Home / Learn / Exam prep / IIT JAM Physics / Lesson 1.10
All lessons

Module 1: Mathematical methods · Lesson 1.10 · about 45 min

Fourier series

How any periodic signal is built from sines and cosines: the coefficient formulas and where they come from, even and odd shortcuts, half-range series, what happens at jumps (and the Gibbs overshoot), Parseval's theorem and power, the complex form, and summing famous series.

Key points, meanings and formulas on one page: choose "Save as PDF" in the print box.

Module 1 · about 45 minutes. By the end you will be able to find the Fourier series of the standard waves quickly, use symmetry to skip most of the integrals, say what the series does at a jump, explain the Gibbs overshoot, use Parseval’s theorem for power and RMS values, switch to the complex form, and evaluate sums like ∑1/n2 from a series.

Any repeating signal, however jagged, is a sum of pure tones at whole-number multiples of its frequency. The Fourier coefficients say how loud each tone is.

1. Building a wave from sines

A function with period 2π can be written as

f(x)=a02+∑n=1∞(ancos⁡nx+bnsin⁡nx)

Each term repeats every 2π (and also at shorter intervals), so the sum does too. The term n=1 is the fundamental; n=2,3,… are the harmonics. A guitar note, the voltage out of a rectifier, a square clock pulse: all of them are sums like this. Choose a wave below and add harmonics one at a time.

A handful of terms already gives the shape. Near a jump, the partial sums overshoot by about 9% of the jump however many terms you keep (Gibbs). The triangle wave has no jumps and converges smoothly.

2. The coefficients and why they work

The whole method rests on orthogonality: over one period, the product of two different members of the set {1,cos⁡nx,sin⁡nx} integrates to zero, while a square gives π:

∫−ππcos⁡mxcos⁡nxdx=πδmn,∫−ππsin⁡mxsin⁡nxdx=πδmn,∫−ππsin⁡mxcos⁡nxdx=0

So to pick out one coefficient, multiply the series by that function and integrate: every other term dies.

Start from the series. f(x) = a₀/2 + Σ(an cos nx + bn sin nx).

Multiply by cos mx. Integrate both sides from −π to π.

Orthogonality kills the rest. Every sine term and every cosine with n ≠ m integrates to zero. Only am cos² mx survives, giving π am.

Read off the coefficient. am = (1/π)∫ f(x) cos mx dx. The same with sin mx gives bm.

The constant. With m = 0, cos 0 = 1 integrates to 2π, which is why the series is written with a₀/2: the formula for an then works for n = 0 too, and a₀/2 is the average of f.

an=1π∫−ππf(x)cos⁡nxdx,bn=1π∫−ππf(x)sin⁡nxdx

For period 2L replace nx by nπx/L and 1π∫−ππ by 1L∫−LL.

Symmetry does half the work. If f is even, every bn=0 and an=2π∫0πfcos⁡nxdx. If f is odd, every an=0 and bn=2π∫0πfsin⁡nxdx. Check symmetry before integrating anything.

Worked example: the square wave. f=+1 on (0,π), −1 on (−π,0). It is odd, so

bn=2π∫0πsin⁡nxdx=2nπ(1−cos⁡nπ)={4/(nπ)n odd0n even

and f=4π(sin⁡x+sin⁡3x3+sin⁡5x5+…). Only odd harmonics: a symmetric square wave has no even ones.

Worked example: f=x. Odd again. bn=2π∫0πxsin⁡nxdx=2(−1)n+1/n, so x=2(sin⁡x−12sin⁡2x+13sin⁡3x−…) on (−π,π). Likewise x2=π23+4∑(−1)nn2cos⁡nx.

Finite series. A product of sines and cosines is already a finite Fourier series; use identities, not integrals. sin3⁡x=14(3sin⁡x−sin⁡3x) and 4cos2⁡x=2+2cos⁡2x.

3. What happens at a jump

Dirichlet’s theorem. If f is piecewise smooth, its Fourier series converges to f(x) wherever f is continuous, and to the average 12[f(x−)+f(x+)] at a jump. The ends of the interval count: for ex on (−π,π) the periodic copy jumps from eπ to e−π at x=π, so the series gives cosh⁡π there.

Smoothness and decay. The smoother the periodic function, the faster its coefficients fall: jumps give 1/n (square wave, sawtooth), corners give 1/n2 (triangle wave, |sin⁡x|), a jump only in the second derivative gives 1/n3. Fast decay means few terms are enough.

“With enough terms the Fourier series of a square wave matches it perfectly everywhere.” Not quite. At every fixed point it does converge, but right next to each jump the partial sums always overshoot by about 9% of the jump (the Gibbs phenomenon). Adding terms only squeezes the overshoot closer to the jump. For a jump from −1 to +1 the peak stays near 1.18. Continuous waves such as the triangle wave have no overshoot at all.

Half-range series. A function given only on (0,L) can be extended either evenly (giving a pure cosine series) or oddly (a pure sine series). Both match f inside (0,L), but they behave differently at the ends: the sine series of f=1 gives 0 at x=0, the average of the jump from −1 to 1 in the odd extension. Half-range sine series are exactly the standing-wave modes of a string fixed at both ends.

4. Power: Parseval’s theorem

The mean square of a periodic function is shared among its harmonics:

12π∫−ππf2dx=(a02)2+12∑n=1∞(an2+bn2)

Each harmonic carries half its amplitude squared (the average of sin2⁡ is 12), and cross terms average to zero by orthogonality.

a₀/2
the average (DC) level of the signal
√(an² + bn²)
the amplitude of the n-th harmonic
(an² + bn²)/2
the share of mean-square power the n-th harmonic carries
Physical meaning
The spectrum of a sound is its list of harmonic amplitudes: a clarinet and a flute playing the same note differ only in these numbers. Power and RMS values simply add over harmonics.

Worked example. v=5+4sin⁡ωt+3cos⁡2ωt volts. Mean square =25+8+4.5=37.5, so Vrms=37.5=6.12 V. For the ±1 square wave, the fundamental alone carries 12(4/π)2=8/π2=81% of the power.

Summing series. Evaluating a series at a clever point, or applying Parseval, gives famous sums. The square wave at x=π/2 gives 1−13+15−⋯=π/4. Parseval for the square wave gives 1+19+125+⋯=π2/8. The series of x2 at x=π gives ∑1/n2=π2/6, and Parseval on it gives ∑1/n4=π4/90.

5. The complex form and driven systems

With Euler’s formula (lesson 1.9) the series becomes tidier:

f(x)=∑n=−∞∞cneinx,cn=12π∫−ππf(x)e−inxdx,c±n=12(an∓ibn)

For real f, c−n=cn∗ and c0 is the average. This is the form that grows into the Fourier transform when the period becomes infinite.

Why physicists love it: linear systems. A linear system (a spring, an LCR circuit) responds to each harmonic separately, so you can solve for one sine at a time and add. An oscillator x¨+ω02x=F(t) driven by a periodic force responds to harmonic n with amplitude Fn/|ω02−n2ω2|: it resonates if ω0 matches any harmonic present. A symmetric square wave at 1 Hz drives a 3 Hz oscillator to resonance but not a 2 Hz one. See damped and forced oscillations.

  • f=a0/2+∑(ancos⁡nx+bnsin⁡nx), an=1π∫fcos⁡nx, bn=1π∫fsin⁡nx; orthogonality makes it work; a0/2 is the average.
  • Even f: cosines only. Odd f: sines only. Period 2L: use nπx/L and 1L.
  • Square wave 4π∑oddsin⁡nx/n; x=2∑(−1)n+1sin⁡nx/n; x2=π2/3+4∑(−1)ncos⁡nx/n2.
  • At a jump the series gives the average; next to it the Gibbs overshoot stays near 9% of the jump. Jumps give 1/n coefficients, corners 1/n2.
  • Parseval: mean square =(a0/2)2+12∑(an2+bn2); RMS values add in quadrature over harmonics.
  • Complex form cn=12π∫fe−inx, c−n=cn∗. Linear systems respond harmonic by harmonic.

Practise: exam questions on this lesson

They start easy and climb to top-rank level. Check each answer, then read the full solution. More on the Fourier series practice page.

Before this: lesson 1.9, complex numbers and Euler’s formula. Next: module 2 begins with Newton’s laws, inertial and non-inertial frames. Fourier ideas return in diffraction, where a far-field pattern is the Fourier transform of the aperture, and in the Schrödinger equation, where box states are a half-range sine series. Draw any wave as a sum of turning arrows in fibub’s Fourier visualiser. Going on to GATE? The GATE Physics course continues with Fourier and Laplace transforms.

Check yourself

1The Fourier series of f(x) = x on (−π, π) contains which terms?
2At a jump from −1 to +1, what does the Fourier series converge to?
3As more terms of the series of a square wave are added, what happens to the overshoot next to each jump?

Sources

  • M. L. Boas, Mathematical Methods in the Physical Sciences, 3rd ed. (2006), chapter 7: Fourier series and transforms
  • G. B. Arfken, H. J. Weber and F. E. Harris, Mathematical Methods for Physicists, 7th ed. (2013), chapter 19
  • A. P. French, Vibrations and Waves, MIT Introductory Physics Series (1971), chapter 2 (Fourier analysis of periodic motion)
  • IIT JAM Physics syllabus (Mathematical Methods section), published by the organising IIT each year

Your ticks stay in this browser. Save them as a file on the course home.

Updated 2026-11-22. Not affiliated with IIT or the JAM organising institute. Exam pattern as published for recent years: always check the official JAM website before the exam. All questions are original.