Engineering topics
Publications and source records attributed to Blaes, B. R..
This paper briefly describes an on-line hardware system based on 16x8x4 cascading neural network architecture.
Miniature high-performance low-mass space avionics systems are desired for planned future outer planetary exploration missions (i.e. Europa Orbiter/Lander, Pluto-Kuiper Express). The spacecraft fuel and mass requirements enabling orbit insertion is the driving requirement. The Micro Navigator is an integrated autonomous Guidance, Navigation & Control (GN&C)micro-system that would provide the critical avionics function for navigation, pointing, and precision landing. The Micro Navigator hardware and software allow fusion of data from multiple sensors to provide a single integrated vehicle state vector necessary for six degrees of freedom GN&C. The benefits of this MicroNavigator include: 1) The Micro Navigator employs MEMS devices that promise orders of magnitude reductions in mass power and volume of inertial sensors (accelerometers and gyroscopes), celestial sensing devices (startracker, sun sensor), and computing element; 2) The highly integrated nature of the unit will reduce the cost of flight missions. a) The advanced miniaturization technologies employed by the Micro Navigator lend themselves to mass production, and therefore will reduce production cost of spacecraft. b) The integral approach simplifies interface issues associated with discrete components and reduces cost associated with integration and test of multiple components; and 3) The integration of sensors and processing elements into a single unit will allow the Micro Navigator to encapsulate attitude information and determination functions into a single object. This is particularly beneficial for object-oriented software architectures that are used in advanced spacecraft. Additional information is contained in the original extended abstract.
An avionics system is a platform residing on a flight vehicle that provides the resources (hardware and software) needed to manage the flight mission. Distributing the avionics system components, both functionally and spatially, can provide characteristics that benefit the overall system robustness, reliability, testability, and maintainability. This abstract discusses the concept of a distributed avionics system consisting of a network of micro-avionics nodes. Additional information is contained in the original extended abstract.
Four Free-Flying Magnetometers (FFMs), developed at the Jet Propulsion Laboratory (JPL) for the Enstrophy mission, were successfully deployed from the payload of a sounding rocket launched from Poker Flats, Alaska on February 11, 1999. The FFMs functioned successfully by synchronously measuring the vector magnetic field at 4 points separate from the payload and at relative distances up to 3 km, and communicated their data, in bursts, to the ground. This is the first time synchronized in-situ multipoint measurements of the Earth's magnetic field utilizing miniature spin-stabilized "sensorcraft" have been performed. The data they provided have enabled, for the first time, the direct measure of field-aligned current density and are enabling new science by determining the fine-scale structure of the currents in the Earth's ionosphere involved in the production of aurora. These proof-of-concept "hockey puck" (80 mm diameter, 38 mm height, 250 gram mass) FFMs were built using off-the-shelf commercial, industrial, and military grade surface-mount electronic components. Radiation-hard electronics was not required for the Enstrophy mission's short sub-orbital flight. The successful design, implementation, and flight demonstration of this 1st generation FFM design has provided a solid base for further development of a 2nd generation FFM design for planetary science applications. A reliable ultra-miniature radiation-hard 2nd-generation FFM utilizing System-On-A-Chip (SOAC) technologies is proposed. This design would be targeted for long-term planetary missions to investigate magnetospheric field configurations in regions having small-scale structure and to separate spatial and temporal variations. A fleet of short-lived (expendable) FFMs would be deployed into a targeted region to gather multiprobe vector magnetic field data. The FFMs would be ejected from a parent spacecraft at a speed of a few m/sec and would cover spatial volumes of order tens of kilometers for times of order one hour. The parent spacecraft would carry a sufficient number of FFMs for multiple deployments. Additional information is contained in the original extended abstract.
Nonvolatile memory (NVM) system that could reliably function in extreme environments is one of the most critical components for many spacecrafts being developed for NASA missions to be launched in next four to seven years. NVM supports the computer system in saving and updating critical state data required for a warm restart after power cycling or in case of a power bus failure. It also provides a power independent mass storage capacity for the scientific data gathered by the instruments. In some cases the window for gathering such data is very small and occurs only once in a given mission. Commercially popular and fully developed Flash NVM technology is inappropriate for many reasons such as the limited read write cycles with slower access speeds, radiation intolerance, higher Single Event Upsets (SEU) rates, etc. It is desirable to have an NVM system based upon a robust cell technology making it immune to the SEUs and with sufficient radiation hardness. Availability of such NVM system seems to be still 5 to 10 years in the future. Meanwhile, it is possible to provide an interim hybrid solution by combining the existing rad-hard technologies. Additional information is contained in the original extended abstract.
Data storage systems based on vertical Bloch lines have been fabricated in 2 um bubble garnet material. The input/output line was fully operational, and experiments in reading and writing showed an encouraging correlation between input and output.
Custom proton sensitive SRAM chips are being flown on the BMDO Clementine missions and Space Technology Research Vehicle experiments. This paper describes the calibration procedure for the SRAM proton detectors and their response to the space environment.
An inverter matrix test circuit was designed for the Clementine space mission and is built into the RRELAX (Radiation and Reliability Assurance Experiment). The objective is to develop a circuit that will allow the evaluation of the CMOS FETs using a lean data set in the noisy spacecraft environment.
On-chip p-FETs were developed to monitor the radiation dose of n-well CMOS ICs by monitoring the threshold voltage shifts due to radiation induced oxide and interface charge.
The SEU/SRAM is a 4-kbit Static Random Access Memory (SRAM) designed to detect Single-Event Upsets (SEUs) produced by high energy particles. This device was used to determine the distribution in the memory cell spontaneous flip potential.
On-chip p-FETs were developed to monitor the radiation dose of n-well CMOS ICs by monitoring threshold voltage shifts due to radiation induced oxide and interface charge.
Data from a MOSFET matrix on two JPL (CIT Jet Propulsion Laboratory) CRRES (Combined Release and Radiation Effects Satellite) chips, each behind different amounts of shielding, are presented. Space damage factors are nearly identical to ground test values for pMOSFETs. The results from neighboring rows of MOSFETs show similar radiation degradation. The SRD (Space Radiation Dosimeter) is used to measure the total dose accumulated by the JPL chips. A parameter extraction algorithm that does not underestimate threshold voltage shifts is used. Temperature effects are removed from the MOSFET data.
A methodology is described for predicting the SEU susceptibility of a standard-cell D-latch using an alpha-particle sensitive SRAM, SPICE critical charge simulation results, and alpha-particle interaction physics. Measurements were made on a 1.6-micron n-well CMOS 4-kb test SRAM irradiated with an Am-241 alpha-particle source. A collection depth of 6.09 micron was determined using these results and TRIM computer code. Using this collection depth and SPICE derived critical charge results on the latch design, an LET threshold of 34 MeV sq cm/mg was predicted. Heavy ion tests were then performed on the latch and an LET threshold of 41 MeV sq cm/mg was determined.
A test chip set being developed to aid in the qualification of spaceborne Application Specific Integrated Circuits (ASICs) is described. The chip set consists of a process monitor for process parameter verification, a fault chip for yield analysis, a reliability chip for ASIC failure rate analysis, and total ionizing dose and single event upset chips for radiation effect analysis. The test structures contained in these chips are discussed along with representative test results.
A set of test 'modules' for verifying the quality of a complementary metal oxide semiconductor (CMOS) process at the end of the wafer fabrication is documented. By electrical testing of specific structures, over thirty parameters are collected characterizing interconnects, dielectrics, contacts, transistors, and inverters. Each test module contains a specification of its purpose, the layout of the test structure, the test procedures, the data reduction algorithms, and exemplary results obtained from 3-, 2-, or 1.6-micrometer CMOS/bulk processes. The document is intended to establish standard process qualification procedures for Application Specific Integrated Circuits (ASIC's).
A bench-level test is being developed to evaluate memory-cell upsets in a test SRAM designed with a cell offset voltage. This offset voltage controls the critical charge needed to upset the cell. The effect is demonstrated using a specially designed 2-micron n-well CMOS 4-kb test SRAM and a Po-208 5.1-MeV 0.61-LET alpha-particle source. This test SRAM has been made sensitive to alpha particles through the use of a cell offset voltage, and this has allowed a bench-level characterization in a laboratory setting. The experimental data are linked to a alpha-particle interaction physics and to SPICE circuit simulations through the alpha-particle collection depth. The collection depth is determined by two methods and found to be about 7 micron. In addition, alpha particles that struck outside the bloated drain were able to flip the SRAM cells. This lateral charge collection was observed to be more than 6 micron.
Qualification procedures for complex integrated circuits are being developed under a U.S. government program known as Qualified Manufacturing Lines (QML). This effort is focused on circuits designed by IC manufacturers and has not addressed application specific IC's (ASIC's) designed at system houses. The qualification procedures described here are intended to be responsive to the needs of system houses who design their own ASIC's and have them fabricated at Silicon foundries. A particular focus of this presentation will be the use of the TID (total Ionizing Dose) Chip to evaluate CMOS foundry processes and to provide parameters for circuit simulators. This chip is under development as a standard chip for qualifying the total dose aspects of ASIC's. The benefits of standardization are that the results will be well understood and easy to interpret. Data is presented and compared for 1.6 micron and 3.0 micron CMOS. The data shows that 1.6 micron CMOS is significantly harder than 3.0 micron CMOS. Two failure modes are explored: (1) the radiation-induced degradation of timing delays; and (2) radiation-induced leakage currents.
A timing sampler consisting of 14 four-stage inverter-pair chains with different load capacitances was fabricated in 1.6-micron n-well CMOS and irradiated with cobalt-60 at 10 rad(Si)/s. For this CMOS process the measured results indicate that the rising delay increases by about 2.2 ns/Mrad(Si) and the falling delay increase is very small, i.e., less than 300 ps/Mrad(Si). The amount of radiation-induced delay depends on the size of the load capacitance. The maximum value observed for this effect was 5.65 ns/pF-Mrad(Si). Using a sensitivity analysis, the sensitivity of the rising delay to radiation can be explained by a simple timing model and the radiation sensitivity of dc MOSFET parameters. This same approach could not explain the insensitivity of the falling delay to radiation. This may be due to a failure of the timing model and/or trapping effects.