Application Note 37 Emulating RF Channel Characteristics Wireless communication is one of the most demanding applications for the telecommunications equipment designer. Typical signals at the receiver have delay spreads in the tens of microseconds, and experience multipath fading resulting in large signal power fluctuations. The only way to ensure that a system is able to function properly in an environment with these characteristics is through extensive testing. In the past, most testing has been through on-line methods, i.e., an actual call is placed between the mobile and the base station of a public cellular system. Using an actual RF channel for testing is prone to unreliable and ever-changing results, and can also be very costly. To achieve accurate and repeatable results, the RF channel must be brought into the lab through the use of channel emulation techniques. This application note is divided into two sections. The first section provides a review of the characteristics typically encountered in the RF channel. The second section provides an overview of the TAS 4500 FLEX RF Channel Emulator. The RF channel characteristics reviewed in the first section are: multipath fading, relative path delay, relative path loss, Doppler shift and log-normal shadowing. The TAS 4500 FLEX RF Channel Emulator provides accurate and flexible emulation of these characteristics. There are five characteristics typically encountered in the RF channel. These characteristics include: multipath fading, relative path delay, relative path loss, Doppler shift and log-normal shadowing. One or any combination of these may be found on a typical RF communication channel. The cause of each of these characteristics will be discussed along with their effect on the transmitted signal. All of these characteristics can be demonstrated with a simple transmitter to receiver diagram. Figure 1 is a diagram of a typical mobile receiver (the car) as it drives along a roadway. Rays A, B, C, and D depict just four of the many signal paths from the transmitter to receiver. A B C Figure 1. Transmitter to Receiver Signal Path Diagram D 2001, Spirent Communications of Eatontown, LP (DBA TAS) 1-800-927-2660 www.spirentcom.com
Multipath Fading (Rayleigh Fading) Multipath fading or Rayleigh fading results in a rapid fluctuation of the signal power. Figure 2 shows an example of a Rayleigh faded signal. A Rayleigh faded signal is caused by the summation of a very large number of individual reflected signals. Each of these signals has a random phase and amplitude at the receiver due to differences in path length and path attenuation. Rayleigh fading is also called fast fading since the fluctuations in the signal power occur very rapidly. The fading of the signal envelope can be easily pictured using a simple two path example. At the receiver the signals from each path can be of any amplitude and phase. If the two signals are of the same amplitude, and their phase will be 180 o apart, there will be total destructive interference with no resultant signal. If the two signals are 0 o apart in phase there will be constructive interference and the signal envelope will be 3 db larger than the individual signals amplitudes. The signals rarely combine to greater than 10 db above an individual path s power. The deep fades (destructive interference) range from just a few db to fades of greater than 50 db. The spacing of the fades is a function of the RF carrier frequency. At 900 MHz the fades will occur every few centimeters. The receiver front end must be sensitive enough to function properly over these power fluctuations. Figure 2. Power vs. Time for a Rayleigh Faded Signal Doppler Freq. = 100 Hz Center Freq. - 900 MHz Span = 0 Hz RBW = 100 khz Sweep Time - 75 ms Relative Path Delay Relative path delay is a phenomenon where individual signal paths from the transmitter arrive at different times at the receiver. An example of this is shown in Figure 1 between Paths (A) and (C). Path (C) will arrive at the receiver a finite time after signal Path (A) because the two signals reached the same destination by traveling courses of different lengths. The net effect of the arrival time difference is to spread the signal in time. In a digital transmission system this will cause the received symbols to overlap resulting in intersymbol interference. The amount of relative path delay varies with the terrain and application. In an indoor application, delays could be in the 10s of nanoseconds (ns); 10 ns corresponds to about 10 feet. In outdoor applications, delays of 10 microseconds (ms) or less are typical; 10 ms corresponds to about 10,000 feet. Delays greater than 50 ms are rare in cellular environments. 2
Relative Path Loss Relative path loss is a phenomenon which occurs when individual signal reflections arriving at the receiver are at different absolute power levels. The difference in power levels between signal paths is caused by physical elements in the signal path. In Figure 1, Path (C) will arrive at the receiver at a lower power level then Path (A). This occurs because a fraction of the original power in signal Path (C) is lost when it is reflected off the building. Signal strength will also vary due to the distance the signal travels. Loss in signal strength should follow the 1/d 2 law, where d is the distance between the transmitter and the receiver. In the actual cellular environment the loss is considerably worse between 1/d 3 to 1/d 6, primarily due to variations in the terrain. Doppler Shift Doppler or frequency shift from the carrier frequency occurs when the distance between the receiver and the transmitter is changing. An example of this is when a mobile receiver (car) is traveling away from the transmitter. Path (A) in Figure 1 will have a Doppler shift due to the movement of the car. The amount of frequency shift (Doppler frequency) from the carrier is determined by the following formula: Freq DOPPLER = Velocity x Freq CARRIER C where C = 3 x 10 8 m/s The Doppler frequency can be either positive or negative depending on whether the mobile receiver is moving away from (negative) or towards (positive) the transmitter. Using this equation, a car traveling towards the transmitter at 100 km/hr with a 900 MHz carrier frequency would experience a 83.4 Hz Doppler shift on the signal. The receiver must function properly even though the signal has been shifted from the original RF carrier frequency. Log-Normal Shadowing Log-normal shadowing is the slow variation of the nominal signal power over time. A plot of signal power versus time for log-normal shadowing is shown in Figure 3. Note that the time scale is much larger than that of the Rayleigh fading diagram shown in Figure 2. Log-normal shadowing is a loss in the signal strength at the receiver due to a blockage or absorption of the signal from the transmitter due to elements in the environment. The blockage of the signal could be caused by elements such as buildings or hills. This is often called slow fading since the receiver is passing through the signal shadow of a large object. The fading is represented by a log-normal distribution of the mean signal power. The standard deviation of the lognormal distribution varies depending on the environment in which the receiver is located. A standard deviation of 6-8 db is typical for urban areas, while a deviation of 10-12 db can be observed in rural locations. 3 Figure 3. Power vs. Time for a Log-Normal Faded Signal Log-Normal Standard Deviation - 10 db Log-Normal Rate - 10 Hz Path Loss = 25 db Center Freq. = 900 MHz Span = 0 Hz RBW = 100 khz Sweep Time = 2 sec
RF Channel Emulation Numerous models have been proposed, attempting to simulate the RF communication channel. Two few of these are the Simplified Three-Path Model (Rummler) and the Three-Ray Dispersive Model (Goldman). An RF channel emulator must be able to adapt to these and many other channel models. The channel format of one such emulator, the TAS 4500 FLEX, is shown in this section. An example of how the TAS 4500 FLEX can be used to model a typical RF communication signal will also be discussed. The TAS 4500 FLEX presents a flexible model that can be used to emulate an RF channel. In Figure 4, the channel block diagram of the TAS 4500 FLEX is shown. The TAS 4500 FLEX contains two individual channels each with six impairment paths. Each of the impairment paths is capable of emulating all of the characteristics described earlier in this document. It should be noted that the two channels may be externally combined to form a 12-path model. Since all of the characteristics are available on each path, practically any RF channel model may be realized. A typical RF signal seen at the receiver is one with a Rician distribution. A Rician signal is simply a combination of a Rayleigh faded signal and a Line-Of-Site (LOS) signal. The Rayleigh signal is caused by the multipath effect of the mobile environment, and will have some maximum Doppler frequency (F dray ). The LOS signal is caused by a direct path between the transmitter and the receiver. The LOS signal will have a Doppler frequency shift (F dlos ) that is less than or equal to the F dray, due to the arrival angle of the LOS signal. In GSM (Global System for Mobile Communication), F dlos is specified to be 0.707 F dray, which corresponds to an arrival angle of 45 o. A Rician distributed signal produced using the TAS 4500 FLEX is shown in Figure 5. The Rician distribution in the TAS 4500 FLEX can be created two ways. One way is by utilizing one channel with one active impairment path which is set to the Rician distribution specified by the GSM standards. It is also possible to set up custom Rician distributions with in one impairment path. The TAS 4500 FLEX allow programmable K- factor and LOS arrival angle. The K-factor is the relation between the power levels of the line-of-sight component and the Rayleigh faded component in Rician fading. Another technique for approximating the signal propagation, such as Rician Fading, is with Nakagami fading. Nakagami fading Figure 4. TAS 4500 FLEX Block Diagram describes the time domain characteristics of the envelope power of a faded signal. There are two controllable parameters associated with this fading type, the direct path angle of arrival and the M value. The Probability Density Function of the envelope of the Nakagami faded signal is controlled by the M value. The Nakagami M value provides a means of differentiating the typical fading environments by providing a relative measure of how direct the received paths are. 4
RF Channel Emulation (continued) The programmable M value describes the fading condition as the ratio of the direct signal component to multi-path faded signal components. The discrete range of M values supported in the TAS 4500 offer a progression from a Rayleigh faded distribution (M=1) to a single direct path (M=100) which approaches a pure frequency shifted component. The values between these two limits (M=3, 5, 10, 15, and 25) describe an environment where both a direct path and a multi-path are present, where the M value defines a relative power ratio of the direct path to faded path components. One reference which describes various fading environments in terms of multiple Nakagami paths is the UMTS Code Division Testbed (the propagation model to be used for CODIT). The programmable angle of arrival will then change the position of the direct path relative to the multi-path component. This is accomplished by scaling the static Doppler shift of the direct path appropriately. This Doppler shift is set according to the following equation: Doppler direct component = Doppler faded component x cosine(angle of arrival) Figure 5. Measured Rician Fading Power Spectral Density Doppler Freq. = 100 Hz Center Freq. = 900 MHz Span = 1 khz RBW = 10 Hz Sweep Time = 30 sec Conclusion The rapid expansion of wireless communication has created a need for accurate and repeatable emulation of the RF channel. Multipath (Rayleigh) fading, relative path delay, relative path loss, Doppler shift and log-normal shadowing are the primary RF channel characteristics. The TAS 4500 FLEX accurately emulates these characteristics in a flexible test instrument. This instrument is also capable of emulating combinations of the RF channel characteristics such as Suzuki fading, which is a superposition of Rayleigh fading and log-normal shadowing. This versatility combined with the complete RF channel emulation makes the TAS 4500 FLEX an indispensable tool for testing RF communications equipment. Spirent Communications of Eatontown, LP (DBA TAS) 541 Industrial Way West, Eatontown, NJ 07724, U.S.A. Phone: (732) 544-8700, Fax: (732) 544-8347, www.spirentcom.com 5 Spirent Communications is a trademark and service mark of Spirent plc. All rights reserved Specifications are subject to change without notice, Printed In U.S.A., 2/01 v.2