Communications systems provide reliable performance only after they have been tested under a wide range of operating conditions. Wirelesscommunications systems, for example, must be evaluated for the effects of such factors as path loss, immunity to interference, multipath reflections, and atmospheric losses. Real-world testing at the site of a communications system will yield the best overall information about the system, but this is not practical in most cases. The "communications interoperability" of a system can be evaluated in a controlled laboratory environment with the help of attenuation matrix units.
An attenuation matrix can be used to simulate the multiple paths traveled by transmitted signals in a wireless-communications system, including direct and reflected signal paths. By adjusting the attenuation levels of the different paths through an attenuation matrix, the real-world fading conditions facing a communications system can be emulated in a repeatable, controlled laboratory environment.
Testing with attenuation matrix units has been performed with a variety of commercialcommunications systems, including cellular-communications systems such as wideband-codedivision- multiple-access (WCDMA) systems, IEEE 802.11 wireless local-area networks (WLANs), third-generation (3G) cellular systems, universal-mobile-telecommunicationssystems (UMTS) equipment, Global System for Mobile Communications (GSM) equipment, and a variety of other voice, data, FAX, and video-communications systems. Attenuation matrix systems have also been deployed in laboratory signal simulation situations to demonstrate battlefield communications, missile communications, satellite communications, and even communications between planets (Earth to Mars and Mars to Earth). Attenuation matrix test subsystems provide system engineers with realistic communication path/measurement tools to evaluate and solve the problem of equipment "interoperability."
Attenuation matrix test subsystems can be implemented in a variety of different ways, with many different block diagrams representing test solutions. The structure of these matrix units are all based on a common design platform but can be adjusted to fit specific operational requirements. The most common design approach to an attenuation matrix subsystem is a passive, nonblocking design. It has a certain number (X) of inputs feeding an input power divider. That power divider then feeds multiple programmable attenuators, which attach to an output power divider and to a certain number (Y) of output ports. The typical design is bidirectional and has significant through path loss due to the passive power dividers. If gain is required, then the system will become unidirectional. This format results in an X by Y attenuation matrix, such as the 6 x 2 attenuation matrix shown in Fig. 1.
When performing the system engineering evaluation of an attenuation matrix, a user must evaluate a number of test-system constraints regarding the attenuation matrix, including:
- The number of inputs versus outputs will define the structure of the (X by Y) matrix.
- These test systems are typically designed with the same input and output frequencies.
- The bandwidth of the system is constrained by the bandwidth of available power dividers (usually a few octaves).
- The input power level is usually determined by the power ratings of the power dividers and the programmable attenuator.
- The system may be designed by one of several different types of programmable attenuators, including relay, PIN diode, FET, and phasecompensated types.
- The system dynamic range is greatly influenced by the port-to-port isolation of the components, such as the power dividers.
- The total insertion loss is also a determining factor in the system dynamic range.
- The system dynamic range may appear simple to define, but it actually requires some system-level analysis, starting with the dynamic range of the attenuator.
- It is also often necessary to specify the maximum allowable phase change during any change in attenuation, especially for phase-sensitive measurements.
Matrix attenuators can be used in a variety of different measurement configurations, although they typically have transceivers on both the input and output ports. To perform interference testing, one of the ports is used as a jammer or interference port to introduce real-world noise or system into the system. The variable attenuators are used to simulate signal loss due to movement, path loss, distance, and/or reflections from the transmitter to the receiver. This information is important to the field system designers locating positions of repeater towers used for signal handover. The test system should provide as much dynamic range as possible in order to improve the quality of the information that can be obtained in determining the field performance of a communications system or device under test.
Not all testing requirement involving attenuation matrices deal with attenuation. Both military and commercial systems feature field requirements involving both amplitude and controlled phase testing. Attenuation matrix subsystems developed at Aeroflex/Weinschel offer this capability through the use of phasecompensated programmable attenuators. Most programmable attenuators provide an accurate amplitude adjustment range, but are not specified for controlled phase during amplitude adjustments. As a result, as the amplitude changes, the phase can shift dramatically for each attenuation step. When controlled phase is a requirement, it must be initially designed into the attenuation matrix test subsystem.
A quick evaluation of an attenuation matrix test subsystem would indicate that the attenuation range of the programmable attenuator determines the dynamic range of the test subsystem, or 63 dB for a programmable attenuation with adjustment range of 0 to 63 dB. One of the keys to using these subsystems is that all input ports are connected to all output ports at all times and the programmable attenuators are used to adjust the signal levels. In practical systems, because all of the inputs are connected to all of the outputs, "sneak paths" occur. "Sneak" paths are nonintended signal leakage paths in the matrix that can limit the dynamic range of the attenuation matrix test subsystem.
Figure 2 shows the block diagram of an 8 x 4 attenuation matrix based on 0-to-63-dB programmable attenuators. This type of design is limited in dynamic range by the isolation of the input and output dividers. The worst-case dynamic range occurs if one primary path is used and all other paths have 0 dB attenuation selected. Any increase in attenuation level in secondary paths will improve the dynamic range of the primary path.
The system dynamic range can be found by evaluating the performance contributions of the individual components. For example, the four-way power divider contributes 7-dB insertion loss with 20-dB port-to-port isolation. The eight-way power divider has 10-dB insertion loss with 20-dB port-to-port isolation. The 63-dB programmable attenuator adds 4-dB insertion loss in its 0-dBattenuation position. The fixed attenuator adds another 6 dB insertion loss, and the coaxial cable adds 2 dB to the insertion- loss total. The main path loss through the system is indicated by the green line through Fig. 2.
The insertion loss can be calculated as follows:
4-way divider + 6 dB attenuation + programmable attenuator + 6 dB attenuation + 8-way divider + cable attenuation = total = 7 dB + 6 dB + 4 dB + 6 dB + 10 dB + 2 dB = 35 dB
Continued on page 2
Ideally, a test-system engineer should hope to be able to adjust the attenuation matrix output signal from having 35 dB loss (at 0 dB attenuation for the programmable attenuator) to 98 dB loss when the full 63 dB attenuation from the programmable attenuator is added.
The sneak path loss for the attenuation matrix of Fig. 2 can be calculated by considering the loss contributions of the individual components as signals travel through the system. The total insertion loss is the sum of the losses of the four-way divider + the cables + the 6-dB attenuator + the programmable attenuator + the 6-dB attenuator + the isolation of the eight-way divider + the cables + the 6-dB attenuator + the programmable attenuator + the 6-dB attenuator + the isolation of the four-way divider + the 6-dB attenuator + the programmable attenuator + the 6-dB attenuator + the eight-way divider + the cables. In this case, the total attenuation through one sneak path is 7 dB + 2 dB + 6 dB + 4 dB + 6 dB + 20 dB + 2 dB + 6 dB + 4 dB + 6 dB + 20 dB + 6 dB + 4 dB + 6 dB + 10 + 2 dB = 111 dB
Since there are seven sneak paths in this system, the other paths will add further signal leakage and degradation in the system dynamic range. The extra loss of dynamic range can be found from:
10log(number of sneak paths) = loss in dynamic range
10log(7) = 8.5 dB In order to obtain the full system
dynamic range, the isolation between the sneak paths and the main signal path must be determined. The total required isolation can be found from the sum of the attenuation range (63 dB), the sneakpath leakage , and the margin for achieving less than 0.5 dB system measurement error (10 dB), which is 63 dB + 8.5 dB + 10 dB = 81.5 dB. The sneak path loss minus the main path loss is 111 dB - 35 dB = 76 dB. According to these calculations, there is not enough system isolation to obtain the full 63 dB of dynamic range when all secondary paths are set to 0 dB. As a result, this attenuation matrix system will be limited to 63 dB - 81.5 dB - 76 dB) or 57.5 dB of dynamic range, when all paths are set to 0 dB and one path is being used. The range of attenuation control will be 35 dB + 57 dB = 92 dB.
Three things can be done to improve this situation:
- Always operate the system with secondary paths containing as much attenuation as possible;
- Increase the 6-dB attenuators to add more port-to-port isolation (although this will also increase the net system insertion loss); and
- Install dividers with higher port-toport isolation (which can be difficult to achieve in wideband systems).
In summary, attenuation matrix units based on this design configuration offer a very dynamic test platform for evaluating field interoperability issues. By understanding the internal elements and the limitations of the test subsystem, a solid set of test data can be obtained to determine communications-systems interoperability.