To build an understanding of how digital technology will change testing, it is useful to first contrast digital with analog testing philosophies. In analog wireless systems the information is conveyed by modulation in the form of AM or FM, which changes the amplitude or frequency of a carrier signal in a linear way.
In a digital system, the analog signal representing the voice is passed to a vocoder, which converts the voice signal to a digital data stream using a defined algorithm. This algorithm varies from system to system, but results in a data stream in the order of 4-10kbytes/s being transmitted.
Analog Testing. Measurement technology for analog systems relies on the signal being present for sufficient time to make the measurement. In the case of the receiver SINAD this may be in the order of several seconds. The test equipment required ranges from power meters, modulation analyzers, audio analyzers, spectrum analyzers and signal generators.
When testing the transmitter of analog systems, five measurements are standard: 1) Power Output; 2) Transmitter Spurious; 3) Frequency; 4) Modulation; and 5) Distortion.
In receiver testing, we inject a low level carrier modulated in a manner appropriate to the demodulator under test, e.g. FM, at a known rate, say 1 kHz. We examine the audio output of the receiver in order to separate the original test tone (1 kHz) from any distortion and noise that the receiver added to the original test signal. We can equate the ratio of the two separated signals to some test threshold, and manipulate the receiver’s circuitry to optimize the ratio such that the distortion and noise portion is minimized.
Traditional test strategies have primarily focused on the parametric performance of the radio terminal, where measurements such as power, frequency, modulation and sensitivity are the primary indicators of performance. The open standard concept adopted for Project 25 introduces some new variables into the testing equation which relate to the interoperability of equipment sourced from different manufacturers supporting the standard.
Project 25, like its analog forebears, is an FDMA (Frequency Domain Multiple Access) system and it produces continuous signals when the radio is keyed. Therefore some of the more complex measurement techniques required for TDMA (Time Domain Multiple Access) systems such as TETRA are greatly simplified. Several key differences, however, distinguish the test required to determine system performance.
Digital Transmitter Testing.
A digital modulator imparts information to an assigned carrier by adjusting the carrier’s power, phase, or frequency, among a small dictionary of possibilities. Since there is no quantitative amount involved with digital modulation, no adjustments are generally required in the modulation path. Instead, it is usually sufficient to examine the quality of the modulation, or how good is the modulator at making its adjustments to the carrier. This is measured in percent, or some other quality score against a goal. When the goal is not met, one looks for a defective component that is causing the problem.
Digital Receiver Testing
Except for its decision circuit, digital receivers are remarkably similar to their analog cousins but they are tested in rather different ways. In the analog case, we find the SINAD procedure to be efficient, as it tests the whole receiver path in one pass.
Receiver sensitivity is determined by measuring its ability to recover data transmitted to it. The measurement referred to as receiver BER, or Bit Error Rate, measures the ratio of the bits received correctly as a percentage of the total bits transmitted. BER = 0.02 is an indication of a better receiver then one with BER = 0.2
Channel coding is a baseband process in most digital radio systems that deliberately adds carefully contrived redundancy to a symbol stream, which the receiver can use to repair a limited amount of damage to a recovered symbol stream. For those radios that protect all the user traffic with redundancy bits (which are also encoded into symbols), we loose the ability to detect individual bit errors. Instead, the frames or blocks of bits (symbols) that can not be repaired are marked, by the receiver itself as having suffered an error. FER = 0.10 is worse then FER = 0.00. The second result indicates no frames were received with uncorrected errors.
Increasingly, all of the baseband processes in digital radios are realized in software, and they either work, and work well; or they don’t work at all. It is not wise to spend valuable technician time verifying these functions.
The modulation selected for Project 25, Phase I is C4FM, which is modified four level Frequency Shift Keying (FSK), filtered with a raised cosine filter for minimizing inter-symbol interference. The modulation can be measured using conventional measuring techniques as long as standard test signals are used. These signals are designed to provide a data stream of all low deviation symbols or all high deviation symbols thus enabling the high + 1.8 kHz and low +0.6 kHz deviation to be measured. This is not practical on a working transmitter where the data content cannot be controlled without removing it from service; thus a new measurement technique has to be used. This requires that the transmitted signal is sampled and the data demodulated. The demodulated data is used to compute the instantaneous deviation from a "perfect" modulator. This deviation is then compared with the actual measured deviation value and a RMS Error Magnitude is calculated. This error is expressed as a % of 1kHz.
For the more complex systems such as that proposed for Project 25 Phase II, the data are represented by the instantaneous phase and amplitude of a vector. Measuring the fidelity of this type of modulation is much more difficult. One method takes a sample of the transmitted waveform over a large number of symbols (typically 100-200). From this information the demodulated data can be determined. This enables the test equipment to calculate the phase trajectory created by the data and then compare this with the sampled data. The errors in the measured data can then be determined and resolved into their various components.
The carrier frequency of the transmitter again can be easily measured if the transmitter is used in a conventional fm mode or using a test pattern. The Project 25 transmitters can be measured in this way. For more complex digital modulated signals such as those proposed for Project 25 II, the measurement of carrier frequency is more difficult, particularly when the system uses a bursting signal such as TETRA. Here the carrier frequency can be obtained from the same analysis that is used to determine the modulation accuracy. The carrier frequency offset from the required channel frequency results in a fixed phase error for all the symbol point measurements. This can be subtracted from the measured phase angles and used to calculate the frequency with knowledge of the symbol period. Transmitter frequency is also important during the key up process. Here excessive frequency error can interfere with other users.
The Sensitive Issue of Bit Error Rate Measurement.
Receiver sensitivity for digital systems is measured by determining the BER (Bit Error Rate) which is defined as the number of bits received in error expressed as a % of the total number of bits received. This measurement is not quite as simple as it sounds because it depends where in the radio system you choose to make the measurement. Of the total throughput of a Project 25 channel of 9600 bit/s, only 4400 bit/s are associated with the digital voice. The remaining 2800 bit/s are used for error correction on the voice signal and 2400 bits/s are devoted to signaling overhead. Thus, a key issue arises: Do you measure the errors before or after correction? Obviously it is the corrected performance which is important to the user because this determines the intelligibility of the speech. To overcome the uncertainty of this measurement a test signal has been defined with a known bit pattern. Comparing the received signal with the expected test signal enables a BER to be calculated. The nominal bit error rate for a Project 25 receiver is 5%. An alternative to the conventional BER is to use a pre-recorded speech pattern, which gives an audible indication of the receiver sensitivity. Because of the high amount of error correction used, the point at which it fails is quite abrupt and the sensitivity can be determined accurately.
Additional measurement such as Adjacent Channel Power and emission spectrum measurements are specified to ensure that Project 25 equipment does not interfere or degrade the performance of the coexistent analog channels. While these measurements are important to overall system integrity, they are unlikely to be required in the system maintenance environment and require performance levels that can only be obtained from specialized test equipment.
Working Together: How Interoperability Impacts Test.
Objective is to ensure terminals of different types and manufacturers operate correctly on the system.
While the open standard creates many benefits for the user in the long term (i.e. reduced equipment costs and greater customer choice) it does introduce another uncertainty. Will the equipment from different suppliers work together seamlessly?
The standardization process is designed to create a standard that defines all aspects of the system operation. It is inevitable however, that equipment suppliers will interpret the documents in differing ways. This may result in a terminal and repeater from different manufacturers not interoperating as expected. Additionally as new features become available on a system, and the equipment is updated, the need to re-assess interoperability may occur. This aspect of system test will gain in significance as more equipment manufacturers enter the Project 25 marketplace.
Two test strategies can be used to determine interoperability. The first method is to test each new equipment type with all of the existing equipment used in the network. It is obvious that not all of the terminal features on all of the user channels can be tested, as the testing time would escalate very rapidly. Even making some very basic interoperability tests, when the number of different units in a system is large produces a tremendous number of combinations to test. This approach does ensure the equipment interoperates but does not ensure adherence to the standard, nor does it necessarily indicate which equipment is non-compliant.
An alternative approach is to use a reference system or device against which all the equipment is checked. If it were practical to use only one reference system for all tests, then this system itself would become the standard. Typically a radio test set could be used as the reference, but this again is based on one manufacturer's interpretation of the standard. Also the number of functions that can be tested will be restricted by the reference system capability. This method significantly reduces the number of tests required although interoperability will only be inferred, not guaranteed.