07-Class.10.Solutions

07-Class.10.Solutions - 1 ECE155A –Vlad Dorfman Class 10...

This preview shows pages 1–4. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1 ECE155A –Vlad Dorfman Class 10 - Solutions Problem 1. Noise Enhancement Introduction: The problem illustrates the noise enhancement for a continuous ZF filter using the example of the Lorentzian channel. The channel s( t ), AWGN noise and the equalizer g ( t ) can be represented as shown in Figure 1a. Figure 1a. ZF Equalization Block Diagram. In the model shown in Figure 1a, 2 / 50 2 50 ) ( ω π ω PW e PW S − = . (1) and the equalizer G(w) is calculated as in (4). Note, that the inverse of S(w) is 2 / 50 50 2 ω π PW e PW , ( 2 ) and contains an exponential term that increases extremely rapidly with frequency. As the density of recording D increases, the range of frequencies |w| ≤π /T in (4) containing the non-zero "inverted" S ( w ) increases as well causing strong noise increase ("enhancement") illustrated in Figure 1b. At the same time, when D (density) is small and T is large, or when the AWGN noise is insignificant and ISI is dominant, ZF is an acceptable technique. + Noise n ( t ) Equalizer g ( t ) Signal + Enhanced noise Channel s ( t ) 2 a) Plot G ( e j ω ) power for D=1.0,1.6,2.0, and 2.5 -10-5 5 10 2 4 6 8 10 12 14 w,rad Magnitude Equalizer Frequency Response, |G(w)| D=1.0, Noise: Analyt.=2.86 σ 2 Sim.=2.86 σ 2 D=1.6, Noise: Analyt.=7.63 σ 2 Sim.=7.64 σ 2 D=2.0, Noise: Analyt.=17.2 σ 2 Sim.=17.3 σ 2 D=2.5, Noise: Analyt.=53.1 σ 2 Sim.=53.2 σ 2 Figure 1. Noise Enhancement vs. Density of Recording "D". b) Calculate the noise power for D=1.0,1.6,2.0, and 2.5 The output noise power is calculated based on that for a linear system that has a FT magnitude |H( ω )|, and input random process X(t) with power spectral density S X ( ω ), the output power spectral density (not total power) will be S Y ( ω ) = |H( ω )| 2 S X ( ω ). Consider that in our case the power spectral density is σ 2 , and the linear system H( ω ) = G( ω ). Then to find the total output noise power one needs to integrate ∫ ∞ ∞ − = ω ω π σ σ d G eq 2 2 2 ) ( 2 , (3) where σ 2 is the input noise power. Input noise is assumed uncorrelated. 3 Since G(w) is calculated as ⎧ T/S(w), |w| ≤π /T G( w )= ⎨ ( 4 ) ⎩ 0 , |w|> π /T Then, substituting (4) into (3), ( ) ( ) 1 50 4 1 50 4 50 4 50 2 50 2 2 ) ( 2 / 50 3 2 / 50 3 3 2 2 50 2 3 2 2 50 2 3 2 2 2 2 / 50 2 2 2 2 − ⋅ = − = = = = = = ∫ ∫ ∫ ∫ − − − T PW T PW T PW T T PW T T PW T T out e PW D e PW T d e PW T d e PW T d e PW T d S T π π π ω π π ω π π ω π π π σ π σ ω π σ ω π σ ω π π σ ω...
View Full Document

This note was uploaded on 11/11/2009 for the course ECE 670377 taught by Professor Wolf during the Winter '07 term at UCSD.

Page1 / 13

07-Class.10.Solutions - 1 ECE155A –Vlad Dorfman Class 10...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online