Why Special Signals Are Needed for Digital Communications Testing
The demand for mobile or personal telecom equipment, as well as wireless data communications systems, is growing worldwide. To accommodate additional customers and new services, spectrum conservation is essential.
To achieve better frequency spectrum utilization, new speech encoding, signal processing and digital modulation techniques have been devised. Some of these are already in use; others are in prototype stages. Unfortunately, existing implementations differ widely in frequency allocation, allowed bandwidth and protocols, and use diverse types of modulation (Table 1).1
To test these diverse systems, new signal generators must furnish–in addition to all the normally expected functions–special signal formatting and digital modulation facilities matching UUT requirements. To better understand the reasons for their unique features and how they are implemented, it is important to know more about some of the underlying technical principles.
About the Technology
Digital Modulation
Ideally, digital signals transition from one state to another within zero time. But this is neither achievable nor desirable when bandwidth is limited. So in digital communication systems, signals are sent through shaping filters before and after modulation. By using this process, carrier frequency deviation limits are not exceeded and interference between successively transmitted symbols is minimized.
Since any signal conveyed by a carrier may be represented by consecutive amplitude and angular position values (frequency or phase), it may be generated by conventional analog amplitude, phase or frequency modulation, or a combination of these. But an I/Q modulator is used in most digital communications systems for more versatile implementations.
AM and FM on an I/Q Diagram
AM may be represented by a stationary vector whose length varies in correspondence with the amplitude of the modulation waveform. Since the phase of a pure AM signal ideally does not vary, the angular position of this vector remains constant (Figure 2a). Conversely, an ideal FM signal will always be of constant amplitude; but since f=df /dt, its phase (angular position) will be continually changing (Figure 2b).
Any point on the I/Q plane in Figure 2 could be reached by applying a combination of conventional amplitude and angular modulation. Alternatively, an I signal and a Q signal could be applied to reach any point on the plane.
I/Q Modulators
Two separate mixers are required to generate the I in-phase and the Q quadrature modulated signals. One of these mixers is presented with one coded and filtered data stream and an in-phase (I) carrier. The other receives the second processed data stream and mixes it with a 90° phase-shifted (Q) carrier (Figure 3).
The two resultant signals are summed, filtered and up-converted to become the transmitted modulated RF signal. At the receiver, the carrier is recovered, applied to one of two demodulators, and a 90° phase-shifted signal (derived from the same recovered carrier) is applied to the other demodulator.
Constellation Diagrams
FSK is one of the oldest forms of digital modulation. In FSK systems, the digital signal is filtered, then used to frequency modulate the carrier. Most FSK systems use Gaussian filters to restrict the bandwidth. Examples of FSK systems include POCSAG and ERMES pagers, DECT and CT2 portable phones.2
Another simple modulation scheme is bi-phase-shift keying (BPSK or 2-PSK). Here the carrier magnitude is constant and, to transmit either a 0 or a 1, the phase is keyed or switched between 0° and 180°. In this simple scheme, only one bit of information is associated with each state so the carrier phase is keyed at the bit rate.
Quadrature phase shift keying (QPSK or 4-PSK) uses four constant phase values: 45°, 135°, 225° and 315°. In this case, 2 bits of binary data, defining any of four symbols, are encoded at each of the states. As serial data is taken 2 bits at a time to form the symbol, the symbol rate is half the bit rate. As a result, QPSK requires only half the bandwidth of BPSK for the same symbol rate.
The quest for greater bandwidth efficiency has led to even more complex modulation schemes. 16 QAM encodes 4 bits of serial data as a single state, reducing the symbol rate to one-quarter of the bit rate. To generate this type of modulation, the I and Q carriers must take on four different levels of amplitude (+3, +1, -1, -3).3
QAM systems are not the only ones employing multiple amplitude levels. While some PSK systems traverse from one state to the next along a circular orbit, using constant amplitudes, others do not. In a BPSK system, for instance, traversing from 0° to 180° may occur along the I axis. In this case, amplitude is decreasing, reaching zero at the I/Q-plane origin, then becoming negative.
In most QPSK systems, such as in the p /4 DQPSK implementations used in the NADC and PDC systems, state-to-state traversals take place via the most direct (straight line) path. Consequently, these signals contain FM and AM components. To simplify demodulation, zero-amplitude crossings are avoided by applying consecutive differential phase offsets or coding. A state-traversal diagram for p /4 DQPSK modulation is shown in Figure 1a.
Figure 4
shows the state diagrams for several common digital modulation schemes. Since the state clusters resemble groups of stars, they are commonly referred to as constellation diagrams.Eye Patterns and Constellations
Digital signal transitions take place at predetermined, fixed time intervals. In a noiseless, infinite-bandwidth environment, there would be ample space between transitions to sample the received signal and determine its binary value.
In the real world, however, signals have passed through bandwidth-limiting filters and do not have square transitions (Figures 1b and 1c). Noise and intersymbol interference usually cause more distortion.
When looking at one symbol period with an oscilloscope, consecutive signal transitions will trace overlapping patterns resembling one or several eyes. The more noise and distortion the signal encounters, the smaller will be the eye opening, introducing uncertainties or errors in the digital-signal recovery process.
The signal quality may not only be impaired by external influences, but also by internal causes such as I/Q modulator misalignment, amplifier distortion or internal interference. While an examination of a closing eye pattern shows the degree of signal degradation, it is usually easier to determine the underlying causes by analyzing the constellation diagram. The 45° oval shape of the clusters indicates crosstalk between the I and Q channels.
Burst and Spread-Spectrum Transmission
TDMA is the most common wireless digital transmission method today. The modulated signal is assigned to a given time slot and emitted sequentially as a series of transmission bursts.
CDMA, an evolving technology, allows groups of modulated signals to be transmitted simultaneously. Each signal is given a code and all signals are transmitted concurrently over a wide frequency spectrum.
Signal generators can simulate TDMA or CDMA signals by using amplitude, pulse or frequency modulation in addition to digital modulation capabilities, or by employing system-specific facilities.
Specific Features
Signal generators must provide a plethora of features to provide the test facilities demanded by the new wireless communications systems. Not only do these include digital modulation and filtering, but also mandated data rates, framing, frequency hopping, bursting and, perhaps, coding.
They must be capable of emulating the specific requirements of each applicable standard. “For example, for testing GSM, a signal generator must provide GMSK modulation with a data rate of 270.833 kb/s, frequency hop and emulate the bursting nature of a TDMA mobile unit,” said Jim Hebert, Product Marketing Manager at the Tektronix RF/Wireless Group. “The signal generator should also be easily configurable for inserting the appropriate data into the modulator to achieve desired effects.”
Signals must be free of spurious emissions and exhibit a low noise floor, just as expected for analog applications. But generators must also provide facilities to alter signal characteristics and introduce controllable impairments needed to evaluate performance limitations of the UUT. In addition, some tests may demand auxiliary signal sources to produce controllable interference.
“There are three classes of signal generators for this market, each with advantages and disadvantages,” commented Mr. Hebert. “The first class is the vanilla generator used as an interferer in much the same way that it is used with analog systems.
“The second contains a general-purpose I/Q modulator which can be used with an arbitrary waveform generator to provide virtually any modulation signal. The Rohde & Schwarz SMHU58 + ADS is an example. The third class of generators supports specific standards, or many specific standards, such as the Rohde & Schwarz SME family,” Mr. Hebert concluded.
The SME generator does not contain an I/Q modulator. Nevertheless, it accommodates nine distinct digital modulation schemes and provides signal formats matching signal characteristics of 34 wireless communications systems using different modulation schemes, frequencies, data rates or information structure.
Most multipurpose digital signal generators do, however, employ I/Q modulators. The Marconi 2050, which emulates NADC, PDC, TETRA and APCO 25 signals, has a 10-MHz I/Q modulation bandwidth to accommodate spread spectrum signal requirements, and provides controlled I and Q offsets and I/Q imbalances.
An implementation approach that provides multifunctionality with an almost limitless expansion potential employs a base unit and application-targeted expansion modules. The Anritsu MG3670B/3671A is available with four expansion units to cover p /4 DQPSK, GMSK and GFSK modulation formats plus a burst function unit to create TDMA frames.
A different approach has been taken by Hewlett-Packard with the HP 8657 Economical Synthesized Signal Generator Series. Distinct models or factory-installed options are tailored to fulfill the needs of specific communications systems.
For example, the D model of the HP 8657 Series provides p /4 DQPSK signals as required by the NADC and JDC systems, while the J model features p /4 DQPSK signals demanded by PHP systems. Similarly, the A model with options 022 or H46 provides 0.3 GMSK modulation as needed for the GSM Pan-European Digital Cellular Radio System or 0.5 GMSK modulation as needed for the CDPD systems.
“For R&D, multifunctional signal generators are usually the best choice,” said Ray Beers, Product Manager at Anritsu Wiltron. “The flexibility and multifunctional capabilities enable the development engineer to define and implement new designs.
“However, the manufacturing or service technician shouldn’t be concerned about modulation details, guard bands or other design concerns. What the manufacturing technician wants to know is: Can I ship this device?” Mr. Beers concluded.
Typical Applications
Here are a few test applications for which the signal generator’s digital modulation, signal formatting or vector control capabilities are essential:
Measuring Receiver Sensitivity
Digital receiver sensitivity is defined as the signal level at which a standard-encoded modulated signal produces a stipulated error rate. The test setup consists of the signal generator, the UUT and a bit error rate tester or a data transmission analyzer.4
Quadrature Modulator Evaluation
Many of the digital communication signal generators supply pristine I and Q outputs, providing pure constellations. Alternatively, controlled signal impairments, such as imbalances, may be introduced for I and Q. These signals are applied to the modulator to be evaluated and its output is analyzed with a data transmission test set or constellation analyzer.
Interference Simulation
By combining digital modulation and standard AM and FM facilities, it is possible to produce fading and Doppler frequency shift effects. Some generators include Rician and Rayleigh fading simulation capabilities (needed to verify mobile system performance) as a standard feature.
References
1. Padgett, J.E., et al, “Overview of Wireless Personal Communications,” IEEE Communications Magazine, January 1995, Vol. 33, No. 1, pp. 28-41.
2. Owen, D., “An Introduction to Digital and Vector Modulation,” Technical Information Bulletin, Marconi Instruments.
3. Digital Radio, Theory and Measurements, Application Note 355, Hewlett-Packard Co.
4. Anritsu MG3670B/3671A Digital Modulation Signal Generator, Applications p. 11.
5. RF Signal Generator Application Series, Winter 1994, Tektronix.
Table 1
Modulation
Network
GMSK
CT2, DCS 1800 (PCN), GSM
p /4-DQPSK
APCO 25, NADC, PDC, PHS, TFTS, TETRA
GFSK
DECT, CT2
4FSK
ERMES, FLEX
FSK
POCSAG, FLEX
Sidebar
Glossary of Terms
APCO Association of Public Safety Communications Officers
CDMA Code Division Multiple Access
CDPD Cellular Digital Packet Data (U.S.)
CT Cordless Telephone
DCS Digital Cellular System
DECT Digital European Cordless Telecommunications
DQPSK Differential Quaternary Phase Shift Keying
ERMES European Radio Message System
FDMA Frequency Division Multiple Access
FLEX Flexible High-Speed Paging System
4FSK Four Levels of Frequency Shift Keying
FSK Frequency Shift Keying
GFSK Gaussian Filtered FSK
GMSK Gaussian Minimum Shift Keying
GSM Groupe Special Mobile (originally); currently Global
System for Mobile Communication (ETSI, Europe)
JDC Japanese Digital Cellular
NADC North American Digital Cellular
PCN Personal Communications Network (Europe)
PCS Personal Communications Services (U.S.)
PDC Personal Digital Cellular (Japan)
PHP Personal Handy Phone System
PHS Personal Handyphone System (Japan, formerly PHP)
p /4DQPSK Differential Quadrature Phase-Shift Keying with a p /4 radian phase shift between successive symbols
POCSAG Post Office Code Standard Advisory Group
QAM Quadrature Amplitude Modulation
QPSK Quaternary Phase Shift Keying
TDMA Time Division Multiple Access
TETRA Trans European Trunked Radio System
TFTS Terrestrial Flight Telephone System
Products
Instrument Features Digital,
Vector Modulation, Fading
The 2050 Series Signal Generators cover ranges from 10 kHz to 1.35 GHz, 2.7 GHz and 5.4 GHz. In addition to conventional modulation capabilities, a wide-bandwidth I/Q modulator provides PSK, DPSK, QPSK, offset PSK and QAM formats, including signals required for NADC, PDC, PHS, TFTS and APCO25. PHP and DECT signals can be generated using ancillary equipment. Programmable RF channel filters are provided. Broadband AM, burst and spread spectrum signals can be generated. Rayleigh and Rician fading simulation capabilities help evaluate mobile receiver performance. From $19,200. Marconi Instruments, Inc., (800) 233-2955.
Generator Series Targeted
For Specific Comm Systems
The HP 8657A and B Signal Generators feature AM and FM and cover 100 kHz to 1,040 MHz and 100 kHz to 2,060 MHz, respectively. Option 022 adds a 0.3 GMSK modulation capability for GSM and PCN receiver testing. The HP 8657D and J feature AM, FM, pulse modulation, square-root raised cosine filters and p /4 DQPSK modulation for the NADC and JDC networks and the PHP systems. The HP 8657B, D and J also provide baseband I/Q signal outputs. From $9,760; with option add $4,080. Hewlett-Packard Co., (800) 452-4844.
Expansion Modules Match
Communications System Needs
The MG3671A Digital Modulation Signal Generator, with one to four expansion units each with continuous data generators, produces signals conforming to GSM, DCS1800, PCS1900, NADC, PDC, PHS, PACS, DECT, WCPE, CT2, TETRA and TFTS communications system standards. An internal burst function unit provides frame/slot data formats for all TDMA signals. The MG3671A covers 300 kHz to 2.75 GHz, delivers low adjacent channel power leakage and stable level accuracy to 13 dBm maximum output power, and achieves a vector error of <1.8%. $49,000. Anritsu Wiltron Co., (408) 776-8300.
General-Purpose Simulator
Performs Many Functions
The TSS-2000 covers 1,400 to 2,500 MHz and functions as a signal simulator or signal and sweep generator. Modulation formats include AM, FM, PM, PCM/FM, PCM/PM, PCM/BPSK and TV/FM. Bit error testing is possible since the unit simulates PCM data streams compatible with major BER test sets. I and Q inputs connect directly to a vector modulator. Unique data patterns are stored in built-in memory, down-loaded and sequenced. Maximum output is +20 dBm; Doppler simulation is available to +100 kHz. Dual RF output fade simulation is optional. From $48,000. Microdyne Corp., (904) 687-4633.
Generator Provides Signals
For >30 Communications Systems
The SME Signal Generator provides AM, FM, pulse and digital modulation, including GMSK, p /4 DQPSK, GFSK, FSK, 4FSK and FFSK. Data-synchronous bursts, frequency hopping and power ramping are internally controlled. Data sequences and trigger conditions are programmable. No external modulation or data sources are required. High spectral purity facilitates out-of-channel measurements. Output levels range from -145 dBm to +16 dBm. The SME 02 covers 5 kHz to 1.5 GHz and the SME 03 5 kHz to 3.0 GHz. The SME is custom-configurable and extendible. Starting at $19,300. Rohde & Schwarz, distributed by Tektronix, Inc., Test and Measurement Division, (800) 426-2200.
Copyright 1995 Nelson Publishing Inc.
November 1995
|
|