🧱 What is y(t)?
\( y(t) \) is the thing changing over time — not a single number, but an entire curve. Each solution is a different story of how the system behaves.
Think of \( y(t) \) like the plot of a movie. At any instant \( t \) you can pause and read one value, but the real meaning is the whole arc from start to finish. Each different solution to a differential equation is a different movie obeying the same genre rules.
🔩 Position of a mass on a spring
🌡️ Temperature in a cooling rod
🌊 Amplitude of a water wave
⚡ Voltage across a capacitor
〰️ Current through an inductor
📡 RF signal waveform in time
The geometric picture
Draw \( t \) horizontal, \( y \) vertical. Every solution is a drawable curve. Solving a differential equation means finding which curves are allowed by the rules.
\( y(t) \) is not a number — it's a whole curve. Each solution is one valid life-story of the system.
⚙️ What is a Differential Equation?
A differential equation doesn't give you \( y \) directly — it gives you the law of change. It's a rulebook, not an answer.
Imagine watching a ball roll on a curved surface. You can't see where it is, but you know one rule: wherever it is, gravity pulls it downhill. That rule is the differential equation. Finding the path is solving it.
Read this as: "the curvature of \( y \) equals the negative of its current height." Rearrange to see the geometry:
curvature = negative of height
When \( y \) is positive, the curve bends downward. When negative, it bends upward. A curve that constantly bends back toward zero traces a sine wave. That's why \(\sin(t)\) and \(\cos(t)\) are the solutions.
Think of a ruler held flat, then bent slightly. The rulebook says: "the more you bend me upward, the more I spring back down." That self-correcting tension is \( y'' = -y \).
\( c_1 \) and \( c_2 \) are free constants encoding initial conditions.
A differential equation is a rulebook about curvature and change. Solving it means finding every curve that obeys the rules — the full family, not just one.
🌊 What is Superposition?
If two separate behaviors each obey the rules, then any weighted mix of them also obeys the rules.
Picture two ripples on a pond. Each ripple by itself follows the laws of water waves. When they meet, they add — and the combined shape also follows the same laws. No new physics needed. That's superposition.
Let \( L \) be a linear operator. If:
$$ L(y_1) = 0 \quad \text{and} \quad L(y_2) = 0 $$Then for any constants \( c_1, c_2 \):
$$ L(c_1 y_1 + c_2 y_2) = 0 $$Why the constants matter
You're not just adding solutions — you can scale them first, then add. Doubling one wave and halving another still gives a valid wave. This scaling freedom is what lets you match any initial condition.
Superposition: valid behaviors can be freely scaled and added, and the result is still valid. LEGO pieces that fit always snap together.
🔥 Why Superposition Works: Linearity
Superposition works because the system treats each input independently — solutions never interact or "see" each other.
Process inputs together OR separately — same result.
Scaling the input scales the output by exactly the same factor.
The proof in four lines
Zero absorbs everything. No mixing, no cross-terms, no surprises. The whole argument rests on \( 0 + 0 = 0 \).
Highway analogy: Linearity is like a highway where each lane flows independently. Cars in lane 1 don't affect lane 2. Count each separately, then add. Nonlinearity would be lanes merging and causing pile-ups — the individual counts no longer predict combined traffic.
Linearity = no interaction between solutions. The proof fits in four lines and the entire argument rests on \( 0 + 0 = 0 \).
🚫 When It Breaks: Nonlinear Systems
Nonlinear terms create cross-interactions between solutions. The combined behavior is no longer the sum of the parts.
Try plugging in \( y_1 + y_2 \) and watch the squared term:
This \( 2y_1 y_2 \) is a product of two different solutions — it never appears when processed separately.
Strangers
Solutions never meet. Add freely. Superposition holds.
Interacting
Adding solutions changes the system. Superposition fails.
Mixing analogy: Linear is water + water = exactly 2 cups. Nonlinear is baking soda + vinegar — the result is more and different than the sum of parts because the ingredients react with each other.
Nonlinearity introduces cross-terms. These break additivity, which breaks superposition. The whole is no longer the sum of its parts.
🧠 What Does "= 0" Really Mean?
Zero doesn't mean nothing is happening. It means there is no outside force — all activity comes from within the system.
"= 0 means y = 0, nothing is moving."
"= 0 means no external input is pushing. Motion comes from stored energy."
The swing analogy
Swing still moves. Gravity + momentum keep it going. "=0" means no new push right now.
\( f(t) \) is the external push. Energy is being injected continuously.
Electrical translation 🔌
Capacitor discharging. No source. Stored energy bleeding through R.
External source \(V_s(t)\) connected. Energy injected from outside.
"= 0" means self-contained. The system runs on stored energy. It might be oscillating wildly — it's just alone, with no outside driver.
⚖️ What Does "Balancing" Mean?
The equation is a dynamic balance — internal forces constantly cancel each other, producing motion in the process.
For \( y'' + y = 0 \), i.e., \( y'' = -y \), trace the geometry through one full cycle:
The two terms in \( y'' + y \) represent two competing tendencies: \( y'' \) is inertia (keep going) and \( y \) is restoring force (come back). Setting their sum to zero means these exactly cancel at every instant — producing oscillation.
Balance = internal effects canceling dynamically. Not stillness — active opposition that produces motion.
🔌 Full Circuit Example
Let's ground everything in a circuit you'll actually build — the series RLC. Every concept above lives inside this one equation.
Here \( y(t) = i(t) \) = current. Three terms, three physical effects competing.
Inductor's inertia. Opposes any change in current.
Resistor's drag. Dissipates energy as heat.
Capacitor's restoring force. Pushes current back.
Setting the sum to zero IS KVL. Three competing internal effects balance out with no external source.
Three behaviors from one equation
Oscillations slowly dying. Energy bouncing between L and C, R bleeding it off.
Returns to zero fastest without oscillating. Used in galvanometers, door closers.
Sluggish decay. Energy burned in R before L-C can exchange it.
Each case is a superposition of two basis solutions. The constants \( A, B \) are chosen to match initial conditions.
Now add a source
Right side non-zero. Decompose:
$$ i(t) = \underbrace{i_h(t)}_{\text{natural response}} + \underbrace{i_p(t)}_{\text{forced response}} $$The homogeneous part uses superposition of basis solutions. The particular part matches the source. This decomposition is only possible because the system is linear.
RLC = three terms balancing = \( L(i)=0 \). Add a source → right side \( \neq 0 \). Superposition splits the answer into "what the circuit does naturally" + "what the source forces it to do."
💥 Why Zero is Special
Zero is the only value that survives addition unchanged — and that's exactly why it's the pivot point for superposition.
Zero added to zero stays zero. The solution set is closed.
The combined solution solves a different equation — one with 2× the forcing.
Geometric interpretation
All solutions to \( L(y) = 0 \) form a vector space — a flat subspace through the origin. You can add any two members and scale any member and stay inside. Solutions to \( L(y) = f \) do NOT form such a space — they form a shifted copy, and adding two of them shifts you off it.
Shadows analogy: Two shadows from separate lights can overlap — the combined shadow is still a shadow (= 0 light). But combine two lit regions and you get twice the light. Homogeneous = working with shadows. Forced = working with light.
\( 0 + 0 = 0 \) always. Solutions to \( L(y)=0 \) form a closed vector space. Any combination stays inside it. Solutions to \( L(y)=f \) do not.
🧠 The Full Unified Picture
Everything we've covered is one connected idea. Here's how it locks together.
- 1System — A differential equation defines rules of change. It constrains which \( y \)'s are possible.
- 2y(t) — Each solution is one valid behavior over time — a whole curve, not a number.
- 3Linear vs Nonlinear — Linear = no cross-interaction. Nonlinear = solutions interfere when combined.
- 4Homogeneous vs Forced — \(=0\) = self-contained natural response. \(=f(t)\) = externally driven.
- 5Superposition — Works when BOTH hold: linear AND homogeneous. Then any combination of solutions is still a solution.
Ultimate mental model
No external force. Solutions don't interact. Combine freely.
LEGO pieces — always snap together perfectly.
External input OR interactions. Combining changes the rules.
Pieces melt into something new.
Where this leads
A linear homogeneous system is one where all behavior comes from the system itself, and because nothing interacts or is added externally, solutions can be combined freely without changing the rules.