Lecture 10 Fourier Transform in Circuit Analysis (Part 1) The Derivation of the Fourier Transform We begin the derivation of the Fourier transform, viewed as a limiting case of a Fourier series, with the exponential form of the series: The Derivation of the Fourier Transform Allowing the fundamental period T to increase without limit accomplishes the transition from a periodic to an aperiodic function. In other words, if T becomes infinite, the function never repeats itself and hence is aperiodic. As T increases, the separation between adjacent harmonic frequencies becomes smaller and smaller. In particular, The Derivation of the Fourier Transform and as T gets larger and larger, the incremental separation Δω approaches a differential separation dω. As the period increases, the frequency moves from being a discrete variable to becoming a continuous variable, or The Derivation of the Fourier Transform As the period increases, the Fourier coefficients Cn get smaller. In the limit, Cn→ 0 as T → ∞. This result makes sense, because we expect the Fourier coefficients to vanish as the function loses its periodicity. Note, however, the limiting value of the product CnT; that is, The Derivation of the Fourier Transform The integral is the Fourier transform of f(t) and is denoted (1) We obtain an explicit expression for the inverse Fourier transform The Derivation of the Fourier Transform Let's now derive the Fourier transform of the pulse shown in Figure 1. The Fourier transform of v(t) is The Derivation of the Fourier Transform Figure 1 which can be put in the form of (sin x)/x by multiplying the numerator and denominator by . Then, The Convergence of the Fourier Integral If f(t) is different from zero over an infinite interval, the convergence of the Fourier integral depends on the behavior of f(t) as t →∞. A single-valued function that is nonzero over an infinite interval has a Fourier transform if the integral The Convergence of the Fourier Integral exists and if any discontinuities in f(t) are finite. An example is the decaying exponential function illustrated in Figure 2. The Fourier transform of f(t) is Figure 2 The Convergence of the Fourier Integral The Convergence of the Fourier Integral A third important group of functions have great practical interest but do not in a strict sense have a Fourier transform. For example, the integral in Equation 1 doesn't converge if f(t) is a constant. The same can be said if f(t) is a sinusoidal function, cos ωot, or a step function, Ku(t). These functions are of great interest in circuit analysis, but, to include them in Fourier analysis, we must resort to some mathematical subterfuge. The Convergence of the Fourier Integral First, we create a function in the time domain that has a Fourier transform and at the same time can be made arbitrarily close to the function of interest. Next, we find the Fourier transform of the approximating function and then evaluate the limiting value of F(ω) as this function approaches f(t). Last, we define the limiting value of F(ω) as the Fourier transform of f(t). The Convergence of the Fourier Integral Let's demonstrate this technique by finding the Fourier transform of a constant. We can approximate a constant with the exponential function As ε → 0, f(t) → A. Figure 3 shows the approximation graphically. The Fourier transform of f(t) is The Convergence of the Fourier Integral Figure 3 The Convergence of the Fourier Integral This function generates an impulse function at ω = 0 as ε → 0. The area under F(ω) is the strength of the impulse and is The Convergence of the Fourier Integral In the limit, f(t) approaches a constant A, and F(ω) approaches an impulse function 2πAδ(ω). Therefore, the Fourier transform of a constant A is defined as 2πAδ(ω), or Using Laplace Transforms to Find Fourier Transforms 1. The following rules apply to the use of Laplace transforms to find the Fourier transforms of such functions. If f(t) is zero for t ≤ 0-, we obtain the Fourier transform of f(t) from the Laplace transform of f(t) simply by replacing s by jω. Thus Using Laplace Transforms to Find Fourier Transforms Using Laplace Transforms to Find Fourier Transforms 2. Because the range of integration on the Fourier integral goes from - ∞ to + ∞, the Fourier transform of a negative-time function exists. A negative-time function is nonzero for negative values of time and zero for positive values of time. To find the Fourier transform of such a function, we proceed as follows. First, we reflect the negative-time function over to the positive-time domain and then find its one-sided Laplace transform. Using Laplace Transforms to Find Fourier Transforms We obtain the Fourier transform of the original time function by replacing s with -jω. Therefore, when f(t) = 0 for t ≥ 0+, Using Laplace Transforms to Find Fourier Transforms Both f(t) and its mirror image are plotted in Figure 4. The Fourier transform of f(t) is Using Laplace Transforms to Find Fourier Transforms Figure 4 Using Laplace Transforms to Find Fourier Transforms 3. Functions that are nonzero over all time can be resolved into positive and negative-time functions. We use the previous two cases to find the Fourier transform of the positive - and negative - time functions, respectively. The Fourier transform of the original function is the sum of the two transforms. Thus if we let Using Laplace Transforms to Find Fourier Transforms Using Laplace Transforms to Find Fourier Transforms An example involves finding the Fourier transform of e-a|t| For the original function, the positive- and negative time functions are Using Laplace Transforms to Find Fourier Transforms The Fourier Transform of a Signum Function The signum function, defined as +1 for t > 0 and -1 for t < 0.The signum function is denoted sgn(t) and can be expressed in terms of unit-step functions, or Figure 5 shows the function graphically. To find the Fourier transform of the signum function, we first create a function that approaches the signum function in the limit: The Fourier Transform of a Signum Function Figure 5 The function inside the brackets, plotted in Figure 6, has a Fourier transform because the Fourier integral converges. The Fourier Transform of a Signum Function Figure 6 The Fourier Transform of a Signum Function The Fourier Transform of a Unit Step Function To find the Fourier transform of a unit step function, we do so by recognizing that the unit-step function can be expressed as The Fourier Transform of a Cosine Function Fourier Transforms of Elementary Functions