It's not whether you win or lose, but how you place the blame
17 July 1985
March AFB (KRIV)
Riverside, California
NASA 712, Convair 990-30A-5, N712NA "Galileo II"
4 Minor, 15 Uninjured
Accident sequences are often described as a "tree". That is, a series of events where the individual influences at different points determine the direction the "tree branch" will go. These branches continue based on the events that have occurred, the events that will occur, and the influences that interrelate with said events. Taking an aircraft accident and looking back, different places can be identified where had differing influences acted upon a said event, a different branch direction, or "fork in the road", might have been taken and the accident averted. Like trees, no two accidents are truly the same. The end-result might be the same, as might the causes; but like any tree in nature, no two accident "event trees" are the same; each is a unique chain of direct, indirect, intentional, unintentional, known, unknown, internal, or extenal actions or inactions. In the past few years, there has been renewed discussion of rejected takeoffs (RTO) as they apply to transport-category aircraft, specifically in the areas of RTO philosophy; that is, proper crew performance of RTOs, quick decision-making for an RTO in the limited time available, and a review of the relationship of V1 (commonly known as go/no go speed) and it's application/use in regards to different emergencies. V1 is one of those black and white phenomenas that upon closer inspection, can contain many shades of grey in so far as being a hard and fast number for continuing or rejecting a takeoff based on differing emergencies. Like anything else in aviation, V1 is another guide, or reference point, to help a pilot make a quick decision at a critical point of flight. Our featured accident today will discuss RTOs, and how a successful RTO where the aircraft remained on the runway and was brought to a controlled, successful stop with no injuries, still resulted in the complete destruction of the aircraft.
N712NA was one of four Convair 990 jetliners that were acquired by the National Aeronautics and Space Administration as research aircraft. The Convair 880/990 series had a fairly short life as a passenger airliner, being used by such carriers as Delta and American, but were a very interesting transport-category aircraft development. NASA used their Convairs for a variety of duties, the most significant being atmospheric and space research using onboard telescopic instruments installed on the aircraft. NASA 712 was known as "Galileo 2", having replaced it's sister ship NASA 711 (Galileo) which was lost in a midair collision with a US Navy P-3 Orion patrol craft whilst landing at NAS Moffet Field, California, in 1973. A few of NASA 712s research projects included one known as the Airborne Auroral Expedition. In one part of this project, 712 flew westward from Churchill, Canada to Fairbanks, Alaska against the direction of the earth's rotation, but at nearly the same speed. In doing so, it remained in essentially the same spot relative to the sun on the opposite side of the earth for a period of several hours. This research and photographs attained helped prove that the aurora in the midnight sky sometimes underwent a "quiet phase" in regards to brightening, development, and breakup disproving the then-accepted notion of the particular stage of the aurora always being related to the hour. 712 and it's sister Convair 990s also flew steep descent testing, comprising power-off landing approaches and demonstration of minimum lift-to-drag ratio (L/D) landings. These were directly related to the interest in the use of low L/D lifting bodies for recovery to landing from space, later to become the Space Shuttle. NASA 712 also was equipped with a digital navigation/guidance/automatic-landing system to study energy-management approach/landing profiles for commercial jet transports; in essence, one of the first partial-glass equipped aircraft. All in all, NASA 712 was a very interesting and highly unique aircraft. By 1985, 712 was still assigned to the Ames Research Center as a technical research platform.
On evening of 17 July 1985, NASA 712 was preparing to takeoff from March AFB on a research flight. 712 was cleared for takeoff at 1810L, and taxied onto RW 32, which was 13,300' x 200'. 712 was dispatched at 232,500 lbs, 7500 lbs below it's max ramp weight. V1/Vr/V2 were calculated to be 151, 154, and 167 respectively. The critical field length was calculated to be 10,500', 2800' less than the actual runway length, putting 712 well within Category 1 takeoff ops limits. 712 was unique in that it had an onboard closed-circuit TV system that monitored various parts of the aircraft outside. This day, the ventral camera of the system, located on the aircraft centerline abeam the wing leading edges, just happened to be focused on the right main landing gear assembly, and was being monitored by two of the fifteen scientists riding in the passenger compartment. This video would prove invaluable to the investigation. Midway into 712s takeoff roll, two of the scientists watching the ventral camera noticed the #3 tire (forward-left of 4 tires on the right gear bogie assembly) beginning to deform and bubble, and a few seconds later, completly blow out. Another scientist occupying a window seat on the right side just behind the wing trailing edge noticed black objects "fly by" his window. Ground witnesses located on the parking ramp, as well as the USAF tower controllers, noticed white smoke emanating from underneath 712 at this time. None of this was immediately apparent to the cockpit crew, as the first indication of anything wrong to them was two loud "bang" noises and airframe vibration, sounds so loud they were clearly recorded on the aircraft's CVR. At this time, the aircraft's speed was anywhere from 140 to 143 knots, the difference being what the pilot remembered seeing to what some of the scientists monitoring inertial-nav speed readouts on their equipment remember seeing. The flight data recorder indicated a max speed attained of 143 knots. In any event, the aircraft was still about 10 knots below go/no go speed so the aircraft commander (AC) in the co-pilot's seat called for an abort at the same time the flight engineer (FE) called a blown tire. An RTO was commenced with the AC closing the throttles, deploying the spoilers, and selecting reverse on all four engines. Due to the good excess of runway remaining, the AC elected to commence only very light braking, allowing the aircraft to slow with the thrust primarily. At first touch of the brakes, the plane swerved slightly to the right, and was immediately corrected back to runway centerline. 9 seconds after the first two explosion sounds, the CVR recorded another loud bang, with the FE then commenting "blew another one". 5 seconds later, reverse thrust is heard taking effect as the RTO procedures progressed, and yet another loud bang was heard on the CVR. As the FE called "3000 remaining", referencing the aircraft passing the 3-board on the runway, and unaware of any further problems, the AC closed the throttles and began slowing in order to make the departure-end turnoff and clear the runway. At this time, one of the scientists in the back called "fire on the right side", and the AC immediately began slowing the aircraft straight-ahead, coming to a stop at the 12,700' point from departure, and 600' length remaining. The crew secured the aircraft, and the AC opened the right cockpit window to check the right side of the aircraft, noting the right main landing gear bogie on fire, and raw jet fuel pouring from the underside of the right wing inboard of the #3 engine. All 19 crew successfully evacuated the aircraft as it quickly became a conflagration. By the time base CFR trucks arrived, about a minute and a half, the aircraft's entire aft half was ablaze. Firefighting efforts were hampered by the massive amount of jet fuel remaining and localized, and fueling the fire. This caused extreme temperatures to be encountered, making approach of the aircraft difficult. CFR trucks expended 1,915 gallons of Aqueous Film Forming Foam (AFFF) and 59,000 gallons of water, however the fire was not extinguished and the aircraft was destroyed.
Probable Cause
*Tire Failure-Multiple-Right Main Landing Gear
Secondary Factors
*Takeoff-Aborted-Flightcrew
*Fuel Cell-Compromised-Right Side
*Fire-Right MLG Assembly
Tertiary Factors
None
MikeD says
The analysis of this accident is a very interesting study of investigation techniques as well as RTO philosophy.
Technical Analysis
March AFB was typical of many USAF "heavy" aircraft bases in that it had the concrete taxiways and runways as opposed to the asphalt. Consequently, analysis of NASA 712s tires prior to takeoff roll could be assessed. This was accomplished by noting where 712 had taxied from the runup area onto the runway. In this area, heavy jet exhaust deposits from other aircraft were present, and when 712 taxied over these, there were noted "white spots" where 712s tires "erased" the soot deposits and left exposed concrete. The marks were uniform in width, showing that 712 had no abnormalities of it's tires such as low pressure, flat spots, or other problems (notably, any brake drag) prior to the actual takeoff roll; a fact backed up by the scientists monitoring the right main landing gear assembly via the closed-circuit TV. The first pieces of tire were noted at the 1400' point right of RW centerline, just shy of the approach-end BAK-12B arresting cable for RW 32. From this point, up to 4000', wavy tire rubber marks in the runway, as well as rubber pieces, were noted consistent with the track of the #3 tire (left front of the 4 wheel bogie). From the 4000' to the 5000' point, heavy rubber marks were noted consistent with the line of tire #4 (right front tire of bogie) indicating where it blew out. It was determined that tire #3 began failing first, losing rubber pieces and making tire #4 to run overloaded, causing it to fail. The failure of tire #4 was followed shortly after by tire #3 less than a second later, accounting for the two initial close-succession "bang" noises heard first on the CVR, as well as felt by the flight crew. The parts found from tire #3 showed no evidence of extreme heat, contrasted with the tread recovered from tire #4 which had evidence of extreme heat, likely from the overload condition. At this time, with @ 8000' of runway remaining, the aircraft was supported on it's right bogie by the rims of tires #3 and #4 (front of the bogie), and tires #7 and #8 (aft two on the bogie). This was evidenced by scoring and gouging in the concrete runway surface from this point on. As the aircraft rolled past @6500' remaining, the rims of tires #3 and #4 began shedding parts in all directions, evidently causing the blowout of tires #7 and #8, accounting for the second pair of loud "bang" noises noted on the CVR. It was determined through runway marks that tire #8 blew first, with tire #7 blowing as the reverse thrust began to spool up, masking that sound from the flightcrew. It was determined through the scoring in the runway concrete that at about the 6000' remaining point is where the grounded wheel rims of tires #3 and #4 ground to the hubs of the bogie. This is also where ground witnesses reported seeing intermittent fire develop underneath the aircraft. It is also assessed that this was the point where the right wing underside and fuel cell was compromised. Cabin crew personnel noted fire on the bogie assembly at 3000' remaining, while large burn marks of high-temperature scorching consistent with Class D fires of burning metals were noted @1300' remaining on the runway. The scorching persisted from this point until the aircraft stop point, and were in-line with the right MLG bogie assembly.
Rejected Takeoff discussion
Following the accident, NASA researched 61 RTOs reported from 1975 to 1985 by various sources, both domestic as well as foreign. It was found that 34% of RTOs were for tire failure, 23% for engine failure, and the remaining 43% for a variety of causes. Previous to this, the FAA had published a 1977 report titled Jet Transport Rejected Takeoffs, Final Report, February 1977. In it, the FAA researched 171 RTOs from between 1964 to 1975. It was discovered that 87% of RTOs were from some failure of tire/wheel/brakes, 74% of these for tire blowouts alone. The data gathered in both reports revealed that engine failure were not a primary factor in RTOs. Training-wise Go/No Go speeds (V1) , as created, were generally predicated on engine loss; however many other takeoff emergencies are now routinely practiced that have the pilot or crew having to decide whether to abort or not. Increases in simulator techcology have allowed near-realistic programming of this and many other emergencies in modern-day flight simulators; in the case of tire failures, other problems associated such as brake problems, directional control, and the effects of anti-skid can also be simulated. Also on the subject of training, the accepted method for RTOs is to get the aircraft stopped quickly. The consensus is that an RTO is one procedure, and doesn't really change for differing emergencies (itself nearly being an emergency, depending on the situation). Having one RTO procedure simplifies training by providing standardization.
Real-world speaking, RTOs can be one of the most demanding and dangerous procedures a flightcrew might need to perform, especially closer to V1 where speeds are higher and analysis/reaction times are significantly reduced. This is why RTO procedures are ingrained and simplified, as well as planned for and briefed prior to each takeoff. Severe preparation mitigates the surprise and shock of performing an RTO. In the case of NASA 712, this is where my aforementioned large branches of the "accident tree" begin to take their route of growth. At the point of tire failure, the pilot of 712 had a fork in the road: continue the takeoff, or reject the takeoff. Each decision had it's ups and downs:
Continue: Though the aircraft was 10 knots below V1, had the pilot decided to continue the takeoff, he'd possibly have been able to burn down/dump fuel, and recover with a lighter-weight aircraft rather than face a high-speed, heavyweight abort. On the downside, who knows what configuration the aircraft would have actually gotten airborne in, since it's unknown when precisely the right wing fuel cells were punctured. It may have been a crapshoot whether the plane would've gotten airborne with just mangled landing gear, or with leaking fuel that could possibly ignite. Also, even landing lighter weight after burning/dumping fuel, there's no guarantee that the right MLG bogie wouldn't have punctured the wing fuel cell after landing, possibly resulting in a fire too, though of much reduced severity.
Reject: In this situation, the crew would be faced with a high-speed and heavyweight abort, like they were. In this situation, the decision "tree" comes to another fork: max braking or rollout:
Max Braking: In RTO procedures, it's generally accepted that max braking be applied as part of the procedure, consistent with maintaining directional control. This provides for stopping quickly and allowing evacuation, if needed, to commence quickly. IMO, this is situation-dependant and part of "making the call" that pilots get paid for. Would max braking have prevented the destruction of the wheel rims and puncturing of the right wing fuel cell in this accident? It's anyone's guess. That's another crap shoot. What's to say that the same thing wouldn't have happened with the deceleration and brake forces on the bogie assembly as opposed to rolling out? But as before, RTO procedures need to be standardized, however IMO, like any emergency procedures, only those steps that are necessary to be taken should be taken, and this again comes back to pilot/crew decision making based on the situation at hand.
Rollout: The rollout decision was made based on the ACs experience as well as his desire to not affect directional control with braking, combined overall with the excellent amount of runway remaining. Again, IMO, we as pilots are paid to make the decision when the situation dictates it. As it was, the tire/rim assemblies came apart, punctured the wing fuel cell, and ignited a fire. Again, a crap shoot. The case of NASA 712 was a rarity, since not often are conditions such where critical field length is so much less than actual runway length, and a flightcrew has so much excess runway to play with.
As mentioned before, V1 philosophy has always been predicated on engine loss go/no go. V1 has evolved to become an abort speed for nearly any emergency that may require an abort. The biggest reasons for standardized abort procedures is simply due to the fact that a rote response is needed for times when the split-second decision needs to be made at high speed with little time, and the crew must respond in coordinated action. But should V1 training be modified to become somewhat situationally dependant for below-V1 continuations? Who knows. My opinion still stands that many RTOs are a gut call based on the situation at hand and the training/experience of the crew. There are many factors that the crew can control, yet many possibilities that the crew has no control over, again situational-dependant. For example, there are some emergencies where I'll easily abort past refusal speed. One example would be an engine failure in my former aircraft when the WX is low IFR and there's little chance I'd be able to work the problem in VMC. Any type of engine fire would be another one. Another example in my former aircraft would be dual generator failure with a takeoff that would put me in IMC (simply due to the myriad of problems that it would cause in-cockpit, and especially the fact that I'd be facing flight controls that have a limited lifetime left from battery-only; approximately 10-15 minutes before complete loss of FBW). But again, those are factors for my particular airplane and may or may not apply to another aircraft or another pilot; so as I often say, it really does depend. My decisions on these was also based on the fact that I had a drag chute and runway arresting gear available that I could use, but my point is that it's sometimes very situational-dependant.
As an aside to this accident, the 25 July 2000 crash of the Air France Concorde on takeoff from Charles DeGaulle airport cannot go unmentioned. In that crash, the crew elected to continue the takeoff, not knowing that they had a massive fire underneath the left wing. Like the crew of NASA 712, the first indication of fire was near end-game.....in the case of 712, near the end of their rollout; and in the case of the Concorde, likely just as they were getting airborne. In this pilots very humble opinion, this is a gut check of the toughest kind. My personal desire is to keep a bad aircraft on the ground if at all possible.....ie, not take a bad plane airborne. But that decision must be tempered with many other factors: where will the airplane end up if I abort and don't stop at the end of the runway? Where will it end up if I takeoff and can't keep it airborne? Is there excess runway available? These and many other questions are all part of my "accident tree" of decision and indecision, of things you can control and things you can't. As I said, every accident is different in specifics, though they may be similiar in generalities.
It is those times when Mr. Murphy is riding in your jumpseat, that all you can do is the best your training and experience provide you, and even that is little insurance-coverage against the multitude of monkey-wrenches Mr. Murphy can and will throw your way. The most anyone can do is give it their best, and hope the decisions they make at the moment, will take them down the correct fork in the "accident tree."
MikeD
The above is not intended to be an undue criticism of the person or persons involved in the incident described. Instead, the analysis presented is intended to further the cause of flight safety and help to reduce accidents and incidents by educating pilots through the sacrifices of others in our profession.
17 July 1985
March AFB (KRIV)
Riverside, California
NASA 712, Convair 990-30A-5, N712NA "Galileo II"
4 Minor, 15 Uninjured
Accident sequences are often described as a "tree". That is, a series of events where the individual influences at different points determine the direction the "tree branch" will go. These branches continue based on the events that have occurred, the events that will occur, and the influences that interrelate with said events. Taking an aircraft accident and looking back, different places can be identified where had differing influences acted upon a said event, a different branch direction, or "fork in the road", might have been taken and the accident averted. Like trees, no two accidents are truly the same. The end-result might be the same, as might the causes; but like any tree in nature, no two accident "event trees" are the same; each is a unique chain of direct, indirect, intentional, unintentional, known, unknown, internal, or extenal actions or inactions. In the past few years, there has been renewed discussion of rejected takeoffs (RTO) as they apply to transport-category aircraft, specifically in the areas of RTO philosophy; that is, proper crew performance of RTOs, quick decision-making for an RTO in the limited time available, and a review of the relationship of V1 (commonly known as go/no go speed) and it's application/use in regards to different emergencies. V1 is one of those black and white phenomenas that upon closer inspection, can contain many shades of grey in so far as being a hard and fast number for continuing or rejecting a takeoff based on differing emergencies. Like anything else in aviation, V1 is another guide, or reference point, to help a pilot make a quick decision at a critical point of flight. Our featured accident today will discuss RTOs, and how a successful RTO where the aircraft remained on the runway and was brought to a controlled, successful stop with no injuries, still resulted in the complete destruction of the aircraft.
N712NA was one of four Convair 990 jetliners that were acquired by the National Aeronautics and Space Administration as research aircraft. The Convair 880/990 series had a fairly short life as a passenger airliner, being used by such carriers as Delta and American, but were a very interesting transport-category aircraft development. NASA used their Convairs for a variety of duties, the most significant being atmospheric and space research using onboard telescopic instruments installed on the aircraft. NASA 712 was known as "Galileo 2", having replaced it's sister ship NASA 711 (Galileo) which was lost in a midair collision with a US Navy P-3 Orion patrol craft whilst landing at NAS Moffet Field, California, in 1973. A few of NASA 712s research projects included one known as the Airborne Auroral Expedition. In one part of this project, 712 flew westward from Churchill, Canada to Fairbanks, Alaska against the direction of the earth's rotation, but at nearly the same speed. In doing so, it remained in essentially the same spot relative to the sun on the opposite side of the earth for a period of several hours. This research and photographs attained helped prove that the aurora in the midnight sky sometimes underwent a "quiet phase" in regards to brightening, development, and breakup disproving the then-accepted notion of the particular stage of the aurora always being related to the hour. 712 and it's sister Convair 990s also flew steep descent testing, comprising power-off landing approaches and demonstration of minimum lift-to-drag ratio (L/D) landings. These were directly related to the interest in the use of low L/D lifting bodies for recovery to landing from space, later to become the Space Shuttle. NASA 712 also was equipped with a digital navigation/guidance/automatic-landing system to study energy-management approach/landing profiles for commercial jet transports; in essence, one of the first partial-glass equipped aircraft. All in all, NASA 712 was a very interesting and highly unique aircraft. By 1985, 712 was still assigned to the Ames Research Center as a technical research platform.
On evening of 17 July 1985, NASA 712 was preparing to takeoff from March AFB on a research flight. 712 was cleared for takeoff at 1810L, and taxied onto RW 32, which was 13,300' x 200'. 712 was dispatched at 232,500 lbs, 7500 lbs below it's max ramp weight. V1/Vr/V2 were calculated to be 151, 154, and 167 respectively. The critical field length was calculated to be 10,500', 2800' less than the actual runway length, putting 712 well within Category 1 takeoff ops limits. 712 was unique in that it had an onboard closed-circuit TV system that monitored various parts of the aircraft outside. This day, the ventral camera of the system, located on the aircraft centerline abeam the wing leading edges, just happened to be focused on the right main landing gear assembly, and was being monitored by two of the fifteen scientists riding in the passenger compartment. This video would prove invaluable to the investigation. Midway into 712s takeoff roll, two of the scientists watching the ventral camera noticed the #3 tire (forward-left of 4 tires on the right gear bogie assembly) beginning to deform and bubble, and a few seconds later, completly blow out. Another scientist occupying a window seat on the right side just behind the wing trailing edge noticed black objects "fly by" his window. Ground witnesses located on the parking ramp, as well as the USAF tower controllers, noticed white smoke emanating from underneath 712 at this time. None of this was immediately apparent to the cockpit crew, as the first indication of anything wrong to them was two loud "bang" noises and airframe vibration, sounds so loud they were clearly recorded on the aircraft's CVR. At this time, the aircraft's speed was anywhere from 140 to 143 knots, the difference being what the pilot remembered seeing to what some of the scientists monitoring inertial-nav speed readouts on their equipment remember seeing. The flight data recorder indicated a max speed attained of 143 knots. In any event, the aircraft was still about 10 knots below go/no go speed so the aircraft commander (AC) in the co-pilot's seat called for an abort at the same time the flight engineer (FE) called a blown tire. An RTO was commenced with the AC closing the throttles, deploying the spoilers, and selecting reverse on all four engines. Due to the good excess of runway remaining, the AC elected to commence only very light braking, allowing the aircraft to slow with the thrust primarily. At first touch of the brakes, the plane swerved slightly to the right, and was immediately corrected back to runway centerline. 9 seconds after the first two explosion sounds, the CVR recorded another loud bang, with the FE then commenting "blew another one". 5 seconds later, reverse thrust is heard taking effect as the RTO procedures progressed, and yet another loud bang was heard on the CVR. As the FE called "3000 remaining", referencing the aircraft passing the 3-board on the runway, and unaware of any further problems, the AC closed the throttles and began slowing in order to make the departure-end turnoff and clear the runway. At this time, one of the scientists in the back called "fire on the right side", and the AC immediately began slowing the aircraft straight-ahead, coming to a stop at the 12,700' point from departure, and 600' length remaining. The crew secured the aircraft, and the AC opened the right cockpit window to check the right side of the aircraft, noting the right main landing gear bogie on fire, and raw jet fuel pouring from the underside of the right wing inboard of the #3 engine. All 19 crew successfully evacuated the aircraft as it quickly became a conflagration. By the time base CFR trucks arrived, about a minute and a half, the aircraft's entire aft half was ablaze. Firefighting efforts were hampered by the massive amount of jet fuel remaining and localized, and fueling the fire. This caused extreme temperatures to be encountered, making approach of the aircraft difficult. CFR trucks expended 1,915 gallons of Aqueous Film Forming Foam (AFFF) and 59,000 gallons of water, however the fire was not extinguished and the aircraft was destroyed.
Probable Cause
*Tire Failure-Multiple-Right Main Landing Gear
Secondary Factors
*Takeoff-Aborted-Flightcrew
*Fuel Cell-Compromised-Right Side
*Fire-Right MLG Assembly
Tertiary Factors
None
MikeD says
The analysis of this accident is a very interesting study of investigation techniques as well as RTO philosophy.
Technical Analysis
March AFB was typical of many USAF "heavy" aircraft bases in that it had the concrete taxiways and runways as opposed to the asphalt. Consequently, analysis of NASA 712s tires prior to takeoff roll could be assessed. This was accomplished by noting where 712 had taxied from the runup area onto the runway. In this area, heavy jet exhaust deposits from other aircraft were present, and when 712 taxied over these, there were noted "white spots" where 712s tires "erased" the soot deposits and left exposed concrete. The marks were uniform in width, showing that 712 had no abnormalities of it's tires such as low pressure, flat spots, or other problems (notably, any brake drag) prior to the actual takeoff roll; a fact backed up by the scientists monitoring the right main landing gear assembly via the closed-circuit TV. The first pieces of tire were noted at the 1400' point right of RW centerline, just shy of the approach-end BAK-12B arresting cable for RW 32. From this point, up to 4000', wavy tire rubber marks in the runway, as well as rubber pieces, were noted consistent with the track of the #3 tire (left front of the 4 wheel bogie). From the 4000' to the 5000' point, heavy rubber marks were noted consistent with the line of tire #4 (right front tire of bogie) indicating where it blew out. It was determined that tire #3 began failing first, losing rubber pieces and making tire #4 to run overloaded, causing it to fail. The failure of tire #4 was followed shortly after by tire #3 less than a second later, accounting for the two initial close-succession "bang" noises heard first on the CVR, as well as felt by the flight crew. The parts found from tire #3 showed no evidence of extreme heat, contrasted with the tread recovered from tire #4 which had evidence of extreme heat, likely from the overload condition. At this time, with @ 8000' of runway remaining, the aircraft was supported on it's right bogie by the rims of tires #3 and #4 (front of the bogie), and tires #7 and #8 (aft two on the bogie). This was evidenced by scoring and gouging in the concrete runway surface from this point on. As the aircraft rolled past @6500' remaining, the rims of tires #3 and #4 began shedding parts in all directions, evidently causing the blowout of tires #7 and #8, accounting for the second pair of loud "bang" noises noted on the CVR. It was determined through runway marks that tire #8 blew first, with tire #7 blowing as the reverse thrust began to spool up, masking that sound from the flightcrew. It was determined through the scoring in the runway concrete that at about the 6000' remaining point is where the grounded wheel rims of tires #3 and #4 ground to the hubs of the bogie. This is also where ground witnesses reported seeing intermittent fire develop underneath the aircraft. It is also assessed that this was the point where the right wing underside and fuel cell was compromised. Cabin crew personnel noted fire on the bogie assembly at 3000' remaining, while large burn marks of high-temperature scorching consistent with Class D fires of burning metals were noted @1300' remaining on the runway. The scorching persisted from this point until the aircraft stop point, and were in-line with the right MLG bogie assembly.
Rejected Takeoff discussion
Following the accident, NASA researched 61 RTOs reported from 1975 to 1985 by various sources, both domestic as well as foreign. It was found that 34% of RTOs were for tire failure, 23% for engine failure, and the remaining 43% for a variety of causes. Previous to this, the FAA had published a 1977 report titled Jet Transport Rejected Takeoffs, Final Report, February 1977. In it, the FAA researched 171 RTOs from between 1964 to 1975. It was discovered that 87% of RTOs were from some failure of tire/wheel/brakes, 74% of these for tire blowouts alone. The data gathered in both reports revealed that engine failure were not a primary factor in RTOs. Training-wise Go/No Go speeds (V1) , as created, were generally predicated on engine loss; however many other takeoff emergencies are now routinely practiced that have the pilot or crew having to decide whether to abort or not. Increases in simulator techcology have allowed near-realistic programming of this and many other emergencies in modern-day flight simulators; in the case of tire failures, other problems associated such as brake problems, directional control, and the effects of anti-skid can also be simulated. Also on the subject of training, the accepted method for RTOs is to get the aircraft stopped quickly. The consensus is that an RTO is one procedure, and doesn't really change for differing emergencies (itself nearly being an emergency, depending on the situation). Having one RTO procedure simplifies training by providing standardization.
Real-world speaking, RTOs can be one of the most demanding and dangerous procedures a flightcrew might need to perform, especially closer to V1 where speeds are higher and analysis/reaction times are significantly reduced. This is why RTO procedures are ingrained and simplified, as well as planned for and briefed prior to each takeoff. Severe preparation mitigates the surprise and shock of performing an RTO. In the case of NASA 712, this is where my aforementioned large branches of the "accident tree" begin to take their route of growth. At the point of tire failure, the pilot of 712 had a fork in the road: continue the takeoff, or reject the takeoff. Each decision had it's ups and downs:
Continue: Though the aircraft was 10 knots below V1, had the pilot decided to continue the takeoff, he'd possibly have been able to burn down/dump fuel, and recover with a lighter-weight aircraft rather than face a high-speed, heavyweight abort. On the downside, who knows what configuration the aircraft would have actually gotten airborne in, since it's unknown when precisely the right wing fuel cells were punctured. It may have been a crapshoot whether the plane would've gotten airborne with just mangled landing gear, or with leaking fuel that could possibly ignite. Also, even landing lighter weight after burning/dumping fuel, there's no guarantee that the right MLG bogie wouldn't have punctured the wing fuel cell after landing, possibly resulting in a fire too, though of much reduced severity.
Reject: In this situation, the crew would be faced with a high-speed and heavyweight abort, like they were. In this situation, the decision "tree" comes to another fork: max braking or rollout:
Max Braking: In RTO procedures, it's generally accepted that max braking be applied as part of the procedure, consistent with maintaining directional control. This provides for stopping quickly and allowing evacuation, if needed, to commence quickly. IMO, this is situation-dependant and part of "making the call" that pilots get paid for. Would max braking have prevented the destruction of the wheel rims and puncturing of the right wing fuel cell in this accident? It's anyone's guess. That's another crap shoot. What's to say that the same thing wouldn't have happened with the deceleration and brake forces on the bogie assembly as opposed to rolling out? But as before, RTO procedures need to be standardized, however IMO, like any emergency procedures, only those steps that are necessary to be taken should be taken, and this again comes back to pilot/crew decision making based on the situation at hand.
Rollout: The rollout decision was made based on the ACs experience as well as his desire to not affect directional control with braking, combined overall with the excellent amount of runway remaining. Again, IMO, we as pilots are paid to make the decision when the situation dictates it. As it was, the tire/rim assemblies came apart, punctured the wing fuel cell, and ignited a fire. Again, a crap shoot. The case of NASA 712 was a rarity, since not often are conditions such where critical field length is so much less than actual runway length, and a flightcrew has so much excess runway to play with.
As mentioned before, V1 philosophy has always been predicated on engine loss go/no go. V1 has evolved to become an abort speed for nearly any emergency that may require an abort. The biggest reasons for standardized abort procedures is simply due to the fact that a rote response is needed for times when the split-second decision needs to be made at high speed with little time, and the crew must respond in coordinated action. But should V1 training be modified to become somewhat situationally dependant for below-V1 continuations? Who knows. My opinion still stands that many RTOs are a gut call based on the situation at hand and the training/experience of the crew. There are many factors that the crew can control, yet many possibilities that the crew has no control over, again situational-dependant. For example, there are some emergencies where I'll easily abort past refusal speed. One example would be an engine failure in my former aircraft when the WX is low IFR and there's little chance I'd be able to work the problem in VMC. Any type of engine fire would be another one. Another example in my former aircraft would be dual generator failure with a takeoff that would put me in IMC (simply due to the myriad of problems that it would cause in-cockpit, and especially the fact that I'd be facing flight controls that have a limited lifetime left from battery-only; approximately 10-15 minutes before complete loss of FBW). But again, those are factors for my particular airplane and may or may not apply to another aircraft or another pilot; so as I often say, it really does depend. My decisions on these was also based on the fact that I had a drag chute and runway arresting gear available that I could use, but my point is that it's sometimes very situational-dependant.
As an aside to this accident, the 25 July 2000 crash of the Air France Concorde on takeoff from Charles DeGaulle airport cannot go unmentioned. In that crash, the crew elected to continue the takeoff, not knowing that they had a massive fire underneath the left wing. Like the crew of NASA 712, the first indication of fire was near end-game.....in the case of 712, near the end of their rollout; and in the case of the Concorde, likely just as they were getting airborne. In this pilots very humble opinion, this is a gut check of the toughest kind. My personal desire is to keep a bad aircraft on the ground if at all possible.....ie, not take a bad plane airborne. But that decision must be tempered with many other factors: where will the airplane end up if I abort and don't stop at the end of the runway? Where will it end up if I takeoff and can't keep it airborne? Is there excess runway available? These and many other questions are all part of my "accident tree" of decision and indecision, of things you can control and things you can't. As I said, every accident is different in specifics, though they may be similiar in generalities.
It is those times when Mr. Murphy is riding in your jumpseat, that all you can do is the best your training and experience provide you, and even that is little insurance-coverage against the multitude of monkey-wrenches Mr. Murphy can and will throw your way. The most anyone can do is give it their best, and hope the decisions they make at the moment, will take them down the correct fork in the "accident tree."
MikeD
The above is not intended to be an undue criticism of the person or persons involved in the incident described. Instead, the analysis presented is intended to further the cause of flight safety and help to reduce accidents and incidents by educating pilots through the sacrifices of others in our profession.
Last edited:
