TUT/ICE 1 ELT-44006 Receiver Architectures and Signal Processing Exam Requirements and Model Questions 2018 General idea of these Model Questions is to highlight the central knowledge expected to be known in the exams. Actual exam questions may have completely different formulation, they may require combining knowledge from different model questions, and they may require somewhat broader understanding of the general concepts than what would be required for answering the specific model questions. The scope of the exam includes lectures 1, 2, 3, 4, 6, and 7 of this year s schedule. Filter bank section of Sampling and Multirate Techniques for Complex and Bandpass Signals is excluded from the scope. Introduction to Multirate Digital Signal Processing: Practical knowledge, as highlighted by the Model Questions of that area, is sufficient. Don t worry about the complicated looking math formalism. All exercises related to lecture 1. Receiver Architectures may be considered as exam quastions. Some of the exercises of lectures 3. Introduction to Multirate Digital Signal Processing and 4. Multirate Techniques for Complex and Bandpass Signals are related to some of questions in the list below. No calculator is needed/allowed in the exam: You are expected to be able to do very basic calculations and db conversions (e.g., for values like 6 db, 12 db, 30 db, 49 db) by pen and paper. For example, some of the exercise problems would be presented in numerically simplified form. 1.1 Describe the principle of superheterodyne receiver. What are the requirements for the RF filter and IF filters? What are the advantages and disadvantages of this receiver principle, considering also possibilities for highly integrated solutions? 1.2 A GSM mobile terminal is receiving a signal at 940.2 MHz center frequency. The receiver utilizes the superheterodyne principle and the first IF frequency is chosen to be 71 MHz. What is the needed local oscillator frequency? Assuming that the signal bandwidth is 200 khz, where is the image band located in the frequency axis? If the downlink frequency band is 935 960 MHz, what is the needed tuning range for the local oscillator? (Actually, there are two choices for all the questions, give both of them.) 1.3 Describe the principle of direct conversion receiver. What are its advantages? What are the most serious problems and non-idealities from the implementation point of view? 1.4 Explain the reasons/mechanisms why large DC-offsets are easily produced in direct-conversion receivers. 1.5 What are the effects of I/Q gain mismatch and/or phase imbalance in direct-conversion receivers? Consider also the case where quadrature mixing is used for implementing a frequency translation to a (low) IF frequency instead of baseband, what are the effects in this case? What is the essential difference in these two cases?
TUT/ICE 2 1.6 Which different techniques can be used for providing/enhancing the image rejection in different receiver architectures? Describe each technique briefly. 1.7 Describe the principle of image-reject mixer, i.e., how quadrature mixing helps in providing the image rejection, e.g., in superheterodyne receivers. 1.8 Which factors determine the dynamic range of a low-noise amplifier stage? 1.9 Why the third-order intermodulation distortion is an important consideration in receivers? What kind of interference and how it may produce? What is IP3? In which cases the second-order intermodulation becomes a problem? 1.10 Consider a communications receiver which has an input IP3 value of 10 dbm. Assume that two strong sinusoidal signals (blockers) at the level of -20 dbm enter the receiver. What is the power level of 3 rd -order intermodulation distortion? 1.11 The table (pages RxArch/35-36) shows the small-signal noise figure analysis for a receiver. Assume that the duplexer is removed (or noise figure of one of the stages is changed to some other value), what would be the small signal noise figure of the receiver? Alternatively: Given the component parameters, calculate the noise figure of a simple receiver chain. 1.12 What are the effects of phase errors and phase noise in digital transmission systems? 1.13 Describe the process how local oscillator phase noise may produce interference on top of the wanted signal in a receiver. 1.14 Be prepared for phase noise calculation examples, like the one on page RxArch/39. 1.15 Assume that the thermal noise floor measured at a GSM receiver input is -174 dbm/hz, the equivalent noise bandwidth is 200 khz, the minimum S/N ratio is 9 db (including implementation margin), and the required receiver sensitivity is -102 dbm. Then what should be the noise figure of the receiver? 2.1 How does the duplexing principle (FDD vs. TDD, or WCDMA vs. GSM) effect on the overall mobile terminal design? Discuss the choice between duplexer and switch to connect the antenna to the receiver and transmitter. 2.2 Discuss the transmitter power amplifier requirements in different mobile communication systems, e.g., constant envelope modulation, GSM/GMSK, CDMA, OFDM. Describe the main distortion effects due to the nonlinearity. 2.3 Explain the meaning of back-off. 2.4 OFDM systems are known to suffer from the problem of high PAPR. Explain what this is about and explain its implications in the transceiver design. 3.1 Represent alternative specifications (no aliasing / transition band aliasing / don t care bands) for decimation (interpolation) by a factor of 6, assuming that the passband edge is 0.4 times the lower
TUT/ICE 3 sampling rate. Which are the possible multistage decimation cases? Explain why multi-stage structures are efficient in sampling rate conversion applications. 3.2 Describe the idea and structure of half-band FIR filters. Explain how they are used in sampling rate conversion applications and why they are particularly effective in these applications. 3.3 Show (a) the four-branch non-decimated polyphase structure and (b) the efficient polyphase decimator structure for a 4 th -band FIR decimator with impulse response [a 0, a 1, 0, a 3, a 4, a 5, 1, a 7, a 8, a 9, 0, a 11, a 12 ]. 3.4 What is a CIC filter? Show the structure and characteristic frequency response of a 2 nd -order CIC filter with decimation factor 8. Why are CIC filters commonly used in digital channel selection and decimation filtering chains? 4.1 What are the essential differences between the spectra of real and complex (I/Q) signals, considering also bandpass and discrete-time cases? 4.2 The figure shows the spectrum of a real bandpass signal (only positive frequencies shown). Show the resulting spectrum after quadrature (or real) bandpass sampling at the rate of 8 khz. 32 khz 40 khz f 4.3 Assume that a bandpass signal with a bandwidth of 200 khz and center frequency of about 12 MHz is to be sampled at a rate of 541.667 khz. What are the possible choices for the center frequency of the signal in real bandpass sampling? What is the best choice, considering the analog bandpass filter requirements needed to attenuate the aliasing to the wanted signal band? 4.4 Consider sampling a bandpass signal with center frequency f c desired signal bandwidth W overall signal bandwidth B (>W) sampling rate f s. What are the requirements for these four parameters in order to avoid aliasing/imaging of other frequency bands on top of the desired signal using (a) real sampling and (b) quadrature sampling? 4.5 Consider a complex bandpass filter obtained from a real lowpass FIR with impulse response ho, h1,..., hn through a frequency shift by f s / 4. What is its impulse response? Possible variations: frequency shift of f s /2 some arbitrary frequency shift up to f s /2
TUT/ICE 4 4.6 A half-band IIR filter has a transfer function of the form 2 H( z) = A ( z ) + z - 1 2 1 A 2 ( z ). (The scaling factor of 0.5 is ignored here for simplicity.) Using such a filter as a prototype, a complex bandpass filter (e.g., a Hilbert-transformer or phase splitter) with a centre frequency of f s / 4 can be designed through a proper frequency shift. What is the transfer function of the resulting filter? 4.7 Show the basic structure and spectral behaviour for implementing frequency translation by mixing with (a) real sinusoidal LO-signal (b) complex exponential signal. Discuss also their practical limitations/non-idealities. 4.8 How does oversampling, i.e., utilizing N-times higher sampling rate than necessary effect on the SNR of and A/D converter? What is the meaning of SFDR? 4.9 When sampling a bandpass signal, what kind of effects are produced by the aperture jitter of the sampling clock? How do these effects depend on the parameters of the system? 4.10 (a) In a certain bandpass sampling case, both the quantization noise and the apperture jitter noise are 60 db below the maximum signal level. What is the dynamic range in this case? (b) Assume that the same bandpass signal would be sampled at a center frequency which is ten times as high as in the first case (other affecting parameters would remain the same as in the first case). What would be the quantization noise and apperture jitter noise levels in this case? (c) Assume further that the sampling rate would also be made ten times as high as in the first case. Then what would be the quantization and jitter noise levels if the processing gain due to oversampling is taken into consideration? 6.1 Consider receiver architectures where the sampling is done from the RF or high IF signal. What are the main problems from the implementation point of view? Why is it important to have some gain and selectivity before sampling? Which factors determine the key requirements for the T&H and A/D blocks and sampling clock? Explain the differences in the A/D-converter requirements considering the cases where selectivity is implemented before A/D-conversion or after A/D-conversion. 6.2 Assume that in an IF sampling receiver architecture, the A/D converter input signal may include a sinusoidal blocking signal with a power of -23 dbm and the desired signal is at a -99 dbm level. The SNR needed for detection is 9 db. Then what should be the dynamic range of the A/D converter? (To simplify the exercise, you can ignore other noise & interference sources except for the quantization noise.) 6.3 Discuss the properties of different receiver structures from multi-mode receiver implementation point of view, considering especially approaches where the receiver selectivity is mostly implemented with DSP. 6.4 Show the IF-sampling receiver structure using real bandpass sampling. With a special choice of the IF center frequency and sampling rate, it is possible to do the final digital down-conversion in a very simple way. Explain this idea. 6.5 Show an efficient structure for down-converting to baseaband (a) real, (b) complex signal with center frequency at quarter of the sampling rate. 7.1 Describe the general principle of frequency synthesizer and its purpose in communications transceivers. What are the main differences in the synthesizer requirements between FDD and TDD based communications systems?
TUT/ICE 5 7.2 How does the phase noise spectrum of a frequency synthesizer depend on the following parameters: - phase noise spectrum of the reference source - VCO phase noise - PLL transfer response - frequency divider in the loop N - reference frequency divider R Consider frequencies both inside and outside the PLL loop bandwidth. What are the reference spurioses in a frequency synthesizer?