Amplitude, Phase, and Angular Frequency

Before every orchestra concert, the oboe plays a single sustained note — the A above middle C — and a hundred instruments bend their tuning toward it. The oboist can play that note loudly or softly, start it a beat early or a beat late, and none of it matters: the pitch is the same 440 cycles per second every night. Loudness, timing, and pitch are three independent dials, and only one of them belongs to the instrument itself. Simple harmonic motion has exactly the same three dials. Every solution of $\ddot x = -\omega^2 x$ is a sinusoid $x(t) = A\cos(\omega t + \varphi)$ , pinned down by three numbers: the amplitude $A$ (how far it swings), the phase constant $\varphi$ (where in its cycle it was caught at $t=0$ ), and the angular frequency $\omega$ (how fast the cycle turns). The deep result of this article is how the three split up: $\omega$ is welded to the apparatus — the stiffness and the inertia — while $A$ and $\varphi$ are yours to choose, set entirely by how the motion was launched. Strike a tuning fork hard or gently and you change $A$ ; strike it now or later and you change $\varphi$ ; but its pitch $\omega$ you cannot touch without a file and a workbench. Mastering this three-parameter language pays off everywhere oscillations appear. It is the vocabulary of every oscilloscope trace, every AC circuit, every radio symbol your phone decodes — and the two initial-condition formulas derived here, $A = \sqrt{x_0^2 + (\dot x_0/\omega)^2}$ and $\tan\varphi = -\dot x_0/(\omega x_0)$ , are the standard bridge from "how it started" to "what it does."

Key takeaways

A sinusoidal oscillation $x(t) = A\cos(\omega t + \varphi)$ is fully specified by three parameters: amplitude $A$ , phase constant $\varphi$ , and angular frequency $\omega$ .
The angular frequency $\omega$ is fixed by the physical system (restoring stiffness over inertia, e.g. $\omega=\sqrt{k/m}$ ) and is independent of how the motion is started.
The amplitude $A$ and phase $\varphi$ carry the two integration constants of the second-order equation of motion — they are set by $x_0 = x(0)$ and $\dot x_0 = \dot x(0)$ : $A = \sqrt{x_0^2 + (\dot x_0/\omega)^2}$ , $\tan\varphi = -\dot x_0/(\omega x_0)$ .
Angular frequency $\omega$ (rad/s), ordinary frequency $f = \omega/2\pi$ (Hz), and period $T = 2\pi/\omega$ are three units for the same rate — the $2\pi$ between $\omega$ and $f$ is the most common error in the subject.
The amplitude–phase form and the linear-combination form $x = B\cos\omega t + C\sin\omega t$ are the same motion: $A = \sqrt{B^2 + C^2}$ , $\tan\varphi = -C/B$ .
Velocity runs a quarter cycle ahead of displacement and acceleration runs half a cycle ahead: $v$ peaks where $x$ crosses zero, $a$ is $x$ upside down.
Key idea: for $\ddot x = -\omega^2 x$ , the frequency $\omega$ belongs to the system while the amplitude $A$ and phase $\varphi$ belong to the initial conditions: $x(t) = A\cos(\omega t + \varphi)$ with $A = \sqrt{x_0^2 + (\dot x_0/\omega)^2}$ and $\tan\varphi = -\dot x_0/(\omega x_0)$ .

Where this lives today

The electric grid is the world's largest amplitude–phase–frequency experiment. Every generator on a continental grid spins in near-lockstep at $f = 50$ or $60\ \mathrm{Hz}$ , and the phase difference between voltage sinusoids at two ends of a transmission line is what pushes power through it. Grid operators watch this live with phasor measurement units — GPS-timestamped sensors that report the local complex amplitude $\mathcal{A} = Ae^{i\varphi}$ dozens of times per second — because a drifting phase angle is the earliest warning of a blackout in progress. As renewable inverters replace heavy spinning machines in the 2020s, holding those phases synchronized without mechanical inertia has become one of the defining control problems of the energy transition; "grid-forming" inverters are, quite literally, devices whose firmware votes on $\varphi$ . The same triple is how information reaches your phone. Wi-Fi and 5G encode bits by transmitting a carrier wave and modulating its amplitude and phase together: each transmitted symbol is one point $(A, \varphi)$ chosen from a constellation of up to 4096 allowed pairs (4096-QAM), redrawn billions of times per second. Meanwhile GPS receivers track the phase of satellite carriers to a hundredth of a cycle — millimetres of equivalent path length — which is how surveyors and self-driving tractors localize to centimetres. When engineers say "the receiver locked on," they mean it solved, in real time, exactly the problem of this article: extracting $A$ , $\varphi$ , and $\omega$ from a measured sinusoid.

The physical picture

Every formula in this article is hiding inside one picture. Take a point moving uniformly around a circle of radius $A$ , sweeping angle at a steady rate $\omega$ radians per second, and shine a light on it from the side: its shadow on a straight line executes exactly simple harmonic motion. The radius of the circle is the amplitude. The angle the point has reached is the phase $\omega t + \varphi$ — the full argument of the cosine, growing steadily forever. The angle it started at, $\varphi$ , is the phase constant. And $\omega$ is called angular frequency for the most literal reason imaginable: on this reference circle it is an actual angular velocity, radians of cycle swept per second, even when the real mass is sliding back and forth on a perfectly straight track.

Figure

The reference circle — a uniformly rotating arm whose shadow is the oscillation

The picture instantly explains the division of labor between the three parameters. To change $A$ you must move the point to a different circle — that takes energy, and it is exactly what striking harder does. To change $\varphi$ you only redefine when you looked: catch the same rotation a moment earlier or later and the starting angle shifts. But to change $\omega$ you must change how fast the arm turns — and for a mass on a spring the turn rate is dictated by $\sqrt{k/m}$ , by the machine itself. Amplitude and phase are the "how you started it" knobs; angular frequency is the knob welded in place at the factory. One more thing the circle makes obvious: the phase constant on its own has no absolute meaning, because it moves when you move your $t=0$ . Only phase differences are physical — between two oscillators driven side by side, or between displacement and velocity within a single oscillator (a fixed quarter turn, as we will see). Get comfortable reading $\omega t + \varphi$ as "the angle around the cycle," and every sign convention in this subject becomes geometry instead of memorization.

The three dials, precisely

Everything starts from the canonical form of the motion. It answers the question: what is the most general position-versus-time curve a harmonic oscillator can follow?

x(t) = A\cos(\omega t + \varphi) \qquad (1)

where: $A \ge 0$ — the amplitude, the maximum excursion from equilibrium; $\omega$ — the angular frequency in rad/s; $\varphi$ — the phase constant in radians, fixing where in the cycle the motion sits at $t = 0$ ; the whole argument $\omega t + \varphi$ is called the phase. Note what equation (1) claims: every harmonic oscillator — spring, pendulum at small angle, LC circuit — traces this one universal shape, and systems differ only in their three numbers. The frequency is not free. It is imposed by the equation of motion itself,

\ddot x = -\omega^2 x \qquad (2)

where: $\omega^2$ — the ratio of restoring stiffness to inertia: $k/m$ for a mass on a spring, $g/L$ for a small-angle pendulum, $1/LC$ for a charge oscillating in a circuit. The stronger the pull-back per unit displacement and the lighter the thing being pulled, the faster the cycle turns. This is the precise sense in which frequency "belongs to the system."

Derivation AFrom the equation of motion to the three-parameter sinusoid

We are establishing that equation (1) is the general solution of equation (2) — not just one solution among many — and locating exactly where each of the three parameters comes from. The strategy is direct: substitute the candidate form into the ODE, see what survives, and then invoke the counting theorem for linear differential equations to certify that nothing has been missed. The point to watch is the logical asymmetry that falls out: the substitution will force the value of $\omega$ but leave $A$ and $\varphi$ completely unconstrained — the mathematical origin of the physical split between system and initial conditions.

Step 1. Propose $x(t) = A\cos(\omega' t + \varphi)$ with $\omega'$ not yet assumed equal to the $\omega$ in the ODE, and differentiate twice: $\dot x = -A\omega'\sin(\omega' t + \varphi)$ , then $\ddot x = -A\omega'^2\cos(\omega' t + \varphi) = -\omega'^2 x$ . Why a cosine guess is legitimate: we are not assuming the answer — any candidate is allowed, because the uniqueness theorem below will certify whether the surviving family is the whole solution space.

Step 2. Substitute into (2): $-\omega'^2 x = -\omega^2 x$ must hold at every instant, so $\omega'^2 = \omega^2$ , i.e. $\omega' = \omega$ . The equation dictates the frequency. Why $A$ and $\varphi$ escape: they cancelled out of the comparison entirely — the ODE is satisfied identically for every amplitude and every phase constant. The equation has no opinion about them.

Step 3. Count the freedom. Equation (2) is linear, homogeneous, and second order, so its solution space is exactly two-dimensional: the general solution carries precisely two arbitrary constants, no more, no fewer. The pair $(A, \varphi)$ supplies exactly two. Therefore (1) is not merely a solution — it is the general solution, and every motion the oscillator can ever perform is somewhere in this family. Why second order means two constants: integrating $\ddot x$ back to $x$ takes two integrations, each introducing one constant — physically, nature needs to be told both where the mass started and how fast it was moving.

Step 4. Read off the physics. The two constants of integration live in $(A, \varphi)$ and must be fixed by initial data $x_0, \dot x_0$ ; the frequency $\omega$ was consumed by the equation itself in Step 2. The three dials split exactly as advertised: two for the launch, one for the machine. The equivalent two-constant form $B\cos\omega t + C\sin\omega t$ (equation 4) spans the same solution space; substituting it into (2) confirms it satisfies the ODE for all $B, C$ .

One rate, three units

Because the phase must advance by $2\pi$ radians to complete one cycle, the three standard measures of "how fast" are locked together. If a cycle takes time $T$ , the phase advances $2\pi$ in time $T$ :

\omega = \frac{2\pi}{T} = 2\pi f \qquad (3)

where: $T$ — the period, seconds per cycle; $f = 1/T$ — the ordinary frequency, cycles per second (hertz). All three describe one rate. Lab instruments and musicians speak in $f$ ; equations speak in $\omega$ , because $\cos$ and $\sin$ take radians, not cycles. Concert A is $f = 440\ \mathrm{Hz}$ but $\omega = 2\pi \times 440 \approx 2765\ \mathrm{rad/s}$ — same note, different unit, and confusing the two silently injects a factor of $6.28$ into everything downstream. All three dials can be read directly off a position–time trace, which is worth internalizing as a picture — it is exactly what you do at an oscilloscope:

Figure

Anatomy of a sinusoid — where A, T, and the phase constant live on the waveform

Height of the crests above the centerline: $A$ . Crest-to-crest spacing: $T$ , hence $\omega = 2\pi/T$ . And the phase constant is a time shift in disguise: the crest nearest $t=0$ sits at $t = -\varphi/\omega$ , so positive $\varphi$ means the oscillation was caught already past its peak — $\varphi/\omega$ is how long ago the last crest happened.

The other outfit: cosine plus sine

The same motion wears a second, equally common outfit, preferred whenever initial conditions must be imposed, because its two constants appear linearly:

x(t) = B\cos\omega t + C\sin\omega t \qquad (4)

where: $B = x_0$ — the initial displacement (set $t=0$ and the sine dies); $C = \dot x_0/\omega$ — the initial velocity in disguise (differentiate, set $t=0$ ). Equation (4) is what you reach for when solving problems; equation (1) is what you reach for when interpreting them, because it displays the physical amplitude and the timing of the peaks directly. Converting between the outfits is one trig identity. Expanding $A\cos(\omega t + \varphi) = A\cos\varphi\cos\omega t - A\sin\varphi\sin\omega t$ and matching coefficients with (4) gives $B = A\cos\varphi$ and $C = -A\sin\varphi$ , which invert to

A = \sqrt{B^2 + C^2}, \qquad \tan\varphi = -\frac{C}{B} \qquad (5)

where: the quadrant of $\varphi$ is fixed by the pair of signs $(\cos\varphi, \sin\varphi) = (B/A,\, -C/A)$ , never by the tangent alone — the tangent repeats every $\pi$ , so it cannot distinguish $\varphi$ from $\varphi + \pi$ . Geometrically, (5) is the Pythagorean theorem on the reference circle: $B$ and $C$ are the two legs, $A$ the hypotenuse.

Launch conditions to motion

The workhorse formulas of this article translate "how it started" into "what it does":

A = \sqrt{x_0^2 + \left(\frac{\dot x_0}{\omega}\right)^2}, \qquad \tan\varphi = -\frac{\dot x_0}{\omega x_0} \qquad (6)

where: $x_0$ — initial displacement; $\dot x_0$ — initial velocity; $\dot x_0/\omega$ — the initial velocity converted into "displacement units" by the natural rate. Read the amplitude formula physically: position and velocity contribute to the swing size as the two perpendicular legs of a right triangle, so an oscillator released from rest has $A = |x_0|$ , one kicked through equilibrium has $A = |\dot x_0|/\omega$ , and any mixed launch lands in between.

Go deeper · Derivation B · 1 minAmplitude and phase from the launch conditions

Derivation BAmplitude and phase from the launch conditions

The goal is equation (6): the two integration constants $A$ and $\varphi$ , expressed through the measurable launch data $x_0 = x(0)$ and $\dot x_0 = \dot x(0)$ . The strategy is to evaluate the canonical form and its derivative at $t = 0$ , producing two equations in the two unknowns, then eliminate in each direction — squaring and adding kills $\varphi$ to isolate $A$ ; dividing kills $A$ to isolate $\varphi$ . The step to watch is the last: the tangent determines $\varphi$ only up to an additive $\pi$ , and picking the wrong branch flips the sign of the initial velocity — the single most common error in oscillation homework.

Step 1. Evaluate position at $t = 0$ :

x_0 = A\cos\varphi

Why: set $t = 0$ in $x(t) = A\cos(\omega t + \varphi)$ ; the phase collapses to the phase constant.

Step 2. Evaluate velocity at $t = 0$ . Differentiating, $\dot x(t) = -A\omega\sin(\omega t + \varphi)$ , so

\dot x_0 = -A\omega\sin\varphi

Why the minus sign: the derivative of cosine is minus sine — this innocent-looking sign is precisely what will make velocity "lead" position by a quarter cycle in equation (7).

Step 3. Square and add to eliminate $\varphi$ . Divide the second equation by $\omega$ first so both have units of length:

x_0^2 + \frac{\dot x_0^2}{\omega^2} = A^2\cos^2\varphi + A^2\sin^2\varphi = A^2

Why this works: $\cos^2 + \sin^2 = 1$ is the Pythagorean identity — on the reference circle, Steps 1 and 2 say the launch data $(x_0,\, \dot x_0/\omega)$ are the two coordinates of the starting point, and $A$ is its distance from the center.

Step 4. Divide Step 2 by Step 1 to eliminate $A$ :

\frac{\dot x_0}{\omega x_0} = \frac{-A\sin\varphi}{A\cos\varphi} = -\tan\varphi

Step 5. Resolve the quadrant. From Steps 1 and 2 directly: $\cos\varphi = x_0/A$ and $\sin\varphi = -\dot x_0/(\omega A)$ . The pair of signs picks the unique $\varphi$ in $(-\pi, \pi]$ ; the tangent alone cannot. In practice: compute $\mathrm{atan2}(-\dot x_0/\omega,\ x_0)$ , or take the calculator's arctangent and add $\pi$ whenever $x_0 < 0$ . As a check: the released-from-rest case $\dot x_0 = 0,\ x_0 > 0$ gives $A = x_0$ , $\cos\varphi = 1 \Rightarrow \varphi = 0$ , recovering $x = x_0\cos\omega t$ ; the pure-kick case $x_0 = 0,\ \dot x_0 > 0$ gives $A = \dot x_0/\omega$ and $\sin\varphi = -1 \Rightarrow \varphi = -\pi/2$ , recovering $x = (\dot x_0/\omega)\sin\omega t$ .

The phase ladder: displacement, velocity, acceleration

Differentiate the canonical form twice and a beautiful pattern appears — each derivative is another sinusoid of the same frequency, stepped forward a quarter cycle and scaled by one power of $\omega$ :

v(t) = -A\omega\sin(\omega t + \varphi) = A\omega\cos\!\left(\omega t + \varphi + \tfrac{\pi}{2}\right), \qquad a(t) = -\omega^2 x(t) \qquad (7)

where: $v_{\max} = \omega A$ and $a_{\max} = \omega^2 A$ — each rung of the ladder multiplies the amplitude by $\omega$ . The phase shifts say when each quantity peaks within the cycle: velocity runs $\pi/2$ ahead of displacement (it peaks at the equilibrium crossing, a quarter period before the next crest), and acceleration runs $\pi$ ahead — it is the displacement flipped upside down, largest and opposite at the turning points, exactly as $F = -kx$ demands.

Figure

The phase ladder — velocity peaks a quarter period before displacement, acceleration is displacement inverted

The compact notation: complex amplitude

The pair $(A, \varphi)$ is clumsy to carry through algebra — two numbers, one of them trapped inside a trig function. There is a famous repackaging that fuses them into a single complex number:

x(t) = \mathrm{Re}\!\left[\mathcal{A}\,e^{i\omega t}\right], \qquad \mathcal{A} = A\,e^{i\varphi} \qquad (8)

where: $\mathcal{A}$ — the complex amplitude (or phasor), whose magnitude is the physical amplitude and whose argument is the phase constant. The rotating arm of the hero picture is $\mathcal{A}e^{i\omega t}$ drawn in the complex plane, and taking the real part is taking the shadow. This notation turns quarter-cycle phase shifts into multiplications by $i$ and makes damped and driven oscillators almost effortless — the complex exponential article develops it fully.

Energy sits in the amplitude

The three dials also partition the energy bookkeeping. Because speed peaks at $v_{\max} = \omega A$ when the displacement is zero, the total mechanical energy — all kinetic at that instant — is

E = \tfrac{1}{2}m\,v_{\max}^2 = \tfrac{1}{2}m\,\omega^2 A^2 \qquad (9)

where: $m$ — the oscillating mass. The energy depends on the amplitude squared and not at all on the phase: $\varphi$ says when the peaks arrive, not how big they are. Doubling the strike quadruples the stored energy while the pitch stays fixed — the loudness/pitch independence of the concert hall, now as a formula.

Go deeper · Derivation C · 1 minIsochronism, or why amplitude cannot change the clock rate

Derivation CIsochronism, or why amplitude cannot change the clock rate

We are proving the statement that gives pendulum clocks and quartz watches their reason to exist: in the harmonic regime, the period is exactly independent of the amplitude. The strategy is disarmingly short — it rests entirely on the linearity of the equation of motion, so the real content of the derivation is understanding what linearity buys and where real systems forfeit it. Watch the logic direction: we show that rescaling a solution produces another solution with the same $\omega$ , which is precisely the statement that swinging bigger does not mean swinging slower.

Step 1. Let $x(t)$ be any solution of $\ddot x = -\omega^2 x$ , and rescale it by an arbitrary constant: $y(t) \equiv \lambda\,x(t)$ .

Step 2. Substitute: $\ddot y = \lambda\ddot x = \lambda(-\omega^2 x) = -\omega^2(\lambda x) = -\omega^2 y$ . The rescaled motion satisfies the same equation with the same $\omega$ — nothing about the time structure changed. Why linearity is the whole story: both differentiation and multiplication-by- $-\omega^2$ pass through the constant $\lambda$ untouched. A force law with an $x^2$ or $x^3$ term would trap $\lambda$ ( $(\lambda x)^3 = \lambda^3 x^3 \ne \lambda x^3$ ), and the rescaled motion would obey a different equation.

Step 3. Interpret. Doubling the launch displacement doubles $A$ (and the energy quadruples by equation 9), but the period $T = 2\pi/\omega$ is untouched: big swings and small swings tick at identical rates. This property — isochronism — is what lets an oscillator serve as a clock even while friction slowly eats its amplitude.

Step 4. Mark the boundary. For a real pendulum the restoring torque goes as $\sin\theta$ , not $\theta$ ; the linearity is only an approximation near the bottom of the arc. The period acquires an amplitude dependence $T(\theta_0) \approx T_0\left(1 + \theta_0^2/16\right)$ for release angle $\theta_0$ — a $0.05\%$ effect at $\theta_0 = 10°$ , ruinous for navigation-grade timekeeping, negligible for a playground swing. The anharmonic pendulum article makes this correction systematic. The harmonic period is exactly amplitude-independent, and the leading anharmonic correction to the pendulum period is the $\theta_0^2/16$ term noted in Step 4.

Which form answers which question

you want	use	notes
physical amplitude and peak timing	$x = A\cos(\omega t + \varphi)$ , eq. (1)	the interpreting form
to impose initial conditions fast	$x = B\cos\omega t + C\sin\omega t$ , eq. (4)	$B = x_0$ , $C = \dot x_0/\omega$ , linear in the constants
to convert between the two	eq. (5)	quadrant from the signs of $(B, -C)$ , never from tan alone
$A, \varphi$ straight from launch data	eq. (6)	the workhorse
peak speed / peak acceleration	$v_{\max} = \omega A$ , $a_{\max} = \omega^2 A$ , eq. (7)	one power of $\omega$ per derivative
algebra with many phase shifts	$\mathcal{A} = Ae^{i\varphi}$ , eq. (8)	phase shifts become multiplications
energy stored in the motion	$E = \tfrac12 m\omega^2 A^2$ , eq. (9)	$\propto A^2$ , blind to $\varphi$

Limiting cases

Case	Result	Meaning
released from rest ( $\dot x_0 = 0$ )	$A = \lvert x_0\rvert$ , $\varphi = 0$ (or $\pi$ if $x_0 < 0$ )	motion starts at a turning point; pure cosine
kicked through equilibrium ( $x_0 = 0$ )	$A = \lvert\dot x_0\rvert/\omega$ , $\varphi = \mp\pi/2$	pure sine; all launch energy is kinetic
very stiff or very light system	$\omega = \sqrt{k/m} \to \infty$ , $T \to 0$	the cycle blurs; same launch speed buys vanishing amplitude $A \to \dot x_0/\omega$
shift the clock: $t \to t + t_1$	$\varphi \to \varphi + \omega t_1$ , $A$ unchanged	phase constant is convention; only phase differences are physical
large-amplitude pendulum	$\omega$ acquires $A$ -dependence, $T \approx T_0(1 + \theta_0^2/16)$	linearity fails; the clean three-parameter split is a small-oscillation privilege

Common trapA 2 Hz oscillator does not have ω = 2 rad/s

The factor of $2\pi$ between ordinary and angular frequency is the most reliably lost constant in physics homework. The rules: $\omega = 2\pi f$ always; the argument of $\cos$ and $\sin$ must be in radians; $T = 2\pi/\omega = 1/f$ . Feeding $f$ where $\omega$ belongs makes every derived quantity wrong by a power of $2\pi$ — periods off by $6.28\times$ , peak accelerations off by $39.5\times$ — while the algebra looks perfectly clean. The second classic trap lives in equation (6): $\tan\varphi = -\dot x_0/(\omega x_0)$ has two solutions per cycle, and the calculator hands you the wrong one whenever $x_0 < 0$ . Always resolve $\varphi$ from the sign pair $\cos\varphi = x_0/A$ , $\sin\varphi = -\dot x_0/(\omega A)$ .

Worked examples

1 · Reading off the dials of a struck cart (confidence builder)

A $0.50\ \mathrm{kg}$ cart on a spring of stiffness $k = 200\ \mathrm{N/m}$ is at $x_0 = +3.0\ \mathrm{cm}$ when it is struck, giving it velocity $\dot x_0 = -0.80\ \mathrm{m/s}$ (toward equilibrium). Find $\omega$ , $f$ , $T$ , the amplitude, the phase constant, and write $x(t)$ .

Step 1. The system dial first: $\omega = \sqrt{k/m} = \sqrt{200/0.50} = 20\ \mathrm{rad/s}$ . Then $f = \omega/2\pi = 3.18\ \mathrm{Hz}$ and $T = 2\pi/\omega = 0.314\ \mathrm{s}$ . These were fixed before the strike and will survive any strike.

Step 2. Amplitude from equation (6). The velocity leg is $\dot x_0/\omega = -0.80/20 = -0.040\ \mathrm{m}$ , so $A = \sqrt{(0.030)^2 + (0.040)^2} = \sqrt{0.0025} = \boxed{0.050\ \mathrm{m} = 5.0\ \mathrm{cm}}$ — a 3-4-5 triangle on the reference circle.

Step 3. Phase from the sign pair, not the bare tangent: $\cos\varphi = x_0/A = 0.030/0.050 = 0.60 > 0$ and $\sin\varphi = -\dot x_0/(\omega A) = 0.80/(20 \times 0.050) = 0.80 > 0$ , so $\varphi$ sits in the first quadrant: $\varphi = \arctan(0.80/0.60) = \boxed{0.927\ \mathrm{rad} = 53.1°}$ .

Step 4. Assemble: $x(t) = (5.0\ \mathrm{cm})\cos(20\,t + 0.927)$ with $t$ in seconds. Sanity checks: $x(0) = 5.0\cos(0.927) = 3.0\ \mathrm{cm}$ ✓, and the peak speed the cart will ever reach is $v_{\max} = \omega A = 20 \times 0.050 = \mathbf{1.0\ \mathrm{m/s}}$ , comfortably above the launch speed $0.80\ \mathrm{m/s}$ , as it must be — the cart still had potential energy to cash in. Total energy: $E = \tfrac12 m\omega^2 A^2 = \tfrac12(0.50)(400)(0.0025) = \mathbf{0.25\ \mathrm{J}}$ .

2 · The quadrant trap, and a timing question (exam level)

A $250\ \mathrm{g}$ glider on an air track oscillates with measured period $T = 0.500\ \mathrm{s}$ . At $t = 0$ it is at $x_0 = -6.0\ \mathrm{cm}$ moving at $\dot x_0 = -0.90\ \mathrm{m/s}$ (away from equilibrium, into the negative side). Find the spring constant, the amplitude, the phase constant, and the first time the glider passes through equilibrium moving in the $+x$ direction.

Step 1. From the period: $\omega = 2\pi/T = 4\pi = 12.57\ \mathrm{rad/s}$ , so $k = m\omega^2 = 0.250 \times (12.57)^2 = \boxed{39.5\ \mathrm{N/m}}$ .

Step 2. Amplitude: the velocity leg is $\dot x_0/\omega = -0.90/12.57 = -0.0716\ \mathrm{m}$ , so $A = \sqrt{(0.060)^2 + (0.0716)^2} = \boxed{0.0934\ \mathrm{m} = 9.34\ \mathrm{cm}}$ .

Step 3. Phase — here is the trap. The naive route gives $\tan\varphi = -\dot x_0/(\omega x_0) = -(-0.90)/(12.57 \times (-0.060)) = -1.194$ , and a calculator returns $\arctan(-1.194) = -0.874\ \mathrm{rad}$ . That answer is wrong: it implies $\cos\varphi > 0$ , i.e. a positive starting position. Use the sign pair instead: $\cos\varphi = x_0/A = -0.642$ (negative) and $\sin\varphi = -\dot x_0/(\omega A) = +0.767$ (positive) put $\varphi$ in the second quadrant: $\varphi = \pi - 0.874 = \boxed{2.268\ \mathrm{rad} = 130.0°}$ . Why the wrong branch is physically absurd: with $\varphi = -0.874$ the motion would be $x(0) = A\cos(-0.874) = +6.0\ \mathrm{cm}$ — the glider on the wrong side of the track. The tangent cannot see the difference because flipping both $x_0$ and $\dot x_0$ leaves it unchanged.

Step 4. Timing. "At equilibrium moving in $+x$ " means $\cos(\omega t + \varphi) = 0$ with $v = -A\omega\sin(\omega t + \varphi) > 0$ , i.e. $\sin(\omega t + \varphi) = -1$ : the phase must reach $3\pi/2$ (mod $2\pi$ ). First occurrence: $\omega t = 3\pi/2 - 2.268 = 2.444$ , so $t = 2.444/12.57 = \boxed{0.194\ \mathrm{s}}$ — about $0.39$ of a period after launch, which fits the story: the glider first travels out to the far turning point at $-9.34\ \mathrm{cm}$ , turns around, and comes flying back through the middle at its top speed $v_{\max} = \omega A = 1.17\ \mathrm{m/s}$ . Energy check: $E = \tfrac12 m\omega^2 A^2 = \tfrac12(0.250)(157.9)(0.00873) = \mathbf{0.172\ \mathrm{J}}$ .

Origins · 1 minJoseph Fourier and how this was discovered

Joseph Fourier

1768–1830 · École Polytechnique and Grenoble

Fourier — orphaned son of a tailor, briefly destined for the priesthood, nearly guillotined in the Revolution, scientific advisor on Napoleon's Egyptian expedition, then prefect of the Isère — did his great physics in stolen hours while running a French province. Studying how heat spreads, he made the audacious claim (1807) that any periodic signal, however jagged, is a sum of sinusoids, each fully described by the three numbers of this article: an amplitude, a phase, and a frequency. Lagrange and Laplace objected — surely a corner cannot be built from smooth cosines? — and held up publication for years; Fourier answered by computing coefficient after coefficient and won. His theorem is why the amplitude–phase–frequency triple is not merely a fact about masses on springs but the universal coordinate system for signals: your phone decodes Wi-Fi, compresses music, and renders this page's images in the basis Fourier built. He also, incidentally, gave the first scientific description of what is now called the greenhouse effect.

History

1602

— Galileo, timing swinging lamps against his pulse, reports the pendulum's isochronism: the period does not care about the swing size — the first hint that frequency belongs to the system, not the launch.

1673

— Huygens' Horologium Oscillatorium turns that fact into the pendulum clock and derives the period formula, founding precision timekeeping on amplitude-independence.

1748

— Euler's Introductio treats sine and cosine as functions of a continuously growing angle — the reference-circle picture — and his formula

e^{i\theta} = \cos\theta + i\sin\theta

seeds the phasor.

1807

— Fourier announces that arbitrary periodic signals decompose into sinusoids labelled by amplitude, phase, and frequency; the trio becomes the universal language of oscillation.

1893

— Steinmetz introduces complex phasors to electrical engineering; power grids are computed as arithmetic on amplitude-and-phase pairs ever since.

Going further

References. Morin, Introduction to Classical Mechanics, ch. 4 (Oscillations) · Goldstein, Classical Mechanics, 3rd ed., ch. 6 · Landau & Lifshitz, Mechanics, Vol. 1, §21 · Taylor, Classical Mechanics, ch. 5 Key results used throughout this article: (i) $A=\sqrt{x_0^2+\dot x_0^2/\omega^2}$ and $\tan\varphi=-\dot x_0/(\omega x_0)$ follow from $x=A\cos(\omega t+\varphi)$ evaluated at $t=0$ ; (ii) $\omega=2\pi f=2\pi/T$ ; (iii) the seconds-pendulum length $L=g/\omega^2\approx0.994\,\mathrm{m}$ with $\omega=\pi\,\mathrm{rad/s}$ and $g=9.81\,\mathrm{m/s^2}$ ; (iv) $E=\tfrac12 m\omega^2 A^2$ . Worked-example numbers on this page ( $A = 5.0$ cm, $\varphi = 0.927$ rad, $E = 0.25$ J for the struck cart; $k = 39.5$ N/m, $A = 9.34$ cm, $\varphi = 2.268$ rad, $t = 0.194$ s for the glider) are illustrative arithmetic recomputed from the boxed formulas.

Builds on: Simple Harmonic Motion · Small Oscillations About Equilibrium. Leads to: Energy in Simple Harmonic Motion · Complex Exponential Representation of Oscillations · Anharmonic Pendulum and Amplitude Dependence · Driven Oscillations.

Why is ω called an angular frequency when the mass moves in a straight line? Where is the angle?

The angle lives on the reference circle. Any SHM can be generated as the shadow of a point moving uniformly around a circle of radius $A$ , and $\omega$ is that point's genuine angular velocity — radians swept per second. The "angle" $\omega t + \varphi$ is how far around the cycle the motion has progressed, with $2\pi$ radians meaning one full oscillation. This is also why $\omega$ must be fed to $\cos$ and $\sin$ in radians: the trig functions are geometrically functions of arc angle. If you prefer, think of $\omega$ as "cycle-progress rate measured in radians" and $f$ as the same rate measured in whole cycles; the circle is what makes the radian version natural.

The formula tan φ = −v₀/(ωx₀) gives two possible angles. How do I pick the right one without memorizing rules?

Go back one step from the tangent. The launch conditions give you the two components separately: $\cos\varphi = x_0/A$ and $\sin\varphi = -\dot x_0/(\omega A)$ . That sign pair points to exactly one quadrant — no ambiguity, nothing to memorize. The tangent is only the ratio of these, and a ratio forgets the two signs individually: $(-6, +7)$ and $(+6, -7)$ have the same tangent but are opposite motions. Practical recipe: compute $\mathrm{atan2}(-\dot x_0/\omega,\, x_0)$ on anything that has it, or take the calculator's arctan and add $\pi$ whenever $x_0 < 0$ . Worked example 2 shows the trap sprung and disarmed.

Can the amplitude be negative? My algebra produced A = −5 cm.

By convention, no: $A$ is defined as the maximum excursion, so $A \ge 0$ . What your algebra produced is a perfectly valid pair — a negative prefactor is just a phase shift in disguise, because $-\cos\theta = \cos(\theta + \pi)$ . So $-5\cos(\omega t + \varphi)$ is the same motion as $+5\cos(\omega t + \varphi + \pi)$ : fold the sign into the phase constant and move on. This is also why the pair $(A, \varphi)$ is unique only once you fix conventions ( $A \ge 0$ , $\varphi \in (-\pi, \pi]$ ). Nothing physical hangs on the choice; energy, for instance, sees only $A^2$ .

If φ changes whenever I restart my stopwatch, how can it be a physical quantity at all?

On its own, it isn't — and recognizing that is the point. Shifting your time origin by $t_1$ shifts $\varphi$ by $\omega t_1$ , so the phase constant is a bookkeeping convention, like where you put the origin of your x-axis. What is physical: phase differences. The difference between two oscillators (do they peak together or in opposition?) survives any clock change, and it is measurable — it decides whether two speakers reinforce or cancel, how much power flows down a transmission line, and the shape of a Lissajous figure on an oscilloscope. Within one oscillator, the quarter-cycle offset between $x$ and $v$ is likewise clock-proof. Physicists say phase is "defined up to a gauge": fix the convention once, then only differences matter.

What does "velocity leads displacement by π/2" actually mean? Velocity can't know the future.

Nothing is predicting anything — "leads" is a statement about where peaks sit within a repeating cycle, not about causality. Watch one full period: the velocity hits its maximum as the mass whips through equilibrium, and a quarter period later the displacement hits its maximum at the turning point (where the velocity has died to zero). Since the pattern repeats forever, every velocity peak precedes the next displacement peak by $T/4$ — that is all "leading" means. The algebra says the same: $v = A\omega\cos(\omega t + \varphi + \pi/2)$ has the same shape as $x$ with its phase advanced by $\pi/2$ . For a periodic signal, "a quarter cycle early" and "three quarters late" are the same statement.

Why do physicists insist on ω when every instrument, tuner, and datasheet reports f in hertz?

Because the mathematics runs on radians. The derivative of $\cos(\omega t)$ is $-\omega\sin(\omega t)$ only when the argument is in radians; write the motion with $f$ and every derivative sprouts a $2\pi$ : $x = A\cos(2\pi f t)$ , $v = -2\pi f A\sin(2\pi ft)$ , $a = -(2\pi f)^2 x$ . Using $\omega$ hides the constant once and for all, so equations stay clean: $v_{\max} = \omega A$ , $E = \tfrac12 m\omega^2 A^2$ , $\ddot x = -\omega^2 x$ . The two communities simply optimized differently: instruments count whole cycles (easy to measure), theory differentiates (easy in radians). The tax is one conversion, $\omega = 2\pi f$ , paid at the border — and forgetting to pay it is the subject's most common error.

Is the frequency really independent of amplitude in the real world, or only on paper?

It is exact only while the restoring force is exactly linear, and no real force is. A pendulum's torque goes as $\sin\theta$ , so its period grows with swing angle — by about $\theta_0^2/16$ fractionally, a $0.05\%$ effect at $10°$ that pendulum-clock makers fought for centuries. A guitar string bent hard goes audibly sharp; a MEMS resonator driven too hard detunes. The honest statement: every stable equilibrium looks harmonic close enough to the bottom of its potential well, so $\omega$ is amplitude-independent in the small-amplitude limit, with corrections growing as powers of the amplitude. That limit is astonishingly useful — quartz watches sit in it to parts per million — but it is a limit. The anharmonic pendulum article computes the corrections systematically.

How do I read A, ω, and φ off an oscilloscope trace or a position-time graph?

Three rulers, in order. Amplitude: measure peak-to-peak height and halve it — using the full swing by mistake doubles $A$ and quadruples your energy estimate. Frequency: measure the time between successive crests (or, more accurately, across ten crests, divided by ten); that is $T$ , and $\omega = 2\pi/T$ . Phase: locate the crest nearest your $t = 0$ line; if it sits at time $t_{\mathrm{peak}}$ , then $\omega t_{\mathrm{peak}} + \varphi = 0$ , so $\varphi = -\omega t_{\mathrm{peak}}$ — a crest before zero means positive $\varphi$ (the cycle was caught past its peak). Cross-check with the intercept: the trace should cross $t = 0$ at height $A\cos\varphi$ with slope of sign $-\sin\varphi$ . Those two checks catch nearly every sign error.

Amplitude, Phase, and Angular Frequency

Where this lives today

The physical picture

The three dials, precisely

One rate, three units

The other outfit: cosine plus sine

Launch conditions to motion

The phase ladder: displacement, velocity, acceleration

The compact notation: complex amplitude

Energy sits in the amplitude

Which form answers which question

Limiting cases

Worked examples

1 · Reading off the dials of a struck cart (confidence builder)

2 · The quadrant trap, and a timing question (exam level)

Going further

Before this

Where this leads