UWB antenna measurement methods

The methods for measuring antenna transfer functions with a network analyzer presented in the following use typically available measurement equipment for far-field antenna measurements with enhanced calibration procedures. The latter are necessary in order to obtain stable and consistent phase information for the transfer function in co- and cross-polarization. These methods complement the standardized antenna measurements known from [68]. A validation of the methods is presented with the prediction of a time domain impulse transmission, which is measured independently with a fast-pulse generator and an oscilloscope.

A prerequisite for the measurements is a stable measurement setup inside an anechoic chamber, which provides low reflections from the walls within the frequency range of operation. If multiple reflections cannot be avoided, the possibility of time gating for valid signals of the additional delay – due to the path lengths of wall or floor reflected signals – needs careful consideration. If the expected τr or even τFWHM overlap with the multiple reflections, the impulse response hAUT of the antenna under test (AUT) can no longer be measured accurately. Furthermore, far-field conditions are necessary in order to obtain a distance-independent impulse response estimation hAUT. The frequency domain far-field criterion r > (2D2)/λ (according to [24]) then applies for the highest frequency, i.e. the smallest wavelength.

UWB is an umbrella term that mainly indicates that a very large absolute bandwidth (Ba > 500 MHz) or a very large relative bandwidth (Br = 2[fu – fl]/[fu + fl] > 0.2) in the RF spectrum is used instantaneously by the system. With this definition, no special purpose or application and no special modulation is defined but it implies that the components of the system must be capable of handling this wide spectrum. As already mentioned in the previous chapter for RF frontends, on the whole it is the relative bandwidth that poses new challenges, so system aspects for a very large relative bandwidth are mainly discussed here. This chapter provides a mathematical description of the UWB radio channel including the antennas and measures to characterize the UWB performance of the analog frontend, including the radio channel in the frequency domain (FD) and in the time domain (TD). The chapter presents two methods to exploit an ultra-wide bandwidth: the transmission of short pulses in the baseband (impulse radio transmission), and the transmission by a multi-carrier technique called orthogonal frequency division multiplexing (OFDM). For impulse radio, the most common pulse shapes are introduced together with methods to generate them. Finally, modulation and coding techniques are considered as well as basic transmitter and receiver architectures. The coordinate system is given in Fig. 2.1.

For many scientists and engineers working in ultra-wideband technology, it seems that the idea of using signals with such a wide instantaneous bandwidth was spread by the US FCC with the accreditation of the frequency band from 3.1 to 10.6 GHz. But, if we look back in history, we find that even the first man-made electromagnetic waves were generated by sparks. Especially famous for electromagnetic research was Heinrich Hertz who, in the 1880s, verified the speed of propagation of electromagnetic waves, their polarization and interaction with objects, and the correct description of these waves by Maxwell's equations at our university in Karlsruhe, Germany. Before this time, electromagnetic waves could only be generated by the aforementioned sparks and were thus ultra-wideband.

Ultra-wideband was banned in the 1920s because it occupied too great a portion of the spectrum and from this point was primarily limited to military applications. This was until 1992 when Leopold Felsen, Lawrence Carin, and Henry Bertoni organized a conference on ultra-wideband, short-pulse electromagnetics in Brooklyn. Our institution, the Institut für Höchstfrequenztechnik und Elektronik (now the Institut für Hochfrequenztechnik und Elektronik) had the privilege of participating in this first conference on ultra-wideband. The topics at the conference were so fascinating that we decided to step into this area. The first research topics were in ground penetration radar, with the idea of detecting anti-personnel mines.

In this chapter typical UWB applications are presented and discussed, particularly regarding their RF system design.

UWB communication

Modeling of the system components

The system model for impulse radio transmission consists of the transmitter, the UWB indoor channel and the receiver, see Fig. 6.1. At the transmitter side, the pulse generator creates pulses at moments defined by the pulse position modulation (PPM) scheme and the time-hopping (TH) code. Without modulation and coding, pulses appear at multiples of the pulse repetition time T. To modulate the signal by PPM, the binary values of the bit stream determine whether a constant delay TPPM is introduced or not with respect to the pulse repetition time.

The TH code adds a further pseudo-random delay to smooth the spectrum. This code must also be known at the receiver to demodulate the signal. To ensure that the radiated signal does not violate the spectral mask, an analog transmit bandpass filter is placed before the transmit antenna. The radiated signal propagates via an indoor radio channel including AWGN towards the receive antenna. It is then amplified, filtered and demodulated; different demodulation schemes are available. The following subsections describe the modeling of the non-ideal system components.

Pulse shape generator

Classical considerations emphasize the idea that impulse radio is a low-cost technique because of its simplified hardware architecture. For example, there is no necessity to use up-converters, since signals are radiated in the baseband. To remain low cost, the associated pulse shapes result from low complexity devices.

A well-known method of increasing the gain and lower the half-power beamwidth of the radiation pattern is to replace the single radiator by an antenna array. Its application in communications is interesting if, for example, a point-to-point connection is to be established. It is of special interest in MIMO systems, where a channel capacity might be increased if multiple radiators, either on the transmit or the receive side (or both) are used [75]. In radar systems an application of arrays is more common in order to achieve lower half-power beamwidths, which in general are used to increase the angular resolution.

In this chapter, specific design issues of UWB antenna arrays are described. According to the methods previously described, the frequency and time domain models are explained and used for the practical array design. In the second part of the chapter an extension of the monopulse technique to UWB systems is described, based on the array theory.

Array factor in UWB systems

The resulting array radiation pattern depends on the following parameters:

1. • number of array elements N

2. • distance between the elements d

3. • frequency f

4. • excitation coefficients – amplitude and phase

5. • radiation pattern of a single array element EF(f, ψ).

The great interest in UWB systems is well-motivated by the huge signal bandwidth. As shown in the previous chapters, the ultra-wide signal bandwidth has many benefits, such as very high data transmission rates and very fine time resolution. Either or both of these potentials would be very attractive for short-range communication systems, if a substantial advantage in terms of power consumption or performance could be achieved compared to existing radio solutions. In particular, the possible time resolution well below 1 ns is a unique feature, which seems to be the enabler for precise indoor localization. Unfortunately, the huge signal bandwidth demands an appropriate circuit design, which in many cases differs from traditional narrowband RF circuit design. The same requirement holds for the sub-circuits in the digital domain in order to obtain the desired timing accuracy. Having indoor localization applications in mind, this chapter shows why the impulse radio UWB circuits are promising in terms of localization accuracy in time-of-arrival (TOA)- and time-of-flight (TOF)-based localization systems. Different pulse generation and pulse detection principles will be introduced and explained. Important design considerations will be discussed on the way to a fully monolithic integrated circuit. Finally, there are a few examples for illustration.

Pulse radio transceiver requirements

Indoor channel requirements

Why did recent short-range radio systems like Bluetooth, WLAN, ZigBee, not accomplish a breakthrough in indoor localization? Summarizing the indoor channel properties from Chapter 2, the answer lies in their limited signal bandwidth which does not allow them to resolve the multipath propagation dominating in indoor environments. Figure 5.1 illustrates again a typical situation in a room.

Preface

The need for a rigorous, yet practical, framework for characterization, modeling, and design of nonlinear electronic components at high frequencies has never been more urgent. The communications revolution is inexorably forcing active devices into more and more strongly nonlinear regimes of operation. This is a consequence of the relentless drive for more efficiency in order to save power, extend battery life, and minimize cooling. The price for efficiency is nonlinearity. Dealing with nonlinearity means that new measurement instrumentation and new modeling and design methodologies are required that go far beyond linear S-parameters. Fortunately, there is an overarching, interoperable paradigm combining all these pieces of the nonlinear puzzle together, seamlessly. The new paradigm is called X-parameters, and that is what this book is about.

The book is intended as a comprehensive introduction to X-parameters. It is aimed at a diverse audience with a wide range of backgrounds. This is quite a challenging undertaking! We are targeting professional microwave engineers, device modeling engineers and scientists, RF and microwave circuit designers, electronic and communications engineers, CAE professionals developing simulator algorithms, and microwave and RF professionals developing new high-speed instrumentation for a wide range of nonlinear characterization applications. The inherent interdisciplinary nature of X-parameters is the prime reason we seek to appeal to this broad audience. The practical solutions based on X-parameters deployed by industry over the past several years depend on contributions in all of these areas.

One of the key features that led to the wide adoption of S-parameters was the availability of hardware and calibration techniques capable of making quick, accurate, and repeatable S-parameter measurements. S-parameters can also be easily extracted in simulation from device or circuit models. In either case, the resulting S-parameters can immediately be used in simulation or design tools. In order to achieve similar success in the nonlinear domain, X-parameters must be easily measured and also easily extracted from simulation.

Measurement hardware

X-parameters, like S-parameters, represent the steady-state behavior of a device in the frequency domain. The linearity assumption of S-parameters, however, greatly simplifies the measurement requirements. The additional capabilities of X-parameters come at the cost of additional complexity in the measurement system as well as the modeling paradigm.

Hardware requirements

X-parameters include cross-frequency terms that capture the distortion products generatedby device nonlinearities. In order to measure these terms, the measurementhardware must be capable of measuring coherent cross-frequency phase. BecauseS-parameters include no cross-frequency interaction (a consequence of the linearityassumption), each frequency may be measured independently with no need for aconsistent time base or cross-frequency phase. Since all S-parameters are ratios ofwaves, there is also no need for accurate measurement of absolute power – only therelative power is needed. As a result, the hardware and calibration techniques developedfor S-parameter measurement generally are not sufficient for X-parameter measurementand must be extended to include cross-frequency phase and calibrated absolute power.

Introduction

This chapter presents a concise treatment of S-parameters, meant primarily as an introduction to the more general formalism of X-parameters. The concepts of time invariance and spectral maps are introduced at this stage to enable an easier generalization to X-parameters in the ensuing chapters. The interpretations of S-parameters as calibrated measurements, intrinsic properties of the device under test (DUT), IP-secure component behavioral models, and composition rules for linear system design are presented. The cascade of two linear S-parameter components is considered as an example to be generalized to the nonlinear case later. The calculation of S-parameters for a transistor from a simple nonlinear device model is used as an example to introduce the concepts of (static) operating point and small-signal conditions, both of which must be generalized for the treatment of X-parameters.

S-parameters

Since the 1950s, S-parameters, or scattering parameters, have been among the most important of all the foundations of microwave theory and techniques.

S-parameters are easy to measure at high frequencies with a vector network analyzer(VNA). Well-calibrated S-parameter measurements represent intrinsic properties of theDUT, independent of the VNA system used to characterize it. Calibration procedures [1]remove systematic measurement errors and enable a separation of the overall values intonumbers attributable to the device, independent of the measurement system used tocharacterize it. These DUT properties (gain, loss, reflection coefficient, etc.) are familiar,intuitive, and important [2]. Another key property of S-parameters is that theS-parameters of a composite system are completely determined from knowledge ofthe S-parameters of the constituent components and their connectivity. S-parametersprovide the complete specification of how a linear component responds to an arbitrarysignal. Therefore designs of linear systems with S-parameters are predictable withabsolute certainty. S-parameters define a complete behavioral description of the linearcomponent at the external terminals, independent of the detailed physics or specifics ofthe realization of the component. S-parameters can be shared between componentvendors and system integrators freely, without the possibility that the component implementation can be reverse engineered, protecting IP and promoting sharing andreuse. Indeed, one may ask the question, “are S-parameters measurements, or do theyconstitute a model?” The answer is really “both.”