basic 0 the transmission of analog information-bearing signals by a digital communi- 235 ion system, and which can indeed do better. NOTES AND REFERENCES NOTES AND REFERENCES 1. ora detailed discus Bennett, and Stein (1966, pp. 129-163). has been confi of the threshold effect in FM receivers, see Rice (1963) and Schwartz, Figure 6.12 is adapted from Rice (1963). The valtity ofthe theoretical curve I inthis figure el experimentally; see Schviartz, Benet, an Stein (1966, . 153). 3. The treatment of the FMFB demodulator presented in Section 65 is based on the paper by Enloe (1962); see also Roberts (1977, pp. 166-181), 4, For a full discussion of threshold effecis in phase-locked loops, see Gardiner (1979, pp. 178- 196) and Roberts (1977, pp. 200-202). 5, For a detailed description of Dolby systems mentioned in the latter part of Section 6.7, see Swemer (1990, pp.732-734). PROBLEM S RSME ARE a 6.1. The sample function x isapplied to the low-pass RC ite of Figure P6.1.The amplitude A, and frequency f-of the sinusoidal component are constants, and w() isa white Gaussian noise of zero mean and power spectral density ‘No/2. Find an expression for the output signal-to-noise rato with the sinusoidal component of (regarded as the signa of intrest \-cos(2eft) + (1) Ly Input somal Output signal Figure P6.1 6.2 _ Suppose next the sample function w() of Problem 6.1 is ap- plied to the band-pass LCR filter of Figure P62, which is tuned to the fiequency f, of the sinusoidal component. Assume that the factor ‘of the filter is high compared with unity. Find an ‘expression for the output signal-to-noise ratio by teating the sino soidal component of x) as the signal of interest. Figure P6.2 6.3 A DSB-SC modulated signal is transmitted over a noisy channel, with the power spectral density of the noise being as shown in Figure P6.3. The message bandwidth is 4 KHz and the carer frequency is 200 kHz. Assuming thatthe average power of the modulated wave is 10 watts, determine the output signl-o: noise ratio ofthe receiver. sup wie ke) 6.4 Evaluate the autocorrelation functions and cross-corelation functions of the in-phase and quadrature components of the narrowband noise atthe coherent detector input for the DSB-SC system, 6.5. Ina receiver using coherent detection, the sinusoidal wave generated by the local oscillator suffers from a phase error Ot), with respect tothe carrier wave cos(2f). Assuming that 0) isa sample function of s zero-mean Gaussian process of variance of, and that. most of the time the maximum value of 6() is smal compared with unity, find the mean-square error of the receiver ‘output for DSB-SC modulation. The mean-square errr is defined as the expected value of the squared difference between the re ceiver output and the message signal component of the receiver output6.6. The average noise power per unit bandwidth measured at the frontend of an AM receivers 10" watt per Hertz:"The modii- lating wave is sinusoidal, with-a cartier power of 80 kilowatts, and a sideband power of 10 kilowatts per sideband. The message bandwidth ig 4 kHz, Assuming the use ofan envelope detector in the receiver, determine the output signal-to-noise ratio of the system. By how many decibels i this system inferior to a DSB ‘SC modulation system? 6.7. Covisider the output of an envelope detector defined by Bq, (6.14), which is reproduced here for convenience: WO) = (he + Acme) +0 +0) (@) Assume thatthe probability ofthe event Ira(t)| > eAelt + kam(e)| is equal to or less. than 8, where ¢<< 1. What isthe prob- ability thatthe effet of the quadrature component nt) is negligible? (b) Suppose that hy is adjusted relative to the message signal __ (suc tha the probability ofthe event ACL + kam] +m(2)<0 is equal to 5. What isthe probability that the approximation . J) = ALL + kom) + m(t) is vali? (©) Comment on the significance of the result in part (b) for the case when 6, and 5, ate both small compared with unity. 6.8 An unmodulated carrier of amplitude A, and frequency and band-limited white noise are summed and then passed through an idcal envelope detector. Assume the noise spectral density to be of height No? and bandwidth 2, centered about the carrer frequency f.. Determine the output sigualto-noise ratio forthe «ase when the cartie-o-noise ratio is big, 6.9 An AM receiver, operating with sinusoidal modulating signal and 80 percent modulation, hasan output signal o-noise falioof30<8, (a) What isthe coresponding caterto-noise ratio? (©) By how many decibels can we decrease the carrier-to-noise rato so that the system is operating just above threshold? 6.10 Consider a phase modulation (PM) system, with the modulated wave defined by s()= Aceos2ft + kam()] where fis «constant and m() is the message signal. The additive noise n() atthe phase detector input is 1n() = nu (0) cos(2nft) ~ no()sin(2nfr) Assuming that the cartier-to-noise ratio at the detector input is high compared with unity, determine (a) the ouput signal-to-noise ratio and (b) the figure of mest ofthe sytem, Compare your r- sults with the FM system forthe cae of sinusoidal modulation, 6.11 An FDM system uses single-sideband modulation to combine 12 independent voice signals and then uses frequency modulation 236 to transmit the composite baseband signal, Each voice signal has an average power P and occupies the frequency band 0.3-3.4 kHz; the system allocates it a bandwidth of 4 KHr. For each voice signal, ‘only the lower sideband is transmitted, The sub-catier waves used for the first stage of modulation are defined by ex(t)=Accos(2rife), OSk< 11 ‘The received signal consists of the transmited FM signal plus white Gaussian noise of zero mean and power spectral density No/2 (@) Sketch the povier spectral density ofthe signal produced at the frequency discriminator output, showing both the signal and noise components. (0) Find the relationship between the-subcarier amplitudes Ay 50 that the modulated voige signals have equal signl-1o- noise ratios 6.12 Inthe discussion on FMihreshold etext presentedin Section 65, we described the conditions for positive-going and negative going click in terms of the envelope rt) and phase () of the narrowband noise (), Reformulate these conditions in terms ofthe {in-phase component n(?) and quadrature component nt) fn). 6.13 By using the pre-efuphass filer shown in Figure 6.172 and witha voice signal asthe modulating wave, an FM transmitter proves a signal tha is essentially frequency-moduated by the lower audio frequencies and phase-modulated by the higher audio frequencies, Explain the reasons for his phenomenon, 6.14 Suppose thatthe transfer functions of the pre-emphasis ‘nd de-emiphasis filters of an FM system are scaled as follows: and wit iran) ‘The scaling factor kis to be chosen so that the average power of ‘the emphasized message signal isthe same as that ofthe orginal message signal mt). (@) Find the value of& that satisfies this requirement forthe case ‘when the power specral density ofthe message signal m() is So Sulf=5 1+" : 0, -Wsfsw elsewhere (©) Whats the corresponding value of the improvement factor produced by using this pair of pre-emphasis and deemphasis filers? Compare this ato wth that obtained in Example 6.3. ‘The improvement facto is defined by Eq, (6.50) 6.15 A phase modulation (PM) system uses a pair of pre- ‘emphasis and de-emphasis filters defined by the transfer functions if Hala EAras af al and Half) = OO tl) ~ Show that the improvement in output signal:to-noise ratio pro- duced by using this pair of filters is - Wifo tan"(W/fo) ‘where W is the message baridwidth. Evaluate this improvement for the case when W = 15 kHz and fo = 2.1 kHz, and compare your result with the corresponding value for an FM system. Computer Problems. 6.16 Using the Matlab model of an AM modulator and envel- cope detector developed in Problem 3.25, do the following: (a) Simulate the modulation and detection ofa 400 Hz modulat- ing wave having a modulation index of 5O percent. Keap a copy ofthis noise-free demodulator outpat asa reference. (b) Add narrowband noise tothe signal prior to demodulation such that the channel SNR is 30 dB. (© Compare the output of the demodulator with the noisy input to the noise free output. Compute the mean-square error between the two and estimate the output SNR using this, Repeat the problem for channel SNRs of 20 dB and 10 dB. At ‘hat channel SNR does envelope detection break down? 6.17 Using the Matlab model of an FM modulator and dlscri- minator developed in Problems 4. 26 and 4.28, do the following: (@) Simulate the modulation and detection ofa 1 KEz modula ig Wave having & modulation index of 2. Keep a copy of. this noise-free demodulator output asa reference. (b) Add narrowband noise to the signal prior to demodulation such thatthe channel SNR is 30 dB. (©). Compare the output of the demodulator with the noisy input to the noise free output. Compute the mean-square ‘error between the (6 and estimate the output SNR using Repeat for channel SNRs of 20, 15 and 10 4B. At whet channel ‘SNR does frequeney discrimination break down? 237