On 12 August 1998 at 7:30:01 Eastern Daylight Time, Titan IVA-20 lifted off from Space
Launch Complex 41, Cape Canaveral Air Station. The booster was a Titan IVA rocket equipped
with a Centaur Upper Stage. The vehicle carried a classified National Reconnaissance
Office payload. Approximately forty seconds after liftoff, the rocket pitched down and
yawed right of its planned trajectory. When the course of the rocket deviated to an angle
of attack approximately 11 to 13 degrees from its planned path, aerodynamic stresses on
the vehicle exceeded its structural design. At this point, the northern-most Solid Rocket
Motor (SRM #1) separated from the core booster, initiating the Inadvertent Separation
Destruct System. At 45.529 seconds, approximately 3 seconds after the automatic destruct
sequence, Mission Flight Control Officers sent command destruct signals to the vehicle.
The Titan IVA had attained an altitude of 17,047 feet, a downrange distance of 4,422 feet,
and was traveling at a velocity of 1,007 feet/second. The accident resulted in loss of the
classified satellite and the launch vehicle. There were no injuries or damage to property
on the ground as a result of the mishap.
The Accident Investigation Board concludes, by clear and convincing evidence, that
pre-launch wire insulation damage on the Vehicle Power Supply existed somewhere in the
Titan Stage II which left at least one powered conductor with exposed wire not detected
during the pre-launch inspections and tests. After liftoff, the exposed wire
intermittently shorted as vehicle vibration increased in the transonic flight region. The
intermittent shorts momentarily caused a power outage of the Missile Guidance Computer,
resulting in the loss of the synchronization signal to the Inertial Measurement Unit
(IMU). This electrical interruption caused drift in pitch and yaw attitude reference
angles due to uncontrolled drive by torque motors within the IMU. When power was restored,
the guidance computer responded to an incorrect attitude reference and issued a maximum
pitch down and yaw right command. The resulting pitch caused an aerodynamic angle of
attack in excess of the structural design limits. At this point, Solid Rocket Motor #1
began to separate from the core vehicle, initiating its automatic self-destruct system,
and leading to the breakup and subsequent failure of the Titan IVA-20 Mission.
DISCLAIMER
Under 10 U.S.C. 2254(d), any opinion of the accident investigators as to the cause or
causes of, or the factors contributing to, the accident set forth in the accident
investigation report may not be considered as evidence in any civil or criminal proceeding
arising from a launch vehicle accident, nor may such information be considered an
admission of liability by the United States or by any person referred to in those
conclusions or statements.
ROBERT C. HINSON
Major General, USAF
Accident Investigation Board President
ACRONYM LISTING
AF Air Force
AFB Air Force Base
AFSPC Air Force Space Command
AGE Aerospace Ground Equipment
AIB Accident Investigation Board
ARO Accelerometer Read-Out
APS Accessory Power System
AS Air Station
AUV Avionics Upgrade Vehicle
CCAS Cape Canaveral Air Station
CDS Command Destruct System
CST Combined Systems Test
CUS Centaur Upper Stage
DCMC Defense Contract Management Command
EDT Eastern Daylight Time
EOD Explosive Ordnance Disposal
GIE Ground Instrumentation Equipment
GMT Greenwich Mean Time (or Zulu = EDT + 5 hours)
Hz Hertz (cycles per second)
IGS Inertial Guidance System
IGPS Inertial Guidance Power System
IMU Inertial Measurement Unit
IRT Initial Response Team
ISDS Inadvertent Separation Destruct System
ITL Integrated Transfer and Launch (includes VIB, warehouse, SMAB, SMARF, Launch Pads,
etc.)
LDA Launch Decision Authority
LDCG Launch Disaster Control Group
LMA Lockheed Martin Astronautics
LOX Liquid Oxygen
MARS Martin Anomaly Reporting System
MEOP Maximum Expected Operating Pressure
MET Mission Elapsed Time (time after liftoff, also "T+" time)
MFCO Mission Flight Control Officer
MGC Missile Guidance Computer
ms Millisecond
NASA National Aeronautics and Space Administration
PLF Payload Fairing
PSPICEÒ Personal Computer Simulation Program with Integrated Circuit Emphasis
ACRONYM LISTING (Continued)
RI Remote Instrumentation
RMIS Remote Multiplexed Instrumentation System
SIB Safety Investigation Board
SLC Space Launch Complex
SRM Solid Rocket Motor
SRMU Solid Rocket Motor Upgrade
SW Space Wing
TPA Turbopump Assembly
TVC Thrust Vector Control
U.S.C. United States Code
USAF United States Air Force
VAFB Vandenberg Air Force Base
VIB Vehicle Integration Building
VPS Vehicle Power Supply
WIS Wideband Instrumentation System
I. AUTHORITY AND PURPOSE
At the direction of the Commander, Air Force Space Command, an investigation of the 12
August 1998 Titan IVA-20 launch vehicle accident was initiated at Cape Canaveral Air
Station, Florida. The purpose of this accident investigation was to gather and preserve
evidence for claims, litigation, disciplinary and adverse administrative actions, and for
any other purposes in accordance with AFI 51-503, Aircraft, Missile, Nuclear, and Space
Accident Investigations and AFSPC SE 001, AFSPC Class A Space Launch Mishap Investigation
Plan. The investigation was required to present a summary of the facts and a statement of
opinion regarding the accident. The investigation team consisted of the following:
Accident Investigation Board:
Major General Robert C. Hinson Colonel Daniel A. Dansro President Vice President HQ
AFSPC/DO SMC/TEB Peterson AFB CO Kirtland AFB NM
Both Major General Hinson and Colonel Dansro possess knowledge and expertise relevant
to space launch accident investigations. Both attended the Air Force Safety Center Board
President's Course and Colonel Dansro has prior experience in accident investigations.
Technical Advisors:
Mr Lance M. Killoran Mr David W. Whittle, NASA National Reconnaissance Office Johnson
Space Center, Houston TX Mr Chester L. Whitehair, Aerospace Corp Mr Gregg W. Kraver, SMC
Det 8 El Segundo CA Cape Canaveral AS FL
Operations
Legal Advisor
Lt Col Mark E. Dowhan, HQ AFSPC/DOSL Major Richard K. Johnson, HQ AFSPC/JA Peterson AFB
CO Peterson AFB CO
Technical Assistant
Public Affairs
Maj Deirdre Healey, SMC Det 8 Capt LeWonnie E. Belcher, HQ AFSPC/PA Cape Canaveral AS
FL Peterson AFB CO
Recorder
2Lt Matthew W. Alexander, SMC Det 8 Cape Canaveral AS FL
II. SUMMARY OF FACTS
A. TITAN PROGRAM SUMMARY
The Titan launch vehicle is a proven booster, having performed more than 330 successful
military and commercial missions. The Titan program began in 1955 with the two-stage Titan
I and Titan II rockets. Beginning in 1961, the Titan program progressed through several
variations of the newer Titan III model. The first Titan IVA launch occurred on 14 June
1989, successfully delivering a Defense Support Program satellite into orbit. Mission A-20
was Titan IVA's twenty-second launch, and the final Titan IVA launch vehicle mission.
The Titan IV launch vehicle is a heavy class expendable launch vehicle and the largest
space booster used by the Air Force. Manufactured by Lockheed Martin Astronautics (LMA),
Titan IV is the latest in the Titan family of launch vehicles. The Titan IVA is capable of
placing over 10,000 pound payloads into geosynchronous orbit above the Earth. A newer
version of Titan IV, the Titan IVB, uses two Solid Rocket Motor Upgrades (SRMUs) and has
upgraded avionics. The Titan IVB has increased lift capability, able to launch payloads of
approximately 12,900 pounds to a geosynchronous orbit.
B. TITAN IVA VEHICLE DESCRIPTION
The two basic propulsion elements of the Titan IVA are the two SRMs and a two-stage
liquid core. A Centaur liquid upper stage can be added when required. The liquid core and
SRMs provide propulsion to low earth orbit. The Centaur Upper Stage (CUS) is used to
deliver the payload to highly elliptical or equatorial geosynchronous orbit. The CUS is
also used to put payloads into transfer orbits in preparation for interplanetary missions.
Each SRM is 112.9 feet long, weighs 687,615 pounds, and can deliver 1.48 million pounds of
thrust. They nominally provide power for the first 110 seconds of flight and then separate
approximately 8 seconds after Stage I ignites. The SRMs provide attitude control for the
launch vehicle through a liquid injection Thrust Vector Control (TVC) system. The system
is controlled by electrical signals from the core vehicle Missile Guidance Computer (MGC)
where the guidance and navigational units are located. Stage I consists of two liquid
propellant rocket engines which together deliver 548,000 pounds of thrust. It operates for
approximately 188 seconds. Stage II is a single liquid propellant engine which delivers
105,000 pounds of thrust. Stage II is shut down by command upon achievement of payload
target velocity. Both Stage I and Stage II use storable, hypergolic propellants consisting
of the fuel (Aerozine 50) and the oxidizer (nitrogen tetroxide). The CUS uses liquid
propellant consisting of liquid hydrogen and liquid oxygen and delivers 33,000 pounds of
thrust, taking the space vehicle to its final orbit. The overall length of the CUS is
29.45 feet with a diameter of 14 feet. It has its own guidance, navigation, and control
system.
Titan IV-A Instrumentation Telemetry Systems
The vehicle
telemetry system collects performance data from the launch vehicle and transmits a
continuous stream of flight data to ground receivers during the launch and flight. This
transmission consists of signals from end instruments sensing various physical parameters
and status of the vehicle's subsystems. Two instrumentation systems of interest to this
investigation are the Wideband Instrumentation System (WIS) and Remote Multiplexed
Instrumentation System (RMIS). The WIS measures characteristics of the external
environment while the RMIS measures health and status of the on-board components. The
Centaur has its own telemetry and guidance systems.
Titan IV-A Guidance
The guidance system for the Titan IVA rocket consists of the IMU, the MGC, and the rate
gyro. The IMU and MGC are located in a compartment identified as Compartment 2A. The rate
gyro is located in Compartment 1B. Each stage has three compartments; they are located at
the top of the stage, at a mid-point in the stage between the fuel and oxidizer tanks, and
at the bottom of the stage. The compartments are labeled A, B, and C, with the A
compartment at the top of the stage and the C compartment at the bottom. The IMU is a
multi-gimbaled gyro system which senses the motions of the vehicle relative to fixed axes
in space. The MGC contains the mission software to calculate the vehicle's attitude. It
computes the vehicle's position using data from the IMU and the rate gyro. As the flight
progresses the computer compares this computed position and attitude with the planned
trajectory and issues an error signal to correct any difference between the two. The error
signal will initiate steering commands to bring the vehicle back onto the planned
trajectory.
The Vehicle Power Supply Wiring Harness
The Vehicle Power Supply (VPS) battery, located in Compartment 2A, supplies power to
the VPS bus, which supplies the Accessory Power System (APS) bus, discrete arming bus, and
the Inertial Guidance Power Supply (IGPS) bus. The power is distributed to other
compartments, components, and sensors by a series of terminal boards and wire harnesses
which are routed inside the vehicle compartments and down cable conduits. The wiring is
tied into bundles and secured to the vehicle by clamps. To pass from one compartment to
another, the wiring bundles are passed through cutouts to cable conduits which run the
length of the stages on the outside of the booster. The conduit is protected by a hinged
metal cover and phenolic headers. The engineering design standards allow the
instrumentation wire bundle to be routed with the VPS cabling. The electrical and
instrumentation cables are bundled in accordance with engineering directions. The bundles
are secured to the stringers or other vehicle hardware with clamps, and may be tied to
existing cabling with string-ties. Two types of clamps were used in the Titan IVA-20
vehicle. One cushioned clamp (45D5) had no wedge and its rubber cushioning only covered
the inside of the clamp and its edges. The other clamp (45D82Z) added a wedge to the
design to prevent pinching of the wires and also extended the cushion to entirely cover
the clamp's surface. To protect against abrasions as the wiring is routed through the
vehicle, scuff wrap is required at doors, cutouts, and cross-overs at the edge of frames
and stringers or other wires bundles. Wiring passing through an access hole must be
protected by four layers of taping.
C. TITAN IVA-20 HISTORY
Mission
The mission of the Titan IVA-20 launch was to deliver a classified National
Reconnaissance Office payload into near-geosynchronous orbit.
Fabrication and Assembly
Construction of the core vehicle for the Titan IVA-20 launch began in early 1991. The
main core vehicle, Stage I and Stage II, were constructed between April 1991 and February
1992 at the Lockheed Martin Astronautics (LMA) Waterton plant near Denver, Colorado.
Construction of the Stage II fuel tank, oxidizer tank, and the Stage I oxidizer tank were
completed in April 1991. The various electrical wiring harnesses for the vehicle were
assembled in October 1991, and installed and tested between December 1991 and February
1992. After construction of the core vehicle and installation of the electrical wiring
harnesses, the vehicle was stored at the Waterton plant from 11 March 1992 to 24 July
1997. Additional rework occurred during the vehicle's storage. LMA build documentation
indicates 14 separate items were corrected on Stage II between 11 November 1994 and 18 May
1995. This work corrected or repaired damaged insulation, defective crimping of the
wiring, and addressed a previous failure to check bonding on connectors, among other
defects. There was also rework, which included the rerouting and clamping of the wiring
harness in Compartment 2A. Prior to shipping the vehicle to CCAS, Stage II was reinspected
in July 1997. During this inspection, a motor driven switch showed a dead short; a
defective shim on a stringer was replaced; and gouges on the instrumentation truss and
damaged wiring were discovered, among other items. It was determined that the truss tubing
had been gouged during the rework and repair work performed in 1995. Overall, Titan IV
A-20 was noted to have 44 wiring defects with shorting potential, the highest defect rate
of all Titan IVA launch vehicle. This naturally resulted in more than normal
out-of-position work.
The vehicle was shipped by air and arrived at CCAS on 24 July 1997. At the CCAS Titan
facilities, the vehicle was assembled for launch. The VPS, electrical system, and related
areas were also reinspected and reworked. Some of the more significant work included:
eight work items involving the RMIS wiring between January and June 1998; four work items
related to the installation of WIS wiring in June 1998, which involved routing the WIS
harness with the VPS harness through the gas plug in the Stage II conduits, and numerous
electrical items, involving repositioning the harness at various locations, and work on
the electrical systems in compartments 2A, 2B, 2C and 1A to include replacement of clamps
and hardware, scuff wrap rework, and wrapping nick areas. When the WIS wiring was
installed, the openings in the gas plugs had to be enlarged to accommodate the WIS wiring
installation. The Titan IVA-20 core was moved to the pad at Launch Complex 41 on 3 March
1998 and mated to the Centaur upper stage on 30 March. Tests were conducted on the booster
between March and August. It was fueled on 4 and 5 August, and launched on 12 August.
D. ACCIDENT DESCRIPTION
Pre-launch
Processing for Titan IVA-20 proceeded on schedule, with the Mission Readiness Review
conducted on 4 June 1998. During processing, LMA, reported that a three foot tear in the
CUS insulation blanket developed during a normal test procedure. After review, it was
decided that repair of the blanket was not sufficient, and that the conical section of
insulation blanket needed to be removed and replaced. This unplanned maintenance caused
the launch date to slip from 22 July to 12 August. LMA received the new conical blanket
and began fit checks on 24 July. Repair was completed and the Centaur passed pressure
testing on 30 July. The Combined Systems Test (CST) was successfully completed on 31 July.
The VPS battery was installed on 7 August 1998. The VPS battery open circuit voltage was
periodically monitored from an Aerospace Ground Equipment (AGE) Van and Ground
Instrumentation Equipment (GIE) station from battery installation through launch. The VPS
battery readings were nominal throughout the monitoring process. The Launch Readiness
Review was accomplished on 10 August.
Weather
All weather factors were well within constraints at the time of liftoff. There were
scattered clouds, and the temperature was in the low 80's. Winds were predominantly out of
the west at 6 knots. Pre-launch weather conditions subsequent to the CST on 31 July until
launch were also well within vehicle weather constraints and there were no ground wind or
lightning strike violations.
Launch
On 11 August, the terminal countdown clock was started at 4:35 p.m. Eastern Daylight
Time (EDT) toward the planned 12 August 6:02 a.m. launch. Countdown proceeded smoothly
until 5:18 a.m., when a problem occurred during CUS tanking. A failure in a ground station
electrical relay system caused engineers to discontinue automatic tanking, and use
approved procedures for manually finishing the tanking process, thereby delaying the
planned launch time to 7:30 a.m. With the CUS tanking process finally complete at 6:49
a.m., countdown resumed with no further delays. Titan IVA-20 lifted off from Cape
Canaveral at 7:30:01 a.m.
Flight
Total vehicle weight at liftoff was just over 1.9 million pounds. The vehicle ascended
as predicted, except that 20 feet of the Turbopump Assembly (TPA) drainline did not
release at liftoff. The drainline is ½ inch diameter plastic tubing. Except for the TPA
drainline, the vehicle performed as expected until T+39.416 seconds. At this point, a
series of anomalous electrical power and guidance subsystem events occurred which resulted
in maximum steering commands being issued to the SRM TVC system. These commands placed the
launch vehicle in an angle of attack which exceeded structural limits and caused the
aerodynamic breakup of the vehicle.
The overall sequence of anomalous events are as follows:
· Beginning at T+39.416 seconds, a series of intermittent electrical shorts caused
power fluctuations to occur within the VPS.
· At T+39.463 seconds, an alarm was issued indicating that voltage to the MGC had
fallen below the 21.8 volt threshold. When input voltage to the MGC drops below 21.8
volts, an internal interrupt is generated that starts a power down sequence of the MGC.
The power down causes the MGC to stop sending timing reference signals to the IMU. Loss of
this signal to the IMU caused a loss of gimbal torque motor control. This led to an
uncontrolled physical motion of the inertial platform and loss of the inertial reference
frame.
· At T+39.650 seconds, the MGC recovered power and reinitiated the timing reference
signal to the IMU. The IMU then presented a false indication to the MGC that the vehicle
had pitched up approximately 26 degrees and yawed left 5 degrees from its planned
attitude.
· At T+39.818 seconds, in an attempt to compensate for these perceived changes in
vehicle attitude, the MGC control system commanded the vehicle into a full pitch down and
yaw right.
· When the vehicle's pitch angle reached approximately 13 degrees off planned
trajectory, aerodynamic stresses on the launch vehicle exceeded its structural design
limits.
· At T +41.545 seconds, SRM No. 1 (the north SRM) separated from the core vehicle,
triggering the Inadvertent Separation Destruct System (ISDS) of that SRM. That explosion
impacted the core vehicle causing its destruction.
· At T +41.709 seconds SRM No. 2 was destroyed by its ISDS.
According to telemetry data, the vehicle had attained an altitude of 17,047 feet, a
downrange distance of 4,422 feet, and was traveling at a velocity of 1,007 feet/second. At
T+45.529 seconds, the Mission Flight Control Officers sent command destruct signals to the
vehicle.
During the period of anomalous events between 39.461 seconds and 41.205 seconds, signal
errors were also detected on the WIS wiring, RMIS data line, and RMIS address line.
Further analysis showed these errors to be caused by either electromagnetic induction, a
short circuit, or an open circuit.
Space Vehicle Activity
The mission failed before the satellite was successfully deployed. The satellite was
dormant at the time of failure, except for the activation of heaters and some transducers.
This investigation determined that space vehicle activity did not cause this mishap.
Impact
See "Assessment of Debris Impact Locations" in Tab J and JEAB Report,
Recovered Hardware Report.
Accident Response
When the launch vehicle exploded, several thousand pieces of solid propellant and
fragments of the launch vehicle were scattered over an area extending approximately five
miles downrange and three miles north-south off the coastal boundary of Cape Canaveral AS.
In addition, a cloud containing unspent rocket propellant was generated. The winds were
measured to be approximately four knots from the southwest (241 degrees) and took the
cloud harmlessly out to sea. All toxic materials remained within the established predicted
toxic corridor. At no time during this incident was there a toxic risk to any person.
Additionally, no debris impacted the land. For several minutes after the explosion, Range
radar reported tracking debris into the area bordered by the impact limit lines.
At liftoff, members of the Launch Disaster Control Group (LDCG) were at the Titan
fallback area located near the Titan Integrated Transfer and Launch (ITL) area north of
the Vehicle Integration Building (VIB). Immediately after seeing the explosion, the LDCG
determined visually that neither debris nor the cloud would reach them. This observation
was confirmed by coordination through the safety technical advisor. At 7:50 a.m. EDT, the
LDCG initiated a reconnaissance of the beach, working from SLC 41 to the beach and then
south. By this time, the LDCG made contact with camera crews, who verified that no debris
landed in their area and that they were safe. The LDCG then cleared the camera crews from
the areas. At the same time, the LDCG contacted the U.S. Coast Guard to close the affected
area to marine traffic and search for floating debris. At 8:25 a.m., SLCs 37 through 40
were reported clear of debris. At 10:20 a.m. the LDCG contacted NASA to request assistance
in collecting the floating debris. By 11:00 a.m., the Explosive Ordnance Team (EOT) had
covered the beach up to 5 miles north of SLC 47 and found no debris or propellant. At
about 3:00 p.m., the LDCG made arrangements for the Freedom Star and Liberty Star,
operated through NASA's contractor, United Space Alliance, to depart the next morning at
6:30 a.m. to begin recovery of the floating debris. The highest recovery priority at this
time was debris that would be a threat to public safety. The Coast Guard was asked to
recover the debris that was hazardous to marine traffic and to patrol the area overnight.
In the early afternoon of 12 August, the LDCG began to receive reports of large pieces of
debris floating close to shore or washing ashore. At 4:46 p.m., the LDCG turned the
accident site over to the Interim Board President.
For several days after the incident, recovery teams collected floating debris as well
as debris that had washed ashore. In addition, local television and radio stations
broadcast public service announcements asking the public not to retrieve debris themselves
but to notify Air Force officials if any was discovered. The recovery and search effort
was divided into three teams: the beach patrol team, the floating debris team, and the
subsurface debris team. The beach patrol combed local beaches searching for debris and
responded to calls from the public. The surface team coordinated the efforts of the NASA
vessels and Florida Air National Guard helicopters, which were responsible for recovering
ocean surface debris. On 17 August, the Navy Salvage Group was contracted to assist in the
recovery effort. Over the following several weeks, Navy experts mapped the debris field
and recovered numerous pieces from the ocean bottom, diving in water ranging between 12
and 50 feet. The undersurface recovery effort for purposes of this investigation was
terminated on 15 October 1998.
Media Interaction
Under AFI 51-503 and the Space Launch Vehicle Mishap Investigation Policy, public
affairs released information about the mishap to local, national, and international media
via press conferences and status reports. The coverage was balanced and positive in
nature. There were no reported cases of damage or injury in any of the news reports.
Selected press coverage excerpts can be found in Tab U.
E. INVESTIGATION ANALYSIS
Investigation Team Description
The investigation of the Titan IV A-20 launch mishap was the first investigation
conducted under the Space Launch Vehicle Mishap Investigation Policy adopted by the Acting
Secretary of the Air Force on 16 January 1998. The AIB and SIB were appointed concurrently
and ran a dual track investigative process. Under this process, a joint board, called the
Joint Engineering Analysis Board (JEAB), composed of Lockheed Martin Astronautics, U.S.
Air Force, and Aerospace Corporation personnel, conducted the technical analysis of the
mishap. The Safety Investigation Board (SIB) and Accident Investigation Board (AIB)
oversaw the analysis to ensure it was thorough and impartial. The engineering analysis of
the JEAB was not binding on the SIB or AIB and either board was permitted to supplement
the JEAB analysis, order or conduct additional testing, or conduct whatever further
investigation was deemed necessary to fulfill its responsibilities. This investigative
process was implemented to release as much information as possible regarding the mishap to
the public, as quickly as possible. In fulfilling their separate responsibilities and in
their role of oversight, the SIB and AIB were given unfettered access to all meetings of
the JEAB, the contractor's work sites, and to the contractor's personnel. The JEAB
performed the technical evaluation by various methods, specifically described below, and
by thoroughly reviewing their design, manufacturing, assembly, and launch processes. They
questioned their personnel throughout the process and had them review their individual
actions to determine their possible contribution to the mishap. Members of the AIB could
be present during any part of the technical analysis. The mishap analysis by the JEAB was
extensive, but in the end, had to rely on modeling, simulations, and failure analyses to
reach its conclusions.
Investigative Process
The technical investigation consisted of an expanded Ishikawa cause-and-effect
analysis, more commonly referred to as a "fishbone" analysis. This analysis was
supported by factual data developed through the inspections, simulations and analyses
outlined below. All of the areas on the fishbone analysis were eventually closed as
unlikely causes of the mishap, with the exception of the Core Vehicle fishbone.
Telemetry - A detailed analysis of the telemetry data revealed that at least 30
electrical shorts ranging from 1 to 20 milliseconds (ms) occurred within the various
systems of the core vehicle. Multiple shorts on the VPS eventually caused the chain of
events leading to mission failure. To determine the cause of the intermittent electrical
shorts, numerous tests, analyses, and experiments were conducted to recreate the telemetry
signature of the A-20 mishap. These methods included failure analysis, modeling, PSPICE â
analysis, random vibration testing, pedigree reviews, manufacturer testing, and controlled
experiments. A description of these methods and their factual determinations relevant to
this investigation are provided below.
Failure Analysis - Approximately 30% to 40% of the booster was recovered during
recovery and salvage operations discussed earlier. Of the 22 critical components, only
eleven were recovered. Nondestructive evaluations and destructive physical analyses of the
recovered hardware revealed that the recovered components were not defective and none
failed to function as designed. One item of note was an unexplained burned contact on the
Babcock Relay.
IGS Modeling - The goal of IGS modeling was to emulate the guidance system using the
same type of components installed on Titan IVA-20. Shorts and opens of varying duration,
ranging from 1 to 20 ms, were imposed on the model in an attempt to recreate the failure
signature shown in the telemetry of Titan IVA-20. The failure signature was consistently
duplicated only when a minimum of three shorts, approximately 5 ms in duration each, and
separated by 50 ms each, was introduced. The induced shorts resulted in random gimbal
drift ranging from minus 39 degrees to plus 25 degrees. The test results compare favorably
to the theoretical maximum random shift of plus or minus 55 degrees.
PSPICE â - A computer software program called Personal Computer Simulation Program
with Integrated Circuit Emphasis (PSPICE) â was used to simulate the electrical circuitry
of the rocket. The goal of PSPICE â was to recreate the failure signature shown in the
telemetry of Titan IVA-20 by introducing shorts at various locations with various
durations and voltages. PSPICE â showed that this failure signature was duplicated only
when multiple transient shorts were introduced, either on the wire harness on Stage II or
one of 13 components in the VPS located in Stage II. Opens would not duplicate the
telemetry signature observed on Titan IV A-20. Shorts in the components and wiring harness
of Stage I also would not duplicate the telemetry signature.
A further refined run on the PSPICE â, using 18.765 volts as the lowest voltage
achieved during the flight of A-20, closely matched the telemetry signature only when
shorts were introduced at any one of nine locations on Stage II VPS wiring. These
locations included two sites in the 2A compartment, one in the Stage II oxidizer conduit,
three in the 2B compartment, and three in the 2C compartment. Only one of the sites had
RMIS, WIS, and VPS circuits collocated in the same bundle. Telemetry showed data
corruption on these circuits which could be related to the shorting on the VPS wiring.
This site, located in Compartment 2C, was also a site where three separate Martin Anomaly
Reporting System (MARS) rework activities were performed at CCAS.
Random Vibration Test - A random vibration table test was conducted on various wire
bundles with the goal of characterizing wire damage and shorts under simulated Titan
IVA-20 flight dynamic conditions. The test showed that multiple shorts require greater
than 20 gauge wire to duplicate the signature of Titan IVA-20 telemetry data. A wire of
less than 16 gauge would consistently weld or burn through under this test duplicating an
open circuit. In addition, a cooling effect was necessary to duplicate the telemetry
signature of A-20. This testing also substantiated the durability of the wiring, proving
that abrasion against the vehicle structure for less than 40 seconds would not create a
shorting condition in the wiring. The use of kapton wiring on Titan IV vehicles was tested
during this investigation and its durability and life proved to be adequate for the Titan
IV applications, if properly protected. No premature aging effects were found.
Manufacturing Process Review - A review of the entire manufacturing process for Titan
IVA-20 revealed that three damaged conductors were discovered and corrected at the LMA
factory near Denver. An additional 41 defects discovered and corrected on Titan IVA-20 had
shorting potentials. The 44 defects constituted the highest defect rate on any Titan IVA
configuration. As a result, Titan IVA-20 had higher than normal out-of-position work.
Since the mid-1980's, there were 113 documented cases of wiring damage that could
result in a short on the vehicle's primary power wiring system. Ten of these cases were
not discovered until the booster was at the launch site. There have been in excess of
1,000 wiring defects over the course of the Titan program, which did not directly impact
the vehicle's primary power systems.
After the 12 August mishap and during the course of this investigation, the wiring of
Titan IVB and Titan II vehicles which were being processed for launch were given
additional inspections. Several instances of damaged wires were discovered, which included
"open" wires, damaged insulation, and exposed conductor wire.
ROBERT C. HINSON Major General, USAF Accident Investigation Board President
III. STATEMENT OF OPINION
DISCLAIMER Under 10 U.S.C.
2254(d), any opinion of the accident investigators as to the cause or causes of, or the
factors contributing to, the accident set forth in the accident investigation report may
not be considered as evidence in any civil or criminal proceeding arising from a launch
vehicle accident, nor may such information be considered an admission of liability by the
United States or by any person referred to in those conclusions or statements.
There is clear and convincing evidence to show that the 12 August 1998 failure of Titan
IV A-20 was due to a series of electrical shorts in the Vehicle Power Supply (VPS). These
shorts caused the Missile Guidance Computer (MGC) to reset. The resetting of the MGC
resulted in loss of the Accelerometer Read Out (ARO) timing signal to the Inertial
Measurement Unit (IMU). The ARO signal is the source for all timing reference signals
within the IMU, including the circuitry which controls the inertial platform gimbal torque
motors. This timing signal is sent by the MGC whenever the MGC is in the
"compute" mode. When the MGC is in the "reset" mode, the loss of
gimbal torque motor control allowed uncontrolled rotation of the inertial platforms
causing a loss of the vehicle's true reference. When the MGC restarted, it used the
incorrect positions of the inertial platforms to compute the vehicle's attitude in space.
The MGC read signals indicating the vehicle was approximately 26 degrees pitch up and 5
degrees yaw left from the expected flight path. To return to the expected flight path the
MGC ordered the vehicle into a severe pitch down and yaw right. This led to an angle of
attack which eventually caused the structural breakup of the vehicle, activation of the
Inadvertent Separation Destruct System (ISDS), and the failure of the Titan IVA-20
mission.
The Accident Investigation Board believes the following facts are supported by clear
and convincing evidence: pre-launch VPS wire insulation damage existed somewhere in Stage
II which left at least one powered conductor with exposed wire not detected during the
pre-launch inspections and tests. After liftoff, the exposed wire intermittently shorted
as vehicle vibration increased in the transonic flight region resulting in the sequence of
events related above. The cause of the damaged wire was not determined in this
investigation. Any attempt to further isolate the cause of the accident to a particular
location or to a causal activity, such as rework on the vehicle, cannot be supported by
clear and convincing evidence.
During the course of the investigation various possible causes for the damage were
considered. One cause for the damage may be the vehicle's design (sharp metal edges and
abrasion sites), which make it difficult to install, modify and maintain wire harnesses
without damage. Another potential cause is workmanship damage during the harness
fabrication and installation, out-of-position work, or during the handling,
transportation, or storage of the booster. A review of the Titan program's historical
records since 1989 revealed hundreds of wiring faults or defects produced at the factory
and at the launch base which were later discovered by inspection and testing. Of these
faults, 113 were related to a primary power supply. Review of other Titan flights revealed
incidents of in-flight shorting on the flight instrumentation system, giving credence to
the existence of undetected wiring problems on Titan vehicles. Photographs at Tab V
reflect wire damage incidents discovered in other Titan vehicles being prepared for launch
during the course of this investigation. The history of wiring faults and the wiring
defects discovered in recent Titan II and Titan IV vehicle inspections created concerns in
the Accident Investigation Board regarding the quality of workmanship and the need for a
zero defect approach which treats the wiring harnesses as critical components. It is
acknowledged the damaged wires represent a small percentage of the miles of installed
wiring and the wiring activity on the Titan vehicles, but the mishap of Titan IV A-20
highlights the significance of mistakes in this area, particularly in the vehicle's
powered lines. Still the Accident Investigation Board was unable to identify the specific
evidence which would remove any doubts as to the root cause of this accident. We cannot
state with clear and convincing evidence that the mishap was the result of poor
workmanship, handling, transportation, or simple carelessness in the entry or exit of a
compartment.
In the course of this investigation, several areas and components of the vehicle were
reviewed in an effort to determine the cause of the mishap. While not causal, these areas
could be considered contributing factors to the mishap. As currently designed, the vehicle
has minimal tolerance to power interrupts which could occur from a fault anywhere on the
vehicle. Compounding this problem is the fact that the vehicle is especially sensitive to
power interrupts in the guidance system (MGC, IMU; etc.) and other critical electrical
components and circuits. It is clear from this investigation that the Booster Integrity
Testing (BIT) and the harness visual inspections done at the factory as well as the
harness visual inspections and pre-launch testing done at the launch bases were not
perceptive enough to prevent this failure.
Additionally, several design and or build problems were discovered that, while
eventually determined to not be causal or contributing factors, deserve comment because
they were investigated as potential causes of this mishap. The ARO timing signal from the
MGC to the IMU is currently unable to withstand power transients that can reset the MGC.
The Babcock Relay Assemblies were assembled with wrong style washers or the washers may
have been missing. The 40AH Batteries as well as other vehicle batteries were found to
have uncontrolled sense wire routings with the potential to cause internal battery
shorting. Stage I and Stage II conduit covers which protect critical vehicle wiring from
aerodynamic forces and heating have one quarter turn quick disconnect fasteners which can
only be verified to have been properly installed by the installing technicians
"feel". They have the potential of failing the covers in flight if several
fasteners in a row were not properly locked. Also, the conduit cover piano hinges do not
have a positive locking feature should the piano wire slip, potentially allowing the cover
to open during flight. Further, the Stage II conduit gas plug has the potential to damage
critical circuits. The vehicle also has a large number of sharp surfaces that could lead
to wire damage during installation, rework and normal vehicle processing and potentially
go undetected by the technicians and engineers. While all of these items were investigated
as to possible cause of the mishap, they have been determined to not be causal to the
mishap.
The Accident Investigation Board concludes that pre-launch wire insulation damage
existed in the Vehicle Power Supply somewhere in Stage II which left at least one powered
conductor with exposed wire which was not detected during the pre-launch inspections and
tests. After liftoff, the exposed wire intermittently shorted as vehicle vibration
increased in the transonic flight region, leading to the sequence of events related above
and resulting in the failure of the TitanIVA-20 mission. Uncertainty of the exact location
of the damaged wire, whether there was only one shorting site or multiple sites, and
exactly what activity caused the damage, leave the Accident Investigation Board with doubt
as to the root cause of the accident.
ROBERT C. HINSON Major General, USAF Accident Investigation Board President
IV. INDEX OF TABS Tab A AF Form 711, USAF Mishap Report
Tab B Preliminary Message Report
Tab J Technical and Engineering Evaluations
1. Executive Summary of JEAB Failure Analysis Report
2. JEAB Failure Analysis Report and Supporting Documentation on CD ROM
