1. Radio Receiver Systems Chapter 1 Radio Receiver Systems 1.1 Receiver Design The input to a wireless transmitter may be voice, video, data, or other information to be transmitted to one or more distant receivers. These data are usually referred to as the baseband signal. A basic function of the transmitter is to modulate the baseband information onto a high frequency sine wave carrier signal for the purpose of efficient transmission over a noise-filled radio channel, than the direct radiation of the baseband signal. The purpose of a radio frequency (RF) receiver is to process incoming energy into useful information while maintaining a sufficient signal-to-noise ratio (SNR), adding a minimum of distortion. This must be done for widely varying RF power level and in the presence of noise and interferers. Performance depends: - system design - internal circuitry - operating environment Receiver performance specifications: - high gain to restore the lower power of the received signal to a level near its original baseband value. - receiver sensitivity quantifies the ability of a receiver to respond to weak signal levels. Analogue receivers – maximum RF level to ensure a certain demodulated S/N ratio. Digital receivers – maximum BER at a certain RF level as a measure of performance. - receiver selectivity usually refers to a receiver’s ability to reject unwanted signals on adjacent channel frequencies. Typical ranges from 70 to 90dB which is difficult to achieve that many systems do not allow simultaneous active adjacent channels in the same geographical area. - spurious response rejection is required as all receivers have the potential for responding to frequencies other than the desired channel. Minimised by the proper choice of the IF and use of RF filters. Spurious response rejection of 70 to 100dB can be achieved in practical receivers. - isolation from transmitter to avoid saturation of the receiver. - intermodulation rejection measures the receiver tendency to generate its own onchannel interference from two or more strong off-channel signals. 70dB is easily achievable. - receiver self-quieting refers to reduced receiver sensitivity on some channels due to internally generated signals that capture the detector and thus prevent or inhibit the reception of desired, weak signal. Other receiver parameters prominent in many regulatory requirements: Distortion of the demodulated signal S/N ratio for strong signals Frequency stability, cochannel rejection, cross modulation, radiated emission, susceptibility to high RF levels 1 1. Radio Receiver Systems 1.2 Receiver Architecture The radio signal experiences many transformations in its migration from a baseband signal at the transmitter to its replication at the receiver. Primary motivation for selecting certain receiver architecture is the required performance. It may be instructive to walk through Fig.1.1 block by block and summarise the important decisions for each block before tackling the somewhat iterative procedure for determining individual stage specifications. Antenna: receives electromagnetic waves radiated from many sources over a relatively broad frequency range and transfer the energy through the input circuits Filter 1 or the preselector: a highly selective, cavity tuned filter, cascaded with a lowpass filter: to limit the bandwidth of spectrum before reaching the RF amplifier and mixer in order to minimise IM distortion to attenuate receiver spurious response to suppress local oscillator energy originating in the receiver RF amplifier or low noise amplifier (LNA): to amplify the possibly very weak received signal while minimising the noise power that is added to the received signal to attenuate local oscillator energy to isolate filter 1 from filter 2 so that the overall selectivity is not destroyed Filter 2 or image filter: to attenuate receiver spurious response frequencies to attenuate direct IF frequency pickup to attenuate noise at the image frequency originating in or amplified by the RF amplifier to suppress second harmonics originating in the RF amplifier to suppress local oscillator energy leaking back into the antenna 1st mixer: active, passive, unbalanced, singly balanced, doubly balanced, tuned, broadband needs to have high intercept point to down-convert the received RF signal to a lower frequency signal called the intermediate frequency (IF) Injection filter: to attenuate wideband noise around the LO frequency and its harmonics to attenuate the second harmonics in order not to degrade mixer second-order intercept point (IP2) 1st local oscillator: its single sideband (SSB) phase noise determine the receiver’s adjacent channel selectivity performance LO signal must have low spurious signals as they will cause receiver spurious responses must oscillate despite temperature and power supply variations low susceptibility to microphonics where external mechanical or acoustic stress could modulate the LO frequency or amplitude 2 1. Radio Receiver Systems Figure 1.1 Typical dual-conversion receiver. 1st IF filter: protects its following stages from close-in IM signals provides adjacent channel selectivity attenuates the second image equivalent noise bandwidth of the IF chain determines how much noise reaches the detector and determines the modulation bandwidth that can be received IF amplifier: high gain stage raises the power level of the signal so that the baseband information can be recovered easily its intercept point must be high if it directly follows the mixer if it follows one stage of IF filtering, the intercept point requirements can be relaxed because the IF filter offers some protection against high-level, off-channel signals The steps required for deciding the specifications for individual circuit blocks shown in Fig.1.1 can be summarised as follows: 1. Allocate approximate gains and losses as needed to meet the required receiver sensitivity specification and IM distortion requirements 2. Select the first IF frequency 3. Select the first LO injection side 4. Investigate the mixer 5. Based on mixer performance, design the injection filter and select LO technology 6. Investigate filter topologies 7. Design the RF amplifier This type of receiver (Fig.1.1) is known as a superheterodyne receiver because it uses frequency conversion, converting the relatively high RF carrier frequency to a lower IF frequency before final demodulation. Since large in-band interferers accompany the received signal even after the front-end BPF, the nonlinearity of the following stages, particularly that of the low-noise amplifier and the mixer becomes important. As illustrated in Fig.1.2, odd-order nonlinearities yield intermodulation products that fall in the same desired channel. As third-order distortion is usually dominant, the IP3 of each stage must be sufficiently high to avoid corruption of the signal by the intermodulation products. Although distorting the amplitude, this effect is important even if the signal carries information only in its phase or frequency, because the zero-crossing points of the desired signal are corrupted by the intermodulation product. The BPF must exhibit minimal in-band loss while adequately suppressing the harmonics and out-of-band spurious components of the transmitted signal. 3 1. Radio Receiver Systems Figure 1.2 Effect of nonlinearity in the front end. The out-of-channel intermodulation products created by the power amplifier cannot be suppress by the BPF and must be acceptably small by design. Another important concern in the design of transceivers is the dynamics range of the signals. With multipath fading and path loss, the required dynamics range for the received signal is typically greater than 100dB. As the minimum detectable signal is in the microvolts range, not only the input noise of the receiver but also cross-talk become critical. (see Fig.1.3). At the other extreme of the dynamic range, the receiver may experience large signals, for example, amplitudes as high as several hundred millivolts, if it is close to a transmitter. While the signal amplitude is not critical in PM and FM systems, the receive path must still process the signal correctly. This issue leads to the use of automatic gain control (AGC). 1.2.1 Antenna The function of an antenna is to convert an RF signal from a transmitter to a propagating electromagnetic wave or, conversely, convert a propagating wave to an RF signal in a receiver. In a transceiver, where a transmitter and a receiver are co-located for full-duplex communications, the same antenna may be used for both transmit and receive. Some of the more obvious characteristics of an antenna include operating frequency range, size, and pattern coverage. The radiating pattern of an antenna is a plot of the transmitted or received signal strength versus position around the antenna. Wireless systems that provide broadcast-type service, such as television and AM/FM radio require antennas with pattern coverage that is uniform in all directions. Such patterns are called omnidirectional, and can be obtained with wire dipole and monopole antennas. Others systems such as point-to-point radio and DBS receivers, require antennas that radiate (or receive) power preferentially in one direction. The measure of the directionality of an antenna pattern is provided by the directivity or gain of the antenna; an omnidirectional antenna has low gain while a highly directive antenna has high gain. Because of the nature of the electromagnetic operation of an antenna, effective radiation of a signal requires that the antenna have minimum physical dimensions on the order of the electrical wavelength (λ=c/f) at the operating frequency. This means that antenna size decreases with an increase in frequency, so that antennas at low frequencies will be very large, while antennas at microwave frequencies and higher may be very small. 4 1. Radio Receiver Systems Figure 1.3 Desensitisation of LNA by PA output leakage. In additional, it can be shown that the gain of an antenna is proportional to its crosssectional area divided by λ2, so that high antenna gain requires an electrically large antenna. More sophisticated antennas, such as phase array antennas, are able to change the direction of their main beam electronically. Phase array antenna technology can be very useful in commercial wireless systems because the antenna beam can be directed at a given user while avoiding interference from other users. Such systems are called adaptive arrays or smart antennas and may lead to increased channel capacity for cellular and PCS telephone systems if cost reductions can be achieved. 1.2.2 Filters Filters are two port components that are used to selectively pass or reject signals on the basis of frequency. An ideal low-pass filter (LPF) will pass all frequency components below its cutoff frequency, while rejecting higher frequency components. Similarly, a high-pass filter (HPF) will pass frequency components above its cutoff frequency, while rejecting lower frequencies. A bandpass filter (BPF) passes frequency components within a narrow passband, while rejecting frequency components outside the passband. Filters are key components in all wireless transmitters and receivers. They are used to reject interfering signals outside the operating band of receivers and transmitters, to reject unwanted products from the outputs of mixers and amplifiers, and to set the IF bandwidth of receivers. Important filter parameters include the cutoff frequency, insertion loss and the out-ofband attenuation rate, measured in dB per decade of frequency. Filters with sharper cutoff responses provide more rejection of out-of-band signals. Insertion loss, measured in dB, is the amount of attenuation seen by signals through the passband of the filter. Another important consideration is size and integrability with other circuit components. At the present time, it is not possible to construct higher performance bandpass filters in integrated circuits form. The inherent losses associated with RF and microwave integrated circuits leads to filters having relatively high insertion losses and low out-of-band attenuation rates. 5 1. Radio Receiver Systems For this reason, most wireless systems today use individual “off-chip” filters that are located on the circuit board, rather than fully integrated filters. This results in a larger and more costly assembly, but critical filtering performance is optimised. 1.2.3 Amplifiers There are three main categories of amplifiers used in wireless system: low-noise amplifiers (LNAs), used in the input stage of a receiver: power amplifiers (PAs), used in the output stage of a transmitter; and IF amplifiers used in the IF stages of both receivers and transmitters. Important specifications for amplifiers include the power gain (in dB), the noise figure, and the intercept points. The noise figure of an amplifier is a measure of how much noise is added to the amplified signal by the amplifier circuitry. This is most crucial in the front end of a receiver, where the input signal level is very small, and it is desired to minimise the noise added by the receiver circuitry. The noise power in a receiver is affected more by the first few components than by later components. Thus it is imperative that the first amplifier in a receiver has as low a noise figure as possible. Because transistors are nonlinear devices, transistor amplifiers exhibit nonlinear effects. Two important phenomena that occur in amplifiers because of these effects are saturation and harmonic distortion. At low signal levels the output power of an amplifier is linearly proportional to the input power. But because the output voltage of an amplifier cannot exceed the bias voltage level, output power gradually reaches a saturation point as input power increases. Saturation is usually only an issue with power amplifiers. A more prevalent problem is related to the fact that harmonics of input signals are generated at the output of an amplifier, and in the case of multiple input signal frequencies, some of these harmonics will lie within the passband of the amplifier. These harmonics lead to signal or harmonics distortion. Generally the power level of these distortion harmonics is very low but the power level of some of these distortion products increases as the cube of the input signal level. The implication of this effect is that distortion power can be significant even for input power levels well below the saturation point of an amplifier. 1.2.4 Mixers A mixer is a three-port component that ideally forms the sum and difference frequencies from two sinusoidal inputs. This allows the important function of frequency conversion to be performed in superheterodyne transmitters and receivers. In the case of a transmitter, the modulated baseband signal is up-converted in frequency by mixing with a high frequency local oscillator signal. In a superheterodyne receiver, the received signal is down-converted in frequency by mixing with a local oscillator to produce a difference frequency (the IF frequency). In both cases, filters are required to select the desired frequency products, while rejecting undesired frequencies that are produced as a by-product of the mixing operation. 6 1. Radio Receiver Systems Modern mixers generally use diodes or transistors and produce many frequencies, based on the harmonics of the input signals and their combinations, in addition to the desired sum and difference frequencies. A passive mixer (one that uses diodes) always produces an output signal (IF) of less power than the input (RF) signal, because of dissipative losses in the mixer as well as inherent losses in the frequency conversion process. This loss is characterised by the mixer conversion loss. Mixers that use active components (e.g. transistors) generally have lower conversion loss, and may even have conversion gain. As in the case of amplifiers, harmonic distortion and noise are also important considerations in mixer performance. Overall, mixer design usually involves trade-offs between noise performance and conversion loss. 1.2.5 Oscillators Oscillators are required in wireless receivers and transmitters to provide frequency conversion, and to provide sinusoidal sources for modulation. Often these sources must be tunable over a set of frequency range, and must provide very accurate output frequencies. The simplest oscillator uses a transistor with an LC network to control the frequency of oscillation. Frequency can be tuned by adjusting the values of the LC network electronically with a varactor diode. Such oscillators are simple and inexpensive, but suffer from the fact that the output frequency is very susceptible to variations in supply voltage, changing load impedances and temperature variations. Better frequency control can be obtained by using a quartz crystal in place of the LC resonator. A crystal-controlled oscillator (XCO) can provide a very accurate output frequency, especially if the crystal is in a temperature controlled environment. Crystal oscillators, however, cannot easily be tuned in frequency. A solution to this problem is provided by the phase-locked loop (PLL), which uses a feedback control circuit and an accurate reference source (usually a crystal-controlled oscillator) to provide an output that is tunable with very high accuracy. Phase-locked loops and other circuits that provide accurate and tunable frequency outputs are called frequency synthesizers. Important parameters that characterise frequency synthesizers are tuning range, frequency switching time, frequency resolution, cost and power consumption. 1.2.6 Heterodyne Receivers The most popular type of receiver used today is the superheterodyne circuit as shown in Fig.1.4. The IF frequency is generally selected to be between the RF frequency and baseband. At microwave and millimetre wave frequencies, it is often necessary to use two stages of down conversion to avoid problems due to LO stability. The dual-conversion superheterodyne receiver of Fig.1.4 employs two local oscillators and mixers to achieve down conversion to baseband with two IF frequencies. 7 1. Radio Receiver Systems Figure 1.4 Dual-IF heterodyne receiver. Filtering a narrow channel that is centred at high frequencies and is accompanied by large interferers demands prohibitively high Q’s. In heterodyne architectures, the signal band is translated to much lower frequencies so as to relax the Q required of the channel-select filter. This translation is carried out by means of a mixer as shown in Fig.1.5a. To bring the centre frequency from ω1 to ω2, the signal is first mixed with a sinusoid A0cosω0t, where ω0=ω1−ω2, thereby yielding a band around ω2 and another around 2ω1−ω2. A low-pass filter then removes the latter. This operation is called down conversion. Because of its typically high noise, the down conversion mixer is preceded by a low-noise amplifier as shown in Fig.1.5b. For a heterodyne architecture with two input frequencies, x1(t)=A1cosω1t and x2(t)=A2cosω2t, the bands symmetrically located above and below the LO frequency are down-converted to the same centre frequency (Fig.1.6). If the received band of interest is centred around ω1 (=ωLO−ωIF), then the image is around 2ωLO−ω1 (=ωLO+ωIF) and vice versa. This problem of image is a serious one. While each wireless standard imposes constraints upon the signal emissions by its own users, it may have no control over the signals in other bands. The image power can therefore be much higher than that of the desired signal, requiring proper image rejection. 8 1. Radio Receiver Systems Figure 1.5 (a) Simple heterodyne down conversion, (b) inclusion of an LNA to lower the noise figure. Figure 1.6 Problem of image in heterodyne reception. The most common approach to suppressing the image is through the use of an imagereject filter, place before the mixer. As depicted in Fig.1.7, the filter is designed to have a relatively small loss in the desired band and a large attenuation in the image band. These two requirements can be simultaneously met if 2ωIF is sufficiently large. Fig.1.8 shows two cases corresponding to high and low values of IF to illustrate the tradeoff. A high IF leads to substantial rejection of the image whereas a low IF allows great suppression of nearby interferers. The choice of IF therefore depends on trade-offs among three parameters: the amount of image noise, the spacing between the desired band and the image, and the loss of the image-reject filter. Therefore, the heterodyne architecture exhibits a trade-off between image rejection and channel selection. And since the image degrades the sensitivity of the receiver, the choice of IF entails a trade-off between sensitivity and selectivity. The selectivity and sensitivity of the heterodyne architecture have made it the dominant choice in RF systems for many decades. Despite the complexity and the need for a large number of external components, heterodyning is still viewed as the most reliable reception technique. 9 1. Radio Receiver Systems Figure 1.7 Image rejection by means of a filter. Figure 1.8 Rejection of image versus suppression of interferers for (a) high IF and (b) low IF. Pros: Superior sensitivity and selectivity performance Cons: Require high Q components Low level of monolithic integration 10 1. Radio Receiver Systems 1.2.7 Direct Conversion (Homodyne or Zero-IF) Receivers The direct conversion uses a mixer and local oscillator to perform frequency down conversion with a zero IF frequency. The LO frequency is equal to the input carrier frequency as shown in Fig.1.9, which is then converted directly to baseband. The channel selection requires only a low-pass filter with relatively sharp cut-off characteristics. The circuit of Fig.1.9(a) operates properly only with double sideband AM signals because it overlaps positive and negative parts of the input spectrum. For frequency and phase modulated signals, the down conversion must provide quadrature outputs (see Fig.1.9(b)) so as to avoid loss of information, This is because the two sides of FM or QPSK spectra carry different information and must be separated into quadrature phases in translation to zero frequency. The simplicity of the homodyne architecture offers two important advantages over a heterodyne receiver. First, the problem of image is circumvented because ωIF=0. As a result, no image filter is required. The sum frequency is twice the LO and easily filtered. Second, the IF SAW filter and subsequent down conversion stages are replaced with lowpass filters and baseband amplifiers that are amenable to monolithic integration. Direct conversion receivers are simpler and less costly than superheterodyne receivers since there is no IF amplifier, IF bandpass filter or IF local oscillator required for the final down conversion. If the homodyne architecture is so simple, why has it not become popular in RF systems? Direct translation of the spectrum to zero frequency entails a number of issues that do not exist or are not serious in a heterodyne receiver. Examples are: Channel selection – Rejection of out-of-channel interferers by an active low-pass filter is more difficult than by a passive filter, fundamentally because active filters exhibit much more severe noise-linearity-power trade-offs than do their passive counterparts. DC offsets – Since in a homodyne topology the down-converted band extends to zero frequency, extraneous offset voltages can corrupt the signal and more importantly, saturate the following stages. Consider the receiver shown in Fig.1.10, where the LPF is followed by an amplifier and an A/D converter. First, the isolation between the LO port and the inputs of the mixer and LNA is not infinite; that is, a finite amount of feedthrough exists between the LO port to points A and B (Fig.1.10a). This is called LO leakage. The leakage signal appearing at the inputs of the LNA and the mixer is now mixed with the LO signal, thus producing a DC component at point C. This phenomenon is called self-mixing. A similar effect occurs if a large interferer leaks from the LNA or mixer input to the LO port and is multiplied by itself (Fig.1.10b). Second, the total gain from the antenna to point X is typically around 80 to 100dB so as to amplify the microvolt input signal to a level that can be digitised by a low cost, low power ADC. Of this gain, typically 25 to 30dB is contributed by the LNA/mixer combination. In addition to introducing DC offsets, leakage of the LO signal to the antenna and radiation therefrom creates interference in the band of other receivers using the same wireless standard. 11 1. Radio Receiver Systems Figure 1.9 (a) Simple homodyne receiver, (b) homodyne receiver with quadrature down conversion. Figure 1.10 Self-mixing of (a) LO signal, (b) a strong interferer. I/Q mismatch – For phase and frequency modulation schemes, a homodyne receiver must incorporate quadrature mixing. This requires shifting either the RF signal or the LO output by 90°. Shifting the RF signal generally entails severe noise-power-gain tradeoffs, making it more desirable to use the topology in Fig.1.11b. In either case, the errors in the nominally 90° phase shift, and mismatches between the amplitudes of the I and Q signals corrupt the down-converted signal constellation, thereby raising the bit error rate. 12 1. Radio Receiver Systems Figure 1.11 Quadrature generation in (a) RF path, (b) LO path. Even-order distortion – Even order nonlinearity becomes problematic in homodyne down conversion. This is illustrated in Fig.1.12 with two strong interferers close to the channel of interest experience a nonlinearity such as y(t)=α1x(t)+α2x2(t) in the LNA. If x(t)=A1cosω1t+A2cosω2t, then y(t) contains a term, α2A1A2cos(ω1−ω2)t, indicating that two high frequency interferers generate a low frequency beat in the presence of evenorder distortion. Flicker noise – Since the down-converted spectrum extends to zero frequency, the 1/f noise of devices substantially corrupts the signals. This is a severe problem in MOS implementation. For this reason, it is desirable to achieve a relatively high gain in the RF range through the use of active mixers rather than passive mixers. As the stages following the mixer operate at relatively low frequencies, they can incorporate very large devices to minimise the magnitude of the flicker noise. Pros: Simpler and less costly than superheterodyne receivers No image problem, thus no need for image reject filter Use low pass filter, not band pass filter Eliminate many off-chip components High level of monolithic integration Cons: LO spurious leakage at antenna DC offsets – LO self mixing Strong interferers self mixing Require offset cancellation Flicker noise (1/f noise) 1.2.8 Digital-IF Receivers Before the transformation from baseband to a RF channel, the waveform is digitised to utilise the advantages of digital modulation. Coding is applied to the signal to more efficiently use the available bandwidth and to minimise the effects of noise and interference that will be introduced by the channel. The coded signal is filtered, modulated and changed back to an analogue wave that is converted to the desired frequency of transmission. Finally, the RF signal is filtered and amplified before it is transmitted from the antenna. At the receiver, the first IF signal is digitised, mixed with the quadrature phases of a digital sinusoid, and low-pass filtered to yield the quadrature baseband signals as shown in Fig.1.13. Digital processing avoids the problem of I and Q mismatch. 13 1. Radio Receiver Systems Figure 1.12 Effect of even-order distortion on interferers. Figure 1.13 Digital-IF receiver. After down conversion to the IF, the signal is separated into two distinct paths. To convert to baseband, each path is mixed with an LO whose frequency equals the IF frequency. The upper-path signal (I) is simply mixed with the LO and then filtered. In the lowerpath, a 90° phase shift is introduced in the mixing signal. This lower-path (Q) is converted to baseband by mixing with the phase-shifted LO signal, and then filtered. This process produces the I and Q baseband components of the data stream. The principal issue in this approach is the performance required for the A/D converter: - since the signal level at point A is typically no higher than a few hundred microvolts, the quantization and thermal noise of the ADC must not exceed a few tens of microvolts - if the first IF bandpass filter cannot adequately suppress adjacent interferers, the nonlinearity of the ADC must be sufficiently small to minimise corruption of the signal by intermodulation - the ADC dynamic range must be wide enough to accommodate variations in the signal level due to path loss and multipath fading - the ADC must achieve an input bandwidth commensurate with the value of IF while consuming a reasonable amount of power The ADC performance limitation can be partially alleviated as most ADCs incorporate sample-and-hold circuits and hence can perform down conversion. The sampling IF configuration is depicted in Fig.1.14 where the ADC samples the signal at a rate slightly below fIF. The spectrum of the down-converted, digitised signal thus lies around fIF−fS. This operation is followed by quadrature mixing and filtering to translate the spectrum to the baseband. In this receiver, the IF signal is digitalised. The sampled data stream from the ADC is digitally demodulated into its I and Q components and the original signal is reconstructed. Digital IF and sampling IF architectures have not been used in portable terminals because of the ADC performance limitations. 14 1. Radio Receiver Systems Figure 1.14 Sampling IF architecture. 1.2.9 Subsampling Receivers RF input can be sampled at a much lower rate because narrowband signals exhibit only a small change from one carrier cycle to the next. The idea is that a bandpass signal with bandwidth ∆f can be translated to a lower band if sampled at a rate equal to or greater than 2∆f. This operation creates replicas of the spectrum with no aliasing as shown in Fig.1.15. Due to the large reduction in the down conversion rate, the use of subsampling can simplify the design of the local oscillator and its associated synthesiser loop. Despite these features, subsampling suffers from aliasing of noise. Pros: Digital implementation at IF Digital programmable IF filter Support multi-standard Cons: Large dynamic range of ADC Aliasing of noise Require high IP3 RF components High power consumption 1.3 System Specifications 1.3.1 System Noise and Noise Floor Assuming that a system’s gain is sufficient, the weakest signal that may be processed satisfactorily, a figure of merit referred to as noise floor or minimum detectable signal (MDS). For a given system noise power, the MDS determines the minimum signal-tonoise ratio (SNR) at the demodulator of the receiver. The minimum SNR at the output of a receiver is sometimes expressed in terms of the signal plus noise and distortion-to-noise and distortion ratio (SINAD). The SINAD is related to SNR as SINAD = S+N S = 1+ N N Knowing the minimum SNR or SINAD and the noise characteristics of the receiving system allows us to calculate the minimum detectable signal power. 15 1. Radio Receiver Systems Figure 1.15 Subsampling in (a) time and (b) frequency domains. The noise power is equal to Pn = kTB where Pn is the noise power, k is the Boltzmann’s constant (1.38x10-23 watts per Kelvin), T is the temperature in Kelvins, and B is the bandwidth (in Hertz) in which the noise appears. If the noise figure and bandwidth are known, the noise floor can be calculated using the equation Noise Floor = −174dBm + NF + 10 log B 16 1. Radio Receiver Systems 1.3.2 Signal-to-Noise Ratio and Sensitivity Successful radiocommunication depends on the achievement of a specified minimum ratio of signal power to noise power, at the output of the receiver. The input voltage, expressed in absolute units or decibels relative to a microvolt (dBµV) necessary to achieve a particular signal-to-noise ratio in a particular bandwidth may be specified as a figure of merit called sensitivity. To calculate the sensitivity, we write NF = = SNRin SNRout Psig / PRS SNRout where Psig denotes the input signal power and PRS the source resistance noise power, both per unit bandwidth. Since the overall signal power is distributed across the channel bandwidth, B, the two sides of the above equation must be integrated over the bandwidth to obtain the total mean square power. Thus, for a flat channel, Psig ,tot = PRS ⋅ NF ⋅ SNRout ⋅ B The above equation predicts the sensitivity as the minimum input signal that yields a given value for the output SNR. We have Pin , min (dB ) = PRS (dBm / Hz ) + NF (dB ) + SNRmin (dB ) + 10 log B = −174dBm + NF + 10 log B + SNR min where Pin,min is the minimum input level that achieves SNRmin and B is expressed in Hertz. Since Pin,min is a function of the bandwidth, a receiver may appear very sensitivity simply because it employs a narrowband channel (at the cost of low information rate). 1.3.3 Noise Factor and Noise Figure The noise floor predicted by Pn=kTB cannot be achieved and maintained in any real network or system of networks because all real networks generate noise. Determining how closely the SNR achieved at a given input level approaches the SNR achievable at that input level in a noiseless network is therefore of high interest to the circuit and system designer. The degree to which a network’s noise contribution degrades the noise floor is evaluated by its noise factor (F), which is expressed as the ratio N + N added F = in N in where Nin is the noise power available from the source, and Nadded is the noise power added by the network, with both powers determine in the same bandwidth. Expressing this ratio in dB returns the noise figure (NF=10logF), a bandwidth independent figure of merit of great value in evaluating the noise performance of networks and communication systems. It can be viewed as a measure of the degradation in the signal-to-noise ratio between the input and output of the component. We can also express NF as the ratio of the network’s input SNR to its output SNR, S / N in NF = in S out / N out 17 1. Radio Receiver Systems where NF is the noise figure in dB, S is the signal power, N is the noise power, and all powers determined in the same bandwidth. Noise figure is a measure of how much the SNR degrades as the signal passes through a system. Noise figure can also be defined for antennas and antenna systems, N / N ant NFant = t Nt where NFant is the antenna system noise figure in dB, Nt is the antenna system’s thermal noise power and Nant is the total noise power picked up by the antenna system. 1.3.4 Intermodulation The new signals produced through intermodulation distortion (IMD) can profoundly affect the performance even of systems operated far below gain compression. IMD products of significant power can appear at frequencies remote from, in, and/or near the system passband, resulting in demodulation errors (in reception) and interference to other communications (in transmission). Where an IMD product appears relative to the passband depends on the passband width and centre frequency, the frequencies of the signals present at the system input, and the order of the nonlinearity involved. These factors also determine the strength of an IMD product relative to the desired signal. If a weak signal accompanied by two strong interferers experiences third-order nonlinearity, then one of the IM products falls in the band of interest, corrupting the desired component, as shown in Fig.1.16. To understanding how intermodulation arises, assuming, x(t ) = A1 cos ω1t + A2 cos ω 2 t For a system whose output does not depend on the past values of its input, y (t ) = α 1 x(t ) + α 2 x 2 (t ) + α 3 x 3 (t ) = α 1 (A1 cos ω1t + A2 cos ω 2 t ) + α 2 (A1 cos ω1t + A2 cos ω 2 t ) + α 3 (A1 cos ω1t + A2 cos ω 2 t ) Expanding the left side and discarding dc terms and harmonics, we obtain the following intermodulation products: 2 ω1 ± ω 2 = α 2 A1 A2 cos(ω1 + ω 2 )t + α 2 A1 A2 cos(ω1 − ω 2 )t 3α 3 A12 A2 3α A 2 A cos(2ω1 + ω 2 )t + 3 1 2 cos(2ω1 − ω 2 )t 4 4 2 3α A A 3α A 2 A 2ω 2 ± ω1 = 3 2 1 cos(2ω 2 + ω1 )t + 3 2 1 cos(2ω 2 − ω1 )t 4 4 And these fundamental components 3 3 3 3 3 2 3 2 α 1 A1 + α 3 A1 + α 3 A1 A2 cos ω1t + α 1 A2 + α 3 A2 + α 3 A2 A1 cos ω 2 t 4 2 4 2 The components 2ω1−ω2 and 2ω2−ω1 appear in the vicinity of ω1 and ω2. 2ω1 ± ω 2 = All the second-order products are undesired in an amplifier, but in a mixer the sum or difference frequencies form the desired outputs. In either case, if ω1 and ω2 are close, all the second-order products will be far from ω1 or ω2, and can easily be filtered (either passed or rejected) from the output of the component. 18 3 1. Radio Receiver Systems Figure 1.16 Corruption of a signal due to intermodulation between two interferers. The corruption of signals due to third-order intermodulation of two nearby interferers is so common and so critical that a performance metric, has been defined to characterise this behaviour. Called the “third intercept point” (IP3), this parameter is measured by a twotone test in which A (=A1=A2) is chosen to be sufficiently small so that higher-order nonlinear terms are negligible and the gain is relatively constant and equal to α1. As A increases, the fundamentals increase in proportion to A, whereas the third-order IM products increase in proportion to A3 (Fig.1.17a). Plotted on a logarithmic scale (Fig.1.17b), the magnitude of the IM products grows at three times the rate at which the main components increase. The third-order intercept point is defined to be at the intersection of the two lines. 1.3.5 Dynamic Range Dynamic range (DR) is generally defined as 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. Thermal noise sets the lower limit of the power span over which a network can operate. The upper limit of a network’s power span is the level at which the power of one IM product of a specified order is equal in power to the network’s noise floor. The ratio of the noise floor power to the upper limit signal power is referred to as the network’s dynamic range (n − 1)(IPn,in − MDS ) DRn = n where DR is the dynamic range in dB, n is the order, IPin is the input intercept power in dBm and MDS is the minimum detectable signal power in dBm. The receiver dynamic range depends on the noise characteristics of the receiver as well as the type of modulation being used, and the required minimum SNR. 1.3.6 Receiver Selectivity Selectivity is a parameter that quantifies the tendency of a receiver to respond to channels adjacent to the desired reception channel. As international regulations are gradually moving to narrower channel spacings, receiver selectivity assumes greater importance since it frequently limits system performance and places restrictions on frequency allocation and system utilisation. 19 1. Radio Receiver Systems Figure 1.17 Growth of output components in an intermodulation test. [ ] Selectivity = −CR − 10 log 10 − IFsel / 10 + 10 − Spurs / 10 + BW ⋅ 10 SBN / 10 where Selectivity = amount of adjacent channel selectivity relative to nominal receiver sensitivity (dB) CR = capture ratio, or cochannel rejection (dB) IFsel = IF filter rejection at the adjacent channel (dB) Spurs = LO spurious signals present in the IF bandwidth at a frequency offset equal to the channel spacing (dBc) BW = IF noise bandwidth (Hz) SBN = SSB phase noise of LO at a frequency offset equal to the channel spacing (dBc/Hz) 1.3.7 Receiver Spurious Responses Receiver spurious responses are frequencies that are different from the desired receive frequency, yet that still produce demodulated output in the receiver, if encountered at a sufficiently high level. This situation is obviously undesirable and particularly troublesome in receivers capable of tuning over a wide frequency range, because the RF filters need to be wide to accommodate the wide frequency coverage. Most receiver spurious responses are actually mixer spurious responses, which may or may not be further attenuated by RF selectivity in the preceding stages. Most receiver spurious responses result from harmonic mixing of the RF and LO signals. Any RF frequency that satisfies the following relationship is a potential receiver spurious response: ± mf RF ± nf LO = ± f IF where fRF = any incoming frequency into the mixer RF port fLO = local oscillator frequency fIF = desired IF frequency m = integer multiplier of RF frequency n = integer multiplier of LO frequency 1.3.8 Receiver Sensitivity The minimum detectable signal power can be converted to a minimum detectable signal voltage for a given receiver input impedance. This quantity is called the receiver sensitivity. Assume receiver sensitivity is limited by thermal noise. 20 1. Radio Receiver Systems Must consider device gains and noise figures, image noise, and local oscillator wideband noise separately, and then combine them to produce an overall equivalent input noise factor, FT = FIN + F ' IN + F ' ' IN FT = total equivalent input noise factor (linear) FIN = total equivalent input noise factor derived from on-channel stage noise figures and gains (linear) F’IN = total equivalent input noise factor derived from image frequency stage noise figures and gains (linear) F’’IN = total equivalent input noise factor derived from local oscillator wideband noise (linear) The overall receiver sensitivity for a desired detector S/N ratio is given as S e = FT kTB RG N e = receiver sensitivity (V) FT = total equivalent input noise factor (linear) k = Boltzmann’s constant, 1.38x10−23 (J/K) T = temperature (K) B = equivalent noise bandwidth of system (Hz) S/N = required S/N at detector output (linear) RG = system impedance (Ω) 1.3.8.1 Contribution from Stage Gains and Noise Figures In a typical wireless receiver, the input signal travels through a cascade of several different components such as filters, amplifiers, mixers and transmission lines. Each of these stages will progressively degrade the signal-to-noise ratio, so it is important to quantify this effect to evaluate the overall performance of the receiver. Noise figures and gains of all stages must be known to calculate the equivalent input noise factor. Friis formula for calculating cascaded noise figure is used to combine the stage contributions and can be generalised for n stages, n F −1 FIN = 1 + ∑ i −1i i =1 ∏G j j =0 FIN = equivalent input noise figure (linear) Fi = stage noise factor (linear) Gj = stage gain (linear, not in dB), G0=1 This relation proves especially useful if a receiver employs various off-the-shelf building blocks that are characterised independently by manufacturers. These results show that the noise characteristics of a cascaded system are dominated by the first few stages, since the effect of the later stage is reduced by the product of the gains of the preceding stages. Thus for best overall system noise performance, the first stage of a receiver should have a low noise figure and at least moderate gain. Expense and effort are most rewarded when 21 1. Radio Receiver Systems applied to improving the noise characteristics of the first or second stage, as opposed to later stages, since later stages have a diminished impact on overall noise performance. 1.3.8.2 Contribution from Image Noise Image noise (Fig.1.18) is simply noise contained at the receiver’s image frequency which is present at the first mixer’s RF port. This noise will be added to the down-converted signal produced by an on-channel signal on a power basis. Thus, the analysis needs to be carried up to the mixer stage that combines the two sidebands into one IF signal. Image noise is down-converted to IF with similar conversion loss as the desired signal. Must carefully sort out the different noise factors and gains that apply to the image frequency and the on-channel frequency. Carry out the normal analysis at the image frequency and normalise it by the ratio of the overall image frequency gain to the on-channel frequency gain. Filtering deliberately introduced to eliminate image noise by reflection mismatched case. In the match case, the noise figure of a passive stage is equal to its loss. In the mismatched case, we must allow zero noise figure to be assigned to a passive stage, which provides image attenuation by reflection. Image noise contribution to equivalent input noise factor is (assuming that the mixer conversion loss is the same at the desired frequency as at the image frequency) n ∏ G' j n F ' −1 j =1 i 1 + F ' IN = n i −1 ∑ ∏ G j i =1 ∏ G' j j =1 j =0 F’IN = contribution to overall input noise factor by image noise (linear) F’i = stage noise factor at image frequency (linear) F = 1 for stage that provides image attenuation by reflection (ideally the stage immediately preceding the mixer) G’j = stage gain at image frequency (linear), G’0 = 1 Gj = stage gain on-channel (linear), G0 = 1 n = number of stages up to but excluding the mixer 1.3.8.3 Contribution from Wideband LO Noise Wideband noise separated from the LO frequency and its harmonics by ±mfIF spacings will mix to produce noise at the IF frequency. This is shown in Fig.1.18. Noise at frequencies of ±fIF spacing from the LO harmonics also contributes and may be dominant in some cases. In contrast to image noise, wideband LO noise is down-converted to IF with a much higher conversion loss than the desired signal. The noise level in dBc/Hz has to be measured separately at nfLO+fIF and nfLO−fIF for each n. This conversion of noise sidebands into IF noise is called mixer noise balance. It is actually a conversion loss for the relevant sideband from the LO port to the IF port. 22 1. Radio Receiver Systems Figure 1.18 Image noise and wideband LO noise contributions to overall noise figure degradation. There may be a bandpass injection filter between the LO and the mixer, its loss at the appropriate noise sideband needs to be taken into account. m F ' ' IN = ∑ s =1 10 (PLO +Ws − Ls − M s ) / 10 n 1000kT0 ∏ G j j =0 F’’IN = contribution to overall noise factor by wideband LO AM noise (linear) PLO = local oscillator power (dBm) Ws = wideband noise level of sideband s (dBc/Hz) Ls = loss of injection filter at frequency of sideband s (dB) Ms = mixer noise balance for sideband s (dB) k = Boltzmann’s constant, 1.38x10−23 (J/K) T0 = 290K or 16.85°C Gj = gain of stage j (linear) s = index for summation of noise powers at all sidebands of interest m = number of sidebands taken into account j = index to calculate gain up to and including the mixer n = number of stages up to and including the mixer The factor of 1000 comes from converting decibels relative to 1milliwatt to Watts. 1.3.9 Receiver Intercept Point Intercept point is a measure of circuit or system linearity that allows us to calculate distortion or IM product levels from the incoming signal amplitudes. The input intercept point represents a fictitious input amplitude at which the desired signal components and the undesired components (in this case the third-order products) are equal in amplitude as shown in Fig.1.19. The order of the intercept point refers to how fast the amplitudes of the distortion products increase with an increase in input level. For example, for the third-order intercept point (IP3), the IM product will increase in amplitude by 3dB when the input signal is raised by 1dB. 23 1. Radio Receiver Systems Figure 1.19 Third-order intercept diagram for a nonlinear component. When a system is analysed for its overall equivalent intercept point resulting from the combination of its constituent device intercept points, the most frequent assumption is that the distortion products of the various stages are uncorrelated and add on a power basis. 1.3.9.1 Second-Order Intercept Point Second-order intercept point (IP2) is used to predict mixer performance with respect to a particularly troublesome spurious response, called the half-IF (½IF) spurious response. This receiver spurious response is separated from the desired channel by one-half IF frequency. The mechanism for ½IF generation is 2fRF±2fLO, where both harmonics are internally generated, not fed in from the outside. 1 1 IF rejection = ( IP 2 − S − CR ) 2 2 IP2 = equivalent input second-order intercept point at receiver input (dBm) S = receiver sensitivity (dBm) CR = capture ratio, or cochannel rejection (dB) This analysis only needs to include the mixer intercept point because it is the stage that generates the ½IF spurious response. 1.3.9.2 Third-Order Intercept Point For small input powers, the third-order products must be very small, but will increase quickly as input power increases. 24 1. Radio Receiver Systems The output power of the first-order, or linear, product is proportional to the input power and so the line describing this response has a slope of unity (before the onset of compression) as shown in Fig.1.19. The line describing the response of the third-order products has a slope of 3. Both the linear and third-order responses will exhibit compression at high input powers, the extensions of their idealised responses are shown with dotted lines. They will intersect at a point above the onset of compression, as shown in the figure. This is the third-order intercept point. Third-order intercept point (IP3) is another important measure of system linearity. It is the theoretical point at which the desired signal and third-order distortion products are equal in amplitude. IP3 determines the amount of IM distortion produced in the receiver itself when subjected to high-level interference. The following procedure is used to calculate equivalent system input intercept point: 1. Draw block diagram of system with associated gains and IP3s. 2. Transfer all intercept points to system input, subtracting gains and adding losses decibel for decibel. 3. Convert input intercept points to powers (dBm to milliwatts). 4. Assuming all intercept points are independent and uncorrelated, add powers “in parallel”: IPINPUT = [mW] 1 1 1 + + IP1 IP2 1 + IPN 5. Convert IPINPUT to dBm (milliwatts to dBm). IP3 INPUT 1 = 10 log 1 1 + + IP1 IP2 1 + IPN IP3INPUT = equivalent system input intercept point (dBm) IP1 = IP3 of first stage transferred to input (mW) IPN = IP3 of last stage transferred to input (mW) An example is shown in Fig.1.20. IP3 INPUT 1 = 10 log 1 1 1 1 1 + + + + ∞ 15.85 ∞ 19.95 100 = 10 log(8.12) = 9.01 dBm The equivalent system intercept point is 9.01dBm; the amplifier is the dominant contributor because is has the lowest input equivalent intercept point. For an ideal linear amplifier a plot of the output power versus input power is a straight line with a slope of unity, and the gain of the amplifier is given by the ratio of the output power to the input power. The amplifier response of Fig.1.19 tracks the ideal response over a limited range, then begins to saturate resulting in reduced gain. This is the 1dB compression point as the power level for which the output power has decreased by 1dB from the ideal characteristics. 25 1. Radio Receiver Systems Filter Amplifier Filter G=-2dB IP3=∞ G=12dB IP3=10dBm G=-3dB IP3=∞ Mixer G=-7dB IP3=20dBm Amplifier G=22dB IP3=20dBm IP1 = ∞ IP2 = 10+2 = 12dBm=15.85mW IP3 = ∞-12+2 = ∞ IP4 = 20+3-12+2 = 13dBm=19.95mW IP5 = 20+7+3-12+2 = 20dBm=100mW Figure 1.20 Example of system IP3 calculation. The 1dB compression point (the level at which the gain drops by 1dB due to device saturation) is about 11dB below the IP3 for amplifiers and 15dB below the IP3 for mixers. The output power of an amplifier at the 1dB compression point will therefore be about 11dB below the output IP3. The output intercept point equals the input intercept point plus device gain. Intermodulation distortion is a property of all systems that exhibit a nonlinear transfer function, such as amplitude compression at sufficiently high levels. Third-order IM distortion is most often produced when two signals, separated by ∆f and 2∆f from the desired carrier beat together and produce on-channel interference. Intermodulation rejection is the difference in decibels between sensitivity and input signal level sufficient to produce interference can be calculated from the intercept point and receiver sensitivity by use of 1 IM = (2 IP3 − 2 S − CR ) 3 IM = intermodulation rejection (dB) IP3 = equivalent input third-order intercept point (dBm) S = receiver sensitivity (dBm) CR = capture ratio, or cochannel rejection (dB) 26 1. Radio Receiver Systems Example 1.1 Cascaded Noise Figure st st Filter 1 Filter 2 1 mixer 1 IF filter 2nd IF filter 2nd IF amplifier Detector 2nd mixer 1st IF amplifier RF amplifier Injection filter 2nd LO 1st LO Stage Filter 1 RF amplifier Filter 2 1st mixer 1st IF filter 1st IF amplifier 2nd IF filter 2nd mixer 2nd IF amplifier NFtotal Gain (dB) -0.5 12 -3 -7.5 -1.5 22 -5 -8 12 Noise figure (dB) 0.5 2 3 7 1.5 3 5 8 20 Gain (linear) 0.891 15.849 0.501 0.178 0.708 158.489 0.316 0.158 15.849 Noise figure (linear) 1.122 1.585 1.995 5.012 1.413 1.995 3.162 6.310 100.000 n F −1 = 10 log 1 + ∑ i −1i i =1 Gj ∏ j =0 F −1 F −1 F −1 F −1 F −1 F − 1 F3 − 1 F −1 = 10 log F1 + 2 + + 4 + 5 + 6 + 7 + 8 + 9 G1 G1G 2 G1G 2 G3 G1 ...G 4 G1 ...G5 G1 ...G6 G1 ...G7 G1 ...G8 = 10 log[1.122 + 0.657 + 0.07 + 0.567 + 0.328 + 1.116 + 0.015 + 0.119 + 14.032] = 12.56dB or 18.026 Total gain = 20.5dB Equivalent input noise figure = 18.02 or 12.56dB Equivalent input noise temp = (NF−1)*290 = 4935.8K 27 1. Radio Receiver Systems Example 1.2 Receiver Sensitivity Calculation st st Filter 1 Filter 2 1 mixer 1 IF filter 2nd IF filter 2nd IF amplifier Detector 2nd mixer 1st IF amplifier RF amplifier Injection filter 2nd LO 1st LO Stage Filter 1 RF amplifier Filter 2 1st mixer 1st IF filter 1st IF amplifier 2nd IF filter 2nd mixer 2nd IF amplifier Detector Gain (dB) -2.5 12 -2 -8 -1.5 20 -4 -8 20 Noise figure (dB) 2.5 3.5 2 8.3 1.5 4 4 8 10 15 Gain (linear) 0.562 15.849 0.631 0.158 0.708 100.000 0.398 0.158 100.000 Noise figure (linear) 1.778 2.239 1.585 6.761 1.413 2.512 2.512 6.310 10.000 31.623 10dB 12kHz -165dBc/Hz (flat with frequency) 23.5dBm 0dB at fLO±fIF offset 10dB at 2fLO±fIF 20dB at 3fLO±fIF Mixer noise balance 30dB at fLO±fIF 25dB at 2fLO±fIF 20dB at 3fLO±fIF 6dB or 3.981 Required S/N at detector output System impedance 50Ω 290K Ambient temperature Assume that the stage gains and the noise figures are the same at the image frequency as onchannel for all stages, except for filter 2, which has a 10dB loss and a noise figure of 0dB at the image frequency. Neglect any contribution from the second image in the second IF. Filter 2 image attenuation Equivalent noise bandwidth First LO wideband noise First LO power Injection filter attenuation 1. Total equivalent input noise factor derived from on-channel stage noise figures and gains n F −1 FIN = 1 + ∑ i −1i i =1 ∏G j j =0 = 1.778 + 2.205 + 0.066 + 1.025 + 0.465 + 2.405 + 0.024 + 0.212 + 2.276 + 0.077 = 10.533 28 1. Radio Receiver Systems 2. Total equivalent input noise factor derived from image frequency stage noise figures and gains before the mixer The stage gains and the noise fiqures are the same at the image frequency as on-channel for all stages, except for Filter 2 which has a 10dB loss and a noise figure of 0dB at the image frequency. Stage Gain Noise figure Gain Noise figure (dB) (dB) (linear) (linear) -2.5 2.5 Filter 1 0.562 1.778 12 3.5 RF amplifier 15.849 2.239 -10 0 Filter 2 0.100 1.000 F ' i −1 j= 1+ = n i −1 ∑ i =1 G G ' ∏1 j ∏0 j j= j= n F ' IN ∏1 G ' j n 0.562 * 15.849 * 0.1 [1.778 + 2.205 + 0] 0.562 * 15.849 * 0.631 = 0.631 = 3. Total equivalent input noise factor derived from local oscillator wideband noise up to and including the first mixer The LO wideband noise has 6 components, thus m=6. m F ' ' IN = ∑ s =1 10 (PLO +Ws − Ls − M s ) / 10 n 1000kT0 ∏ G j [ j =0 = 10 (23.5−165−0 −30 ) / 10 + 10 (23.5−165− 0−30 ) / 10 + 10 (23.5−165−10− 25 ) / 10 + 10 (23.5−165−10 − 25 ) / 10 + 10 (23.5−165− 20− 20 ) / 10 ][ ] + 10 (23.5−165− 20 − 20 ) / 10 / 1000 * 1.38 *10 − 23 * 290 * (0.562 *15.849 * 0.631 * 0.158) = 5.642 Total input equivalent noise factor, FT = FIN + F ' IN + F ' ' IN = 10.533 + 0.631 + 5.642 = 16.806 Receiver sensitivity, e = FT kTB S RG N = 16.806 * 1.38 * 10 − 23 * 290 * 12000 * 3.981 * 50 = 0.401µV Converting to dBm, ( 0.401 * 10 −6 e = 10 log 50 = −144.93dB = −114.93dBm ) 2 Note: the LO wideband noise contributes significantly to sensitivity degradation while contributions from image noise may be neglected. 29