Solution of ECE 503 Test #3 Su03 6/20/03 1. (2 pts) Why is there no CTFS integration property for periodic signals with non-zero average values? Because, if the average value is non-zero, the integral is not periodic and cannot, therefore, be represented for all time by a CTFS. 2. (10 pts) Let a signal be defined by x( t) = rect (2 t) ∗ comb( t) = ∞ ∑ rect (2(t − m)) . m =−∞ Using a representation time, TF = 2, find the numerical values of the harmonic function, X[ k ], for k = 0,1, 2, 3, 4 . Using the fundamental period of x( t) ( TF = 1) as the representation time, we get directly from the table, 1 k F rect (2 t) ∗ comb( t) ← → sinc . 2 2 Changing the period to TF = 2 changes the harmonic function to k 1 k an integer sinc , 4 2 X[ k ] = 2 0 , otherwise Therefore, X[0] = 1 1 1 1 , X[1] = 0 , X[2] = sinc = , X[ 3] = 0 , X[ 4 ] = 0 2 π 2 2 Alternate Solution: 1 X[ k ] = TF X[ k ] = ∫ x(t)e TF − j 2π ( kf F ) t 1 dt = ∫ x{ (t) cos (πkt ) − j sin(πkt )dt 4 4 3 1424 3 2 2 even 12 even odd 1 1 1 1 x( t) cos(πkt) dt = ∫ x( t) cos(πkt) dt = ∫ x( t) cos(πkt) dt ∫ 2 2 2 −1 0 For k = 0, X[ k ] = 1 1 4 1 1 1 1 x( t) dt = ∫ x( t) dt = ∫ dt + ∫ dt = ∫ 2 −1 2 3 0 0 4 For k ≠ 0, πk 3πk sin − sin 4 4 sin(πkt) sin(πkt) X[ k ] = ∫ cos(πkt) dt + ∫ cos(πkt) dt = + = πk πk 0 πk 3 3 0 1 4 1 4 1 4 4 For any even k (except k = 0), X[ k ] = 0. For any odd k, and 1 πk 3πk sin = sin 4 4 πk 2 sin 4 1 k X[ k ] = = sinc 4 πk 2 and these solutions agreee with the previous ones. X[0] = 1 1 1 1 , X[1] = 0 , X[2] = sinc = , X[ 3] = 0 , X[ 4 ] = 0 2 π 2 2 2πn 2πn (14 pts) Find the DTFS harmonic function of x[ n ] = 2 cos − 5 sin using the 8 12 fundamental period of x[ n ] as the representation time, N F . 3. N F = 24 Using the tables and the change-of-period property, X[ k ] = (comb 24 [ k − 3] + comb 24 [ k + 3]) − j 4. 5 (comb24 [k + 2] − comb24 [k − 2]) . 2 1 1 A DT signal has a DTFT, X( F ) = rect 5 F − + rect 5 F + ∗ comb( F ) . 4 4 (a) (8 pts) Find the signal’s total signal energy, E x . 2 1 1 E x = ∫ rect 5 F − + rect 5 F + ∗ comb( F ) dF 1 4 4 1 1 − + 4 10 1 1 + 4 10 Ex = ∫ dF + 1 1 − − 4 10 1 1 − 4 10 (b) ∫ dF = 2 5 (7 pts) Find the signal, x[ n ] , and express it as a combination of real-valued DT functions (no j’s in the expression). n F → wrect ( wF ) ∗ comb( F ) sinc ← w n F sinc ← → 5rect (5 F ) ∗ comb( F ) 5 1 n F sinc ← → rect (5 F ) ∗ comb( F ) 5 5 πn 1 1 n + j 2 F ←→ rect 5 F − ∗ comb( F ) sinc e 5 4 5 πn 1 1 n − j F → rect 5 F + ∗ comb( F ) sinc e 2 ← 5 4 5 πn −j 1 + j π2n n F → rect 5 F − + e 2 sinc ← e 5 5 2 πn n F cos sinc ← → rect 5 F − 5 5 2 1 1 + rect 5 F + ∗ comb( F ) 4 4 1 1 + rect 5 F + ∗ comb( F ) 4 4 5. F e −πt ← → e −πf , 2 (8 pts) Given the transform pair, x( t) = e −π ( 3( t − 2)) 2 find the CTFT, . What is the numerical maximum value of X( f ) ? e e −π ( 3 t ) 2 −π ( 3( t − 2)) 2 f X( f ) , of 2 1 −π ←→ e 3 3 F f 2 1 −π ←→ e 3 e − j 4 πf 3 F Maximum value occurs where f = 0. Value is 6. 2 1 . 3 ( 2 pts) What is the fundamental reason that ideal filters cannot be actually built? Their impulse responses are all infinite in extent and cannot, therefore, be the impulse responses of causal systems.