High Temperature (N05-012)
1 Identification and significance of the opportunity
The current generation of air vehicle and propulsion systems can be characterized as centralized control systems in which a (redundant) central computer and centrally located analog signal interfacing circuitry is used to interface with sensors and actuators located throughout the aircraft and the propulsion system. In most instances, the digital computer implements the “control laws” while the analog circuitry is used for conditioning the inputs, outputs and for performing actuator loop closures etc. The types and quantities of sensors and actuators are unique to each system and therefore lead to an implementation that is new and customized for each new aircraft and engine. A block diagram of this centralized architecture is shown in fig. 1.0-1. The cost of designing and maintaining these new and customized systems is enormous. Numerous studies and concept papers {1-2} have shown the advantages of distributed open system architectures in which centralized computers communicate via standardized serial data buses with “smart” boxes, which are co-located with the sensors and actuators. The studies also suggest that such a distributed architecture would greatly reduce not only the total ownership cost (i.e. development costs plus acquisition and maintenance cost) but also significantly reduce the weight and improve the fault diagnosis (i.e. detection and isolation) capability due to the simplification and standardization of the wire harnesses. For example, the Navy estimates that it spends more than a million man-hours per year diagnosing wiring faults {3}. In any case, the “smart” boxes would provide for an open architecture and “standardized” bus oriented method for interfacing by the host computer regardless of the signal conditioning and loop closure electronics specific to the actuator or sensor involved. Furthermore, new systems could then be easily and affordably assembled around a central processor by re-using these “smart” boxes. These benefits would be greatly enhanced if these “smart” boxes were somehow made reconfigurable or more “generic” and could therefore be reusable for multiple types of actuators and sensors.
Figure 1.0-1 Centralized Architecture with Custom Elements
There are two obstacles to the realization of such open and distributed air vehicle and propulsion system architectures containing “generic” Electronic Interface Modules for actuators. First, this open and distributed solution requires the “Generic” Electronic Modules (GEM) to co-locate and operate in harsher environments including the high temperature environments (~ 200 deg. C) associated with actuators. The thermal environment of future super-cruise air vehicle propulsion systems and flight control actuators is likely to be even more demanding. However, most electronic components, even industrial grade ones, only operate up to 125 deg. C. The lack of high temperature (~200 deg. C) parts was a major driver for the DARPA/US Air Force HiTeC (High Temperature distributed Control) program {4}, to be launched in 1996. The HiTeC program has led to the commercial development of several Silicon-on-Insulator based High Temperature component families (e.g. Honeywell, Cree) with a basic set of analog and digital components that can continuously operate at High Temperatures (225 deg. C). Prominent among them is the HTMOS family from Honeywell (http://www.ssec.honeywell.com/hightemp/) which offers several basic essential building blocks, such as microprocessors and memories, for program, data and parameter storage, as well as Gate Arrays and Op. Amps., all of which are essential for building a generic electronic module. However the HTMOS family from Honeywell does lack, at least at this point, some key components such as an A/D converter.
The second obstacle to the realization of “generic” electronic modules is the diversity of the quantity and types of sensors (position, speed, pressure difference, temperature etc.), which makes it difficult to develop a “one size fits all” generic analog signal conditioning interface that can be “reused”. In the absence of such a capability, each electronic module would have to be custom to the specific actuator and sensors it was designed for. Even if such a custom electronic box was “open”, i.e. it had standards based interfaces, the cost of developing, procuring and maintaining such custom boxes would once again be significant. The proliferation and lack of reuse for such custom boxes would make the distributed open architecture approach very unattractive. Therefore, the potential benefit of low cost of ownership offered by open systems architectures has created a great incentive for developing Generic Electronic Modules (GEM) which can operate in High Temperature and which can be reused for multiple applications.
The key innovation in the proposed solution is the maximal digital implementation of the GEM, wherein all possible analog functions are digitally implemented making them reconfigurable. This emphasis on digital implementation is driven by two factors. First, digital implementations, especially software driven digital implementations are more re-usable because the software permits the same hardware, without changes, to be re-used for performing different functions (e.g. interfacing with different signal types). Second, digital components are inherently more tolerant of high operating temperatures and are therefore easier to manufacture. This tolerance is primarily due to the resistance to increased leakage currents in the ICs. The proposed GEM therefore is mostly digital with the minimum necessary analog “front ends” consisting basically of Op. amps, that can be reused for interfacing with a diverse group of analog sensor signals. The basic elements of this concept are sketched in fig. 1.0-2 in which the custom hardware components shown in fig. 1.0-1 are replaced with standard data buses and software driven digitally reconfigurable interfaces.
Figure 1.0-2 Reconfigurable GEM Architecture
A key feature of the proposed digital implementation is that it represents a low risk extension of the proven design of a “Smart Terminal” module to the High Temperature environment. The “Smart Terminal” was designed as a reconfigurable module for implementing the closed loop control of hydraulic actuators under the command of a remote host processor via a serial bus. (See section 4.1 for a description of the Smart Terminal design.) The project successfully demonstrated, in the laboratory, several concepts of reconfigurability and digital implementations involving AC and DC sensors and sensor excitations. Therefore, the extension of this Smart Terminal design to the present High Temperature actuator Generic Electronic Module (GEM) requires, primarily, a mapping of that design to a high temperature components based design and thus represents a relatively low risk. Furthermore, a review of the HTMOS parts library reveals that all of the building blocks needed for the High Temperature GEM are either available or can be built using the available HTMOS components. For instance, a 12 bit, accurate, A/D converter, very difficult to implement as a monolithic single device in high temperature, can be constructed using digital glue logic around a 1 bit A/D converter, i.e., an op. amp, and a comparator both of which are readily available from the HTMOS family. Other functions can also be suitably mapped over to a combination of op. amps, digital logic and the microprocessor. We are therefore confident that a Generic Electronic Module (GEM) for High Temperature can be demonstrated during Phase I of this program and will therefore pave the way for an affordable, open, distributed architecture air vehicle and propulsion control systems in Phase II.
2 Phase I Technical Objectives
The stated technical objective of the Phase I program is to develop and demonstrate the feasibility and capabilities of an open systems solution to the general problem of lack of generic electronic interfaces for actuators in extreme environments in air vehicles and propulsion systems.
The proposed program builds on the concepts of reconfigurability and digital implementation of analog functions developed during the “Smart Terminal” program mentioned earlier (see section 4.1 for details of the Smart Terminal design). The reconfigurability or reusability in this concept is derived from the use of different software algorithms, which permit the same analog “front end” hardware to be re-used for interfacing with DC or AC analog signals in order to “create” different analog functions. The list of analog “front end” functions implemented digitally in that project included generation of A/C excitation and AC demodulation needed for LVDT position sensors. These functions were digitally implemented in the microprocessor with the help of an op. amps., and A/D, D/A converters.
The primary difference between the “Smart Terminal” implementation and the requirements of the GEM is the High Temperature environment. Because the earlier project had no High Temperature requirements, it was built using industrial grade ICs. As such there was no need to construct basic building blocks such as an A/D or a UART or Bus Transceivers needed by the serial bus. Since these building blocks are not available in the library of HTMOS parts, they will have to be built using available components from the library of HTMOS and other families. In addition to these front end functions, the GEM design will require additional building blocks for frequency and pulse type input signals associated with pressure and temperature sensors, or torque motor position signals. Additionally, the design will need to build the bus interface logic functions and transceivers for use with serial buses such as RS-485.
In view of this past experience base with the design of the Smart Terminal, the program objectives will be appropriately tailored to focus on the design and testing of the High Temperature building blocks before integrating them into the design of the GEM.
3 Phase I Work Plan
3.1 Interface Definition:
This task involves defining the interface characteristics, including electrical characteristics, accuracy and resolution etc. for a representative Generic Electronic Module (GEM). These interfaces will include selected classes of inputs signals from actuator position sensors and associated equipment, output signals for actuators drives and associated equipment, along with interfaces for a data bus and for power received from the host. The input signal class will include the following signals:
1. DC and AC actuator position signals from hydraulic and/or EM actuators
2. Frequency signals from pressure and temperature sensors
3. Speed and frequency signals from torque motors and EM actuators
4. Analog or discrete switch position signals from contactors and solenoids
The output signal class will consist of the following signals:
1. Current drive signals for hydraulic actuators
2. H Bridge signals (with or without PWM) for torque motors and EM actuators
3. Shutoff and engage discrete signals for solenoids and switches
It should be pointed out that the proposed GEM does not include the final power drive stages for high power actuator drives for several reasons. First, these drive tend to consume a lot of power and generate lots of heat, which can adversely affect the temperature inside a GEM enclosure. Second, the interfaces tend to be quite specific to the type of actuator and including them in a GEM would not be consistent with its “generic” nature. Finally, the EMI treatment needed for some of these power stages tends to be quite elaborate (e.g. Transorbs) and tailored to the type of drive involved. It is believed that the actuator housing is therefore the best possible location for these power stages because the large mass and relatively steady temperature profiles at these sites provide significant reliability benefits for the electronics in spite of their high temperature environment. The best example of this reliability benefit is the FADEC engine controller which has an excellent reliability in spite of its harsh environment because it is fuel cooled. In any case, the output(s) built into the GEM will be generic in design so that they can directly be connected to the power drive stages located at the actuators.
The serial data bus for the GEM will be the ubiquitous RS-232. This choice is primarily driven by the constraints of the HTMOS High Temperature chipsets available for implementation at this time. However, the choice is not much of a performance constraint because the commands to be sent over this bus from the host control computer are generated at no more than 100 Hz. This update rate is not likely to go up for super-cruise air vehicles because of inherent mechanical considerations. This update rate is distinctly different from the actuator inner loop closure rates which can be much higher (> 1000 Hz) especially for DDV actuators and the latest EM actuators on newer aircrafts such as JSF. In any case the RS-232 bandwidth achievable will depend on the speed and functions available from the High Temperature chips in the market. It is anticipated that the available chipsets will support development of a 9600 Baud RS-232 which can easily support the above command update rates.
This task will also define the aircraft power requirements for the GEM. The power requirements of the actuator power drive stages are too specific to the type of actuators involved and will not be addressed in this task.
3.2 High Temperature Electronics Building Blocks:
This task will focus on building and testing the following major new building blocks in a standalone fashion:
1. A/D Converter: In most actuator applications, actuator positional accuracy requirements are adequately met by a 12 bit A/D converter. The Honeywell HTMOS library does not offer any A/D converter. A 10 bit A/D converter offered by Cissoid Inc., a fab-less High Temperature Silicon-on-Insulator component development company (http://www.cissoid.com/), is still under development. Hence the 12 bit A/D converter will have to be built up from the library of parts that are available. An A/D converter can be constructed in one of several ways. An A/D based on the Successive Approximation Register (SAR) method can be constructed (see fig. 3.2-1) using an Op amp., plus a comparator, precision D/A converter and a precision capacitor. Each ADC conversion is divided into two distinct phases as defined by the position of the switches shown in Fig. 3.2-1. During the sampling phase (with SW1 and SW2 in the “track” position), a charge proportional to the voltage on the analog input is developed across the input sampling capacitor. During the conversion phase (with both switches in the “hold” position), the capacitor DAC is adjusted via the SAR logic until the voltage on node A is zero, indicating that the sampled charge on the input capacitor is balanced out by the charge being output by the capacitor DAC. The digital value finally contained in the SAR is then latched out as the result of the ADC conversion. Control of the SAR, and timing of acquisition and sampling modes, is handled by the ADC control logic. Clearly the acquisition and conversion times of the SAR method will require fine tuning in order to meet the rapid sampling requirements of the A/D within the limitations of the HTMOS family.