Transcript p1008_sanders_s

Radiation-Hardened
re-programmable FieldProgrammable Gate Array
(RHrFPGA)
A.B. Sanders1, K.A. LaBel1, J.F. McCabe1,
G.A. Gardner2, J. Lintz2, C. Ross2, K. Golke2,
B. Burns2, M.A. Carts3, and H.S. Kim4
1. NASA/GSFC, Code 561.4 Greenbelt, MD 20771
2. Honeywell, Defense and Space Electronics Systems,
Clearwater, FL 33764-7290
3. Raytheon/ITSS, Lanham, MD 20706
4. Jackson and Tull, Chartered Engineers, Seabrook, MD
20706
Sanders
1
2004 MAPLD, Paper 1008
OUTLINE
•
•
•
•
•
•
•
•
Sanders
Introduction
Radiation Test Suite
Test Configuration
Program Test Methods
Test Procedure
Test Results
Summary
Acknowledgements
2
2004 MAPLD, Paper 1008
INTRODUCTION
•
•
•
•
•
Radiation-Hardened Reprogrammable FieldProgrammable Gate Array
(RHrFPGA) by Honeywell
Configurable Logic Blocks
provide functional elements for
constructing user’s logic
I/O Cells provide the interface
between the package pins and
internal signal lines
Programmable Interconnect;
Resources provide routing paths
to connect the inputs and
outputs onto the appropriate
networks
Customized configuration is
established by programming
internal static memory cells that
determine the logic functions
and internal connections
implemented in the FPGA
INTERNAL FPGA
Sanders
3
2004 MAPLD, Paper 1008
DEVICE CHARACTERISTICS
• Characteristics:
Silicon Epi thickness is 0.2 m
Gate length is 0.35 m
Material thickness over the Silicon Epi equivalent to 8.3 m(Si)
Physical cross-sections of the memory and flip-flop cells are
0.45 m2 and 2.1 m2, respectively
– SOI technologies offer greater speed and power reduction
compared to conventional bulk CMOS
–
–
–
–
• Motivation: Why Develop the RHrFPGA?
– Reconfigurable FPGAs offer a significant advantage over onetime programmable Antifuse FPGAs
– Cache Logic design allows part of the FPGA to be
reprogrammed without loss of register data, while the
remainder continues to operate without disruption
– Rad hard need for space and military applications without
complex external mitigation circuits (single chip solution)
Sanders
4
2004 MAPLD, Paper 1008
BACKGROUND
• BACKGROUND:
Development of the
RHrFPGA
– In July 1998, Honeywell and
Atmel announced the signing
of a license for Honeywell to
develop a radiation hardened
version of Atmel’s 30,000
gate, 6400 register Field
Programmable Gate Array
(FPGA), the AT6010
– Honeywell developed a
CMOS Silicon-On-Insulator
(SOI) version of the AT6010 to
meet the radiation hardness
levels required for
commercial and military
space and missile systems
– The radiation hardened FPGA
development was funded and
managed by NASA Goddard
Space Flight Center
Sanders
5
2004 MAPLD, Paper 1008
RADIATION TEST SUITE
Interface
Cable
PC
RTU Test
Board
• ‘Suite’ includes: PC, Cables, Custom Remote Terminal Unit
(RTU) Interface Dongle (in Cable), and RTU Test Board
(Sockets for RHrFPGA DUT, foreground)
Sanders
6
2004 MAPLD, Paper 1008
RTU Test Board Layout
Clock Input
‘Control’ RHrFPGA
Socket (exercises DUT)
RHrFPGA
‘Device Under Test’ Socket
(under Beam)
Program &
Command
Input
Error
Output
Power Input
Parallel & Serial
EEPROM Sockets
for (opt.) On-card Config.
Sanders
Configuration
Select Switches
7
2004 MAPLD, Paper 1008
ION BEAM CHARACTERISTICS
Ion
Energy
Angle
Range (m)
Effective LET (MeV/(mg/cm2))
Xe
1955
0
93
53.2
60
42
109
0
92
85.9
60
41
174
Au
2955
RHrFPGA Heavy Ion Testing at Room Temperature at TAMU
• Orientation: Test fixture was oriented so angular rotation was
parallel to the gate width of the test devices’ transistors
• Angular rotation was limited to 60 degrees due to fixture
shadowing of the die at higher angles of incidence
Sanders
8
2004 MAPLD, Paper 1008
RHrFPGA Heavy Ion SEU Test Configuration at TAMU
Sanders
9
2004 MAPLD, Paper 1008
RHrFPGA TEST PROGRAMS (Vectors)
Test Program Name
Flip-Flops
Tested
Configuration Bits
Tested
Application with I/O (Demodulator)
1506
131152
Load/Verify -Application
100
131152
Load/Verify – Boot Zeros
100
131152
Full Shift Register Vertical – Dynamic
1450
131152
Shift Register with Logic – Dynamic
297
131152
Shift Register with Xbus – Dynamic
184
131152
Shift Register with Lbus
670
131152
Shift Register w/Other – Dynamic
975
131152
RHrFPGA Heavy Ion SEU Test Programs at TAMU
Sanders
10
2004 MAPLD, Paper 1008
TEST PROCEDURE
• Establish the correct test conditions
• Load the RHrFPGA controller and test device with
the proper configurations and verify test set
functionality
• Irradiate the test device to the desired effective
fluence while monitoring the device for SEU and
monitoring for proper health
• Read the controller status registers to determine
the number of upsets or test anomalies
• Read the test device configuration to check for
configuration SRAM upsets
• Record all relevant test data from exposure run
Sanders
11
2004 MAPLD, Paper 1008
HEAVY ION TESTING RESULTS
• The test evaluated the RHrFPGA using eight different test
programs and configurations
– Seven were optimized for SEU testing to evaluate specific
internal memory elements within the device and one test
program represented a current RHrFPGA application
• Nominal supply voltage is 3.3V and the devices were tested
at the worst case voltage of 3V
• Tests well below nominal (1.8V and 2.1V) were utilized to
validate error detection
• The RHrFPGA test devices did not experience SEU or other
SEE to the maximum available test LET of 174 MeVcm2/mg
– Performed at minimum rated supply voltage of 3.0V
– Applied all eight tests for fluences of  1.0x107 ions/cm2 per
test
Sanders
12
2004 MAPLD, Paper 1008
PROTON TEST MOTIVATION AT IUCF
• Memory Elements: The RHrFPGA memory elements use
SEU hardening techniques similar, but harder than those of
Honeywell’s HX6408 SRAM
– The HX6408 SRAM proton sensitivity was shown to be
dependent on the proton angle of incidence on the die
– Sensitivity attributed a single secondary heavy ion hitting two
transistors within a memory cell
– The SEU cross-section is highest for a proton angle of
incidence parallel to the path between the two sensitive
transistors in a cell
– A grazing parallel proton beam oriented normal to this path is
thus expected to produce an SEU cross-section similar to a
normal incidence beam.
• The test irradiated the RHrFPGA with proton beam nearly parallel
to the die surface
• The die was oriented so that the beam was parallel to the sensitive
path directions
Sanders
13
2004 MAPLD, Paper 1008
POSITIONING ASSEMBLY AT IUCF
•
•
•
The rotation angle was
limited to 70 degrees due
to alignment constraints
and concerns of
irradiating the control
device
All exposures were
performed at a 70 degree
angle of incidence
Different axes of rotation
for the two cell types were
used because the
configuration RAM and
application flip-flops are
orthogonal to each other
Positioning Assembly at IUCF
Sanders
14
2004 MAPLD, Paper 1008
RAM CELL TEST
Test Board in horizontal position with rotation about the vertical axis
Sanders
15
2004 MAPLD, Paper 1008
FLIP-FLOP TEST
Test Board in vertical position with rotation about the vertical axis
Sanders
16
2004 MAPLD, Paper 1008
RHrFPGA Proton SEU Test Configuration at IUCF
Sanders
17
2004 MAPLD, Paper 1008
RHrFPGA TEST PROGRAMS (Vectors)
Test Program Name
Flip-Flops
Tested
Configuration RAM Bits
Tested
Application with I/O (Demodulator)
1506
131152
Full Shift Register Vertical
1450
131152
Shift Register with Lbus
670
131152
Shift Register with Logic
297
131152
Shift Register with Xbus
184
131152
5760
5790
131152
RHrFPGA Proton SEU Test Programs at IUCF
Sanders
18
2004 MAPLD, Paper 1008
PROTON TESTING RESULTS
• The test evaluated the RHrFPGA using six
different test programs and configurations
• The test was optimized to evaluate the
RHrFPGA’s two unique types of memory
elements.
• The RHrFPGA test devices were irradiated to a
proton fluence of 3.4x1013 p/cm2 with 203 MeV
protons
• Test parts did not exhibit SEU or any other SEE,
demonstrating that the RHrFPGA is essentially
immune to proton-induced SEU
Sanders
19
2004 MAPLD, Paper 1008
SUMMARY
 Heavy Ion Testing
 Two RHrFPGA devices did not upset to the maximum
available test LET of 174 MeVcm2/mg and an ion
fluence of  1.0x107 ions/cm2
 The test consisted of seventeen exposure runs at the
minimum specified operating voltage of 3.0V
 The test results were consistent with analytical
predictions indicating a much higher minimum SEU
LET threshold than could be obtained from a heavy ion
SEU test
 Proton Testing
 Three RHrFPGA devices did not experience SEU or
other SEE to a proton fluence of 3.4x1013 p/cm2 per test
device
 Test results were also consistent with analytical
predictions indicating that the device is not sensitive
to proton induced SEE
Sanders
20
2004 MAPLD, Paper 1008
ACKNOWLEDGEMENTS
The Authors would like to acknowledge the
sponsors of this effort: NASA Electronic Parts and
Packaging Program (NEPP), NASA Flight Projects,
and the Solar Dynamic Observatory (Project).
The Authors wish to acknowledge the following
individuals for their contribution to this publication:
Kenneth A. LaBel, Martha V. O’Bryan, Gary Gardner,
and John Lintz.
Sanders
21
2004 MAPLD, Paper 1008