The frictionless idealization
In the last guide you met the spring-mass-damper system and its electrical twin, the RLC circuit, tied together by the mechanical-electrical analogy. The full mechanical law was m x'' + c x' + k x = 0: mass times acceleration, plus a damping term, plus the spring's restoring pull. Now we make one bold simplification — we set the damping c to zero. No air resistance, no friction in the coils, no resistor bleeding off energy. What survives is the cleanest oscillator in all of physics.
What remains is m x'' + k x = 0, or after dividing by the mass, x'' + (k/m) x = 0. This is a homogeneous second-order equation with constant coefficients and no first-derivative term at all. The motion it describes is called simple harmonic motion — 'simple' because nothing complicates it, 'harmonic' because the answer turns out to be a single pure tone, one sine wave with no overtones. It is the bare skeleton on which the next three guides will hang friction, forcing, and resonance.
Solving it: two imaginary roots
You already know the machinery from the second-order rung: guess x = e^(rt) and the equation x'' + (k/m) x = 0 collapses to the characteristic equation r^2 + (k/m) = 0. Solving for r gives r^2 = -(k/m), and since k and m are both positive, the right side is negative — there is no real square root. The roots are a pure imaginary pair, r = +-i*omega, where omega = sqrt(k/m). This is the complex-conjugate-roots case in its purest form, with a real part of exactly zero.
A real-part of zero is exactly what makes the motion neither grow nor decay. When the roots are alpha +- i*beta, the general solution carries a factor e^(alpha t) out front: a positive alpha would blow the oscillation up, a negative one would damp it down. Here alpha = 0, so e^(0*t) = 1, and that envelope is flat. The oscillation rings at constant amplitude, the same height forever — the unmistakable fingerprint of an undamped system.
Feeding r = +-i*omega through the standard recipe for complex roots gives the general solution x(t) = C1 cos(omega t) + C2 sin(omega t). Both pieces are genuine solutions, and by the superposition principle every combination of them is too. The two free constants C1 and C2 are the two you expect from a second-order equation; you pin them down with an initial value problem — typically the starting position x(0) and starting velocity x'(0).
One wave, two faces: amplitude and phase
The form C1 cos(omega t) + C2 sin(omega t) is correct but it hides the picture. A sum of a sine and a cosine at the same frequency is, remarkably, just a single shifted sine. There is an identity that lets you fold the two into one: C1 cos(omega t) + C2 sin(omega t) = A cos(omega t - phi). The same motion, re-dressed. This second form is the amplitude-phase form, and it is the one your eye actually wants.
C1 cos(wt) + C2 sin(wt) = A cos(wt - phi)
A = sqrt(C1^2 + C2^2) <- the amplitude
tan(phi) = C2 / C1 <- the phase angle
A : how far it swings (the peak displacement)
phi : where in the cycle it starts (a time shift)
w : how fast it cycles (w = omega = sqrt(k/m))Now each symbol carries plain meaning. The amplitude A = sqrt(C1^2 + C2^2) is how far the mass swings from rest to its farthest point — the size of the motion. The phase phi tells you where in its cycle the oscillation was when the clock started: it slides the whole wave left or right in time without changing its shape. Together amplitude and phase are the two numbers that distinguish one swing from another. Crucially, neither of them touches the frequency: change how hard you pluck the spring or when you let go, and A and phi change, but the rate of oscillation does not budge.
The natural frequency: a property of the machine
That stubborn quantity omega = sqrt(k/m) is the natural frequency, the single most important number an oscillator owns. It is the rate at which the system wants to ring when you disturb it and step back. Notice what it depends on: the stiffness k and the mass m, and nothing else — not the amplitude, not the phase, not how you started it. A stiffer spring (larger k) rings faster; a heavier mass (larger m) rings slower. The system has its own pitch built in, exactly as a tuning fork or a wine glass does.
It pays to keep two related words straight. The omega in cos(omega t) is the angular frequency, measured in radians per second; the everyday frequency f = omega/(2*pi) counts full cycles per second (hertz), and the period T = 2*pi/omega is the time for one complete swing. They are three names for the same rhythm. Through the mechanical-electrical analogy this very same omega = 1/sqrt(LC) is the natural frequency of an LC circuit — the rate at which charge sloshes between the capacitor and the inductor with no resistor to stop it. One formula, two physical worlds.
Where the energy goes — and why it never leaves
There is a deeper way to see why the amplitude stays fixed forever, and it does not even need the formula for x(t). Multiply m x'' + k x = 0 by the velocity x' and you can recognize the result as the time derivative of E = (1/2) m (x')^2 + (1/2) k x^2 being zero. That quantity E — kinetic energy plus spring potential energy — is a conserved quantity: it does not change as the motion runs. With no damping there is nowhere for energy to go.
Picture the energy sloshing back and forth. At the extremes of the swing the mass is momentarily still — all the energy is stored in the stretched or squeezed spring as potential. Flying through the centre the spring is relaxed but the mass is at top speed — all the energy is kinetic. Between these the total never changes; it merely trades costume. That endless, lossless exchange is precisely what holds the amplitude A constant. The moment you add even a whisper of damping, E starts to drain, A shrinks, and you have left simple harmonic motion for the underdamped world of the next guide.