Visual walkthrough — Normal modes — coupled oscillators, normal coordinates

---

Step 1 — Draw the machine and name every arrow

WHAT. Two identical blocks, each of mass $m$, sit on a frictionless line. A wall spring on the left, a wall spring on the right, and a middle spring joining the two blocks. Left and right springs have stiffness $k$; the middle one has stiffness $k_c$ (the c is for "coupling").

WHY. Before any algebra we must fix what each symbol means as a distance. If we skip this, every later sign is a coin-flip. So we mark $x_1$ = how far block 1 slid from its resting spot, $x_2$ = how far block 2 slid, both measured rightward-positive.

PICTURE. In the figure, the blue block is block 1, the orange block is block 2. Green arrows are the displacements $x_1,x_2$ from the dashed equilibrium lines.

---

Step 2 — How much is each spring stretched?

WHAT. A spring's stored energy depends on its stretch — how much longer (or shorter) it is than its relaxed length. We read three stretches off the picture:

• Left spring stretch $=x_1$ (block 1 pulling away from the wall).
• Right spring stretch $=x_2$.
• Middle spring stretch $=x_2-x_1$.

WHY that middle one? The middle spring's two ends are the two blocks. Its new length changes by (how far the right end moved) minus (how far the left end moved) $=x_2-x_1$. If both blocks slide the same way by the same amount, $x_2-x_1=0$: the middle spring is relaxed. Hold that thought — it is the entire secret of the slow mode.

PICTURE. The red bracket highlights $x_2-x_1$. Watch: when the arrows point the same way and equal length, the bracket shrinks to zero.

---

Step 3 — Write the total energy, then let the machine choose its motion

WHAT. Add up the moving energy (kinetic) and the stored spring energy (potential).

$T=\underset{\text{block 1 moving}}{\underbrace{\frac{1}{2}m{\dot{x}}_1^2}}+\underset{\text{block 2 moving}}{\underbrace{\frac{1}{2}m{\dot{x}}_2^2}}$

Here ${\dot{x}}_1$ means "the speed of $x_1$" — how fast that distance is changing per second. (The dot is Newton's shorthand for a rate of change; one dot = speed, two dots = acceleration.)

$V=\underset{\text{left spring}}{\underbrace{\frac{1}{2}k{x}_1^2}}+\underset{\text{right spring}}{\underbrace{\frac{1}{2}k{x}_2^2}}+\underset{\text{middle spring}}{\underbrace{\frac{1}{2}{k}_c(x_2-x_1{)}^2}}$

WHY energy, not forces? Forces need you to track which way each spring pushes — easy to flip a sign under pressure. Energy is just a positive scalar; the machinery of Lagrangian Mechanics then extracts the correct forces automatically, signs and all. We use the rule $\frac{d}{dt}\frac{\partial L}{\partial {\dot{x}}_i}=\frac{\partial L}{\partial x_i},L=T-V.$ Read it as: "rate of change of momentum $=$ the force," where the force is $-\partial V/\partial x_i$.

Turning the crank (differentiate $V$, watch the chain rule bite the $(x_2-x_1{)}^2$ term):

$m{\ddot{x}}_1=\underset{\text{wall pulls back}}{\underbrace{-k{x}_1}}+\underset{\text{middle spring pull}}{\underbrace{k_c(x_2-x_1)}}$ $m{\ddot{x}}_2=\underset{\text{wall pulls back}}{\underbrace{-k{x}_2}}-\underset{\text{same spring, opposite end}}{\underbrace{k_c(x_2-x_1)}}$

PICTURE. The middle-spring term appears in both lines with opposite signs — that is coupling drawn as a single red spring tugging the two blocks in opposite directions.

---

Step 4 — The bold guess: everybody vibrates at ONE frequency

WHAT. We guess a solution where both blocks do simple harmonic motion at the same frequency $\omega$, differing only in amplitude: $x_j(t)=A_j\cos (\omega t+\varphi ).$

• $A_j$ = the amplitude (max swing) of block $j$ — a number, possibly negative.
• $\omega$ = the angular frequency, shared by both blocks.
• $\varphi$ = a common phase (where in the cycle we start).

WHY this shape? A pure cosine is the signature of Simple Harmonic Motion: acceleration proportional to displacement. Differentiating twice gives the beautiful fact ${\ddot{x}}_j=-{\omega }^2{x}_j,$ so acceleration is just $-{\omega }^2$ times position. This turns the messy time-differential equations into plain algebra in the numbers $A_1,A_2$. We are betting such special synchronized motions exist; we will verify by finding them.

PICTURE. Two cosine waves of the same frequency; only their heights (and possibly their up/down flip) differ.

Substituting ${\ddot{x}}_j=-{\omega }^2{x}_j$ and cancelling the common cosine: $(k+k_c-m{\omega }^2)A_1-k_c{A}_2=0$ $-k_c{A}_1+(k+k_c-m{\omega }^2)A_2=0$

---

Step 5 — When does a nonzero motion survive? The secular equation

WHAT. Write the two equations as a matrix squeezing a vector to zero: $\underset{\mathbf{K}-{\omega }^2\mathbf{M}}{\underbrace{\left(\begin{array}{cc}k+k_c-m{\omega }^2& -k_c\\ -k_c& k+k_c-m{\omega }^2\end{array}\right)}}\left(\begin{array}{c}{A}_1\\ A_2\end{array}\right)=\left(\begin{array}{c}0\\ 0\end{array}\right).$

WHY a determinant? A matrix flattens space by a factor called its determinant. If that factor is not zero, the only vector it can send to the origin is the zero vector itself — meaning $A_1=A_2=0$, no motion, boring. To get a real vibration we need the matrix to collapse space flat (determinant $=0$) so that a whole line of nonzero vectors gets crushed to zero. That surviving line is our mode. This is exactly the eigenvalue idea: ${\omega }^2$ is an eigenvalue, $(A_1,A_2)$ its eigenvector.

PICTURE. A unit square getting squashed onto a line — area zero — the moment ${\omega }^2$ hits a special value.

Set the determinant to zero: $(k+k_c-m{\omega }^2{)}^2-k_c^2=0⟹k+k_c-m{\omega }^2=\pm k_c.$

The two sign choices hand us the two frequencies: $\boxed{{\omega }_1^2=\frac{k}{m}}\boxed{{\omega }_2^2=\frac{k+2k_c}{m}}.$

---

Step 6 — What each frequency actually looks like (the two mode shapes)

WHAT. Put each ${\omega }^2$ back to learn the ratio $A_2/A_1$ — the shape of that vibration.

Mode 1, ${\omega }_1^2=k/m$: the bracket equals $+k_c$, so $k_c{A}_1-k_c{A}_2=0\Rightarrow A_2=A_1$. Blocks move together.

Mode 2, ${\omega }_2^2=(k+2k_c)/m$: the bracket equals $-k_c$, so $A_2=-A_1$. Blocks move opposite.

WHY the slow/fast split? In Mode 1 the middle spring's stretch is $x_2-x_1=0$ — it sleeps, contributes no extra force, so the frequency is just the lone-block value $\sqrt{k/m}$. In Mode 2 the middle spring is worked hardest (stretch $=2x_1$), acting like extra stiffness $2k_c$, pushing the frequency up.

PICTURE. Top: symmetric, arrows parallel, middle spring relaxed (drawn gray/limp). Bottom: antisymmetric, arrows opposed, middle spring stretched taut (red).

---

Step 7 — New coordinates that make the equations fall apart (nicely)

WHAT. Invent two combined variables: ${\eta }_1=\frac{1}{\sqrt{2}}(x_1+x_2),{\eta }_2=\frac{1}{\sqrt{2}}(x_1-x_2).$ ${\eta }_1$ = "how much both move together," ${\eta }_2$ = "how much they move apart." (The $\frac{1}{\sqrt{2}}$ is just neat bookkeeping so energy looks clean.)

WHY these two? They are exactly the mode shapes from Step 6. Add the two equations of motion and the middle-spring terms cancel; subtract them and the middle-spring terms double: ${\ddot{\eta }}_1=-{\omega }_1^2{\eta }_1,{\ddot{\eta }}_2=-{\omega }_2^2{\eta }_2.$ Each is a single simple harmonic oscillator — completely independent of the other. The tangled 2-body problem has become two separate 1D springs.

PICTURE. The $x_1$–$x_2$ plane rotated 45° onto the ${\eta }_1$–${\eta }_2$ axes; along each new axis the motion is a lone clean oscillation.

---

Step 8 — Edge and degenerate cases (never leave the reader stranded)

WHAT & WHY — three limits, each a sanity check.

1. No coupling, $k_c\to 0$. Then ${\omega }_2^2\to k/m={\omega }_1^2$: the two frequencies merge. Physically the middle spring vanishes and the blocks become two identical independent oscillators — any motion is allowed, the mode split disappears. This degenerate case is why weak coupling gives near-equal frequencies (Step 9's beats).

2. Rigid coupling, $k_c\to \infty$. ${\omega }_2^2\to \infty$: it costs infinite energy to move the blocks apart, so Mode 2 freezes out. The blocks lock together and only the slow ${\omega }_1$ survives — a single glued mass on the two wall springs.

3. Start exactly on an eigenvector. Release both from $x_1(0)=x_2(0)=a$ with zero speed: then ${\eta }_2(0)=0$ and stays zero forever. Pure single-frequency motion at ${\omega }_1$, coupling spring never touched. Picking an eigenvector as your start excites one mode only.

PICTURE. The two frequencies ${\omega }_1,{\omega }_2$ plotted against coupling $k_c$: they touch at $k_c=0$, and ${\omega }_2$ climbs away as $k_c$ grows while ${\omega }_1$ stays flat.

Recall Check yourself on the limits

As $k_c\to 0$, what happens to the two normal-mode frequencies? ::: They become equal, ${\omega }_2\to {\omega }_1=\sqrt{k/m}$ — the mode split vanishes (degenerate). As $k_c\to \infty$, what happens to Mode 2? ::: Its frequency $\to \infty$; that motion freezes, the blocks lock together and only the slow mode remains.

---

Step 9 — Any real motion is a sum of the two: beats

WHAT. Pull block 1 aside, block 2 at rest ($x_1(0)=a$, $x_2(0)=0$, no initial speed). In normal coordinates ${\eta }_1(0)={\eta }_2(0)=a/\sqrt{2}$, phases zero. Transforming back: $x_1(t)=\underset{\text{slow envelope}}{\underbrace{a\cos \left(\frac{{\omega }_2-{\omega }_1}{2}t\right)}}\underset{\text{fast wiggle}}{\underbrace{\cos \left(\frac{{\omega }_1+{\omega }_2}{2}t\right)}}.$

WHY it looks like breathing. Two nearly-equal cosines added become a fast oscillation inside a slowly pulsing envelope — beats (see Beats and Superposition). Energy pours entirely into block 2, then back to block 1, over and over. Weaker coupling ⇒ closer frequencies ⇒ slower sloshing.

PICTURE. Fast blue oscillation hugged by a slow orange envelope; block 2 (green) fills up as block 1 empties.

---

The one-picture summary

One glance, whole story: two springs + two blocks → energy → the single-frequency guess → determinant $=0$ → two frequencies → two mode shapes (slow same / fast fight) → normal coordinates decouple everything → real motion = their superposition (beats).

Recall Feynman retelling — the whole walkthrough in plain words

Two carts held by springs, with a rubber band tying them. First I just labelled how far each cart slid and noticed one thing: the rubber band only stretches by the difference of the slides. I added up all the bounce-energy and let the energy rule spit out how the carts push each other — and there it was, the rubber band shoving them oppositely. Then I made a wild bet: maybe there are special ways both carts wiggle at one shared speed. Plugging that bet in turned calculus into arithmetic, and a "when does anything survive?" question — the determinant — coughed up exactly two speeds. Looking closer, one speed is the carts moving together (rubber band asleep, so slow) and one is them moving apart (rubber band fighting, so fast). Finally I renamed my variables into "together" and "apart," and the equations magically split into two separate simple springs. So any messy motion I could ever start is just those two simple wiggles added — and if the frequencies are close, you literally watch the energy slosh from one cart to the other and back: beats.

---

See also: Eigenvalues and Eigenvectors · Small Oscillations · Phonons and Lattice Vibrations · Hinglish version