Slide 1: RTG4 FPGA Update: Qualification, Radiation, Sub-QML
RTG4™ FPGA Update: Qualification, Radiation, Sub-QML Minh Nguyen Space Forum 2019
Slide 2: Agenda
Agenda RTG4™ FPGA Update Qualification Radiation Sub-QML FPGAs Space Forum 2019 2
Slide 3: RTG4 FPGA Update
RTG4 FPGA Update Space Forum 2019 3
Slide 4: RTG4 Product Overview
RTG4 Product Overview Resources RT4G150 CG1657 / LG1657 CQ352 Logic Elements (TMR Register + 4-Input C Logic) 151,824 151,824 18x18 Multiply-Accumulate Blocks 462 462 RAM Mbits (1.5 Kbit and 24 Kbit Blocks, with ECC) 5.2 5.2 UPROM Kbits 381 381 DDR2/3 SDRAM Controller (with ECC)hjhkjhkj 2 x 32 0 PCI
Slide 5: Low-Power Grade
RTG4 Low-Power Grade Low-power grade (–L) for RTG4 9 standard speed (-STD) available 8 NOW 7 25% quiescent supply current reduction: Static Power (W) 6 Competing RT FPGA from 4.1 A to 3.1 A at 125 °C 5 Competing RT FPGA -L RT4G150L device setting available 4 (Low Power Grade) RTG4 in Libero So
Slide 6: Qualification Status
RTG4 Qualification Status CG/LG 1657 QML Class Q and Class V qualification completed! SMD 5962-16208 SMD has been approved and is posted on the DLA web site SMD numbers are also posted on the Microsemi web site CQ352 QML Class Q and QML Class V in progress QML-Q qualification is expected
Slide 7: Availability Schedule
RTG4 Availability Schedule RT4G150 PROTO FPGAs: Available to lead time now RT4G150 development kit: Available to lead time now CG1657 daisy chain packages: Available to lead time now CG1657 B, E, V-flow flight units, and CQ352 B, E-flow flight units: Available to lead time now CQ352 EV-flo
Slide 8: Radiation Summary
RTG4 Radiation Summary Total Ionizing Dose Stays within parametric limits > 125 Krad (Si) Single Event Latch-Up No failure at facility limit of 103 MeV-cm2/mg, 100 °C Configuration Upset No failure at facility limit of 103 MeV-cm2/mg, 100 °C Flip-Flop Single Event Upset 2.6E-12 errors/bit-day, GEO s
Slide 9: Radiation Update
RTG4 Radiation Update Test Environment Test Schedule Status Fabric, SRAM and PLL SEE Proton Complete Report available on Microsemi Web SERDES SEE Proton UCD in 10/2018 Less link loss than previous HI testing Report in progress SERDES SEE Heavy Ion LBNL in 10/2018 Report in progress In-Beam Programmi
Slide 10: PLL Results
PLL Results Heavy Ion Testing, GEO solar min PLL Mode Clock Source Recovery Error rate 1 Loss of Lock (upset/PLL/day) every Single PLL External Self 8.28 E-05 33 years Single PLL External PLL Reset 4.00 E-07 6800 years TMR PLL External PLL Reset 1.88 E-05 145 years Proton Testing No upsets det
Slide 11: Power-On Reset
RTG4 Power-On Reset Power-on reset (POR) test in heavy ion beam Average flux = 46 ions/cm2/s LET = 37 MeV.cm2/mg 10 out of 10 attempts were successful Space Forum 2019 11
Slide 12: In-Beam Programming Radiation Test Campaigns
RTG4 In-Beam Programming Radiation Test Campaigns 4 test campaigns were performed to test RTG4 in-beam programming 1. Heavy ion test done at LBNL (2016) with average flux ~1x103 ions/cm2/s 2. Neutron test at LANL (August 2018) 3. Heavy ion test at TAMU (September 2018) with average flux ~35 ions/c
Slide 13: 1st Test Campaign: Heavy Ion at LBNL, 2016
1st Test Campaign: Heavy Ion at LBNL, 2016 Single event effects on reprogramming Reprogramming in-beam performed to maximum LET 30 MeV-cm2/mg with average flux of 1x103 ions/cm2/s Programming failures detected Programming successful after parts were removed from heavy ion beam Programming
Slide 14: Programming SEFI (LBNL, 2016)
Programming SEFI (LBNL, 2016) Weibull Fitting LETth W S σsat 0.1 500 1.2 1.2×10-3 Space Forum 2019 Microchip Confidential ©2019 14
Slide 15: 2nd Test Campaign: Neutron at LANSCE (August 2018)
2nd Test Campaign: Neutron at LANSCE (August 2018) Los Alamos National Laboratory LANSCE - WNR Neutron energies 10 MeV – 800 MeV Programming passed 10 times out of 10 attempts in neutron beam Attempt # 1 2 3 4 5 6 7 8 9 10 Program Pass Pass Pass Pass Pass Pass Pass Pass Pass Pass Verify Pass P
Slide 16: 3rd Test Campaign: Heavy ion at TAMU (Sept 2018)
3rd Test Campaign: Heavy ion at TAMU (Sept 2018) Beam parameters at TAMU: Average flux = 35 ions/cm2/s LET = 37 MeV-cm2/mg In beam 2 out of 10 Programming attempts passed; 2 out of 2 Verify failed If more Verify attempts are tested, it is most likely that Verify will pass Attempt # 1 2 3 4
Slide 17: Comparison Between TAMU & GEO Flux
Comparison Between TAMU & GEO Flux Goal: “real” Programming failure rate GEO Solar Min, 100 Mil Al Beam parameter @ TAMU: • LET = 37 MeV.cm2/mg • Average flux = 35 ions/cm2/s GEO parameters (CREME96), 100 Mil Al shielding: • LET = 37 MeV.cm2/mg • Flux Solar Min ~ 6.41E-10 ions/cm2/s • Flux Sol
Slide 18: 4th Test Campaign: Proton at UCD (Oct 2018)
4th Test Campaign: Proton at UCD (Oct 2018) In beam, 10 out of 10 programming attempts passed; 10 out of 10 verify passed UC Davis, 64MeV proton beam Attempt # 1 2 3 4 5 6 7 8 9 10 Program Pass Pass Pass Pass Pass Pass Pass Pass Pass Pass Verify Pass Pass Pass Pass Pass Pass Pass Pass Pass Pass
Slide 19: In-Flight Programming Guidance
RTG4 In-Flight Programming Guidance Current guidance Highly unlikely that a destructive event will occur during programming in space Probability of first-time success for programming in GEO is estimated > 99%, based on 1st test campaign It is highly likely that in space, no ion will disrupt
Slide 20: Programs Baselining RTG4
Programs Baselining RTG4 Mission Extension Orion WFIRST SBC AIDA Vehicle MEV-1 (NASA) (NASA) (LANL) (ESA, DLR, NASA) Launched Oct 9th, 2019 SLIM JPSS-2 Reconfigurable ALOS-3 Optical ALOS-4 SAR (JAXA) (NOAA) Processor (JAXA) (JAXA) (Commercial) Space Forum 2019 20
Slide 21: Sub-QML FPGAs
SUB-QML FPGAS Space Forum 2019 21
Slide 22: RT FPGAs for New Space
RT FPGAs for New Space Objective Create versions of RT FPGAs that can meet aggressive price targets to support New Space constellations that would not use traditional B / E / V flow FPGAs Requirements Reduction in cost and price Faster lead-time / cycle-time Sub-QML RT FPGAs Ceramic –
Slide 23: Cost and COTS
Cost and COTS Satellite operators seeking lower acquisition cost and faster service entry Commercial Off The Shelf (COTS) components to reduce cost and lead-time The cost of COTS – lower component cost, higher cost of ownership General Industry QML COTS Rad Tolerant Unit Cost Low High Lead
Slide 24: Sub-QML: Bridging the Gap Between QML and COTS
Sub-QML: Bridging the Gap Between QML and COTS Qualification QML Class V Rad Characterization Rad Hardened Traceability QML Class Q By Design Lot Homogeneity Rad Hardened By Design Sub-QML Hermetic Rad Hardened By Design Sub-QML Plastic • Flight Heritage / Baselined Rad Hardened • Radiation Support
Slide 25: Sub-QML Components
Sub-QML Components Reducing or eliminating QML testing and documents removes a lot of cost Elimination of solder columns removes cost and reduces lead times Plastic packaging reduces cost further General Industry QML Sub-QML Sub-QML COTS Rad Tolerant RT Hermetic RT Plastic Unit Cost Lowest
Slide 26: Screening Flows
RTG4 Screening Flows Screening Flow Purpose Package Qualification Burn-In Temp Test Life Test DPA Static -55°C – V NSS, NASA Class1 Hermetic Ceramic QML-V Wafer-Lot Assy Lot Dynamic 125°C Advanced Static -55°C – Generic E Hermetic Ceramic QML-Q Optional Traditional Space Dynamic 125°C Group C Entry
Slide 27: Screening Flows for Traditional Space
RTG4 Screening Flows for Traditional Space Screening Flow Purpose Package Qualification Burn-In Temp Test Life Test DPA Static -55°C – V NSS, NASA Class1 Hermetic Ceramic QML-V Wafer-Lot Assy Lot Dynamic 125°C Advanced Static -55°C – Generic E Hermetic Ceramic QML-Q Optional Traditional Space Dynami
Slide 28: Screening Flows: Hermetic Ceramic for New Space
RTG4 Screening Flows: Hermetic Ceramic for New Space Screening Flow Purpose Package Qualification Burn-In Temp Test Life Test DPA Static -55°C – V NSS, NASA Class1 Hermetic Ceramic QML-V Wafer-Lot Assy Lot Dynamic 125°C Advanced Static -55°C – Generic E Hermetic Ceramic QML-Q Optional Traditional Sp
Slide 29: Reduced Flow and Mil Temp Hermetic
RTG4 Reduced Flow and Mil Temp Hermetic Reduced Flow Same test and screening flow as B-flow (no test elimination) Internal gas analysis screened to 10,000 ppm(v) moisture limit Minimum order quantity (MOQ) and non-cancelable, non-returnable (NCNR) will apply No datapack, no assembly lot gr
Slide 30: Screening Flows – Plastic Packages for New Space
RTG4 Screening Flows – Plastic Packages for New Space Screening Flow Purpose Package Qualification Burn-In Temp Test Life Test DPA Static -55°C – V NSS, NASA Class1 Hermetic Ceramic QML-V Wafer-Lot Assy Lot Dynamic 125°C Advanced Static -55°C – Generic E Traditional Space • Hermetic Plastic Ceramic
Slide 31: Plastic Package
RTG4 Plastic Package Military temperature screening flow Test all parts at hot, cold and room Guarantee TJ -55°C to 125°C No MIL-STD-883 screening – no burn-in, temp cycle, centrifuge, fine and gross leak, etc. Silicon Thermal Interface Die Material Decoupling FCG1657 Package Capacitors
Slide 32: Conclusion
Conclusion RTG4 achieved QML Class V qualification RTG4 achieved first flight heritage on MEV-1 RTG4 FPGA family continues to grow with low-power version, qualified CQ352 package, and new plastic package New RTG4 radiation data available on Microsemi website Sub-QML FPGAs bridging the gap