Reliability

Motivation - 35 Years Ago

Section 2: RELIABILITY AND QUALITY ASSURANCE REQUIREMENTS 2.2 R&QA REQUIREMENTS FOR PHASED HARDWARE DEVELOPMENT 2.2.2 STUDY/DEFINITION PHASE REQUIREMENTS b. Development of preliminary mathematical model and reliability predictions. (NPC 250-1) c. Establishment of reliability and safety goals and other R&QA requirements in preliminary specifications. (NMI 5320.1, NMI 5330.1, NPC 500-1). 2.2.3 DESIGN PHASE REQUIREMENTS e. Development of mathematical models and reliability predictions. (NPC 250-1, RA006-007-1) g. Apportionment of reliability goals to equipments and components. (NPC 250-1)
Office of Manned Space Flight - Apollo Program. NHB 5300.1A, July 1966 Apollo Reliability and Quality Assurance Program Plan

Reliability - 35 Years Ago

LEVEL RELIABILITY ANALYSIS AND MODELING ACTIVITY HARDWARE I........................................... Apollo Mission Reliability Estimates Missions Apollo Program Office - R&QA Model Integration Center Estimates II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SC LV LC GOSS .................. Launch Vehicle/ Spacecraft/Ground Support Systems/ Stage/Module/ Subsystem/ Black Box Mission/Launch Vehicle/Spacecraft/ Ground Support Systems/Stage/ ............................. Module/Subsystem

Apollo Program Office - R&QA Review Stage/Module/ Subsystem/ Black Box/ Component

III . . . . . . . . . . . . . . . . . . . . . . .

Contractor Reliability Estimates Center Review

...........

IV . . . . . . . . . . . . . . .

Subcontractor and Design Group Estimates Contractor Review

...

Subsystem/ Black Box/ Component/Part

Office of Manned Space Flight - Apollo Program. NHB 5300.1A, July 1966 Apollo Reliability and Quality Assurance Program Plan

Reliability

Reliability
Introduction to Reliability Historical Perspective Current Devices Trends

The Bathtub Curve

Failure rate,

Infant Mortality

Useful life Constant

Wear out

Time

The Bathtub Curve (2)

What is the "bathtub" curve? In the 1950s, a group known as AGREE (Advisory Group for the Reliability of Electronic Equipment) discovered that the failure rate of electronic equipment had a pattern similar to the death rate of people in a closed system. Specifically, they noted that the failure rate of electronic components and systems follow the classical bathtub curve. This curve has three distinctive phases: 1. An infant mortality early life phase characterized by a decreasing failure rate (Phase 1). Failure occurrence during this period is not random in time but rather the result of substandard components with gross defects and the lack of adequate controls in the manufacturing process. Parts fail at a high but decreasing rate. 2. A useful life period where electronics have a relatively constant failure rate caused by randomly occurring defects and stresses (Phase 2). This corresponds to a normal wear and tear period where failures are caused by unexpected and sudden over stress conditions. Most reliability analyses pertaining to electronic systems are concerned with lowering the failure frequency (i.e., const shown in the Figure) during this period. 3. A wear out period where the failure rate increases due to critical parts wearing out (Phase 3). As they wear out, it takes less stress to cause failure and the overall system failure rate increases, accordingly failures do not occur randomly in time.

Introduction to Reliability
Failure in time (FIT)
Failures per 109 hours ( ~ 104 hours/year )

Acceleration Factors
Temperature Voltage

Introduction to Reliability (cont'd)

Most failure mechanisms can be modeled using the Arrhenius equation.
ttf = C e ttf C EA k T EA/kT

time to failure (hours) constant (hours) activation energy (eV) Boltzman's constant (8.616 x 10-5eV/K) temperature (K)

Introduction to Reliability (cont'd) Acceleration Factors

A.F. = ttfL -----ttfH

A.F. ttfL ttfH

= acceleration factor = time to failure, system junction temp (hours) = time to failure, test junction temp (hours)

Introduction to Reliability (cont'd) Activation Energies

Failure Mechanism EA(eV) Oxide/dielectric defects 0.3 Chemical, galvanic, or electrolytic corrosion 0.3 Silicon defects 0.3 Electromigration 0.5 to 0.7 Unknown 0.7 Broken bonds 0.7 Lifted die 0.7 Surface related contamination induced shifts 1.0 Lifted bonds (Au-A1 interface) 1.0 Charge injection 1.3 Note: Different sources have different values these values just given for examples.

Acceleration Factor - Voltage Oxides and Dielectrics

Large acceleration factors from increase in electric field strength A.F. = 10 / (MV / cm)
= 0.4 e k T
0.07/kT

- Boltzman's constant (8.616 x 10-5eV/K) - temperature (K)

Acceleration Factor: Voltage

Median-time-to-fail of unprogrammed antifuse vs. 1/V for different failure criteria with positive stress voltage on top electrode and Ta = 25 C.

Integrated Circuit Reliability

Historical Perspective
Application Apollo Guidance Computer Commercial (1971) Military (1971) High Reliability (1971) SSI/MSI/PROM 38510 (1976) MSI/LSI CICD Hi-Rel (1987) Reliability < 10 500 2,000 10,000 44-344 43 FITs Hours Hours Hours FITs FITs

Device and Computer Reliability 1960's Hi-Rel Application

Apollo Guidance Computer
Failure rate of IC gates: < 0.001% / 1,000 hours ( < 10 FITS ) Field Mean-Time-To-Failure ~ 13,000 hours

One gate type used with large effort on screening, failure analysis, and implementation.

Device Reliability:1971
Reliability Level of Parts and Practices Commercial Military High Reliability Representative MTBF (hr) 500 2,000 10,000

(104 hours)

MIL-M-38510 Devices (1976)

Circuit Types 5400 5482 5483 5474 54S174 54163 4049A 4013A 4020A 10502 Description Quad, 2-input NAND 2-bit, full adder 4-bit, full adder Dual, D, edge-triggered flip-flop Hex, D, edge-triggered flip-flop 4-bit synchronous counter Inverting hex buffer Dual, D, edge-triggered flip-flop 14-stage, ripple carry counter Triple NOR (ECL) FITS 60 44 112 72 152 120 52 104 344 80 280

HYPROM512 512-bit PROM

Harris CICD Devices (1987)

Circuit Types HS-6504 HS-6514 HS-3374RH HS-54C138RH HS-80C85RH Package Types Flat Packs (hermetic brazed and glass/ceramic seals) LCC DIP FITS @ 55C, Failure Rate @ 60% U.C.L. 43.0 4k X 1 RAM 1k x 4 RAM Level Converter Decoder 8-bit CPU HS-8155/56 HS-82C08RH HS-82C12RH HS-8355RH 256 x 8 RAM Bus Transceiver I/O Port 2k x 8 ROM

Actel FPGAs
Technology (m) 2.0/1.2 1.0 0.8 0.6 0.45 0.35 RTSX 0.6 0.25 0.22 FITS # Failures Device-Hours

33 9.0 10.9 4.9 12.6 19.3 33.7 88.9 78.6

2 6 1 0 0 0 0 0 0

9.4 6.1 1.9 1.9 7.3 4.8 2.7 1.0 1.2

x x x x x x x x x

107 108 108 108 107 107 107 107 107

Xilinx FPGAs
XC40xxXL
Static: Dynamic: 9 FIT, 60% UCL 29 FIT, 60% UCL

XCVxxx
Static: 34 FIT, 60% UCL Dynamic: 443 FIT, 60% UCL

UTMC and Quicklogic

FPGA
< 10 FITS (planned) Quicklogic reports 12 FIT, 60% UCL

UT22VP10
UTER Technology, 0 failures, 0.3

Antifuse PROM
64K: 19 FIT, 60% UCL 256K: 76 FIT, 60% UCL

RAMTRON FRAMs
Technology 1608 (64K) 4k & 16K Serial FITS 1281 37 # Failures 1 152 # Devices 100 4257 Hours 103 103 Device-Hours 105 4.3 x 106

Note: Applied stress, HTOL, 125C, Dynamic, VCC=5.5V.

1

The one failure occurred in less then 48 hours. The manufacturer feels that this was an infant mortality failure.
2

12 failures detected at 168 hours, 3 failures at 500 hours, and no failures detected after that point.

Actel FIT Rate Trends

Skylab Lessons Learned

58. Lesson: New Electronic Components Avoid the use of new electronic techniques and components in critical subsystems unless their use is absolutely mandatory. Background: New electronic components (resistors, diodes, transistors, switches, etc.) are developed each year. Most push the state-ofthe-art and contain new fabrication processes. Designers of systems are eager to use them since they each have advantages over more conventional components. However, being new, they are untried and generally have unknown characteristics and idiosynchracies. Let some other program discover the problems. Do not use components which have not been previously used in a similar application if it can be avoided, even at the expense of size and weight.