Fourier Series To use the Fourier Series, signal π₯(π‘) must be periodic with period ππ seconds (or seconds/cycle) and fundamental frequency ππ Hertz (cycles/seconds), where ππ = 1/ππ In that case you can write the signal as a sum of cosines and sines, like this ∞ π₯(π‘) = ∑ π΄π πππ (2ππππ π‘) + π΅π π ππ(2ππππ π‘) π=0 Note that k is an integer greater than or equal to zero, π΄π and π΅π are real numbers that can be positive or negative. The cosine terms will give you the even part of x(t), while the sine terms will give you the odd part of x(t). You can find the π΄π and π΅π coefficients using the equations: π΄π = 2 ππ ∫ π₯(π‘)πππ (2ππππ π‘)ππ‘ ππ π‘=0 2 ππ π΅π = ∫ π₯(π‘)π ππ(2ππππ π‘)ππ‘ ππ π‘=0 Those equations are only for positive k. At k = 0 use π΄0 = 1 ππ ∫ π₯(π‘)ππ‘ ππ π‘=0 π΅0 = 0 An alternative way to write x(t) is ∞ π₯(π‘) = ∑ πΆπ πππ (2ππππ π‘ + ππ ) π=0 In this case k is an integer greater or equal to zero, πΆπ is a real number greater or equal to zero, and ππ is a real number we usually write between – π and π radians. You can find πΆπ and ππ from π΄π and π΅π like this: πΆπ = √π΄2π + π΅π2 −π΅π ππ = π‘ππ−1 ( ) π΄π π΄π = πΆπ cos(ππ ) π΅π = −πΆπ sin(ππ ) Usually we use the complex exponential form of the Fourier Series, which looks like this ∞ π₯(π‘) = ∑ ππ π +π2ππππ π‘ π=−∞ Note that k is an integer that can be positive, negative or zero, j is the square root of -1, and ππ is in general complex. When x(t) is real (which is always is in EE 3400) then the negative harmonics have a special relationship with the positive harmonics because ∗ ππ = π−π where the star operator means complex conjugate (change sign of imaginary part). You can find ππ using ππ = 1 ππ ∫ π₯(π‘)π −π2ππππ π‘ ππ‘ ππ π‘=0 You can convert between ππ and the other Fourier Series coefficients like this: πΆπ = 2 |ππ | π΄π = 2 π πππ(ππ ) ππ = πππππ(ππ ) π΅π = −2 πΌππππππππ¦(ππ ) Fourier Transform If x(t) is not periodic, but it is finite energy, then you cannot use the Fourier Series, but you can use the Fourier Transform. The energy of a signal is ∞ πΈπππππ¦ = ∫ π₯ 2 (π‘)ππ‘ −∞ The Fourier Transform equations are ∞ π₯(π‘) = ∫ π(π)π +π2πππ‘ ππ −∞ ∞ π(π) = ∫ π₯(π‘)π −π2πππ‘ ππ‘ −∞ Here f is a real number, that can be both positive and negative. When x(t) is real (as it always is in EE 3400) the negative frequencies will always be the complex conjugate of the positive frequencies π(π) = π ∗ (−π) Combining Fourier Series and Fourier Transform If x(t) is the sum of a finite energy (aperiodic) signal, π¦(π‘), and a periodic signal, π§(π‘), which has a fundamental frequency π0 like this π₯(π‘) = π¦(π‘) + π§(π‘) And suppose the Fourier Transform of π¦(π‘) is π(π) and the Fourier Series of π§(π‘) is ππ then we speak of the Fourier Transform of x(t) as ∞ π(π) = π(π) + ∑ ππ πΏ(π − ππ0 ) π=−∞ is not periodic, but it is finite energy, then you cannot use the Fourier Series, but you can use the Fourier Transform. The energy of a signal is Fourier Transform Theorems Discussed in EE 3400 ∞ Fourier Transform π₯(π‘) π(π) = ∫π‘=−∞ π₯(π‘)π −π2πππ‘ ππ‘ Linearity ππ₯(π‘) + ππ¦(π‘) ππ(π) + ππ(π) Time Delay/Advance π₯(π‘ − π) π −π2πππ π(π) Time Expansion/Compression π₯(ππ‘) 1 π π (π) |π| Duality π₯(π‘) → π(π) → π₯(−π‘) → π(−π) → π₯(π‘) Derivative πΉπ π ππ‘ πΉπ πΉπ π₯(π‘) π‘ πΉπ π2πππ(π) 1 Integral ∫−∞ π₯(πΌ)ππΌ π2ππ π(π) Convolution π₯(π‘) ∗ π¦(π‘) π(π)π(π) Multiplication π₯(π‘)π¦(π‘) π(π) ∗ π(π) Even π₯(π‘) = π₯(−π‘)πππ πππ π‘ π(π) ππ ππππ Odd π₯(π‘) = −π₯(−π‘)πππ πππ π‘ π(π) ππ πππππππππ¦ Real π₯(π‘)ππ ππππ πππ πππ π‘ π(π) = π ∗ (−π) Common Fourier Transform Pairs Used in EE 3400 Square Pulse π(π) = π πππ(π) x(t) is one volt high, 1 second wide, centered at t=0 Triangular Pulse π(π) = π πππ 2 (π) x(t) is one volt high, 2 seconds wide, centered at t=0 One-Sided Decaying Exponential π₯(π‘) = π −ππ‘ π’(π‘) 1 π(π) = π+π2ππ Two-Sided Decaying Exponential π₯(π‘) = π −π|π‘| 1 π(π) = π+π2ππ Bell Shaped Curve π₯(π‘) = π −ππ‘ 2 π π(π) = √π π − π2 π2 π DC π₯(π‘) = 1 π(π) = πΏ(π) Impulse π₯(π‘) = πΏ(π‘) π(π) = 1 Cosine π₯(π‘) = πππ (2πππ π‘) 1 1 π(π) = 2 πΏ(π − ππ ) + 2 πΏ(π + ππ ) Sine π₯(π‘) = π ππ(2πππ π‘) π π π(π) = − 2 πΏ(π − ππ ) + 2 πΏ(π + ππ ) Sinusoid π₯(π‘) = π΄πππ (2πππ π‘ + π) π(π) = π΄π ππ 2 πΏ(π − ππ ) + π΄π −ππ π΄ 2 πΏ(π + ππ ) Sign π₯(π‘) = { −1 πππ π‘ < 0 1 πππ 0 < π‘ −π π(π) = ππ Unit Step π₯(π‘) = { 0 πππ π‘ < 0 1 πππ 0 < π‘ −π 1 π(π) = 2ππ + 2 πΏ(π)