Hildin, John; Arias, Sergio; Teletronics Technology Corporation (International Foundation for Telemetering, 2006-10)
      Today’s data acquisition systems are typically comprised of data collectors connected to multiplexers via serial, point-to-point links. Data flows upstream from the sensors or avionics buses to the data acquisition units, to the multiplexer and finally to the recorder or telemetry transmitter. In a networked data acquisition system, data is transported through the network “cloud”. At the core of the network “cloud” is the network switch. The switch is responsible for distributing and directing data within the network. Network switches are commonplace in the commercial realm. Many businesses today could not function without them. A network-based data acquisition system, however, places additional burdens on the network switch. As in a commercial network, the switch in a data acquisition system must be able to distribute data packets within the network. In addition, it must be able to perform in a harsh environment, occupy a minimal amount of space, operate with limited or no external cooling, be configurable, and deal with the distribution of time information. This paper describes the required features of a ruggedized network switch and the implementation challenges facing its design. As a core component of a network-based data acquisition system, an ideal switch must be capable of operating in a large number of configurations, transporting and aggregating data between data sources and data sinks, with a mixture of devices operating at rates ranging from a few thousand bits per second to several gigabits per second, over twisted pair or fiber optic links. To ensure time coherency, the switch must also facilitate a time distribution mechanism, e.g., IEEE-1588 Precision Time Protocol (PTP). The gigabit switch described here uses the PTP to implement an end-to-end clock synchronization, for distributed acquisition nodes, to within 300 nanoseconds.

      Roach, John; Hildin, John; Teletronics Technology Corporation (International Foundation for Telemetering, 2006-10)
      Traditionally, acquired instrumentation data on a non-destructive test article is recorded to a nonvolatile memory recorder. The data acquisition system usually samples and formats its inputs before transmitting the data to the recorder (also known in this paper as a data sink) via a PCM serial data stream (i.e., clock and data). In a network-based data acquisition architecture, the inclusion of an IP-based recorder adds a new dimension to the data acquisition process. Any IP network inherently allows for the bi-directional exchange of data. In this environment, the IPbased recorder can be treated as both a data sink for parameter recording and a data source for parameter extraction, data rate statistics, and recorder status reporting. The network model recasts the data recorder’s function as a file server to which multiple clients could be simultaneously requesting services. Those clients that represent the data acquisition nodes are requesting storage of their acquired parameters. Clients, such as transmitters or test engineers, are requesting access to archived data or status information for further processing. This paper presents the advantages of using an IP-based recorder in a network-based data acquisition system. The availability of an IP interface along with the intelligence built into the recorder expands its capabilities beyond that of a conventional PCM recorder. These capabilities include real-time health monitoring, support for the Simple Network Management Protocol (SNMP), data mining, reporting of real-time performance and network statistics.

      Berdugo, Albert; Hildin, John; Teletronics Technology Corporation (International Foundation for Telemetering, 2006-10)
      Airborne data acquisition systems have changed very little over the years. Their growth has primarily been in the area of digital filtering and the acquisition of new avionic busses. Communication between data acquisition units operating as a system still employs Time Division Multiplexing scheme. These schemes utilize command and data busses like CAIS and PCM. Although this approach is highly efficient, it has many drawbacks. These drawbacks have resulted in rigid system architecture, system bandwidth limitations, highly specialized recorders to acquire unique avionic busses that would otherwise overwhelm the system bandwidth, and unidirectional flow of data and control. This paper describes a network centric data acquisition system that is Ethernet based. Although Ethernet is known as an asynchronous bus, the paper will describe a deterministic time distribution over the bus per IEEE-1588 that allows the use of a packet network for airborne data acquisition. The acquisition unit within the network system is defined by its MIB (Management Information Base) and operates as a data source unit. Other network components may operate as a data sink unit, such as recorders, or as a data source and sink. The role of different units in the network system will be evaluated. The paper will also describe network gateways that allow the use of traditional PCM systems with a network-based system.