With the growing popularity of mobile communications, and the concurrent increase in existing analog cellular systems, the frequency spectrum is becoming very crowded. Several solutions have been developed to address this situation, but one–code division multiple access (CDMA)–is distinguishing itself with the advantages of high capacity and lower power.
CDMA operational definitions are outlined in the IS-95 standard adopted by the Electronic Industries Association (EIA) and the Telecommunication Industry Association (TIA).1 Specifically, IS-95 promises to increase capacity by 20 times that of the current analog system, to eliminate the need for frequency assignments to cells, and to reduce the required RF transmit power.2,3 Several cellular operators have already committed to CDMA and others are awaiting the outcome of commercial trials before making the same commitment.
As base-station and mobile-station manufacturers work toward developing and delivering the best CDMA solutions, the need to properly test their equipment rapidly increases. Proper testing, in this case, means that CDMA products must be tested as an integral part of the development, production and quality-control cycles.
The recommended minimum performance standards for the base stations and mobile stations are specified in TIA/EIA IS-97 and IS-98, respectively. These standards ensure that the base stations and mobile stations perform satisfactorily under several impairments imposed on the system by the wireless channel environment, such as noise, multipath fading and interference.
Additive White Gaussian Noise
The most prevalent problem in any communication system is noise. Noise with power evenly distributed over all frequencies and amplitude distributed in Gaussian shape is called white Gaussian noise.
The central limit theorem states that if the random variables are independent, then the density of their sum x = x1 + … + xn tends to a normal (Gaussian) curve as n approaches infinity.4 This type of noise best represents the actual noise present in the communications system.
This known noise is added to the communication system to test its bit error rate (BER) performance. BER, as a function of bit energy per noise density (Eb/No), is a quantitative quality factor when evaluating digital communication links.
To provide accurate BER measurements, a parameter called crest factor is very important. Crest factor is the ratio of peak noise voltage to the rms noise voltage. Gaussian noise with an infinite crest factor will always produce a small but finite BER.
But if the crest factor is low, then this will limit the peak noise voltages and will not be able to introduce bit errors. This is illustrated by the voltage distribution of an additive white Gaussian noise (AWGN) source (Figure 1). If the crest factor is low, then the voltage excursions beyond ±x will not exist. This is particularly evident when Eb/No is large and if a very low BER measurement is being made. For example, a noise source with a crest factor of approximately 18 dB is adequate for measuring BER as low as 10-10 in BPSK- or QPSK-modulated systems.
CDMA measures the frame error rate (FER) as a function of Eb/No to verify the performance of its base-station and mobile-station receivers. The FER, rather than BER, is the fundamental performance measure for CDMA.
Eb/No must be set accurately, with excellent repeatability and stability. In fact, to comply with the IS-97 and IS-98 testing requirements, a minimum of 0.2-dB accuracy is required.
A typical setup for CDMA mobile station testing is shown in Figure 2.
Multipath Fading Effects
Due to reflection, refraction and scattering of radio waves by natural or man-made structures, such as hills, ground, buildings or walls, the transmitted signal often reaches the receiver by more than one path, resulting in an effect called multipath fading. The direct- and indirect-path components combine at the receiver to produce a distorted version of the transmitted signal.
The direct-path component, if it exists, exhibits a frequency shift without frequency spreading, also known as a pure Doppler effect. The indirect-path components exhibit random relative phases and this randomness is described by several different statistical distributions depending on the characteristics of the multipath medium. Well-accepted distribution models, such as Rayleigh, Rician and log-normal, along with some more general formulas such as those developed by Nakagami and Suzuki, describe the user’s particular multipath environment. For more detailed description of these different fading statistics, refer to Reference 5.
CDMA uses various diversity techniques to alleviate the multipath fading effect.6 For space diversity, the base stations use two receive antennas and, during soft hand-offs, multiple base stations simultaneously talk to the mobile.
For frequency diversity, which is inherent in spread-spectrum systems, CDMA uses a 1.25-MHz bandwidth. Then only the multipath components that arrive within 1 ms will cause the signal to experience a deep fade. Since most components will be delayed much longer than 1 ms in many environments, only a narrow portion of the signal will be lost.
For time diversity, CDMA uses convolutional encoding in the transmitter for interleaving, and Viterbi decoding with soft decision points in the receiver for forward error correction. CDMA also uses multiple receivers and pairs them with the strongest signals. The base-station receiver uses four receiving elements and the mobile uses three. This multiple correlator is called a RAKE receiver.
In IS-97 and IS-98, several receiver tests involve demodulation of traffic channels in a multipath fading environment and during a soft hand-off. Diversity combining of identical power-control subchannels and selection of the proper power-control bits are also tested. To properly test the receiver and transmitter under multipath conditions, the channel simulator must provide at least:
Three independent Rayleigh faded paths.
Vehicle speed up to 100 km/h.
Path-to-path attenuation and delay.
Two independent input channels for diversity and soft hand-off testing.
One particular note has to be made regarding multipath fading emulators. Appendix A of IS-98 contains figures for determining the confidence level of the error rate.7 The figures provide limits on the allowable measured rate for a 0.95 confidence level for FERs of 0.03, 0.01 and 0.005.
These figures show that if the true error rate is close to the specified error rate, the test time can become increasingly long (Figure 3). That is, if the equipment-under-test just barely meets the specifications, the test time can be very long.
The figures show up to 100,000 frames to be measured. At 20 ms per frame, this equates to about 2,000 s (or about 33 minutes) of test time. During this time, the multipath fading emulator should maintain its randomness in generating fading statistics. If the random pattern is repeated within this time period, measured results are ambiguous at best.
CW Interference
To measure the performance of a base-station or mobile-station receiver in the presence of continuous wave (CW) interference, single-tone desensitization and intermodulation spurious response attenuation tests are conducted. The single-tone desensitization uses a CW generator to apply a tone that is offset by +750 kHz (or +900 kHz) from the center frequency of the assigned channel. By measuring the FER of the base-station receiver, its capability to receive a CDMA signal on the assigned frequency is tested.
The intermodulation spurious-response attenuation test is similar to the single-tone testing, except it uses two interfering tones that are separated from the assigned frequency and from each other to produce third-order harmonics in the band of the desired CDMA signal. With each CW signal at +900 kHz and at +1,700 Hz, or -900 kHz and -1,700 kHz, the FER of the base-station receiver is measured.
Power Control
CDMA controls the mobile-station power with open-loop and closed-loop methods. Open-loop power control uses the received power at the mobile station as a reference and sets the sum of its receive and transmit powers to be constant at -73 dBm.
Closed-loop power control overrides the open-loop by sending power-control bits from the base station every 1.25 ms to force the mobile receiver to increase or decrease its transmit power by 1 dB. These power-control techniques allow the mobile station to transmit minimal power to maintain a communication link, saving battery life, lowering amplifier output power and reducing health hazards from radio waves.
Although the IS-97 and IS-98 standards do not require testing power-control functions under an AWGN environment, it seems necessary to do so for CDMA base-station and mobile-station developers. And since the power-control bits are sent 800 times per second with a feedback loop, an AWGN generator should constantly adjust the new noise power levels to maintain a desired Eb/No.
References
1. TIA/EIA/IS-95, Mobile Station-Base Station Compatibility Standard for Dual-Mode Wideband Spread Spectrum Cellular System, Global Engineering Documents, July 1993.
2. Lee, W.C.Y., Mobile Communications Design Fundamentals, Second Edition, John Wiley & Sons, Inc., 1993, p. 370.
3. Padgett, J.E., Gunter, C.G., and Hattori, T., “Overview of Wireless Personal Communications,” IEEE Communications Magazine, Vol.33, No.1, January 1995, pp.28-41.
4. Papoulis, A., Probability, Random Variables and Stochastic Processes, Second Edition, McGraw-Hill, 1984, p.194.
5. Hashemi, H., “The Indoor Radio Propagation Channel,” Proceedings of the IEEE, Vol. 81, No. 7, July 1993, pp. 943-968.
6. Whipple, D.P., “The CDMA Standard,” Applied Microwave & Wireless, 1995, pp. 58-66.
7. TIA/EIA/IS-98, Recommended Minimum Performance Standards for Mobile Stations Supporting Dual-Mode Wideband Spread Spectrum Cellular Mobile Stations, Global Engineering Documents, December 1994.
About the Author
Alex Kim is the New Business Development Manager at Noise Com. He holds a B.S.E.E. degree from MIT and an M.S.E.E. degree from the University of Southern California. Noise Com, Inc., E. 49 Midland Ave., Paramus, NJ 07652, (201) 261-8797.
Sidebar
Glossary of Terms
Eb/No
–Bit energy-to-noise density ratio. This relates the power attributable to single bit and the associated noise power in a 1-Hz bandwidth. The bit energy is found by dividing the total RF carrier power by the data rate. That is, Eb/No = C .
BW
N fb
Where: C = carrier power
N = noise power
BW = noise bandwidth
fb = bit rate
Soft Hand-off
–At cell boundaries where a mobile station undergoes a hand-off to the next cell, the mobile station maintains the link with multiple base stations until the most suitable base station is determined for the hand-off. This make-before-break concept is called a soft hand-off.
Copyright 1995 Nelson Publishing Inc.
October 1995
|
|