VIEWS: 6 PAGES: 106 POSTED ON: 11/3/2011 Public Domain
RF Microelectronics RFIC Professor: Dr. Ali Fotowat Ahmady Date: July 16, 2007 Lectures: 1-10 (sections 1-5) KAVOSHCOM Contents 1. Introduction 2. Basic concepts 3. Digital modulation, Spectral control, Detection 4. Multiple access standards, TDM, CDM, OFDM 5. TRx architecture 6. LNA and Mixer 7. Oscillator 8. Frequency Synthesizer 9. Power Amplifier November 2, 2011 2 Section One 1. Introduction a. Complexities b. Goals c. Technology November 2, 2011 3 Complexities Simple FM Radio November 2, 2011 4 Complexities Philips GSM phone November 2, 2011 5 IS-19 cellular telephone RF section block diagram A 45 MHz offset frequency oscillator generates the required receiver and transmitter local oscillator frequency. November 2, 2011 6 IS-55 block diagram A narrowband IF filter is required for digital operation, as well as an ADC in the baseband. November 2, 2011 7 Why…? 1. IF frequency 2. 50-ohm impedance matching 3. Shielding 4. All-band filtering 5. I&Q receiver 6. Gain and phase setting in receiver 7. I&Q transmitter 8. LC filter for transmitter 9. SSB mixer 10. Out-of-band noise 11. Output power control 12. … November 2, 2011 8 Disciplines required in RF design As the industry moves toward higher integration and lower cost, RF and wireless design demands increasingly more “concurrent engineering”. November 2, 2011 9 RF design hexagon The trade-offs involved in the design of such circuits can be summarized in the “RF design hexagon”. November 2, 2011 10 Then … In chip Increase frequency Decrease power Change architecture November 2, 2011 11 Technology Three critical factors influencing the choice of technologies in the competitive RF industry: Performance Cost Time to market Issues play an important role in the decisions made by the designers: Level of integration Form factor Prior successful experience November 2, 2011 12 Technology The technologies constitute the major section of the RF market: GaAs Silicon Bipolar BiCMOS SiGe CMOS CMOS technology must resolve a number of practical issues: Substrate coupling of signals that differ in amplitude by 100dB, parameter variation with temperature and process, and devices modeling for RF operation. November 2, 2011 13 Contents 1. Introduction 2. Basic concepts 3. Digital modulation, Spectral control, Detection 4. Multiple access standards, TDM, CDM, OFDM 5. TRx architecture 6. LNA and Mixer 7. Oscillator 8. Frequency Synthesizer 9. Power Amplifier November 2, 2011 14 Section Two 2. Basic concepts 1. Cascaded stage nonlinearity 1. 1 dB compensation point 2. IP3 2. Intersymbol interference 3. Noise figure 4. Sensitivity and dynamic range November 2, 2011 15 Effects of nonlinearity Model nonlinearity as a Taylor series expansion up to its third order term: y (t ) x (t ) x (t ) x (t ) 1 2 2 3 3 If a sinusoid is applied to a nonlinear system: x(t ) A cost A cost A2 cos2 t A3 cos3 t A2 A3 1 cos 2t 3 cos t cos 3t 2 4 2 A2 3 3 A3 2 A2 3 A3 y(t ) 1 A cost cos 2t cos3t. 2 4 2 4 The term with input frequency is called the fundamental and the higher-order terms the harmonics. November 2, 2011 16 1 dB Compression Point The small-signal gain of a circuit is usually obtained with the assumption that harmonics are negligible. 3 3 A2 Gain 1 4 In RF circuit, 1-dB Compression point defined as: The input signal level that causes the small-signal gain to drop by 1 dB. November 2, 2011 17 1 dB Compression Point To calculate the 1-dB compression point, we can write from gain equivalent: 3 20 log 1 3 A12 dB 20 log 1 1dB 4 That is, 1 A1dB 0.145 3 In typical front-end RF amplifiers: The 1-dB compression point occurs around -20 to -25 dBm (63.2 to 35.6 mVpp in a 50 Ohm system). November 2, 2011 18 Desensitization and Blocking If a small signal and a large interferer are applied to a compressive system, the “average” gain for the small signal is reduced: Assume, x(t ) A1 cos 1t A2 cos 2t. The output is 3 3 2 y (t ) 1 A1 3 A13 3 A1 A2 cos 1t ..., 4 2 Which, for A1 A2 , reduces to 3 2 y (t ) 1 3 A2 A1 cos 1t .... 2 November 2, 2011 19 Desensitization and Blocking The gain for the desired signal is equal to 3 Gain 1 3 A2 2 2 A decreasing function of A2 if a3<0 . For sufficiently large A2, the gain drops to zero, and we say the signal is “blocked”. The interferer is called a blocking signal. Many RF receivers must be able to with stand blocking signals 60 to 70 dB greater than the wanted signal. November 2, 2011 20 Cross Modulation When a weak signal and a strong interferer pass through a nonlinear system, Weak signal: x1 (t ) A1 cos 1t Strong interferer: x2 (t ) A2 1 m cos mt cos 2t Then, 3 2 m2 m2 y (t ) 1 A1 3 A1 A2 1 cos 2mt 2m cosmt cos1t 2 2 2 Cross modulation is the transfer of modulation on the amplitude of the interferer to the amplitude of the weak signal. November 2, 2011 21 Cross Modulation If two signals experience nonlinearity, amplitude modulation in one appears in the other. Most important in “multi-carrier” systems. Example include cable TV transmitters and base station transmitters. November 2, 2011 22 Intermodulation If the input sinusoid frequency is chosen such that its harmonics fall out of the passband, The output distortion appears quite small even if the input stage of the filter introduces substantial nonlinearity. November 2, 2011 23 Intermodulation Assume; x(t ) A1 cos 1t A2 cos 2t. Thus, y (t ) 1 A1 cos1t A2 cos2t 2 A1 cos1t A2 cos2t 2 3 A1 cos1t A2 cos2t . 3 Expanding the left side and discarding dc terms and harmonics, obtain the following Intermodulation products: 1 2 2 A1 A2 cos1 2 t 2 A1 A2 cos1 2 t 3 A12 A2 cos21 2 t 3 A12 A2 cos21 2 t 3 3 21 2 4 4 22 1 3 3 A2 A1 cos22 1 t 3 A2 A1 cos22 1 t 2 3 2 4 4 November 2, 2011 24 Intermodulation Intermodulation in a nonlinear system: 1 A1 3 3 A2 A12 4 Corruption of a signal due to Intermodulation between two interferer: November 2, 2011 25 Third Intercept Point “IP3” IP3 is measured by two-tone test A is chosen to be sufficiently small so that higher-order nonlinear terms are negligible and the gain is relatively constant and equal to 1 November 2, 2011 26 IP3 Calculation of IP3: • IIP3|dBm=Pinput|dBm +DPdB/2 . • IIP3|dBm=Poutput|dBm -GaindB+DPdB/2 . • OIP3|dBm=Poutput|dBm +DPdB/2 . November 2, 2011 27 The relationship between 1-dB compression point and iip3 A1dB 0.145 9.6dB AIP3 43 November 2, 2011 28 Cascade nonlinear stages Input-output characteristics of the two stage are expressed: y2 (t ) 1 y1 (t ) 2 y12 (t ) 3 y13 (t ) y1 (t ) 1 x(t ) 2 x 2 (t ) 3 x 3 (t ) Then, y2 (t ) 11 x(t ) 3 1 21 2 2 13 3 x 3 (t ) ... Thus, 4 11 AIP3 . 3 3 1 21 2 2 1 3 3 November 2, 2011 29 Cascade nonlinear stages November 2, 2011 30 Cascade nonlinear stages This equation readily gives a general expression for three or more stages: 1 1 12 12 12 2 2 2 2 ... AIP3 AIP3,1 AIP3, 2 AIP3,3 Typical receiver IP3 is -15 dBm. November 2, 2011 31 Example Stage 1 is a super linear amplifier, Thus AIP3,1=∞ . Super linear G Stage 1 Stage 2 • AIP3,tot= AIP3,2 /11 • AIP3,tot|dBm = AIP3,2|dBm – Gain1st satge|dB. Thus If stage 1 is an amplifier, IIP3tot =IIP32 – GdB If stage 1 is an attenuator, IIP3tot =IIP32 + GdB November 2, 2011 32 Intersymbol Interference Linear time-invariant systems with insufficient bandwidth distort the signal. An example of such behavior: a periodic square wave output with exponential tail Low-pass filter November 2, 2011 33 Intersymbol Interference With a random sequence of ONES and ZEROS as the input : Vin Low-pass Vout filter November 2, 2011 34 Intersymbol Interference Each bit level is corrupted by decaying tails created by previous bits. Called “Intersymbol Interference” (ISI). Leads to higher error rate in the detection of random waveforms transmitted through band-limited channels Particularly troublesome in wireless communications because of narrow bandwidth allocated to each channel November 2, 2011 35 Intersymbol Interference Methods of reducing ISI: In Transmitter: Pulse shaping (Nyquist signaling) In Receiver: Equalization November 2, 2011 36 Input-Referred Noise Representation of noise by input noise generators The correlation between two sources must be taken into account November 2, 2011 37 Input-Referred Noise An example to illustrate the idea: I 2 nD g m Vn 2 2 Vn 8kT /(3g m ) 2 2 gm I 2 2 2 I nD n Z in I n 8kT /(3gm Zin ) 2 2 I 2 nD 4kT (2 g m / 3) November 2, 2011 38 Noise Figure Most of the front-end receiver blocks are characterized in terms of their “noise figure” rather the input-referred noise Noise Figure in dB: Signal-to-noise ratio at the input Signal-to-noise ratio at the output Noise figure is a measure of how much the SNR degrades as the signal passes through a system November 2, 2011 39 Calculation of Noise Figure SNR in is the ratio of the input signal power to the noise generated by the source resistance, R s, modeled by 2 VRS November 2, 2011 40 Calculation of Noise Figure Vin 2 2 SNRin VRS 2 2 Voltage gain from Vin to the input port of the circuit (node P) Av Vin2 2 2 SNRout V RS 2 (Vn I n Rs ) Av 2 2 2 Voltage gain from P to V out V n and I n R s are added before squaring to account for their correlation November 2, 2011 41 Calculation of Noise Figure VRS (Vn I n Rs ) 2 2 NF 2 VRS 4kTRs (for the spot noise figure to emphasize the very small bandwidth ) 2 Vn,out 1 NF 2 2 A 4kTRs Vn,out =Total noise at the output A Av November 2, 2011 42 Calculation of Noise Figure Thus, to calculate the Noise figure, we divide the total output noise power by the square of the voltage gain from V in to V out and normalize it to the noise of R s. As an example: consider the single resistor, R p November 2, 2011 43 Calculation of Noise Figure What is the noise figure of this circuit with respect to a source resistance R s ? Vn ,out 4kT ( Rs llR p ) 2 RP A Rs R p Rs NF 1 Rp November 2, 2011 44 Sensitivity For an RF receiver: The minimum signal level that the system can detect with acceptable signal-to-noise ratio SNRin Psig / PRS NF SNRout SNRout Psig PRS .NF .SNRout The overall power is distributed across the channel bandwidth, B. Thus the two sides of the equation must be integrated over the bandwidth to obtain the total mean square power Psig ,tot PRS .NF .SNRout .B for a flat channel November 2, 2011 45 Sensitivity Pin ,m in / dBm PRS / dBm / Hz NF / dB SNRm in/ dB 10 log B Minimum input level that achieves SNR min 4kTRs 1 PRS 4 Rin The noise power that R s delivers to the receiver PRS kT 174dBm / Hz Assuming conjugate matching at the input November 2, 2011 46 Sensitivity At room temperature: Pin ,m in 174 dBm / Hz NF 10 log B SNRm in Total integrated noise of the system called Noise Floor Since P in,min is a function of the bandwidth, a receiver may appear very sensitive because it employs a narrowband channel November 2, 2011 47 Dynamic Range (DR) Generally is the ratio of the maximum input level that the circuit can tolerate to the minimum input level at which the circuit provides a reasonable signal quality In RF design, this definition is based on the Intermodulation behavior and the sensitivity called : Spurious-free dynamic range November 2, 2011 48 Dynamic Range (DR) The upper end of the dynamic range is: The maximum input level in a two-tone test for which the Third-order IM products do not exceed the noise floor P PIM ,out PIIP3 P in out 2 P P G out in PIM ,out PIM ,in G P PIM ,in 3P PIM ,in PIIP3 P in in in 2 2 And hence, 2 PIIP3 PIM ,in P in 3 November 2, 2011 49 Dynamic Range (DR) The input level for which the IM products become equal to the noise floor is: 2 PIIP3 F P ,m ax in 3 F 174 dBm / Hz NF 10 log B November 2, 2011 50 Dynamic Range (DR) SFDR is the difference (in dB) between P in, max and P in, min 2 PIIP3 F SFDR ( F SNRmin ) 3 2( PIIP3 F ) SFDR SNRmin 3 SFDR represents the maximum relative level of interferers that a receiver can tolerate while producing an acceptable signal quality from a small input level November 2, 2011 51 Contents 1. Introduction 2. Basic concepts 3. Digital modulation, Spectral control, Detection 4. Multiple access standards, TDM, CDM, OFDM 5. TRx architecture 6. LNA and Mixer 7. Oscillator 8. Frequency Synthesizer 9. Power Amplifier November 2, 2011 52 Section 3 3. Digital Modulation, spectral control, detection 1. Quadrature modulation and spectral control 2. Constant and Variable envelope signals November 2, 2011 53 Aspects of Digital Modulation BER Maximizing SNR (receiver) Power Efficiency Avoiding Spectral Regrowth (transmitter) Spectral Efficiency Maximizing The Number of Channels (transceiver) November 2, 2011 54 Digital Modulation Trade-Offs higher BER simple architecture but complex November 2, 2011 55 Quadrature Modulation A binary data stream could be subdivided into pairs of two bits and each pair represented with one of four levels before performing modulation. Bits bm and bm+1 are impressed upon a single carrier x(t ) bm Ac cos ct bm 1 Ac sin c t November 2, 2011 56 Quadrature Modulation This is possible because cosωct and sinωct are orthogonal functions November 2, 2011 57 Quadrature Modulation Called “Quadrature Modulation” or “Qudrature Multiplexing” This operation is illustrated: I (In-phase) Q (Quadrature) XBB November 2, 2011 58 S/P Converter November 2, 2011 59 S/P Converter November 2, 2011 60 Quadrature Modulation To obtain constellation, we assume bits bm and bm+1 are rectangular pulses with a height ±1 and write the modulated signal as : x(t ) 1 cos ct 2 sin ct Where 1 and 2 can each take on a value of +Ac and -Ac November 2, 2011 61 Quadrature Modulation Signal Constellation for Quadrature Modulation November 2, 2011 62 QPSK If the bit waveform is a rectangular pulse, a QPSK signal is obtained. One of four phases of a sinusoid is selected according to the symbol k xQPSK 2 Ac cos(ct ) 4 k 1,3,5,7 November 2, 2011 63 QPSK An important drawback of QPSK is large phase changes at the end of each symbol [-1 -1] [1 1] 180° phase transition in a QPSK waveform November 2, 2011 64 QPSK 180º phase step or, equivalently, a transition between two diagonally opposite points in the constellation Such transitions are undesirable if the waveform is to be filtered and subsequently processed by a nonlinear power amplifier November 2, 2011 65 OQPSK The above drawback is OK with Offset QPSK Modulation with half-the-symbol-period-offset in time November 2, 2011 66 OQPSK Phase transitions in OQPSK : only ±90º November 2, 2011 67 OQPSK Drawback: Differential encoding, which plays an important role in noncoherent receivers, is not executable in OQPSK So /4-QPSK is another variant of QPSK The /4-QPSK signal consists of two QPSK schemes, one rotated by 45º with respect to the other: November 2, 2011 68 /4-QPSK The modulation is performed by alternately taking the output from each QPSK generator November 2, 2011 69 /4-QPSK Possible phase transition in the constellation is 135º November 2, 2011 70 MSK & GMSK Continuous modulation schemes such as MSK: Minimum Shift Keying, and GMSK, Gaussian MSK, avoid abrupt phase changes which; 1. lead to a wide spectrum 2.present difficulties in the design of power amplifiers November 2, 2011 71 Constant-and Variable-Envelope-Signals A modulated waveform x(t ) A(t ) cos[c t (t )] is said to have a constant envelope if A(t) does not vary with time. The modulation schemes described above have constant envelope Constant-and Variable-Envelope-Signals behave differently in a nonlinear system The spectrum “grows” when a variable-envelope signal passes through a nonlinear system QAM and OFDM are examples of variable-envelope signals which are less power efficient November 2, 2011 72 Contents 1. Introduction 2. Basic concepts 3. Digital modulation, Spectral control, Detection 4. Multiple access standards, TDM, CDM, OFDM 5. TRx architecture 6. LNA and Mixer 7. Oscillator 8. Frequency Synthesizer 9. Power Amplifier November 2, 2011 73 Section 4 4. Multiple access standards, TDM,CDM,OFDM a. Duplexing b. FDM/TDM/CDMA c. AMPS/NA-TDMA/CDMA/UMTS d. DECT/BlueTooth/11b/11a/11n/16d/15d November 2, 2011 74 Duplexing Problem: Two-way communication by a transceiver Answer : Duplexing Time-Division Duplexing (TDD) Frequency-Division Duplexing (FDD) November 2, 2011 75 TDD The same frequency band is utilized for both transmit (TX) and receive (RX) paths, but the system transmits for half of the time and receives for the other half. Fast enough to be transparent to the user November 2, 2011 76 TDM Merits: The transmitter is disabled during reception, so the two TX and RX paths do not interfere. Direct (peer-to-peer) communication between two transceivers, an especially useful feature in short- range, local area network applications, is allowed. Drawback: The strong signals generated by all of the nearby mobile transmitters fall in the receive band thus desensitizing the receiver November 2, 2011 77 FDD Employs two different frequency bands for the transmit and receive paths. Merit: The “Duplexer Filter” (the two combined front-end band-pass filters), makes the receiver immune to the strong signals transmitted by other mobile units November 2, 2011 78 FDD Drawbacks: Components of the transmitted signal that leak into the receive band are attenuated by typically only 50dB Owing to the trade-off between the loss and the quality factor of the filters, the loss of the duplexer is typically higher than that of a TDD switch. Spectral leakage to adjacent channels in the transmitter output which occurs: When the power amplifier turns on and off to save energy When the local oscillator driving the modulator undergoes a transient By contrast, in TDD such transients can be timed to end before the antenna is switched to the power amplifier output. November 2, 2011 79 Multiple Access Problem: How to allow simultaneous communication among multiple transceivers Answer: Multiple Access Methods Frequency-Division Multiple Access (FDMA) Time-Division Multiple Access (TDMA) Coded-Division Multiple Access (CDMA) November 2, 2011 80 FDMA User 1 Channel User 2 Number User 3 November 2, 2011 81 FDMA The available frequency band partitioned into many channels each assigned to one user The channel assignment remains fixed until the end of the call Principal access method in early cellular networks because of its relative simplicity Insufficient capacity in crowded areas November 2, 2011 82 TDMA The same band is available to each user but at different times (time-Division multiple access) User 1 User 2 User 3 November 2, 2011 83 TDMA Time assigned to one user time slot :Tsl Time assigned to all users Frame : Tf Every Tf seconds each user finds access to the channel for Tsl seconds Speech data Synchronization data Control data November 2, 2011 84 TDMA Problem: What about the data of all other users when only one user is allowed to transmit? The data stored for (Tf-Tsl) seconds and transmitted as a “Burst” TDMA BURST Data requires to be in Digital form to be buffered Speech compression and Coding would be allowed November 2, 2011 85 TDMA vs. FDMA TDMA is a power saving system since the Power Amplifier of the transmitter is turned on for only one time slot in every frame Even with FDD, proper timing of TDMA bursts prevents simultaneous enabling of the transmit and receive paths in each transceiver TDMA more complex than FDMA because of: A/D conversion, Digital modulation, Time slot and Frame synchronization November 2, 2011 86 CDMA Employs Orthogonal Messages to avoid interference Assigns a certain code to each transceiver User 1 User 2 User 3 Each bit of Data translated to that code before modulation November 2, 2011 87 CDMA A special case of spread spectrum (SS) communication The baseband data spread over the entire available bandwidth Also called Direct Sequence Spread Spectrum (DS-SS) communication The code is called “spreading sequence” or “pseudo random noise” November 2, 2011 88 CDMA In the receiver, the demodulated signal is decoded by multiplying it by the same code Upon multiplication the desired signal is “dispread” with its bandwidth returning to its original value The unwanted signal remains spread Correlation November 2, 2011 89 CDMA Important Feature: Soft Capacity Increasing the number of users only linearly increases the Noise Floor Critical Issue: Power Control High-power transmitter can halt communications among others November 2, 2011 90 CDMA The receiver monitors the signal strength of each transmitter and periodically sends the power adjustment requests to each one Received signal levels are typically within 1 dB of each other Reduction in the average power dissipation of the mobile unit Reduction in the average interference seen by other users November 2, 2011 91 Wireless Standards Details and limits on the design of transceivers Some standards used in Cellular and Cordless systems: AMPS NADC CDMA UMTS November 2, 2011 92 AMPS-1983 Advanced Mobile Phone Service also called (“1.Generation”) Communication method: analog Duplex-Method: FDD Multiple-Access: FDMA Band 824-849/869-894 MHz, 832 „Channels“ with 30 kHz width Modulation: FM November 2, 2011 93 AMPS November 2, 2011 94 NADC North American Digital Cellular System Mobile Phone System USA (“2. Generation”) Communication Method: digital Duplex-Method: FDD Multiple-Access: FDMA 824-849/869-894 MHz, 832 „Channels“ with 30kHz width combined with TDMA with 6 Users, gives 4992 „traffic channels“ Modulation: π/4-DQPSK November 2, 2011 95 NADC November 2, 2011 96 NADC November 2, 2011 97 Qualcomm CDMA (IS-95)-1994 Mobile Phone system USA from 1994 (“2. Generation”) Communication Method : digital Duplex-Method: FDD Multiple-Access: CDMA Band 824-849/869-894 MHz 20 “Channels” with 1.25 MHz width Modulation: Offset-QPSK November 2, 2011 98 CDMA November 2, 2011 99 GSM Global System for Mobile Communication (GSM) Mobile Phone System of Germany from 1991 (2nd Generation) worldwide adoption Communication Method: digital Duplex-Method: FDD Multiple-Access: FDMA Band 890-915/935-960 MHz 125 “Channels” with 200kHz Bandwidth (1 “Channel” unused) combined with TDMA with 8 User November 2, 2011 100 GSM November 2, 2011 101 GSM November 2, 2011 102 UMTS Universal Mobile Telephone Service (3rd generation) Communication Method: digital Duplex-Method: FDD/TDD Multiple-Access: CDMA combined with FDMA/TDMA (hierarchy dependent) Band 1900-2025/2110-2200 MHz, 5 MHz Bandwidth Modulation: QPSK November 2, 2011 103 DECT Digital European/Enhanced Cordless Telephone Applied for wireless Telephone in 1991 Communication: digital Duplex- Method : TDD (Timing not critical in medium ranges ) Multiple-Access: FDMA band 1880-1900 MHz, 10 “Channels” with 2 MHz Bandwidth combined with TDMA with 12 Users gives 120 “traffic channels” Modulation: GMSK November 2, 2011 104 DECT November 2, 2011 105 DECT November 2, 2011 106