ebook img

Recent Single Event Effects Results for Candidate Spacecraft Electronics for NASA PDF

2005·18.7 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Recent Single Event Effects Results for Candidate Spacecraft Electronics for NASA

To be presented by M. OBryan, at IEEE NSREC 2005, Seattle WA, July 15,2005. 1 t Recent Single Event Effects Results for Candidate Spacecraft Electronics for NASA Martha V. O'Bryan, Member, IEEE, Kenneth A. LaBel, Member, IEEE, Scott D. Kniffin, Member, IEEE, Christian Poivey, Member, IEEE, James W. Howard Jr., Member, IEEE, Ray L. Ladbury, Member, IEEE, Stephen P. Buchner, Member, IEEE, Timothy R. Oldham, Member, IEEE, Paul W. Marshall, Member, IEEE, Anthony B. Sanders, Member, IEEE, Hak S. Kim, Donald K. Hawkins, Martin A. Carts, James D. Forney, Tim Irwin, Christina M. Seidleck, Stephen R. Cox, Christopher Palor, David Petick, Wesley Powell, and Barry L. Willits Index Terms-Single Event Effects, spacecraft electronics, Abstracr-Vulnerability of a variety of candidate spacecraft digital, linear bipolar, and hybrid devices. electronics to proton and heavy ion induced single event effects is studied. Devices tested include digital, linear bipolar, and hybrid I. INTRODUCTION devices. As spacecraft designers use increasing numbers of commercial and emerging technology devices to meet stringent performance, as well as economic budgets and schedule requirements, ground-based testing of such devices for This work was supported in part by the NASA Electronic Parts and Packaging Program (NEPP), NASA Flight Projects, and the Defense Threat susceptibility to single event effects (SEE) has assumed ever Reduction Agency (DTRA). greater importance. Martha V. O'Bryan, Muiiiz Engineering Inc., c/o NASA Goddard Space The studies discussed here were undertaken to establish the Flight Center (GSFC), Code 561.4, Bldg. 22, Rm. 062A, Greenbelt, MD sensitivities of candidate spacecraft electronics to heavy ion 20771 (USA), phone: 301-286-1412, fax: 301-286-4699, email: [email protected] and proton-induced single event upsets (SEU), single event Kenneth A. LaBel, NASNGSFC, Code 561.4, Greenbelt, MD 20771 latchup (SEL), and single event transient (SET). Note: For (USA), phone: 301 -286-9936, email: [email protected] proton displacement damage (DD) and total ionizing dose Ray L. Ladbury, MuRiz Engineering Inc., c/o NASNGSFC, Code 561.4, Greenbelt, MD 20771 (USA), phone: 301-286-1030, email: (TID) results see a companion paper entitled "Recent Total [email protected] Ionizing Dose and Displacement Damage Results for James W. Howard Jr., Jackson & Tull Chartered Engineers, Washington, Candidate Spacecraft Electronics for NASA" by Cochran, et D. C. 20018, c/o NASAESFC, Code 561.4, Greenbelt, MD 20771 (USA), phone: 301 -286-1 023, email: [email protected] al. that is also being submitted to IEEE NSREC [ 11. Scott D. Kniffin, Christian Poivey, Christina M. Seidleck, Hak S. Kim, Martin A. Carts, James D. Fomey, and Tim Irwin, MuRiz Engineering Inc., 11. TESTT ECHNIQUAENSD SETUP c/o NASA/GSFC. Code 561.4, Greenbelt, MD 20771 (USA), phone S. Kniffin: 301-286-1 185, email: [email protected]; phone A. Test Facilities C. Poivey: 301-286-2128, email: [email protected]; phone C. Seidleck: 301 -286-1 009, email: [email protected]; phone All SEE tests were performed between February 2004 and H. Kim: 301 -286-5587, email: [email protected]; phone M. Carts: 301- February 2005. Heavy Ion experiments were conducted at the 286-2600, email: [email protected]; phone J. Fomey: 301-286- 9855, email: [email protected]; hone T. Irwin: 301-286-2014, Brookhaven National Laboratories' (BNL) Single Event Upset email: [email protected] Test Facility (SEUTF) [2], at Texas A&M University Stephen P. Buchner, and Timothy R. Oldham, QSS, c/o NASNGSFC, Cyclotron (TAMU) [3], and at the Single-Event Effects Test Code 561.4, Greenbelt, MD 20771 (USA), phone S. Buchner: 301-286-5019, email: [email protected]; phone T. Oldham: 301 -286-5489, Facility (SEETF) at the National Superconducting Cyclotron email: [email protected] Laboratory (NSCL) at Michigan State University (MSU) [4]. Donald K. Hawkins, Anthony B. Sanders, and Stephen R. Cox, The BNL SEUTF uses a twin Tandem Van De Graaf NASNGSFC, Code 561.4, Greenbelt, MD 20771 (USA), phone D. Hawkins: 301 -286-4789, email: [email protected]; phone A. Sanders: 301 - accelerator while the TAMU facility uses an 88'' Cyclotron. 286-1 151, email: [email protected]; phone S. Cox: 301 -286- The NSCL MSU facility uses tandem K500 and K1200 7721, email: [email protected] cyclotrons to deliver on target ions with energies up to 125 Paul W. Marshall, Consultant, 7655 Hat Creek Road, Brookneal, VA MeV/n. All three facilities are suitable for providing a variety 24528 (USA), phone: 434-376-3402, email: [email protected] Christopher Palor, formally of Orbital Sciences Corporation, c/o of ions over a range of energies for testing. At all facilities, test NASA/GSFC, Code 561.4, Greenbelt, MD 20771 (USA), phone: 301-286- boards containing the device under test (DUT) were mounted 1 239, email: [email protected] in the test area. For heavy ions, the DUT was irradiated with David Petrick and Wesley Powell, NASAIGSFC, Code 564, Microelectronics and Signal Processing Branch, Greenbelt, MD 20771 USA, ions with linear energy transfers (LETs) ranging from 0.59 to phone: 301 -286-9727, email: [email protected], and phone: 301- 120 MeV*cm'/mg. Fluxes ranged from 1x10' to 5x10' 286-2823, email: [email protected] particledcm' per second, depending on the device sensitivity. Barry L. Willits, Spectrum Astro Space Systems, Gilbert, AZ 85233 USA, phone: 480-892-8200, email: [email protected] Representative ions used are listed in Table I. LETs between To be presented by M. OBryan, at IEEE NSREC 2005, Seattle WA, July 15,2005. 2 the values listed were obtained by changing the angle of incidence of the ion beam on the DUT, thus changing the path length of the ion through the DUT and the "effective LET" of Laser: 590 nm, 1 ps pulse width, beam spot size -1.2 pm the ion [5]. Energies and LETS available varied slightly from B. Test Method one test date to another. Unless otherwise noted, all tests were performed at room Proton SEE tests were performed at the Indiana University temperature and with nominal power supply voltages. Cyclotron Facility (IUCF) [6]. Proton test energies incident on I) SEE Testing - Heavy Ion: the DUT are listed in Table 11. Proton SEE tests were Depending on the DUT and the test objectives, one or performed in a manner similar to heavy ion exposures. more of three SEE test methods were used: However, because protons cause SEE via indirect ionization of Dynamic - the DUT was exercised continually while being recoil particles, results are parameterized in tern of proton exposed to the beam. The errors were counted, generally by energy rather than LET. Because such proton-induced nuclear comparing DUT output to an unirradiated reference device or interactions are rare, proton tests also feature higher other expected output. In some cases, the effects of clock cumulative fluence and particle flux rates than do heavy-ion speed or device modes were investigated. Results of such tests experiments. should be applied with caution because device modes and Laser SEE tests were performed at the pulsed laser facility clock speed can affect SEE results. at the Naval Research Laboratory (NU) [7] [8]. The laser Static - the DUT was loaded prior to irradiation; data were light had a wavelength of 590 nm resulting in a skin depth retrieved and errors were counted after irradiation. (depth at which the light intensity decreased to l/e - or about Biased - the DUT was biased and clocked while ICC (power 37% - of its intensity at the surface) of 2 pm. A nominal pulse consumption) was monitored for SEL or other destructive rate of 100 Hz was utilized. effects. In some SEL tests, hctionality was also monitored. Ti LE I: HEAVYI ON TEST FACILITIEASN D TESTH EAVYI ONS In SEE experiments, DUTs were monitored for soft errors, I I Surface such as SEUs and for hard errors, such as SEL. Detailed 1 I LET in Si, Range in descriptions of the types of errors observed are noted in the MeV-cm*/mg Si,)Im individual test results. [9] Ion MeV (Normal 1 Incidence) I SET testing was performed using a high-speed oscilloscope. BNL ~ ~ 7I 9 3 05 I 36.9 I 38.7 Individual criteria for SETS are specific to the device being 1127 I 321-370 I 59.7-60.1 I 31-34.3 tested. Please see the individual test reports for details. [9] TAMU 7 Nez0 266-270 2.7-2.8 261 -267 Heavy ion SEE sensitivity experiments include f Ar40 496-497 8.7 - 9 174-17 5 measurement of the saturation cross sections and the Linear Energy Transfer threshold (LETh). The LET&i s defined as the t Cd3 944 17.8-21.5 172 maximum LET value at which no effect was observed at an t 1 I I I Kra4 912-916 25.4-29.3 116-170 effect fluence of 1x10' particles/cm2. In the case where events I I I I f Ag'09 1634 38.5-44.2 96.8-156 are observed at lower fluences for the smallest LET tested, I I I I f 1215-1934 47.3-55.3 94.2-156 LET,,, will either be reported as less than the lowest measured I I I LET or determined approximately as the LET* parameter fiom tAu19711 883-2247 53.9-85 100-155 * I I I I a Weibull fit. Nez0 515 1.7 799 2) Pulsed Laser Facility Testing *xeZ9 3197 37.9 286 The DUT was mounted on an X-Y-Z stage in fiont of a i Ol6 880 0.59 3607 1O Ox lens that produced a spot size of about 1.2 pm hll-width half-maximum (FWHM). The X-Y-Z stage could be moved in steps of 0.1 pm for accurate positioning of SEU sensitive regions in front of the focused beam An illuminator together with a charge coupled device (CCD) camera and monitor were MSU Kr7' 9438 6.8 4440 - used to image the area of interest, thereby facilitating accurate XeIz4 17360 14.1 3300 positioning of the device in the beam. The pulse energy was varied in a continuous manner using a polarizerhalf-waveplate I Particle combination and the energy was monitored by splitting off a Facility I Particle I Etnneerrggyy,, ((MMeeVV)) portion of the beam and directing it at a calibrated energy IInnddiiaannaa UUnniivveerrssiittyy CCyycclloottrroonn FFaacciilliittyy ((IIUUCCFF)) 1 PPrroottoonn I 5544--119977 meter. To be presented by M. OBryan, at IEEE NSREC 2005, Seattle WA, July 15,2005. 3 111. TESTR ESULTOS VERVIEW TABLEV I: ABBREVIATIONS AND CONVENTIONS: H = heavy ion test Abbreviations for principal investigators (PIS) are listed in P =proton test (SEE) TableIV, SEE test result categories are summarized in L = laser test Table V, abbreviations and conventions are listed in Table VI, LET = linear energy transfer (MeV*cm2/mg) LET&= linear energy transfer threshold (the maximum LET value at Device Category Abbreviations are listed in Table VII, SEE which no effect was observed at an effective fluence of 1x10’ results are summarized in Table VIII, and SE: results are particIes/cm*) featured in Table IX. Unless otherwise noted, all LETS are in SEE = single event effect SEU = single event upset MeV*cm2/mg and all cross sections are in cm2/device. This SEL = single event latchup paper is a summary of results. Complete test reports are SET = single event transient available online at http://radhome.gsfc.nasa.gov [9]. SEFl = single event functional interrupt SEB = single event burnout PrTiAnBcLipEaw l I:nLv IeSTsO tiFg PaRtoINrC (IPPAI)L INVESTI IGAAbTbOrReSv iation I SSEESG R= =- s siinnggllee e evveenntt sgnaatpeb raucpktu re BERT = bit error rate test < = SEE observed at lowest tested LET Scott Kniffin SK > = No SEE observed at highest tested LET Ken LaBel 0 = cross section (cm2/device,u nless specified as cm2/bit) OSAT = saturation cross section at LET,, (cm2/device,u nless specified as cm2/bit) Tim Oldham Christian Poive LDC = lot date code DUT = device under test P.I. =principal investigator Samp. = sample TABLEV : LIST OF CATEGORIES HI = Heavy Ion Following ground SEE irradiation, devices generally are P = Proton categorized into “useability” categories for spacecraft interest. ADC = analog to digital converter Recommendations for SEE are color coded according to the ALU =arithmetic logic unit ASIC = application specific integrated circuit following kev: BERT = bit error rate test or tester CCD = charge collection device CLB = configuration logic block CMOS = complementary metal oxide semiconductor DAC = digital to analog converter spaceflight applications, but may require mitigation FET = field effect transistor technioues FPGA = field programmable gate array Category 3: Recommended for usage in some NASNGSFC LApE = NASA Institute of Advanced Microelectronics spaceflight applications, but requires extensive MSB = most significant bits mitigation techniques or hard failure recovery mode NVM = non-volatile memory (may require latent damage screening) Op Amp = operational amplifier PROM = programmable read-only memory Category 4: Not recommended for usage in any NASNGSFC PWM =pulse width modulator spaceflight applications RAM = random access memory RW: Research Test Vehicle - Please contact the P.I. SRAM = static random access memory before utilizing this device for spaceflight SSPC = solid state power controller applications TABLEV II: DEVICEC ATEGORYA BBREVIATIONS: P = Processor Component TC = Test chip MX = Mixed Signal A = Analog M = Memory H = Hybrid L = Logic or 110 - 0 0 v cu cu f d z z co +I 7 z 0 m E m v) z-I e. I 8 8 I 5 0 0 v) E 6 I a 2 0" - e $50s z 2 N0 0d C -N C - e3 .0- Ws1 -LmC --OIr Y m3 8 Nm Q2 (NIn0 08 8 2 a 0 - - I I I I a I a a a a a - - - - - - f R N 0rNmIn3. INNNr0r3.. L3mN Lkz0? 5.2. XI In1 I (\I K I22- m0W02 I*0m0W0-I mI20(00n 1 cW _. m N m m N - 2 In In - r d N m m v2) 35-:1 5$ 8 z m -0 i I Y 1 I I I I - - - - XI -1 -1 5 I - - To be presented by M. OBryan, at IEEE NSREC 2005, Seattle W A, July 15,2005. 6 (b) embedded software running on the PowerPC processor, Iv. TESTR ESULTSA ND DISCUSSION and (c) user interface sohare running on a standalone PC. This PC communicates with the Memec board via a RS-232 As in our past workshop compendia of GSFC test results, interface. Using this application, MGT upset events are each DUT has a detailed test report available online at observed as bit errors or link failures, which are displayed on http://radhome.gsfc.nasa.gov [9] describing in further detail, the user interface log window. Processor errors are detected as test method, SEE conditions/parameters, test results, and software malfunctions, also indicated on the user interface log graphs of data. This section contains a summary of testing window (i.e., corrupted RS-232 communications, software performed on a selection of featured parts. hanging, etc.). I) Virtex II Pro XC2VP7 FPGA porn Xilinx: The Xilinx Virtex-I1 Pro is a SRAM-based platform FPGA that embeds multiple microprocessors within the fabric. The FPGA used was the commercial Xilinx Virtex-I1 Pro XC2VP7-6FG456C device. This device includes a single embedded PowerPC processor, 4.4 million configuration bits, 792 kl3 of BlockRAM, 8 RocketIOTM Multi-Gigabit Transceivers (MGT), 4 Digital Clock Managers (DCM), and 44 dedicated 18x18 multipliers [32]. The package used will be the wire-bond 456-pin ball grid array (FG456). The objective of this coarse Single Event Effect (SEE) test was to determine the suitability of the commercial Virtex-I1 Pro family for use in spaceflight applications. To this end, this test was primarily intended to determine any Single Event Latchup (SEL) susceptibilities for these devices. Secondly, this test was intended to measure the level of Single Event Upset (SEU) susceptibilities and in a general sense, where they occur. The coarse SEE test used a commercial-off-the-shelf (COTS) Virtex-I1 Pro evaluation board provided by Memec. This board is the Memec DS-KIT-2VP7FG456 (Fig. l), which contains one soldered FPGA along with external RAM, PROMS, RS-232 port, JTAG connectors, MGT drivers and connectors, oscillators, power converters, and various user switches. Also included with this board was a prototyping daughter card (DS-KIT-P 16O-PROTO), which was populated with RS-422 line drivers to generate discrete pulses that indicate detected upsets. The FPGA on this board is replaced Fig 1. Memec Development Board with a delidded device and partially covered with a shield. The BlockRAM, multipliers, configuration RAM, and CLBs During SEU testing, this shield was placed on the device to were tested with a dedicated test structure. This test structure, only expose certain portions of the logic, routing, which is depicted in Fig. 2, consists of dual sets of circuitry, configuration memory resources, MGTs, or PowerPC. with one set in the exposed area and the other in the shielded The FPGA circuitry that underwent SEE testing included area. Both are driven with a pseudo-random data generator. the following Virtex-I1 Pro functional elements: (a) the The outputs of these sets of circuitry are then compared. If PowerPC processor, (b) MGTs, (c) BlockRAM, (d) dedicated detected, the comparators generate error pulses indicating multipliers, (e) CLBs, and (0 configuration RAM. These upsets in the exposed circuitry. functional elements were tested using the combination of the This test structure can either be integrated with the BERT BERT reference application and standalone test structures. application, resulting in a single FPGA image, or programmed The Xilinx Bit Error Rate Test (XBERT) reference as a standalone function. During testing, this will enable the ' application, which was modified to accommodate this board flexibility to isolate different hnctionality of the FPGA, or run and test, tested the operation of the processor and the MGTs. all logic at once. By floorplanning the designs accordingly, For testing purposes, the MGT cables were hooked up in different parts of the deviceldesign can be shielded to suppress loopback (i.e. TX -> RX). This allows the transmitted pseudo- any unwanted SEES. random data to be compared at the receiver, and detect any bit Two onboard oscillators were used by the FPGA to derive errors. the internal clock fiequencies using DCMs. A 100 MHz The XBERT application consists of: (a) a FPGA image oscillator was connected to one DCM that used this as a containing processor peripherals and MGT support circuitry, reference to supply the PowerPC with a 200 MHz clock and To be presented by M. OBryan, at IEEE NSREC 2005, Seattle WA, July 15,2005. 7 - the FPGA fabric with a 50 MHz clock. The other DCM was Cyclical Current Ramping connected to a 125 MHz oscillator, which controlled the MGT 4 -- Vcclnt lcclnt reference clock that derived the data rate. The MGTs were set 3.5 - - p to run at 2.5 Gbps. I* 3 - i 2 2.5 - E a 2 - PRS Data Generator I t I I I c I I V I I.. : I -0.5 Logic Block A Logic Block B 0 100 200 300 Time (seconds) I I I Fig 3. Current Ramping Characteristic. I I I I I II II b) Virtex I1 Pro XC2 VP7 - SEFI Testing 1 I I I The purpose of this testing phase was to document any I I I I observed event that would be classified as a Single Event Ii I Error Comp Block I b RS-422 tOPC Functional Interrupt (SEFI) such as a failure in a MGT link or I I I I ’ the PowerPC log data getting corrupted, skipping instructions, I I I I or halting. The main objective was to focus on the PowerPC I I I I operation. For this case, the mask was used to approximately I I ‘-----------------------J cover the PPC. Although the exact location of the PowerPC Fig 2. SEU Comparison Logic. resources are unknown, the mask placement served as a rough estimate for initial testing purposes. a) Virtex I1 Pro XC2 VP7 - SEL Testing Fig. 4. shows the SEFI data collected. When testing with the The main goal of the coarse SEE test for the Virtex-I1 Pro PowerPC core exposed, the flux of the beam was turned do& was to determine if the device will enter a latch-up state under very low as the PPC was extremely sensitive. Therefore, radiation conditions. When testing at the cyclotron facilities at collecting statistically significant SEFI data was very difficult Texas A&M or Michigan State Universities, no destructive and time consuming. SEL event was observed to a LET of 53.9 MeV-cm’Img and a fluence of 10’ ions/cm2. During SEL testing, some interesting observations were c) Virtex II Pro XCZVP7 - Configuration Bit SEU Testing During the SEL testing, a configuration readback was made. While the device was irradiated the internal current performed after each run to determine the number of upset (IccINT) slowly rises at a constant rate, which is a function of configuration bits. This was simply performed by clicking the the radiation characteristics. The increased current is most ‘Verify’ command in the Xilinx ISE 6.li iMPACT tool. This likely the result of the configuration bits turning “on” causing internal bus contentions. However, once a current of - 3.4 A is command counts the number of differences found in the configuration data. Fig. 5 shows the results of this data. Note reached, the current drops to 0 A, jumps back up to nominal, - that there are 4.4 million configuration bits for this particular and then continues to ramp. It was determined that the device part. was being reprogrammed via on-board PROMS. To make an attempt to account for the MGT configuration As shown in Fig. 3, the current cycling is symmetric and the bits, a mask was placed on the FPGA to allow only four MGTs core voltage (VccmT)d oes not sag. An existing 40-pin remote- to be exposed. The cross section was reduced by sensing power cable was extended with custom 12-inch fly- leads to accommodate the Memec board. Although not shown approximately a factor of 10. This agrees with the fact that about 90% of the die was shielded from the radiation. on the plot, due to the modified power delivery and measurement setup, VCClNT actually sagged at higher currents. VcclNT was falling below the minimum voltage required to properly power the device due to a small IR drop across the fly-lead extension. This caused a power-on reset (POR) sequence to occur in the FPGA. To be presented by M. OBryan, at IEEE NSREC 2005, Seattle WA, July 15,2005. 8 For each run, the number of MGT bit errors was recorded. t I 1 1 I I This data was extracted from the PowerPC log file. The run was terminated upon MGT link failure. Fig. 6 shows the cross t No PPC (Average) With PPC (Raw) section data collected with a Weibull fit to that data. The threshold LET was found to be about 0.1 MeV*cmz/mga nd the saturation cross section was approximately 2.6 x 10” cm2. 1 10-4 I I I I I I t t - / \L -I -w 5 v c .g 10-5 t 0 a, cn v) 10-8 v) 9 0 10 20 30 40 50 60 0 Effective LET ( MeV.cm2/mg) 5 10-6 y Fig 4. SEFI Cross Section. I W I- After each test run, the configuration bit errors were counted (z3 and the action required to reestablish the fhctionality of the device was documented. During SEL testing, over 400,000 10-7 I I I 1 I I configuration bit errors (>lo%) were recorded twice and the 0 5 10 15 20 25 30 JTAG link failed twice. Both types of occurrences were Effective LET (MeVwm2/mg) probably due to configuration errors in the JTAG circuitry. Fig 6. MGT Bit-Error Cross Section. e) Viriex I1 Pro XC2 VP7 - Summary The commercial-grade Virtex-I1 Pro did not enter a latch-up state during these tests. However, a preliminary conclusion is I that the MGTs and embedded PowerPC have a high I susceptibility to the heavy ion radiation since SEFIs occurred Evv)) too quickly to collect substantial data. Due to the limitations on isolating the PowerPC with a shield configuration, new test 5 c1 - 0.01 : R methods need to be developed in order to gather more tl conclusive data on its operation. .- - m *_ L The Memec board was a good test bed for the coarse SEE .K0_ b test, however, in order to allow more flexibility for future tests, cs another board more suited for radiation testing is needed. [24] m3 Entire Die Exposed G.K=. ‘ c- Only 4 MGTs Exposed 2) 28 VDC Solid State Power Controller RP2 IO00 series from v0 DDC: 0.001 I I I I I 1 The RP21000 series are 28VDC Solid State Power 0 5 10 15 20 25 30 35 Controller (SSPC) rated from 2 through 25A. They are hybrid Effective LET (MeV*cm2/mg) Fig 5. Configuration Memory Bit-Error Cross Section. devices. Fig. 7 shows a picture of a de-lidded device. The SSPC uses five active integrated circuits that may be d) Viriex II Pro XC2 VP7 - MGT SEU Testing potentially sensitive to Single Event Effects (SEE). All the The goal of this test was to gather upset data on the MGTs. active components are located sufficiently far apart from one The data pattern used to drive the MGT transmit ports was a another that they cannot all be irradiated at the same time. Two pseudo-random pattern of 1+ X6+X7.T he MGTs were running different device areas were irradiated in order to check the at a date rate of 2.5 Gbps. The shield configuration used when SEE sensitivity of all active parts. Fig. 7 shows the two areas testing for MGT SEES masked off the DUT except for the area that were irradiated. The picture shows a 5A device containing one MGT, specifically MGT6 on the XC2VP7. (RP21005). All devices used the same control circuitry. The only difference is that they used a different number of power MOSFETs to draw the rated current. The 2A device (RP21002) only uses one transistor. The 5A devices use 2 transistors, as shown in the figure. The 10A device (RP21010) uses four transistors. To be presented by M. OBryan, at IEEE NSREC 2005, Seattle WA, July 15,2005. 9 - 34 output voltages were applied ranging from 22V to 40V, and -# different load conditions were applied ranging from 10% to - 29 - 100% of maximum load. In the off state, two different voltages, 40V and 50V, were applied to the device output. 2 24- 40V corresponds to the maximum recommended value, and & - 50V corresponds to the absolute maximum rating. 24 19- 8 The output of the DUT was monitored with an oscilloscope. 2 As soon as the DUT output exceeded a given trigger level set l4- below the nominal value, the resulting waveform, termed a 5+ 9- SET, was captured on the oscilloscope and subsequently stored on a PC. 4- No SEE was observed when the area 2 was irradiated up to turned off, the control input should go back to the low level - 10-2 3 and then to the high level to turn the switch on again. Fig 8 shows a typical transient at the device output. Fig 9 shows an .- 10.3 example where the switch was turned off and needs a control 5 input command to be turned back on. The device is sensitive to -0 $ SET down to lowest test LET of 2.8 MeV*cmz/mg.T he SET 10"; v cross section at the maximum tested LET is about IxlO-' c cm2/device. Fig 10 shows the SET cross section curve. No "00 10-5: :t destructive condition was observed up to the maximum tested 2 LET of 77 MeV*cm2/mg in the on mode. In the off mode, vg) 10-6: destructive conditions, output power MOSFET Single Event G Burnout (SEB), were observed down to a LET of 29.9 MeV*cm'/mg. [ 191 I0 -7 I I I I I I I I 0 10 20 30 40 50 60 70 80 90 Effective LET (MeV*cm*/mg) Fig. 10. SET cross section curve. To be presented by M. OBryan, at IEEE NSREC 2005, Seattle WA, July 15,2005. 10 Martha V. O'Bryan, Scott D. Knifin, Raymond L. Ladbury, James W. V. SUMMARY Howard, Jr., Kenneth A. LaBel, Cheryl J. Marshall, Robert A. Reed, Anthony B. Sanders, Christina M. Seidleck, Martin A. Carts, Donald K. We have presented recent data from SEE on a variety of mainly commercial Hawkins, Stephen R. Cox, Mark Walter, Christopher Palor, Moses devices. It is the authors' recommendation that this data be used with McCall, Steve Meyer, Dave Rapchun, Hak S. Kim, James D. Forney, caution. We also highly recommend that lot testing be performed on any Stephen P. Buchner, Timothy R. Oldham, John Sutton, Timothy L. suspect or commercial device. Irwin, Elionex Rodriguez, Dale McMorrow, Paul W. Marshall, John Link, John Rodgers, and Suhail Mohammed, "Current Single Event VI. ACKNOWLEDGMENT Effects Results for Candidate Spacecraft Electronics for NASA," 2004 IEEE Radiation Effects Data Workshop, NSREC04, pp. 10-18, July The Authors would like to acknowledge the sponsors of this 2004. effort: NASA Electronic Parts and Packaging Program Timothy Oldham, Hak Kim, and Suhail Mohammed, "Heavy Ion Testing of the Motorola Nano-crystal Nonvolatile Memory," (NEPP), NASA Flight Projects, and the Defense Threat http://radhome.gsfc.nasa.gov/radhome/papers/TO22204~Motorola.pdf, Reduction Agency (DTRA). Feb. 2004. Christian Poivey, Hak Kim, and Stephen Buchner, "Single Event REFERENCES Testing of the LS2805SlEM DC-DC converter from International Rectifier," http://radhome.gsfc.nasa.gov/radhome/papers/T082604~ Donna J. Cochran, et al., "Recent Total Ionizing Dose Results and LS2805S.pdf, Aug. 2004. Displacement Damage Results for Candidate Spacecraft Electronics for Christian Poivey, Jim Fomey, and Stephen Buchner, "Single Event NASA," submitted for publication in IEEE NSRECO5 Data Workshop, Testing of the 28VDC Solid-state Power Controller RP21005 from July 2005. DDC," http://radhome.gsfc.nasa.gov/radhome/papers/T1128 04- M. Wiplich "Brookhaven National Laboratories' (BNL) Tandem Van de RP21005.pdf, Dec. 2004. Graaf Accelerator Facility", http://tvdglO.phy.bnl.gov/index.html,J uly Christian Poivey and Stephen Buchner, "Single Event Testing of the 2003. 28VDC Solid-state Power Controller 53278 from Micropac," B. Hyman, "Texas A&M University Cyclotron Institute, K500 http://radhome.gsfc.nasa.gov/radhome/papers/T118 20 4-53278.pdf, Superconducting Cyclotron Facility," http://cyclotron.tamu.edu/ Dec. 2004. facilities.htm, July 2003 Stephen Buchner, Jim Fomey, and Christian Poivey, "SEE Testing of National Superconduction Cyclotron Laboratory (NSCL) at Michigan LTC1419 Analog to Digital Converter," http://radhome.gsfc.nasa.gov/ State University, http://www.nscl.msu.edu/, May 2004. radhome/papers/T060304-LTC 1419 .pdf, Jun. 2004. W.J. Stapor, "Single-Event Effects Qualification," IEEE NSREC95 Stephen Buchner and Jim Fomey, "SEE Testing of 7872 14-Bit ND Short Course, sec. 11, pp 1-68, July 1995. Converter (Maxwell)," http://radhome.gsfc.nasa.gov/radhome/papers/ C. C. Foster, S. L. Casey, P. Miesle, N. Sifri, A. H. Skees, K. M. BO12605 7872.pdf, Jan. 2005. Murray, "Opportunities for Single Event and Other Radiation Effects [23] Christian-Poivey, Christopher Palor, and Ray Ladbury "Heavy ion Testing and Research at the Indiana University Cyclotron Facility," Single Event Effects test of Octal, 8-Bit, Serial DAC with Output Buffer IEEE NSREC96 Data Workshop, pp. 84-87, July 1996. MAX529 from MAXIM," http://radhome.gsfc.nasa.gov/radhome/ J. S. Melinger, S. Buchner, D. McMorrow, T. R. Weatherford, A. B. papers/TO60304_MAX529,pdf,J un. 2004. Campbell, and H. Eisen, "Critical evaluation of the pulsed laser method 1241 David Petrick, Wesley Powell, and James Howard, "Virtex I1 Pro SEE for single event effects testing and fundamental studies,'' IEEE Trans. Test Methods and Results," Submission 226 (Session P) 2004 MAPLD Nucl. Sci., vol41, pp. 2574-2584, Dec 1994. lntemational Conference, Sept. 2004. D. McMorrow, J. S. Melinger, and S. Buchner, "Application of a Pulsed 1251 Scott Kniffin, Jim Howard, and Chris Palor, "Single Event Transient Laser for Evaluation and Optimization of SEU-Hard Designs," IEEE and Destructive Single Event Effects Testing of the National Trans. Nucl. Sci., vol47, no. 3, pp. 559-565, June 2000. Semiconductor 54ACTQ16245 16-Bit Transceiver With 3-State NASAIGSFC Radiation Effects and Analysis home page, Outputs," http://radhome.gsfc.nasa.gov/radhome/papers/TO60504~ http://radhome.gsfc.nasa.gov 54ACTQ16245.pdf. June 2004. A. [lo] J. W. Howard-Jr., K. LaBel, M. A. Carts, C. Seidleck, and J. [26] Anthony Sanders, Ken LaBel, and Christian Poivey, "Synopsis V1.0: Gambles, "Validation and Testing of Design Hardening for Single Event Heavy Ion Latch-up Test Results for the Altera Stratix FPGA EPIS25," Effects Using the 8051 Microcontroller as a Test Vehicle," submitted http://radhome.gsfc.nasa.gov/radhome/papers/T122004-Altera-ESP12 5 for publication to 2005 IEEE Trans. on Nucl. Sci., Dec 2005. .pdf, Dec. 2004. [I I] Scott Kniffm and Jim Foumey, "Single Event Effects Test Report for 1271 Anthony Sanders, Ken LaBel, and Christian Poivey, "Synopsis V2.0: Candidate CINDI Devices Test Trip to TAMU, FEB Heavy Ion Latch-up Test Results for the Altera Stratix FPGA EPIS25," 2004,"http://radhome.gsfc.nasa.gov/radhome/papers/T022404~IFN423_ http://radhome.gsfc.nasa.gov/radhome/papers/T1O430 5-Al tera-EP 1S 25 OPAl28.pdf, Feb. 2004. .pdf, Mar. 2005. 1121 Timothy R. Oldham, Hak Kim, and Suhail Mohammed. "Heavy Ion [28] Stephen Buchner, "Single Event Transients UT54LVDS03 1/32," Testing of the Motorola Nano-crystal Nonvolatile Memory," http://radhome.gsfc.nasa.gov/radhome/papers1/ 1B2 304-UT54LVDS03 http://radhome.gsfc. nasa.gov/radhome/papers/TO60304~freescale- 1-32.pdf, Nov. 2004. nvm.pdf, June 2004. 1291 Stephen Buchner and Chris Palor, "SEE Testing of 28C010T EEPROM 1131 Timothy R. Oldham, Hak Kim, Suhail Mohammed, and Kenneth A. (Maxwell)," ht tp://radhome.gsfc.nasa.gov/radhome/papers/ LaBel, "Heavy lon Testing of the Motorola Nano-crystal Nonvolatile B012605-28COIO.pdf, Jan. 2005. Memory," http://radhome.gsfc.nasa.gov/radhome/papers/1T2 1804- 1301 Stephen Buchner, Chris Palor and Jim Fomey, "SEE Testing of freescale-nvm.pdf, Dec. 2004. HN58C1001 TI 5 EEPROM (Renesas)," http://radhome.gsfc.nasa.gov/ 1141 T. R. Oldham, M. Suhail, P. Kuhn, E. Prinz, H. Kim, and Kenneth A. radhome/papers/BOl2605-HN58C1001. pdf, Jan. 2005. LaBel, "Effects of Heavy Ion Exposure on Nanocrystal Nonvolatile [31] Christian Poivey, Hak Kim, and Ray Ladbury, "Single Event Latchup Memory," submitted for publication to 2005 IEEE Trans. on Nucl. Sci.. testing of the 512Mbit SDRAM from Maxwell," Dec 2005. http://radhome.gsfc.nasa.goviradhome/papers/T112 8 04-SDRAM- [ 151 Timothy R. Oldham, Mohammed Suhail, Peter Kuhn, Erwin Prinz, Hak TestChip.pdf, Dec. 2004. Kim, and Kenneth A. LaBel "Effects of Heavy Ion Exposurc on [32] "Virtex-11 Pro Platform FPGA Handbook", Xilinx, Oct. 2002. Nanocrystal Nonvolatile Memory," Proceedings, 2004 Non-Volatile Memory Technology Symposium, Orlando, FL, pp. 39-42, Nov. 2004.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.