Damped Harmonic Oscillator

Flick a wine glass and it sings for seconds — thousands of vibrations before the note fades. Push a screen door with a hydraulic closer and it swings shut once, no wobble, no bounce. Slam a car over a pothole and the chassis dips, recovers, and settles in under a second. All three are the same system — a mass, a restoring force, and friction — living in three different regimes of one differential equation. The only thing that distinguishes a bell from a door closer is a single dimensionless number: the ratio of how fast friction eats motion to how fast the spring wants to oscillate. The frictionless oscillator you have already mastered is a beautiful lie: its energy sloshes between kinetic and potential forever. Every real oscillator leaks. The simplest honest model adds a drag force proportional to velocity, $-b\dot x$ , and that one term transforms the physics: amplitudes decay exponentially, the oscillation frequency shifts downward, and — most surprisingly — if you crank the friction high enough, the system stops oscillating altogether and just oozes home. The boundary between ringing and oozing is a knife-edge called critical damping, and it is one of the most deliberately engineered conditions in technology. This article solves the damped oscillator completely, sorts its three personalities — underdamped, critically damped, overdamped — and introduces the quality factor $Q$ , the single number engineers quote for everything from quartz watches to black holes.

Key takeaways

Adding viscous drag to the spring–mass system gives $m\ddot x + b\dot x + kx = 0$ , or in standard form $\ddot x + 2\gamma\dot x + \omega_0^2 x = 0$ with damping rate $\gamma = b/2m$ and natural frequency $\omega_0 = \sqrt{k/m}$ .
The trial solution $x = e^{rt}$ turns the differential equation into algebra: $r_\pm = -\gamma \pm \sqrt{\gamma^2 - \omega_0^2}$ . The sign of the discriminant sorts all behaviour into three regimes.
Underdamped ( $\gamma < \omega_0$ ): a decaying cosine, $x = A e^{-\gamma t}\cos(\omega_d t + \phi)$ , oscillating at the reduced frequency $\omega_d = \sqrt{\omega_0^2 - \gamma^2}$ .
Critically damped ( $\gamma = \omega_0$ ): $x = (A + Bt)e^{-\gamma t}$ — the fastest return to equilibrium without overshoot.
Overdamped ( $\gamma > \omega_0$ ): two real exponentials; the slow one dominates, making the return slower than critical.
Energy decays as $e^{-2\gamma t}$ ; the quality factor $Q = \omega_0/2\gamma$ counts roughly how many radians of ringing survive — from $Q \sim 0.5$ in a car suspension to $Q \sim 10^6$ in a quartz resonator. On a decay trace, the logarithmic decrement $\delta = \gamma T_d$ turns two peak heights into a measurement of $\gamma$ .
Key idea: one dimensionless ratio, $\gamma/\omega_0$ , decides everything — less than 1 rings, greater than 1 creeps, exactly 1 settles fastest.

The core insight

Substituting $x = e^{rt}$ converts the damped oscillator equation into the characteristic equation $r^2 + 2\gamma r + \omega_0^2 = 0$ ; the sign of the discriminant $\gamma^2 - \omega_0^2$ alone determines whether the system oscillates inside a decaying envelope, returns critically fast, or crawls back without ever crossing zero.

Where this lives today

The most spectacular damped oscillator ever measured is a black hole. When two black holes merge, the remnant is briefly deformed and "rings down" through quasinormal modes — literally damped sinusoids $e^{-\gamma t}\cos(\omega_d t + \phi)$ whose frequency and decay rate are fixed by the hole's mass and spin. LIGO fits templates of exactly this form to its strain data; general relativity predicts the ringdown of a ~62-solar-mass remnant at a few hundred hertz with $Q \sim 3$ , and testing whether both numbers match Kerr-black-hole predictions is a frontier test of gravity itself. At the everyday end, damping regimes are a design language. Vehicle suspensions target slight underdamping ( $\zeta = \gamma/\omega_0 \approx 0.7$ ) as the best compromise between settling speed and harshness; skyscraper tuned mass dampers are sized in $\gamma$ and $\omega_0$ to drink energy out of wind sway; MEMS gyroscopes are vacuum-packaged specifically to remove air damping and push $Q$ into the tens of thousands; and in every RLC circuit, the same equation runs with $L \to m$ , $R \to b$ , $1/C \to k$ , so a pulse either rings, settles critically, or sags — chosen deliberately by the circuit designer.

The physical picture

Take the frictionless spring–block oscillator and lower it into honey. The spring still pulls toward equilibrium with $-kx$ , but now every motion pays a toll proportional to speed. Energy that used to slosh forever between kinetic and potential is bled away as heat, a little on every pass, and the swings shrink. Two clocks now compete inside the system. The spring's clock ticks at the natural frequency $\omega_0 = \sqrt{k/m}$ — how fast the system wants to oscillate. The friction's clock runs at the damping rate $\gamma = b/2m$ — how fast drag erases motion. If the spring's clock is faster ( $\gamma < \omega_0$ ), the mass completes many swings before dying: underdamped. If friction's clock wins ( $\gamma > \omega_0$ ), the mass never gets to swing at all; it is reeled back to equilibrium like a spoon pulled through syrup: overdamped. And when the two clocks tick at exactly the same rate, $\gamma = \omega_0$ , the system sits on the boundary — critically damped — returning home as fast as it possibly can without a single overshoot.

Figure

Three regimes, one release — same spring, same start, different friction

The math: one equation, three personalities

The honest model needs exactly one new term. Keep the spring's $-kx$ , and add a drag force $-b\dot x$ , proportional to velocity — the simplest law that always opposes the motion and vanishes at rest, and the one that honey, oil, and slow air actually obey. Feed both forces into Newton's second law and divide by $m$ to expose the two competing clocks of the physical picture:

m\ddot x + b\dot x + kx = 0 \qquad \Longrightarrow \qquad \ddot x + 2\gamma\dot x + \omega_0^2 x = 0 \qquad (1)

where: $b$ — drag coefficient (kg/s); $\gamma \equiv b/2m$ — damping rate (s $^{-1}$ ); $\omega_0 \equiv \sqrt{k/m}$ — the natural frequency the system would have without friction. The factor of 2 in $\gamma$ is a convention that keeps every later formula clean — and forgetting it is the most common algebra slip in the subject.

Derivation AExponentials turn calculus into algebra

We are deriving the general solution of equation (1) — every damped oscillator, in every regime, captured in one formula. The strategy is the standard opening move for any linear equation with constant coefficients: guess $x = e^{rt}$ , because differentiation only multiplies an exponential by a constant, so the guess converts calculus into algebra. What emerges is a quadratic — the characteristic equation — whose two roots carry all the physics: real parts are decay rates, imaginary parts are oscillation frequencies. The step to watch is the last one: a second-order equation demands two independent solutions, and when the two roots collide at critical damping the recipe momentarily breaks and must hand you a factor of $t$ instead.

Step 1. Try an exponential. Let $x = e^{rt}$ . Each time derivative brings down one factor of $r$ : $\dot x = re^{rt}$ , $\ddot x = r^2e^{rt}$ . Why: for linear equations with constant coefficients, exponentials are the functions that differentiation cannot change in shape — they are the natural alphabet of the problem.

Step 2. Substitute and factor. $(r^2 + 2\gamma r + \omega_0^2)\,e^{rt} = 0$ , and since $e^{rt}$ never vanishes,

r^2 + 2\gamma r + \omega_0^2 = 0 \qquad (2)

Why: the differential equation has become a quadratic — all the physics now lives in two roots.

Step 3. Solve the quadratic:

r_\pm = -\gamma \pm \sqrt{\gamma^2 - \omega_0^2} \qquad (3)

where: the real part of $r$ is a decay rate; any imaginary part is an oscillation frequency. Why: the discriminant $\gamma^2 - \omega_0^2$ can be negative, zero, or positive — and that sign is the whole story.

Step 4. Superpose. The general solution is $x(t) = C_1 e^{r_+ t} + C_2 e^{r_- t}$ , with $C_1, C_2$ fixed by the initial position and velocity. Why: the equation is linear and second order — two independent solutions, two constants. When the roots coincide ( $r_+ = r_-$ ), the pair $e^{rt}$ and $t\,e^{rt}$ takes over as the independent set. *Verified: reduces to $r = \pm i\omega_0$ , pure oscillation, when $\gamma = 0$ ; matches Morin Ch. 4 and Taylor Ch. 5.*

Go deeper · Derivation B · 1 minThe underdamped solution made real

Derivation BThe underdamped solution made real

The goal is the decaying cosine of equation (4) — the formula behind every ringing system from a wine glass to a post-merger black hole. When $\gamma < \omega_0$ the discriminant of the characteristic equation is negative and the two roots are complex conjugates, so the strategy is to massage the complex general solution into an explicitly real displacement: factor out the decay envelope the two roots share, then demand that the physical $x(t)$ be a real number, which locks the superposition constants into a conjugate pair and collapses the complex exponentials into a cosine via Euler's formula. Watch the frequency that emerges — it is not $\omega_0$ but the strictly smaller $\omega_d$ : friction slows the clock even while the oscillator still rings.

Step 1. Let the discriminant go negative ( $\gamma < \omega_0$ ) and pull out the $i$ : $\sqrt{\gamma^2 - \omega_0^2} = i\sqrt{\omega_0^2 - \gamma^2} \equiv i\omega_d$ . Why: the roots become the complex-conjugate pair $r_\pm = -\gamma \pm i\omega_d$ — decay and oscillation in one package.

Step 2. Factor the common envelope: $x(t) = e^{-\gamma t}\left(C_1 e^{i\omega_d t} + C_2 e^{-i\omega_d t}\right)$ . Why: both roots share the real part $-\gamma$ , so the decay factors out of the oscillation.

Step 3. Demand a real displacement. Setting $C_2 = C_1^*$ and applying Euler's formula collapses the bracket to a cosine:

x(t) = A\,e^{-\gamma t}\cos(\omega_d t + \phi), \qquad \omega_d = \sqrt{\omega_0^2 - \gamma^2} \qquad (4)

where: $A, \phi$ — amplitude and phase from initial conditions; $\omega_d$ — the damped frequency, always below $\omega_0$ . Why: friction retards the motion on every swing, so each cycle takes slightly longer than the frictionless one. *Verified: as $\gamma \to 0$ , the envelope $\to 1$ and $\omega_d \to \omega_0$ , recovering simple harmonic motion exactly; matches Landau & Lifshitz §25.*

At the boundary $\gamma = \omega_0$ the two roots collide at $r = -\gamma$ and the recipe breaks — one root can only supply one solution. The missing partner is $t\,e^{-\gamma t}$ (the universal signature of a repeated root), giving the critically damped solution

x(t) = (A + Bt)\,e^{-\gamma t} \qquad (5)

where: $A, B$ — fixed by the initial position and velocity. The $Bt$ factor lets the solution grow briefly before the exponential strangles it, which is why a critically damped system can cross equilibrium at most once and never rings — the fastest possible homecoming, the setting a galvanometer or door closer is tuned to. Push the friction past the boundary and the square root turns real: the roots split apart along the negative real axis, each contributing its own pure exponential decay, and the oscillation is gone for good. For $\gamma > \omega_0$ — the overdamped case:

x(t) = C_1 e^{r_+ t} + C_2 e^{r_- t}, \qquad r_\pm = -\gamma \pm \sqrt{\gamma^2 - \omega_0^2} \qquad (6)

where: $r_+$ is the slow rate (closer to zero) and controls the late-time creep. Here is the counterintuitive punchline: piling on more friction makes the slow rate $\gamma - \sqrt{\gamma^2-\omega_0^2}$ smaller — an overdamped system takes longer to settle than a critically damped one. More honey, slower homecoming. (The full engineering anatomy of the three regimes — settling times, tolerance bands, design case studies — has its own article: Damping regimes.)

Figure

The roots in the complex plane — damping slides them together, collides them, splits them

Energy decay and the quality factor

The displacement solutions say where the mass is; the engineering question is how fast the stored energy drains. Energy is quadratic in amplitude — $\tfrac12 kx^2$ in the spring, $\tfrac12 mv^2$ in the mass — so the amplitude envelope $e^{-\gamma t}$ of an underdamped oscillator enters twice, and the energy law is its square:

E(t) \approx E_0\,e^{-2\gamma t} \qquad (7)

where: $E_0$ — initial mechanical energy; the approximation is excellent whenever many cycles fit inside the decay ( $\gamma \ll \omega_0$ ). Amplitude halves in time $\ln 2/\gamma$ ; energy halves twice as fast. Comparing a wine glass to a car suspension calls for a number that ignores the hardware's size and units altogether. The natural candidate is the ratio of the two competing clocks — how fast the system rings against how fast it forgets. That ratio is the quality factor, the single figure of merit every practitioner quotes:

Q = \frac{\omega_0}{2\gamma} = \frac{\sqrt{mk}}{b} \qquad (8)

where: $Q$ — dimensionless; roughly the number of radians of oscillation completed before the energy falls by $e^{-1}$ , or equivalently $Q/\pi$ full cycles per $e$ -folding of amplitude — and after $Q$ full cycles the stored energy is down by $e^{-2\pi} \approx 1/535$ . A car suspension is engineered near $Q \sim 0.5$ – $0.7$ (settle fast, don't bounce); a church bell rings at $Q \sim 10^3$ ; a quartz watch crystal holds $Q \sim 10^5$ – $10^6$ , which is precisely why it keeps time. Engineers often quote the same information as the damping ratio $\zeta = \gamma/\omega_0 = 1/2Q$ : $\zeta < 1$ rings, $\zeta = 1$ is critical, $\zeta > 1$ creeps. None of this is useful in the lab until $\gamma$ can be measured — and the decay trace itself hands it over, two peaks at a time.

Go deeper · Derivation C · 1 minThe logarithmic decrement and the quality factor

Derivation CThe logarithmic decrement and the quality factor

We are deriving the experimentalist's bridge between theory and a ringdown trace on a screen: the logarithmic decrement, equation (9), which converts the heights of two successive peaks into the damping rate $\gamma$ — no curve fitting, no calibration. The strategy is to compare the underdamped solution with itself exactly one damped period $T_d = 2\pi/\omega_d$ later: the cosine returns to the same value, so everything oscillatory cancels and only the envelope ratio survives. The point to watch is which period appears — it is $T_d$ , not the frictionless $2\pi/\omega_0$ , though for light damping the distinction evaporates and the decrement collapses into the simple relation $Q \approx \pi/\delta$ .

Step 1. Take the ratio of displacements one period apart. At two successive peaks, the cosine factor in $x(t) = A\,e^{-\gamma t}\cos(\omega_d t + \phi)$ is identical, so

\frac{x(t)}{x(t + T_d)} = \frac{e^{-\gamma t}}{e^{-\gamma (t + T_d)}} = e^{\gamma T_d}

Why: the oscillatory part repeats after exactly one period — only the envelope remembers that time has passed.

Step 2. Take the logarithm and name the result the logarithmic decrement:

\delta \equiv \ln\frac{x_n}{x_{n+1}} = \gamma\,T_d = \frac{2\pi\gamma}{\omega_d} \qquad (9)

where: $x_n, x_{n+1}$ — the heights of two successive same-side peaks. Why this matters: $\delta$ is directly readable off any oscilloscope or sensor trace, making it the standard experimental route to $\gamma$ .

Step 3. Relate to $Q$ for light damping. When $\gamma \ll \omega_0$ , $\omega_d \approx \omega_0$ , so $\delta \approx 2\pi\gamma/\omega_0 = \pi/Q$ , i.e.

Q \approx \frac{\pi}{\delta} \qquad (\gamma \ll \omega_0)

Why the approximation is safe: the frequency shift is second order in $\gamma/\omega_0$ , so even at $Q = 5$ the error in $\pi/\delta$ is under a few percent. *Verified: consistent with Landau & Lifshitz §25; for the oscillator of worked example 1 ( $\gamma = 2.0\ \mathrm{s^{-1}}$ , $T_d = 0.641$ s), $\delta = 1.28$ and $\pi/\delta = 2.45$ , closing the loop with the exact $Q = 2.5$ .*

Figure

Ringdown of an underdamped oscillator — successive peaks shrink by a fixed ratio

Limiting cases

Case	Result	Meaning
$\gamma = 0$	$x = A\cos(\omega_0 t + \phi)$	frictionless SHM recovered; $Q \to \infty$
$\gamma < \omega_0$ (underdamped)	decaying cosine at $\omega_d < \omega_0$	rings; amplitude drops by $e^{-\gamma T_d}$ per cycle
$\gamma = \omega_0$ (critical)	$(A+Bt)e^{-\gamma t}$	fastest settle, at most one zero crossing
$\gamma > \omega_0$ (overdamped)	two real decays; slow rate $\gamma - \sqrt{\gamma^2-\omega_0^2}$	no oscillation, slower settling than critical
$\gamma \gg \omega_0$	slow root $\to -\omega_0^2/2\gamma$	motion degenerates to a creeping relaxation; inertia irrelevant

Common trap“More damping always stops the motion sooner”

False — and the failure is exactly where intuition trusts itself most. Past the critical point, extra friction slows the return: the overdamped slow root $\gamma - \sqrt{\gamma^2 - \omega_0^2}$ shrinks toward zero as $\gamma$ grows, so a heavily overdamped door closer takes far longer to shut than a critically damped one. If you want the fastest possible settling without overshoot, you want $\gamma = \omega_0$ exactly — not more. Two companion traps: an underdamped oscillator vibrates at $\omega_d = \sqrt{\omega_0^2 - \gamma^2}$ , below the natural frequency, not at $\omega_0$ ; and the damping rate is $\gamma = b/2m$ — dropping that factor of 2 silently corrupts $\omega_d$ , $Q$ , and every decay time you compute afterwards.

Worked examples

1 · Classify and quantify (confidence builder)

A 1.0 kg block on a spring of $k = 100\ \mathrm{N/m}$ moves through oil with drag coefficient $b = 4.0\ \mathrm{kg/s}$ . Which regime is it in? Find the oscillation frequency, the quality factor, and the time for the amplitude to halve.

Step 1. Compute the two rates: $\omega_0 = \sqrt{k/m} = \sqrt{100} = 10\ \mathrm{rad/s}$ ; $\gamma = b/2m = 4.0/2.0 = 2.0\ \mathrm{s^{-1}}$ .

Step 2. Compare: $\gamma < \omega_0$ , so the system is underdamped — it oscillates.

Step 3. Damped frequency: $\omega_d = \sqrt{\omega_0^2 - \gamma^2} = \sqrt{100 - 4} = \sqrt{96} = \boxed{9.8\ \mathrm{rad/s}}$ — just 2% below the natural frequency, even though the ringing dies visibly fast.

Step 4. Quality factor: $Q = \omega_0/2\gamma = 10/4 = \boxed{2.5}$ — a low- $Q$ system; only a few swings survive.

Step 5. Amplitude halving time, from the envelope $e^{-\gamma t}$ : $t_{1/2} = \ln 2/\gamma = 0.693/2.0 = \boxed{0.35\ \mathrm{s}}$ — about half a period ( $T_d = 2\pi/9.8 = 0.64$ s), consistent with $Q = 2.5$ . The energy halves twice as fast, in $0.17$ s.

2 · Designing a critically damped instrument (exam level)

A galvanometer's needle assembly has effective mass $m = 0.80\ \mathrm{kg}$ and restoring stiffness $k = 45\ \mathrm{N/m}$ . (a) Find the drag coefficient $b$ that makes it critically damped. (b) The manufacturer, "for safety," doubles that $b$ . Find the two decay rates and show the needle now settles several times slower.

Step 1. Natural frequency: $\omega_0 = \sqrt{k/m} = \sqrt{45/0.80} = \sqrt{56.25} = 7.5\ \mathrm{rad/s}$ .

Step 2. Critical damping means $\gamma = \omega_0$ , i.e. $b/2m = \omega_0$ , so $b_c = 2m\omega_0 = 2\sqrt{mk} = 2\sqrt{0.80 \times 45} = 2\sqrt{36} = \boxed{12\ \mathrm{kg/s}}$ .

Step 3. At this setting the needle returns as $(A + Bt)e^{-7.5t}$ : the reading is within 1% of final in roughly $0.9$ s, with zero overshoot.

Step 4. Doubling to $b = 24\ \mathrm{kg/s}$ gives $\gamma = 24/1.6 = 15\ \mathrm{s^{-1}}$ — now $\gamma > \omega_0$ : overdamped.

Step 5. The two rates, from (3): $\sqrt{\gamma^2 - \omega_0^2} = \sqrt{225 - 56.25} = \sqrt{168.75} = 12.99\ \mathrm{s^{-1}}$ , so $r_+ = -15 + 12.99 = \boxed{-2.0\ \mathrm{s^{-1}}}$ (slow) and $r_- = -28.0\ \mathrm{s^{-1}}$ (fast).

Step 6. The late-time approach is governed by the slow rate: $2.0\ \mathrm{s^{-1}}$ versus the critical $7.5\ \mathrm{s^{-1}}$ — the "safer" needle creeps to its reading almost four times slower. Doubling the damping degraded the instrument: critical damping is an optimum, not a direction.

Origins · 1 minLord Rayleigh and how this was discovered

Lord Rayleigh

1842–1919 · Cambridge & Terling Place, Essex

John William Strutt, third Baron Rayleigh, gave dissipation its permanent home in mechanics. His two-volume Theory of Sound (1877–78) treated vibrating strings, plates, and air columns with damping built in from the start, and he introduced the dissipation function — a systematic way to fold friction proportional to velocity into the equations of analytical mechanics that is still taught under his name. Rayleigh worked with legendary breadth: he explained why the sky is blue (Rayleigh scattering), co-discovered argon (Nobel Prize, 1904), and set the resolution limit of optical instruments. The damped oscillator formalism of this article — down to the $-b\dot x$ term — is essentially his housekeeping.

History

1660

— Hooke's linear spring law provides the restoring force; friction remains an unmodelled nuisance.

1743

— d'Alembert and Euler develop the theory of linear differential equations with constant coefficients — the characteristic-equation method used here.

1877

— Rayleigh's Theory of Sound systematizes damped vibration and introduces the dissipation function.

1908

— Frahm patents the tuned vibration absorber; engineered damping enters ship and machine design.

1934

— Hydraulic telescopic shock absorbers tuned near critical damping become standard equipment in automobiles.

2015

— LIGO records the ringdown of a newborn black hole — a damped sinusoid at 250 Hz with

Q \approx 3

, fit by exactly the mathematics on this page.

Going further

Texts. Morin, Introduction to Classical Mechanics, Ch. 4 · Taylor, Classical Mechanics, Ch. 5 · Landau & Lifshitz Vol. 1 §25 · Goldstein, Classical Mechanics, Ch. 6 · French, Vibrations and Waves, Ch. 4.

Provenance. Equation of motion, characteristic roots, $\omega_d=\sqrt{\omega_0^2-\gamma^2}$ , the three-regime classification, energy decay $e^{-2\gamma t}$ , log decrement $\delta=\gamma T_d$ , and $Q=\omega_0/2\gamma$ are standard and appear in Morin Ch. 4, Taylor Ch. 5, Goldstein §6, and Landau-Lifshitz §25. Verifier should recompute: roots of $r^2+2\gamma r+\omega_0^2=0$ ; the worked numbers ( $\omega_0=10$ , $\gamma=2$ , $\omega_d=\sqrt{96}\approx9.80$ , $Q=2.5$ ); and $e^{-2\pi}\approx 1/535$ . The RLC mapping $m\to L,\,b\to R,\,k\to1/C$ is textbook standard.

Prerequisites. Simple harmonic motion · Complex exponential representation of oscillations · Amplitude, phase, and angular frequency · Small oscillations about equilibrium. Leads to. Damping regimes · Driven oscillations · Resonance · Transient and steady-state response.

How can (A + Bt)e^(−γt) be the fastest decay when it contains a growing factor of t?

The factor $t$ grows only linearly while $e^{-\gamma t}$ dies exponentially — the product always goes to zero, and quickly. The $Bt$ piece matters mostly at the start, where it lets the solution absorb any initial velocity. The deeper point is the rate: an underdamped system decays at $\gamma < \omega_0$ , and an overdamped system's slow root is also smaller than $\omega_0$ . Critical damping is the unique arrangement where both independent solutions decay at the full rate $\omega_0$ — no slower mode survives to linger. The polynomial prefactor is just the price of the two roots colliding; it never wins against the exponential.

Why does damping make the oscillation slower instead of faster?

Friction opposes the velocity through the entire swing — on the way out and on the way back — so every cycle takes a little longer than the frictionless one, and the frequency drops to $\omega_d = \sqrt{\omega_0^2 - \gamma^2}$ . The complex-plane picture says the same thing: the roots live on a circle of radius $\omega_0$ , so as the real (decay) part grows, the imaginary (frequency) part must shrink. The effect is second order and usually tiny: even the visibly dying oscillator of worked example 1 ( $Q = 2.5$ ) rings only 2% flat of $\omega_0$ .

Where does the factor of 2 in γ = b/2m actually come from?

Pure bookkeeping — but bookkeeping you must keep straight. Defining $\gamma = b/2m$ makes the characteristic roots come out as $-\gamma \pm \sqrt{\gamma^2 - \omega_0^2}$ and the envelope exactly $e^{-\gamma t}$ , with no stray halves anywhere downstream. If you instead define the damping constant as $b/m$ (some texts call that $\beta$ or $2\beta$ — conventions differ), the same physics appears with factors of $\tfrac12$ scattered through $\omega_d$ , $Q$ , and the decay times. Mixing two books' conventions in one calculation is the classic way to be wrong by exactly a factor of 2.

Which decay do I use for half-life questions — e^(−γt) or e^(−2γt)?

Ask what the question actually measures. Anything read off the displacement trace — peak heights, amplitude, "how long until the swing is half as wide" — follows the envelope $e^{-\gamma t}$ and halves in $\ln 2/\gamma$ . Anything about stored energy or total heat dissipated follows $e^{-2\gamma t}$ and halves twice as fast, because energy is quadratic in amplitude. In worked example 1 the amplitude halves in 0.35 s but the energy in 0.17 s. If an exam answer is off from yours by exactly a factor of 2 in the exponent, this is almost certainly where it happened.

Real sliding friction is roughly constant, not proportional to velocity — is −b·ẋ even realistic?

It is the honest model for viscous drag: slow motion through air, oil, or honey, eddy-current braking, and electrical resistance all give a force close to $-b\dot x$ — and it is the only friction law that keeps the equation linear and exactly solvable. Constant-magnitude (Coulomb) friction is genuinely different physics: it makes the amplitude decay linearly, not exponentially, and leaves the mass stuck slightly off-center in a dead band. Engineers still lean on this article's machinery by defining an "equivalent viscous damping" that dissipates the same energy per cycle — a controlled approximation, not a blunder.

Does an overdamped oscillator ever cross equilibrium?

At most once, just like the critically damped case. The solution $C_1 e^{r_+ t} + C_2 e^{r_- t}$ vanishes when $e^{(r_+ - r_-)t} = -C_2/C_1$ , and an exponential takes each value only once — so one crossing, and only if the initial velocity is directed hard toward equilibrium (enough to shoot through before the drag reels the mass back). Released from rest, it never crosses: the displacement just melts monotonically to zero. Endless back-and-forth through equilibrium is the exclusive signature of the underdamped regime.

Is Q counting cycles, radians, or energy foldings? I keep seeing all three claims

All three are the same number viewed through different windows, and the article uses each where it is most natural. $Q$ is the radians of phase accumulated while the energy falls by $e^{-1}$ ; equivalently the amplitude takes $Q/\pi$ full cycles to fall by $e^{-1}$ ; equivalently after $Q$ complete cycles the energy is down by $e^{-2\pi} \approx 1/535$ . All follow from comparing the phase rate $\omega_d \approx \omega_0$ with the energy rate $2\gamma$ and using $Q = \omega_0/2\gamma$ . When you meet driven oscillations, the same $Q$ will reappear as the sharpness of the resonance peak, $\omega_0/\Delta\omega$ — one number, four faces.

My constants C₁ and C₂ came out complex — did I make an error?

No — in the underdamped regime the roots are complex, so the superposition constants must be complex too. The physics enters when you demand that $x(t)$ itself be real: that forces $C_2 = C_1^*$ , which removes two of the four real numbers hiding in $C_1, C_2$ and leaves exactly two — the amplitude $A$ and phase $\phi$ of the real form (4), fixed by initial position and velocity. A quick sanity check after any solution: evaluate $x(0)$ and $\dot x(0)$ and confirm both come out real and match the given data.

Damped Harmonic Oscillator

Where this lives today

The physical picture

The math: one equation, three personalities

Energy decay and the quality factor

Limiting cases

Worked examples

1 · Classify and quantify (confidence builder)

2 · Designing a critically damped instrument (exam level)

Going further

Before this

Where this leads