Engineering topics
Publications and source records attributed to Smith, L. S..
A fifth set of heavy ion single event effects (SEE) test data have been collected since the last IEEE publications (1,2,3,4) in December issues for 1985, 1987, 1989, and 1991. Trends in SEE susceptibility (including soft errors and latchup) for state-of-the-art parts are evaluated.
JPL and the Aerospace Corporation have collected a fourth set of heavy ion single event effects (SEE) test data. Trends in SEE susceptibility (including soft errors and latchup) for state-of-the-art parts are displayed. All data are conveniently divided into two tables: one for MOS devices, and one for a shorter list of recently tested bipolar devices. In addition, a new table of data for latchup tests only (invariably CMOS processes) is given.
Multiple-bit errors caused by single-ion tracks in a 256-kb DRAM fabricated by a bulk process were observed for different ion species and stopping power values. The results demonstrate the utility of this device for the evaluation of ion-beam uniformity and ion-beam-induced charge collection in IC devices. The data indicate that single-ion-induced charge transport results in multiple-bit error clusters due to lateral diffusion of excess minority carriers (electrons). Charge collection occurred from a depth of up to 35 microns from the surface of the device. An apparent charge loss was observed for very heavy ions with a high stopping power (Au at 350 MeV).
A system using Cf-252 and associated nuclear instrumentation has determined the single-event upset (SEU) cross section versus linear energy transfer (LET) curve for several 2K x 8 static random access memories (SRAMs). The Cf-252 fission fragments pass through a thin-film organic scintillator detector (TFD) on the way to the device under test (DUT). The TFD provides energy information for each transiting fragment. Data analysis provides the energy of the individual ion responsible for each SEU; thus, separate upset cross sections can be developed for different energy and mass regions of the californium spectrum. This californium-based device is quite small and fits onto a bench top. It provides a convenient and inexpensive supplement or alternative to accelerator and high-altitude/space SEU testing.
The spread of charge induced by an ion track in an integrated circuit and its subsequent collection at sensitive nodal junctions can cause multiple-bit errors. The authors have experimentally and analytically investigated this phenomenon using a 256-kb dynamic random-access memory (DRAM). The effects of different charge-transport mechanisms are illustrated, and two classes of ion-track multiple-bit error clusters are identified. It is demonstrated that ion tracks that hit a junction can affect the lateral spread of charge, depending on the nature of the pull-up load on the junction being hit. Ion tracks that do not hit a junction allow the nearly uninhibited lateral spread of charge.
JPL and Aerospace have collected a third set of heavy-ion single-event phenomena (SEP) test data since their last joint IEEE publications in December 1985 and December 1987. Trends in SEP susceptibility (e.g., soft errors and latchup) for state-of-the-art parts are presented. Results of the study indicate that hard technologies and unacceptably soft technologies can be flagged. In some instances, specific tested parts can be taken as candidates for key microprocessors or memories. As always with radiation test data, specific test data for qualified flight parts is recommended for critical applications.
A technique has been developed to support the study of the effects of cosmic rays on integrated circuits. The system is designed to determine the particle distribution across the surface of an integrated circuit accurately while the circuit is bombarded by a particle beam. The system uses photomultiplier tubes, an octal discriminator, a computer-controlled NIM quad counter, and an IBM PC. It provides real-time operator feedback for fast beam tuning and monitors momentary fluctuations in the particle beam. The hardware, software, and system performance are described.
Data for the 9450 I3L bipolar microprocessor and the 80C86 CMOS/epi (vintage 1985) microprocessor are presented, showing single-event soft errors for the full MIL-SPEC temperature range of -55 to 125 C. These data show for the first time that the soft-error cross sections continue to decrease with decreasing temperature at subzero temperatures. The temperature dependence of the two parts, however, is very different.
The accumulation of JPL data on Single Event Phenomena (SEP), from 1979 to August 1986, is presented in full report format. It is expected that every two years a supplement report will be issued for the follow-on period. This data for 135 devices expands on the abbreviated test data presented as part of Refs. (1) and (3) by including figures of Single Event Upset (SEU) cross sections as a function of beam Linear Energy Transfer (LET) when available. It also includes some of the data complied in the JPL computer in RADATA and the SPACERAD data bank. This volume encompasses bipolar and MOS (CMOS and MHNOS) device data as two broad categories for both upsets (bit-flips) and latchup. It also includes comments on less well known phenomena, such as transient upsets and permanent damage modes.
Report presents latest in series of investigations of "soft" bit errors known as single-event upsets (SEU). In this investigation, SEU response of low-power, Schottky-diode-clamped, transistor/transistor-logic (TTL) static random-access memory (RAM) observed during irradiation by Br and O ions in ranges of 100 to 240 and 20 to 100 MeV, respectively. Experimental data complete verification of computer model used to simulate SEU in this circuit.
Modeling of SEU has been done in a CMOS static RAM containing 1-micron-channel-length transistors fabricated from a p-well epilayer process using both circuit-simulation and numerical-simulation techniques. The modeling results have been experimentally verified with the aid of heavy-ion beams obtained from a three-stage tandem van de Graaff accelerator. Experimental evidence for a novel SEU mode in an ON n-channel device is presented.
The authors collected a set of heavy-ion single event upset (SEU) test data since their last publication in December, 1985. Trends in SEU susceptibility for state-of-the-art parts are presented.
A detailed experimental investigation of single-event upset (SEU) in static RAM (SRAM) chips fabricated using a family of high-performance NMOS (HMOS) depletion-mode-load process technologies, has been done. Empirical SEU models have been developed with the aid of heavy-ion data obtained with a three-stage tandem van de Graaff accelerator. The results of this work demonstrate a method by which SEU may be empirically modeled in NMOS integrated circuits.
Heavy secondary ions do not significantly alter device responses. Conclusion that external reaction products cause no significant alteration of single-event-upset response based on comparison of data obtained from both lidded and unlidded devices and for proton beams impinging at angles ranging from 0 degrees to 180 degrees with respect to chip face. Study also found single-event-upset cross section increases only modestly as proton energy increased to 590 MeV, characteristic of maximum energies expected in belts of trapped protons surrounding Earth and Jupiter.
Single-event upsets reduced by use of oversize transistors. Computers made less susceptible to ionizing radiation by replacing bipolar integrated circuits with properly designed, complementary metaloxide-semiconductor (CMOS) circuits. CMOS circuit chips made highly resistant to single-event upset (SEU), especially when certain feedback resistors are incorporated. Redesigned chips also consume less power than original chips.
Document summarizes 15 studies of single-event upsets covering 60 different types of semiconductor devices. Studies discussed in document include verification of basic reactions induced by heavy ions and protons and surveys of latchup and bit-flip susceptibility of several types of devices and device-fabrication technologies.
Single-Event Upset (SEU) response of a bipolar low-power Schottky-diode-clamped TTL static RAM has been observed using Br ions in the 100-240 MeV energy range and O ions in the 20-100 MeV range. These data complete the experimental verification of circuit-simulation SEU modeling for this device. The threshold for onset of SEU has been observed by the variation of energy, ion species and angle of incidence. The results obtained from the computer circuit-simulation modeling and experimental model verification demonstrate a viable methodology for modeling SEU in bipolar integrated circuits.
Tests showed that bipolar chips in the attitude control computer of the Galileo spacecraft would likely cause catastrophic mission failure due to single particle upset. This paper describes the design and testing of CMOS replacements which are speed compatible with the bipolar parts and are immune to upset by 165-MeV krypton ions.