Television system dissect the image and transmit the pixel information serially. The image is divided into a stack of horizontal stripes (“lines”) which are scanned left to right, producing a sequence of pixel (picture element) brightness values. The lines are scanned in order, one after the other, from top to bottom. Brightness values for each pixel are transmitted to the receiver(s). The image is reconstructed by a display device, whose pixels are illuminated according to the received brightness values. This chapter presents television technology in historical order: (1) the electromechanical system that Nipkov patented in 1884 but which was not demonstrated until 1923; (2) all-electronic television, made possible by the development of cathode ray picture tubes and camera tubes; and (3) digital television, which uses data storage and processing in the receiver, allowing the station to update the changing parts of the image, rather than retransmit the entire image for every frame. With the lowered data rate, the bandwidth needed previously to transmit one analog television program can now hold multiple programs.

The Nipkov system

Electronic image dissection and reconstruction were first proposed in the Nipkov disk system, patented in 1884, which used a pair of rotating disks, as shown in Figure 19.1. The camera disk dissected the image while the receiver disk reconstructed it. The receiver screen, a rectangular aperture mask, was covered by an opaque curtain containing a pin hole, illuminated from behind by an intensity-modulated gas discharge lamp.

The idea that radio waves could be used to detect the presence of stationary or moving objects emerged around 1900, almost as soon as radio itself. Christian Hueslmeyer, a German inventor, demonstrated an apparatus in 1904 which, when mounted on a bridge above the Rhine, rang a bell when a ship passed beneath. He used a (now) primitive spark gap RF source and coherer detector. The system may have shown only marginal potential for collision avoidance, as the German Navy demonstrated no interest. Sir Robert Watson-Watt developed meteorological radar in Britain in the 1930s and then a chain of air defense radars during World War II. In the U.S., the MIT Radiation Laboratory, set up to develop military microwave radar systems, had nearly 4000 employees between 1940 and 1945. The acronym RADAR, for Radio Detection And Ranging, has been attributed to U.S. Navy officers F. R. Furth and S. M. Tucker, who introduced it in 1940, though the term remained classified throughout the war.

Today, radar goes beyond aircraft tracking to applications as diverse as space object monitoring, storm tracking, detection of clear air turbulence, and velocity measurements of speeding automobiles and baseballs. In this chapter, we look at some commonly used radars, some general system aspects of radar, and, finally, some RF components and techniques developed specifically for radar.

Some representative radar systems

Classic surveillance radar

Figure 21.1 is a block diagram of the classic radar system used to monitor air traffic.

Modulation means variation of the amplitude or the phase (or both) of an otherwise constant sinusoidal RF carrier wave in order that the signal carry information: digital data or analog waveforms such as audio or video. In this chapter we look at pure amplitude modulation (AM) and pure frequency modulation (FM). Historically, these were the first methods to be used for communications and broadcasting. While still used extensively, they are giving way to modulation schemes, mostly digital, some of which amount to simultaneous AM and FM.

The simplest form of AM is on/off keying. This binary digital AM (full on is a data “1” and full off is a data “0”) can be produced with a simple switch, originally a telegraph key in series with the power source or the antenna. The first voice transmissions used a carbon microphone as a variable resistor in series with the antenna to provide a continuous range of amplitudes. With AM, the frequency of the carrier wave is constant, so the zero crossings of the RF signal are equally spaced, just as they are for an unmodulated carrier. The simplest FM uses just two frequencies; the carrier has frequency f0 for data “zero” and f0 + Δf for data “one.” FM is usually generated by a VCO (voltage-controlled oscillator). For binary FSK (frequency shift keying), the control voltage has only two values: one produces f0 and the other produces f0 + Δf.

It was already discussed in earlier chapters that the amplitude and the phase responses of amplifiers change with frequency, either intentionally or non-intentionally. In some applications, a specific frequency response is desired, for example a band-pass characteristic for an LNA (here, LNA stands for “low-noise amplifier” – the low-noise input stage of a receiver). To shape the frequency response according to our needs, we use reactive components; such as inductors and capacitors. For other applications a flat frequency response is required; however, gain inevitably drops at higher frequencies. The reason for this “non-intentional” change of the frequency response is the “parasitic” components of the circuit; i.e. all non-avoidable reactive (usually capacitive) components related to the devices and the interconnections. To investigate the essential frequency-dependent behaviors of amplifiers, it is necessary to improve the small-signal equivalent circuit of a MOS transistor developed in Chapter 2.

The small-signal equivalent circuit of a MOS transistor biased at a certain operating point in the saturation region was given in Chapter 2, Fig. 2.6. To extend the usability of this equivalent circuit to high frequencies (in other words, to radio frequencies), it is necessary to add the parasitic capacitances and the parasitic resistances that were discussed in Chapter 1, as shown in Fig. 3.1(a). In this equivalent circuit RD, RS and RG are the series resistances of the corresponding regions of the device, where d′, s′ and g′ represent the so-called “internal nodes” that are inaccessible from the device terminals; d, s and g represent the external terminals.

In the earlier chapters of this book, we have introduced and examined the structure and operation of fundamental building blocks, or essential components, of high-frequency integrated circuits. The emphasis has been on transistor-level operation, the influence of device characteristics and parasitic effects, as well as the input–output behavior in time and frequency domains – with the intention to fill the gap between fundamental electronic circuits textbooks and more advanced RF IC design textbooks that mainly focus on the state-of-the-art. In this chapter, we will address a different domain and consider various aspects of “bringing together an entire system”, especially for interfacing high-frequency analog components with corresponding digital processing blocks, using data converters. Instead of considering specific system architectures, the main aspects of system design will be presented with a generic approach in the following, as much as possible. The main philosophy is very similar to that of the earlier chapters: discussing design-oriented strategies and drawing attention to key issues that must be taken into account when combining analog and digital building blocks. For a detailed discussion of the system components and their circuit-level realizations, the reader is advised to consult any one of the excellent texts that already exist in this domain.

General observations

The vast majority of integrated systems used in communication applications today consist of analog as well as digital building blocks, combined within one package in close proximity to each other, or fabricated on a single chip substrate.

During the early days of radio design, the tuned amplifier was one of the most important subjects of electronic engineering. It was used as the input RF amplifier of radio receivers tunable in a certain frequency range and as an intermediate frequency (IF) amplifier tuned to a fixed frequency. From the 1930s to 1950s RF amplifiers using electron tubes were one of the most important and interesting research areas of electronics engineering and were investigated in depth. TV and all other wireless systems were other application areas for tuned amplifiers.

After the emergence of transistor circuits in the 1950s, the knowledge already acquired was sufficient for transistorized tuned amplifiers, because the frequency range was still limited to several hundreds of MHz. Active filters, one of the benefits of analog ICs, largely replaced another class of frequency-selective circuits, i.e. L-C filters. As a result of these developments the importance of inductors and circuits containing inductors decreased and eventually these subjects disappeared from many electronic engineering curricula and textbooks.

The rapid expansion of wireless personal communication and data communication during the recent decade, and the developments in IC technology that extended the operation frequencies into the GHz range, resulted in the “re-birth” of frequency-selective circuits containing inductors.

Basic MOS amplifiers are the main building blocks of a vast array of analog signal processing systems as well as other analog electronic circuits. The overall performance of a complex circuit strongly depends on the performances of its basic building blocks. In this chapter the main properties of these basic circuits will be investigated.

Common source (grounded source) amplifier

The basic structure of a common source amplifier is shown in Fig. 2.1(a). The gate of the NMOS transistor, M, is biased with a DC voltage source VGS, to conduct the appropriate DC (quiescent) current. A signal source (vi) is connected in series with the bias voltage to control the drain current. The load resistor RD helps to convert the drain current variations into output voltage variations (output signal). Since the output of a MOS amplifier is usually connected to the gate of another MOS amplifier, the DC and low-frequency load coming from the subsequent stage is negligible and the only load is RD. In Fig. 2.1(b), this “DC load line” with a slope equal to (−1/RD) is drawn on the output curves of M. For a given value of the gate bias voltage (for example VGS1), the drain current is ID1, which corresponds to the intersection of the load line and the output curve corresponding to VGS1. This intersection point is called as the “operating point” or the “quiescent point” and denoted by Q.

In the first half of the twentieth century the radio was the main activity area of the electronics industry and, correspondingly, RF circuits occupied a considerable part in the electronic engineering curriculum and books published in this period. Properties of resonance circuits and electronic circuits using them, single-tuned amplifiers as the input stages of receivers, double-tuned circuits as IF amplifiers, RF sinusoidal oscillators and high-power class-C amplifiers have been investigated in depth. It must be kept in mind that the upper limit of the radio frequencies of those days was several tens of MHz, the inductors used in tuned circuits were air-core or ferrite-core coils with inductance values in the micro-henries to milli-henries range, having considerably high quality factors, ranging from 100 to 1000, and the tuning capacitors were practically lossless.

The knowledge developed for the vacuum-tube circuits easily adapted to the transistors with some modifications related to the differences of the input and output resistances of the devices. In the meantime, the upper limit of the frequency increased to about 100 MHz for FM radio and to hundreds of MHz for UHF-TV. The values of the inductors used in these circuits correspondingly decreased to hundreds to tens of nano-henries. But these inductors were still wound, high-Q discrete components.

In the second half of the twentieth century, the emergence of integrated circuits drastically increased the reach of electronic engineering.

Oscillators, especially sinusoidal oscillators, are among the main components of all RF systems. They are being used as so-called “pilot” oscillators to generate the carrier frequencies of transmitters and as local oscillators in receivers to generate the signals that are necessary for frequency conversion purposes. The most important features of sinusoidal oscillators are the frequency of oscillation and its spectral purity, i.e., the frequency stability and the phase stability. The RF sinusoidal oscillators are almost exclusively based on the resonance effect.

In certain applications it is necessary to adjust the frequency of oscillation in a certain range, or to fine-tune the frequency to a pre-determined value. Since this feature is usually realized by a varactor in the resonance circuit of the oscillator, these circuits are called VCOs (voltage controlled oscillators). In some other applications, the frequency of oscillation must be fixed at a certain frequency with the maximum possible precision and stability. For these cases, quartz crystal oscillators are preferred. If the target frequency is higher than the range supported by the crystal oscillator, a higher frequency VCO can be “locked” to the frequency of the crystal oscillator.

As already explained in Chapter 4, an L-C circuit with external excitation will swing at a frequency determined by the values of its components. The magnitude of the swing decreases in time, owing to the losses of the system.

MOS transistors

The basic structure of an n-channel metal–oxide–semiconductor (NMOS) transistor built on a p-type substrate is shown in Fig. 1.1. The MOS transistor consists of two disjoint p-n junctions (source and drain), bridged by a MOS capacitor composed of the thin gate oxide and the poly-silicon gate electrode. If a positive voltage with sufficiently large magnitude is applied to the gate electrode, the resulting vertical electric field between the substrate and the gate attracts the negatively charged electrons to the surface. Once the electron concentration on the surface exceeds the majority hole concentration of the p-type substrate, the surface (the channel) is said to be inverted, i.e., a conducting channel is formed between the source and the drain. The carriers, i.e., the electrons in an NMOS transistor, enter the channel region underneath the gate through the source contact, leave the channel region through the drain contact, and their movement in the channel region is subject to the control of the gate voltage.

To ensure that both p-n junctions are continuously reverse biased, the substrate potential is kept lower than the source and drain terminal potentials. Note that the device structure is symmetrical with respect to the drain and source regions; the different roles of these two regions are defined only in conjunction with the applied terminal voltages and the direction of the drain current.

Understanding amplifier phase noise is vital to the comprehension of oscillators. There are two basic classes of noise, additive and parametric, as illustrated in Fig. 2.1. However, their behavior, and the underlying physical mechanisms, are strikingly different. This chapter focuses on amplifiers but includes a few words on the photodetector, which forms part of the resonator–amplifier loop in the emerging photonic (or optoelectronic) oscillators. The same principles apply to other electronic and optoelectronic components.

Random noise is normally present in electronic circuits. When such noise can be represented as a voltage or current source that can be added to the signal, it is called additive noise. Another mechanism can be present, referred to as parametric noise, in which a near-dc process modulates the carrier in amplitude, in phase, or in a combination thereof. The most relevant difference is that additive noise is always present, while parametric noise requires nonlinearity and the presence of a carrier. Additive noise is generally white, while parametric noise can have any shape around ν0, depending on the near-dc phenomenon. Additive noise originates in the region around the carrier frequency ν0 while parametric noise is brought there from near-dc by the carrier. This is made evident by a gedankenexperiment (thought experiment) in which we inspect the amplifier output around ν0 with an ideal spectrum analyzer, switching the input carrier on and off.