta = 3.0 + 1.0(0.25) = 3.25$ kg.

Step 2. (a) $A = \dfrac{mg\sin\theta\cos\theta}{3.25} = \dfrac{1 \times 10 \times 0.5 \times 0.866}{3.25} = \boxed{1.33\ \mathrm{m/s^2}}$ (backward).

Step 3. (b) $N = \dfrac{Mmg\cos\theta}{3.25} = \dfrac{3 \times 10 \times 0.866}{3.25} = \boxed{8.0\ \mathrm{N}}$ — below the fixed-incline $8.66$ N, the escaping-face discount.

Step 4. (c) $a_{\text{rel}} = \dfrac{(M+m)g\sin\theta}{3.25} = \dfrac{4 \times 10 \times 0.5}{3.25} = \boxed{6.15\ \mathrm{m/s^2}}$ — above $g\sin\theta = 5$ : the retreating face lets the block down faster. (d) Along-face slide from rest: $t = \sqrt{2L/a_{\text{rel}}} = \sqrt{2(1.3)/6.15} = \sqrt{0.423} = \boxed{0.65\ \mathrm{s}}$ . Timing note: with the three formulas cached, this is arithmetic off one denominator — under 3 minutes. Without them, derive via the Steps 1–4 of Derivation B (~8 minutes), so cache them, but attach their limits ( $M \to \infty$ sanity) since JEE-Advanced likes perturbing the setup (wedge against a wall: recoil dies, $N$ returns to $mg\cos\theta$ — spot it instantly). Budget 5 min total.

Exam traps

The fixed-incline reflex. $N = mg\cos\theta$ and $a_{\text{rel}} = g\sin\theta$ die the moment the wedge moves — see the pitfall box. Solve for $N$ ; never import it.

Pseudo-force on the wrong body or with the wrong mass. In the wedge frame, the block gets the toll $ma_0$ (its own mass). If you also analyse the wedge in its own frame, the wedge gets $Ma_0$ — but usually the wedge's equation is what the external agent handles; don't mix.

Frame mixing. Wedge-frame pseudo-forces with ground-frame accelerations double-counts — the non-inertial module's cardinal sin, and wedges are where it happens most. Declare the frame in line one.

Constraint written in the ground frame. The block's ground acceleration is NOT along the face (the face itself accelerates). Only the relative acceleration lies along the face: $\vec{a}_{\text{block}} = \vec{A}_{\text{wedge}} + \vec{a}_{\text{rel}}(\text{along face})$ . Writing $a_{\text{block},y}/a_{\text{block},x} = \tan\theta$ is the classic free-wedge wreck.

Momentum shortcut with a rough floor. Horizontal momentum conservation needs the floor smooth. Friction under the wedge is an external horizontal force — the displacement formula (1) is void, and only the force route survives.

Direction of the recoil. The wedge recoils away from the side the block descends toward — the face's push has a horizontal component pointing from block into wedge. Getting the recoil direction backwards flips every sign downstream.

Which $g$ . JEE papers declare $g = 10\ \mathrm{m/s^2}$ almost always; every formula here scales linearly in $g$ , so the declaration passes straight through.

Origins · 1 minJean le Rond d'Alembert and how this was discovered

Jean le Rond d'Alembert

1717–1783 · Paris Academy of Sciences

Abandoned as an infant on the steps of the Saint-Jean-le-Rond chapel and named after it, d'Alembert rose to co-edit the Encyclopédie and to write the treatise this archetype secretly runs on. His Traité de dynamique (1743) proposed a maneuver of enduring slyness: carry the $ma$ of Newton's second law across the equals sign and treat it as a force — the "inertial force" — so that every problem of motion becomes a problem of equilibrium. That is precisely the pseudo-force route of the driven wedge: statics in the accelerating frame, dynamics nowhere in sight. D'Alembert's principle became the hinge on which Lagrange swung mechanics into its analytical form, where wedges, constraints, and normal forces dissolve into generalized coordinates. He also gave the wave equation its first solution and his name to the ratio test's rival in every calculus course — not bad for a man whose formal schooling was intended for the priesthood he politely declined.

History

1743

— d'Alembert's Traité de dynamique: inertia moved to the force side — dynamics as statics.

1788

— Lagrange's Mécanique analytique makes the block-on-wedge a two-coordinate exercise; constraint forces vanish.

1894

— Ernst Mach's Science of Mechanics popularizes the free wedge as a momentum-conservation showcase.

1900

s — the accelerating wedge settles into JEE/olympiad canon as the standard test of frame fluency.

Going further

Texts. HC Verma, Concepts of Physics Vol. 1, Ch. 5–6 (frames and friction) · DC Pandey, Mechanics 1 (the wedge problem set) · Kleppner & Kolenkow Ch. 2, Example 2.16 onward (the honest constraint treatment).

Provenance. All results are standard. A verifier should recompute: $a_0 = g\tan\theta$ and $N = mg/\cos\theta$ (both routes); the free-wedge trio $A, N, a_{\text{rel}}$ with denominator $M + m\sin^2\theta$ and the numbers $1.33\ \mathrm{m/s^2}$ , $8.0$ N, $6.15\ \mathrm{m/s^2}$ for $(1, 3\ \mathrm{kg}, 30°)$ ; Example 2's $0.16$ m; Example 3's window $1.8$ – $20\ \mathrm{m/s^2}$ .

Prerequisites. Non-Inertial Frames & Pseudo-Forces · The Block-on-Incline Family · Connected Blocks, Strings & Constraints. Leads to. Limiting & Extreme Cases (technique) · Man on Boat / Plank on Ice (momentum). Deep dives (UG). Constraints & Constraint Forces · the Lagrangian route to the free wedge.

How do I pick between the pseudo-force route and the ground-frame route?

Ask one question: is the block at rest relative to the wedge? Yes (driven-wedge, frozen-block, "moves as one unit" problems) → ride the wedge, pay the toll $ma_0$ , solve statics: fastest by far. No (free wedge, block sliding while wedge moves) → the wedge frame is legal but its acceleration is unknown, so the toll is an unknown too; the ground frame with the relative-acceleration constraint (or the momentum shortcut) is cleaner. Rule of thumb: pseudo-forces buy the most when the frame's acceleration is given; constraints and momentum buy the most when it is asked.

Why does the free wedge press the block more lightly than a fixed one?

Because the face is running away. The block's vertical dynamics reads $mg - N\cos\theta = m a_{\text{vert}}$ : on a fixed incline the block's vertical acceleration is modest, but on a free wedge the block descends faster (the surface retreats), so $a_{\text{vert}}$ is larger, and the equation forces $N$ smaller. Push it to the limit: a massless wedge ( $M \to 0$ ) squirts out instantly, $N \to 0$ , and the block simply free-falls. The normal force measures how firmly the surface insists on being in the block's way — a fleeing surface insists less.

Where exactly does the wedge's recoil force come from? Gravity on the block points straight down.

From the third-law partner of the normal force. Gravity pulls the block into the face; the face pushes the block perpendicular to itself, and the block pushes back on the face with equal $N$ — a vector perpendicular to the face, which on a tilted face has a horizontal component $N\sin\theta$ . That sideways component is the recoil engine. The chain is worth reciting: gravity (vertical) → block presses face (perpendicular) → horizontal component shoves the wedge. A tilted constraint converts vertical drive into horizontal motion — the same conversion a ship's bow wave or a splitting axe performs.

Why is the driven-wedge answer a₀ = g tan θ independent of the block's mass?

Every force in the frozen-block balance scales with $m$ : gravity's along-face pull $mg\sin\theta$ , the pseudo-force's along-face push $ma_0\cos\theta$ . Newton's second law divides the $m$ away — the equivalence of inertial and gravitational mass, the same cancellation that makes all bodies fall alike and all pendulums in a given car tilt alike. Practical consequence: a feather and an anvil freeze on the same wedge at the same $a_0$ ; and if an exam's driven-wedge answer contains $m$ (without friction or extra forces), a term is misplaced.

For the momentum shortcut, why does the pair's centre of mass stay fixed horizontally but not vertically?

Conservation follows the force ledger, direction by direction. Horizontally: smooth floor, no external horizontal force on the pair — momentum stays zero, COM stays put. Vertically: gravity and the floor's normal act on the pair, and while the block descends they do NOT cancel (the floor pushes less than the total weight — the free-wedge system's COM is accelerating downward), so vertical momentum is not conserved and the COM drops. One system, two directions, two verdicts — always audit them separately.

What changes if the wedge is against a wall, or the floor is rough?

Wall: the wedge cannot recoil; the wall supplies whatever horizontal force is needed. You are back to the fixed incline exactly — $N = mg\cos\theta$ , $a_{\text{rel}} = g\sin\theta$ — and the wall feels $N\sin\theta$ . JEE-Advanced loves this as a sub-part because it tests whether your free-wedge formulas were memorized or understood. Rough floor: friction under the wedge is an external horizontal force — the momentum shortcut dies, and the wedge may not move at all: check whether the required holding friction $N\sin\theta$ (with fixed-incline $N$ ) is within $\mu_s$ (floor) × (total weight); if yes, it is a statics problem; if no, the wedge slides with floor friction in its equation.