An airplane is like an egg. There‘s no such thing as a fairly good one......
20 April 1985
Wilkes-Barre/Scranton International Airport (ICAO: KAVP), Pennsylvania
North American CT-39A Sabreliner, 62-4496, c/s TAC 01
Headquarters, Tactical Air command, USAF, Langley AFB, Virginia
5 Fatal
In this 14th installment of MikeD’s Accident Synopsis series, I discuss an accident that occurs at a time when most aircrews feel the mission is pretty much over and done with. When it comes to In Flight Emergencies (IFEs), there’s never really a “good” time to have one, per se. The general consensus is it’s always preferable to turn an IFE into a ground emergency, if at all possible, the thought being that it will be far easier to take care of said emergency on the ground rather than in-flight. Of course, like any rule of thumb, there are no absolutes; and it is no different with this rule either as accidents such as Saudia 163 in 1980 demonstrated. But in terms of the question of better times for an emergency to occur, especially one that requires quick reaction time, vice one with time to deal with it; being on the ground is generally preferable. Of course there are ground emergencies that can have their own serious challenges, such as a Rejected Take Off (RTO), as was demonstrated in the British AirTours 737 RTO at Manchester, UK, as well as the NASA Convair 990 at March AFB, California [which I‘ve previously written a synopsis on]; both having occurred in 1985. Or even ground emergencies that can happen while taxiing and subsequent runway incursion, as happened in the Northwest 299/1482 ground collision at Detroit in 1990. Landing accidents that occur following a successful touchdown and rollout, barring any outside factors apart from the involved aircraft such as too short of a runway, or a runway incursion by another aircraft or vehicle, etc; aren’t exceptionally common. Far more often than not, once a successful landing has been accomplished, getting slowed, turned off the runway, and taxiing to one’s destination, usually has little risk: the majority of the risk having already passed. But exceptions to the rule have occurred and will continue to, and this was the case that the crew of a USAF CT-39A Sabreliner utility jet faced on landing rollout at the Wilkes-Barre/Scranton International Airport (ICAO: KAVP) on the late afternoon of 20 April 1985, in an accident that killed a USAF 4-star General, his wife, two pilots, and an enlisted crew chief.
At the end of Fiscal Year (FY) 1984, General Jerome F. O’Malley was at the pinnacle of leadership in the U.S. Air Force. A career bomber and reconnaissance pilot, O’Malley had piloted the Boeing B-47 Stratojet strategic bomber, been the first-flight pilot of the Lockheed SR-71A Blackbird supersonic reconnaissance jet, had done two tours in the Vietnam theatre in separate reconnaissance Wings in both South Vietnam and Thailand flying both the McDonnell-Douglas RF-4C and F-4D Phantom II combat jets, had commanded Pacific Air Forces (PACAF) in Hawaii, and was now succeeding General Wilbur Creech as head of Tactical Air Command (TAC), the Major Command (MAJCOM) in the USAF that commanded all continental USA-based fighter/attack/ reconnaissance aircraft, as well as other types. By now, having held this top MAJCOM post for nearly 8 months, General O’Malley was well-placed to move onto becoming the next Chief of Staff of the USAF, the top military officer spot in the organization.
One of the benefits of being a MAJCOM commander, was having one’s own assigned aircraft, for use in getting around the command to the multitude of TAC bases located around the continental United States (CONUS), as well as to and from Andrews AFB (KADW) for visits to the Pentagon and Washington D.C, and to anywhere else the TAC commander would need to go for official functions. During this 1984-1985 timeframe, the USAF was in a slow transition phase in replacing the venerable North-American T-39A Sabreliner jet, a variant of the civilian Sabreliner 40 series, which had been in use since the 1960s; with the C-21A Learjet, a version of the civilian Learjet model 35. While this transition was in the middle of taking place, the TAC Commander’s aircraft was still a T-39A, one of the best-maintained and cleanest T-39s in the entire USAF: a 1962-acquired model, serial number 62-4496, and flying under the callsign “TAC-01“. An additional benefit of being a MAJCOM commander as a pilot-rated officer, was being able to get qualified in and to pilot your own assigned utility aircraft, albeit with an Instructor Pilot (IP) on that type-aircraft flying along as co-pilot (but as Aircraft Commander of record), per USAF regulations.
In April 1985 during a busy month of official travel, General O’Malley had been invited along with his wife, as keynote speaker for a Boy Scouts of America dinner banquet in Scranton, Pennsylvania, O’Malley’s hometown, on Sunday 21 April. The banquet was in honor of childhood acquaintance and United States Congressman Joseph McDade (R-PA). To make the Sunday banquet date, O’Malley planned to depart his home base of Langley AFB (ICAO: KLFI) on the late afternoon of Saturday 20 April, to Wilkes-Barre/Scranton International Airport, where family and officials awaited his arrival. Departing at approximately 2155Z from KLFI, CT-39 TAC-01 had onboard O’Malley in the pilot seat, an IP Captain in the co-pilot seat; and in the passenger compartment rode an additional Lieutenant-Colonel pilot to round out the required 2-pilot flight crew for the CT-39s return trip to KLFI following dropping off General O’Malley and his wife at Scranton, as well as the CT-39s assigned dedicated crew chief, the lead maintenance person responsible for the jet itself.
Flight time was less than one hour to get from Langley to Wilkes-Barre. Takeoff, enroute leg, and arrival into the Wilkes-Barre terminal area was uneventful, and KAVP was reporting 5000 Scattered with 10+ miles visibility, and winds out of the north/northeast at 4 knots. TAC-01 was assigned a visual approach by Wilkes-Barre TRACON to RW 04, which was (at the time) 6001’x 150’ length (now 7501'). RW 4/22 at KAVP has a negligible gradient to it, and is pretty much like any other runway of it’s length and width, with the exception of having fairly steep terrain drop-offs past each runway end. At the computed landing weight of the CT-39, the T.O. 1T-39A-1 flight manual showed an approximately 115 knot final approach speed, with 2180’ of rollout being required to stop, following a touchdown point 1000’-1200’ in the runway’s touchdown zone, resulting in approximately 3200’-3400’ of total runway used from the approach end.
Accordingly, a normal landing should easily have been able to be made using less than half of the available length of RW 4, with a left turn at taxiway D to exit and taxi to the ramp. Handed off to the tower on final, TAC-01 checked in with Wilkes-Barre tower and was cleared to land. TAC-01 touched down on RW 4 at 2248Z in by what witness accounts described as a normal landing, minus speed brakes, of which use was optional for landings. TAC-01 continued its rollout however with no apparent braking action noticed by ground witnesses. In the cockpit of the CT-39, the crew experienced brake failure, and as later evidenced by overhead panel switch positions found in the wreckage, the T-39s hydraulic pump switch was switched to OFF by the crew and the Auxiliary Hydraulic Power Switch was set to ON, with still no braking action apparent [The early T-39 series were not equipped with thrust reversers]. Still rolling at approximately 95 knots, and with over half of the runway consumed already from attempting to stop with both the main as well as the auxiliary braking systems to no avail, the crew pulled the emergency brakes T-handle located on the center console, and effective braking was noted from skid marks on the runway at approximately 750’ remaining and 17’ left of runway centerline. The CT-39s ground path began to correct back to runway centerline, and the aircraft departed the end of RW 4 at approximately 50 knots, rolled through some sod, and became airborne again momentarily as it went over the 125' steep terrain drop off at the end of the runway. Crashing down into the terrain bottom, the CT-39’s fuselage ruptured, the aircraft broke up, and immediately erupted into an intense post-crash fuel fire. The local controller, having witnessed the accident sequence, initiated the crash alarm from the tower as the CT-39 was nearing the end of the runway. At the fire station, located between the General Aviation ramp and the passenger terminal on the field, a medium CFR/ARFF truck with two firefighters responded a little over 30 seconds after the klaxon sounded in their station, and arrived on scene within 2 minutes. However the CT-39 was engulfed in flames, and there were challenges in immediately accessing the wreckage in order to begin firefighting efforts due to the terrain. All 5 persons onboard the jet suffered fatal injuries ranging from blunt force trauma, to thermal burns and smoke inhalation. The aircraft was destroyed.
Probable Cause:
*Brake Failure- Anti-Skid System- Total/All
*Brake Failure- Hydraulic Main/Auxiliary- Partial
*Anti-Skid/Brake Failure Checklist- Incorrect Application- Flight Crew, 62-4496
Secondary Factors:
*Emergency Brake System- Late Initiation- Flight Crew, 62-4496
*Maintenance- Anti-Skid Control Valve- Overhaul Improper- Aircraft Depot
Tertiary Factors:
*Maintenance- Hydraulic Accumulators, Main/Auxiliary- Overhaul Improper- Aircraft Depot
*Maintenance- Recordkeeping- Poor/Faulty- Dedicated Crew Chief 62-4496
*Terrain- Rough/Uneven/Steep
*Runway Arresting System- Unavailable
MikeD says:
This accident brings up four specific areas of discussion related to the causal factors listed above:
1. The CT-39A hydraulic and normal/emergency braking system
2. 62-4496 Flight crew actions on landing rollout
3. 62-4496 Dedicated Crew Chief maintenance recordkeeping deficiencies
4. Maintenance Depot Overhaul Discrepancies
The CT-39A Hydraulic and Normal/Emergency Braking System: The North American T-39A-model series of aircraft utilize a hydraulic system consisting of a 3000 psi main system and a 3000 psi auxiliary accumulator system. The main hydraulic system operates the landing gear, main gear inner doors, speed brake, wheel brakes, and nose wheel steering (NWS) via a hydraulic pump and an accumulator that is pressurized by engine bleed air. The hydraulic pump is electrically actuated whenever system pressure is below 2700 psi, and terminates operation when the system pressure reaches 3000 psi again. The auxiliary accumulator system is a backup system in the event of main system failure and must be manually activated in order to power the speed brake, wheel brakes, and NWS systems only, and provides sufficient system pressure to complete a landing were the main hydraulic system to fail. The auxiliary system pressure should only be used (system switched ON) only as-necessary and required in order to preserve auxiliary accumulator pressure. The ON/OFF switches for both systems are on the center overhead panel, while their respective pressure gauges are located on the bottom-center of the main instrument panel. In normal operation, the Main system is ON, and the Auxiliary system switch is OFF.
The wheel brake system on the T-39 series aircraft is normally run through the main hydraulic system, and can also be run through the auxiliary hydraulic accumulator when that system is activated. Additionally, there is a manual/emergency braking system activated when the emergency brake control T-handle on the center console is pulled out, and this system has its own fluid supply from two small reservoirs located in the nose gear well, with separate brake lines to the wheel brakes. One critical limitation of this emergency system, and a limitation that’s covered by a WARNING in the USAF Technical Order T.O. 1T-39A-1 operations manual, is that upon activation of the emergency braking system, the brakes for both the pilot and co-pilot run through the same brake control valve. Only one pilot should attempt to pump the brakes of the emergency system when activated, since with both pilots pressing the brakes together when this system is in use results in each pedal canceling one-another out, and ends up with no braking action occurring. The use of the emergency braking system therefore, requires some degree of crew coordination in order to prevent this anomaly from occurring.
The antiskid system portion of the wheel brakes is electrically actuated by a switch located just above the hydraulic system switches on the overhead center panel, and hydraulically operated through the main hydraulic system only, and via the left and right antiskid control valves to their respective wheel brakes. In the event of antiskid system power failure or complete electrical failure, normal brakes are still available to the flight crew.
During all ground operations, primarily taxi operations, there is a specific NOTE in T.O. 1T-39A-1 requiring the co-pilot to monitor hydraulic pressure. Ironically, there have been at least two taxiing accidents involving Sabreliner aircraft where hydraulic pressure was lost from the main hydraulic system: one in February 2008 at Fort Lauderdale-Executive (ICAO: KFXE) involving a Sabreliner 80 where the main system failed, the emergency system was improperly operated by the flight crew and the plane collided with a few other aircraft; and one incident in July 2013 involving a Sabreliner 65 at Las Vegas-McCarran (ICAO: KLAS), where the main hydraulic system had failed on final approach, the auxiliary system was activated and an uneventful landing and runway exit was made, but on taxi-in, the brakes and NWS ceased working and the flight crew lost control of the aircraft, with it striking a metal pole.
62-4496 Flight Crew Actions on Landing Rollout: One of the limitations accident investigators faced with the accident of 62-4496, was that there was no Cockpit Voice Recorder (CVR) or Flight Data Recorder (FDR) information available, as these items weren’t required by regulations back then, and thus weren’t installed. So the only verified information regarding flight crew action(s) or inactions(s) can be gleaned by the cockpit switch positions for the main and auxiliary hydraulic systems and antiskid system, by witness testimony, and by tire skid marks at the departure end of RW 4. All else would be conjecture. As will be discussed later on in this synopsis, 62-4496 had a history of brake and antiskid problems which had been known by both maintenance as well as the flight crew, as they recurred here and there. The problem resurfaced during landing at KAVP, and appears to have been the worst occurrence, and ultimately the fatal one. From T.O. 1T-39A-1, the procedure for antiskid/brake failure is described as such:
“Braking action failure may be caused by mechanical malfunction, normal hyraulic system failure, or a malfunction of the antiskid system (T.O 1T-39-888 C/W) [the Tech Order complied with regarding antiskid system installation]. Since diagnosis of particular circumstances may be difficult, the following procedure ensures that, regardless of cause, actions are taken sequentially to provide for the various contingencies:
WARNING: The ANTISKID INOP light indicates only a limited number of failure modes. If abnormal brake operation is encountered at any time, use the following procedure:
1. BRAKES- RELEASE
2. ANTISKID- OFF [items 1 and 2 being boldface/immediate action]
3. Brakes- Apply
If no brakes, proceed to Step 4.
4. Aux Hydraulic- ON
5. Brakes- Apply
If auxiliary system fails, proceed to Step 6.
6. Emergency Brake T-Handle- Pull
7. Brakes- Apply
Apply rapid, intermittent pressure to the top of the brake pedals to pump up the brakes. A few applications will be required to obtain pressure for emergency braking. Do not release the pedals completely, otherwise the pumping action will have to be repeated.
WARNING: When the emergency brake system is used, only one pilot will operate and pump the brakes to ensure effective braking.
NOTE: Do not attempt to taxi on the emergency brake system, other than to clear the runway.”
Based upon the cockpit switch positions found in the wreckage of the overhead panel, the anti-skid switch was in the ON position, the main hydraulic switch was in the OFF/RESET position, and the auxiliary hydraulic switch was in the ON position. The emergency brake T-Handle was pulled. Based on these findings, the boldface critical/immediate action items were not accomplished correctly, specifically with item 2 not being accomplished. As it turns out and will be explained shortly, item 2 was the one item that if accomplished, would’ve allowed the brakes to work again, due to there being a specific malfunction having been found within the anti-skid system of 62-4496, specifically with the right anti-skid valve. The failure to accomplish item 2, prevented items 3 thru 5 from taking effect. Late application of item 6, the emergency brake T-Handle, made application of item 7 late due to the need to pump up the brakes 3 to 5 times before item 7, brake application, could be successfully initiated. And item 6 was additionally dependant on good crew coordination in the heat of the moment, to ensure that both pilot and co-pilot didn’t attempt to pump the brakes concurrently, thus cancelling out the brakes altogether. Whether this occurred or not prior to the runway skid marks at the 750’ remaining point where braking first took effect, is unknown, due to the absence of any CVR/FDR data. As such, braking began far too late to prevent exiting the runway at the departure end with significant speed still, and proceeding down the steep embankment thereafter.
62-4496 Dedicated Crew Chief Maintenance Recordkeeping Deficiencies: Along with the aforementioned privilege of the MAJCOM commander having his own dedicated utility transport jet, he also had his own dedicated crew chief (DCC) for that aircraft. The DCC is the primary enlisted maintenance person responsible for the upkeep and maintenance status of a particular USAF aircraft. The DCC can be a ground crew chief, such as on fighter aircraft, or can be a flying crew chief, such as is found on transport aircraft when they are flying cross-country. The DCC keeps his own tools and small parts supply onboard the jet, and supervises any maintenance done by others on the aircraft, ensuring and being fully knowledgeable of every discrepancy, big or small, for his aircraft. Normally helped by an assistant crew chief (ACC), the DCC takes pride in his aircraft, since on combat aircraft and on some support aircraft, his/her name will be on the aircraft.
The DCC for 62-4496 had been selected by General Creech, and was the DCC who came with the aircraft when it was inherited by General O’Malley. The DCC was very familiar with all the intricacies of 62-4496, and was given a form of carte blanche to do what was needed to keep it in top shape, as well as any priority he needed for parts or equipment. 62-4496 was kept maintained and clean/polished at all times. But there was a significant item that wasn’t being accomplished correctly, that had it been, it could’ve highlighted a trend that had been developing with 62-4496 for a number of months, following an inspection that had been accomplished at an overhaul facility at Andrews AFB (ICAO: KADW) at the beginning of the Calendar Year (CY), which will be discussed more in-depth later in this synopsis.
For any USAF aircraft, the complete set of aircraft forms that every USAF aircraft is required to have, is known as the 781 book. This particular book is divided into sections, denoted by a specific suffix. The 781H, for instance, known as the Aerospace Vehicle Flight Status and Maintenance, is the primary form at the front of the book that shows what the aircraft’s status is: is it up, is it down. Does it have an inspection due, were the daily inspections complied with, and what is the aircraft total time, and individual sortie times filled in below that. Other 781 forms for other purposes also exist within that book.
The DCC for 62-4496 did not maintain a specific form known as a 781A, although he did maintain the other parts of the 781 manual, for reasons unknown. The 781A, known as the Maintenance Discrepancy and Work Document, is a form that is used to record specific problems encountered with the airplane, or any work being done to the plane, as well as how it affects the up/down status of the airplane, what actions were taken to fix/clear the specific discrepancy, and who signed it off. This form is the best to show items that potentially trend, as the normal pilot can thumb back usually a few pages and see what other items have occurred, as well as what was done about them, if anything. What the DCC did maintain was a personal “discrepancy log” of his own for 62-4496 which, while nearly identical to a 781A in terms of what a problem was, what the correction action was, and when it happened; he didn’t keep his in the book with the rest of the 781 forms, therefore no one other than himself really knew what was up with 62-4496 at any given time, nor could they potentially be the “second set of eyes” to see the trending information the DCC was dealing with regarding failures of the braking system over the period of time from when the plane had gone through the depot facility at Andrews AFB, until the week of the fatal crash. Nor could the normal “quality Assurance” (QA) maintenance inspectors take a look at the aircraft and attempt to spot trends, due to the same lack of information in the 781 books. Only the DCC had this information, and he was able to do business this way, directly contrary to USAF regulations at the time, because of the latitude extended to him by both General Creech as well as General O’Malley regarding doing whatever he saw fit in the maintenance of 62-4496.
Some of the specific notes found in the DCCs personal log post-accident, contained the following items, with emphasis on aircraft brake and anti-skid problems that were recurring multiple times over 34 flights following a time-phase inspection conducted in January at the depot at Andrews AFB, are cited below:
- 14 January – Aircraft returned to Langley AFB after Isochronal at Andrews AFB [on 5 January] (notably major work was done on the brake system including change-out of both normal and auxiliary hydraulic accumulators).
- 18 January – Aircraft pulled to the left on landing.
- 30 January – Anti-skid kicked off. (General O’Malley was in the pilot’s seat and had trouble holding the aircraft as he applied power prior to take off.)
- 1 February – Still having problems with anti-skid.
- 15 February – Changed “O” ring on power brake control valve.
- 1 March – Replaced “O” Ring on power brake control valve fitting; it has been leaking again.
- 12 March – Pilot’s brake pedals failed to stop aircraft on touchdown, master caution light illuminated (co-pilot was able to get effective braking action).
Had the DCC been entering these discrepancies into the correct Form 781A, either in addition to, or in lieu of his personal logbook; there’s a good chance that the above trends could’ve been spotted a mechanic working on the aircraft, a QA inspector, a pilot, or nearly any maintenance person who knew how to read the forms. As is, many of these trends, particularly the brake issues, stemmed from problems with shoddy workmanship encountered from the USAF depot where 62-4496 had received a phase inspection in January, 3 months prior to the accident. For reasons unknown, the DCC for 62-4496 either didn’t notice a trending problem, or else didn’t find it significant enough to do anything about it.
Maintenance Depot Overhaul Discrepancies: On 5 January 1985, 62-4496 went to KADW to receive a 6 day time-phased inspection where certain time-phased parts were being replaced as necessary, as well as other inspection items taking place. On 11 January, 62-4496 returned to KLFI, and as short as a few days later, problems began recurring with the braking and anti-skid systems of the aircraft, among other things. Found in the wreckage, many of these parts were recovered, and as they were torn down for inspection by investigators, shoddy workmanship began to be discovered in certain critical parts that are primary causal factors to this accident. The primary causal factor, failure of the main/auxiliary brake system, was traced to the right anti-skid control valve of 62-4496. Teardown analysis of this valve post-accident, revealed that when it was overhauled and rebuilt during it’s inspection at KADW, it was incorrectly reassembled and adjusted such that a pole met an armature in the valve, thus not allowing brake pressure to be achieved within the system. Why this had not manifested itself to the extent it did on the fatal flight, is anyone’s guess; but it did show that there was a trending problem occurring, even though each fix was believed to have fixed the problem, when in reality, those only masked the true culprit. The failure of this anti-skid control valve would’ve still occurred the day of the accident, however had the crew of 62-4496 correctly followed the checklist for anti-skid/brake failure, specifically item 2 regarding turning off the anti-skid switch, that would’ve removed the anti-skid valve from the equation, and allowed hydraulic pressure to build, and pressure would’ve been available when the auxiliary control switch was selected ON by the crew, if not earlier.
Additional factors found regarding shoddy workmanship, but not causal to the accident itself, were the hydraulic accumulators for both the main hydraulic system, as well as the auxiliary system. The main system accumulator had an improperly installed backup O-ring assembly, while the auxiliary system had a damaged piston, determined to have been caused due to damage in the reassembly phase prior to its reinstallation in 62-4496. Neither of these factors were causal to the accident, as when tested, both hydraulic accumulators operated normally. So these are tertiary findings, but still factors nonetheless, even though they were not causal.
MikeDs Final Thoughts: This accident is one of those where although the accident occurred through a series of maintenance discrepancies, there was still a “way out” had the crew executed the emergency checklist appropriately and covered the initial item that would’ve taken care of the rest in one fell swoop: the anti-skid being switched off. Barring that, they still had another trick up their sleeve with the emergency brake handle, but it’s late application didn’t allow time to stop this already fast moving emergency, figuratively and literally speaking. The items in question were critical action items, and were likely so for good reason, what the North American company described in the narrative prior to the procedure as “….actions taken sequentially,” pending they are indeed accomplished “in sequence”.
Once the emergency brakes were activated, it was far too late to stop on the remaining runway. At nearly most other airports, a runway overrun like this would only result in no-damage at best, and maybe serious damage at worse, as the aircraft rolled through the sod. KAVP unfortunately, has the severe terrain drop offs at each end of the runway, a place that the crew of 62-4496 couldn’t avoid going once they found the proverbial square corner they were in following getting the late activation of the emergency brakes. For it’s part, KAVP….due to it’s terrain drop off’s at the end of the runway….became in 2008 one of 45 planned and funded airports to receive the Engineered Material Arresting System (EMAS). Where runway arresting systems have been in use for decades ever since the advent of jets decades ago, at both military as well as joint-use civilian fields for jets both with as well as without tail hooks; these systems haven‘t been prevalent at civilian fields, even the non-tailhook pop-up systems that catch the main landing gear of an aircraft such as the MA-1A. Somewhat understandable as they are expensive to both install, as well as have trained crews to test them and maintain them. EMAS, on the other hand, gives a cheap and easy alternative incorporating collapsible and crushable concrete at the runway overruns for a certain distance that the overrun aircraft can enter and decelerate through as it sinks through the crushable materials. The system works much in the same way that the sand on the runaway truck ramps located at the bottom of highway downhill segments/grades work. Had an EMAS-style system been engineered and available 23 years prior to its installation, CT-39A 62-4496 would’ve been nothing more than a safety hazard footnote occurrence in the history of the USAF, vice the major tragedy it became, and the reminder it gave us that indeed The Runway to Hell is Paved With Good Intentions.
Incidentally, a landing overrun accident of a civil Saberliner 60 occurred 7 November 1992 at Phoenix-Sky Harbor International Airport (KPHX), in nearly identical dynamics as the accident here of CT-39A 62-4496.
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.
USAF CT-39A 62-4496:
20 April 1985
Wilkes-Barre/Scranton International Airport (ICAO: KAVP), Pennsylvania
North American CT-39A Sabreliner, 62-4496, c/s TAC 01
Headquarters, Tactical Air command, USAF, Langley AFB, Virginia
5 Fatal
In this 14th installment of MikeD’s Accident Synopsis series, I discuss an accident that occurs at a time when most aircrews feel the mission is pretty much over and done with. When it comes to In Flight Emergencies (IFEs), there’s never really a “good” time to have one, per se. The general consensus is it’s always preferable to turn an IFE into a ground emergency, if at all possible, the thought being that it will be far easier to take care of said emergency on the ground rather than in-flight. Of course, like any rule of thumb, there are no absolutes; and it is no different with this rule either as accidents such as Saudia 163 in 1980 demonstrated. But in terms of the question of better times for an emergency to occur, especially one that requires quick reaction time, vice one with time to deal with it; being on the ground is generally preferable. Of course there are ground emergencies that can have their own serious challenges, such as a Rejected Take Off (RTO), as was demonstrated in the British AirTours 737 RTO at Manchester, UK, as well as the NASA Convair 990 at March AFB, California [which I‘ve previously written a synopsis on]; both having occurred in 1985. Or even ground emergencies that can happen while taxiing and subsequent runway incursion, as happened in the Northwest 299/1482 ground collision at Detroit in 1990. Landing accidents that occur following a successful touchdown and rollout, barring any outside factors apart from the involved aircraft such as too short of a runway, or a runway incursion by another aircraft or vehicle, etc; aren’t exceptionally common. Far more often than not, once a successful landing has been accomplished, getting slowed, turned off the runway, and taxiing to one’s destination, usually has little risk: the majority of the risk having already passed. But exceptions to the rule have occurred and will continue to, and this was the case that the crew of a USAF CT-39A Sabreliner utility jet faced on landing rollout at the Wilkes-Barre/Scranton International Airport (ICAO: KAVP) on the late afternoon of 20 April 1985, in an accident that killed a USAF 4-star General, his wife, two pilots, and an enlisted crew chief.
At the end of Fiscal Year (FY) 1984, General Jerome F. O’Malley was at the pinnacle of leadership in the U.S. Air Force. A career bomber and reconnaissance pilot, O’Malley had piloted the Boeing B-47 Stratojet strategic bomber, been the first-flight pilot of the Lockheed SR-71A Blackbird supersonic reconnaissance jet, had done two tours in the Vietnam theatre in separate reconnaissance Wings in both South Vietnam and Thailand flying both the McDonnell-Douglas RF-4C and F-4D Phantom II combat jets, had commanded Pacific Air Forces (PACAF) in Hawaii, and was now succeeding General Wilbur Creech as head of Tactical Air Command (TAC), the Major Command (MAJCOM) in the USAF that commanded all continental USA-based fighter/attack/ reconnaissance aircraft, as well as other types. By now, having held this top MAJCOM post for nearly 8 months, General O’Malley was well-placed to move onto becoming the next Chief of Staff of the USAF, the top military officer spot in the organization.
One of the benefits of being a MAJCOM commander, was having one’s own assigned aircraft, for use in getting around the command to the multitude of TAC bases located around the continental United States (CONUS), as well as to and from Andrews AFB (KADW) for visits to the Pentagon and Washington D.C, and to anywhere else the TAC commander would need to go for official functions. During this 1984-1985 timeframe, the USAF was in a slow transition phase in replacing the venerable North-American T-39A Sabreliner jet, a variant of the civilian Sabreliner 40 series, which had been in use since the 1960s; with the C-21A Learjet, a version of the civilian Learjet model 35. While this transition was in the middle of taking place, the TAC Commander’s aircraft was still a T-39A, one of the best-maintained and cleanest T-39s in the entire USAF: a 1962-acquired model, serial number 62-4496, and flying under the callsign “TAC-01“. An additional benefit of being a MAJCOM commander as a pilot-rated officer, was being able to get qualified in and to pilot your own assigned utility aircraft, albeit with an Instructor Pilot (IP) on that type-aircraft flying along as co-pilot (but as Aircraft Commander of record), per USAF regulations.
In April 1985 during a busy month of official travel, General O’Malley had been invited along with his wife, as keynote speaker for a Boy Scouts of America dinner banquet in Scranton, Pennsylvania, O’Malley’s hometown, on Sunday 21 April. The banquet was in honor of childhood acquaintance and United States Congressman Joseph McDade (R-PA). To make the Sunday banquet date, O’Malley planned to depart his home base of Langley AFB (ICAO: KLFI) on the late afternoon of Saturday 20 April, to Wilkes-Barre/Scranton International Airport, where family and officials awaited his arrival. Departing at approximately 2155Z from KLFI, CT-39 TAC-01 had onboard O’Malley in the pilot seat, an IP Captain in the co-pilot seat; and in the passenger compartment rode an additional Lieutenant-Colonel pilot to round out the required 2-pilot flight crew for the CT-39s return trip to KLFI following dropping off General O’Malley and his wife at Scranton, as well as the CT-39s assigned dedicated crew chief, the lead maintenance person responsible for the jet itself.
Flight time was less than one hour to get from Langley to Wilkes-Barre. Takeoff, enroute leg, and arrival into the Wilkes-Barre terminal area was uneventful, and KAVP was reporting 5000 Scattered with 10+ miles visibility, and winds out of the north/northeast at 4 knots. TAC-01 was assigned a visual approach by Wilkes-Barre TRACON to RW 04, which was (at the time) 6001’x 150’ length (now 7501'). RW 4/22 at KAVP has a negligible gradient to it, and is pretty much like any other runway of it’s length and width, with the exception of having fairly steep terrain drop-offs past each runway end. At the computed landing weight of the CT-39, the T.O. 1T-39A-1 flight manual showed an approximately 115 knot final approach speed, with 2180’ of rollout being required to stop, following a touchdown point 1000’-1200’ in the runway’s touchdown zone, resulting in approximately 3200’-3400’ of total runway used from the approach end.
Accordingly, a normal landing should easily have been able to be made using less than half of the available length of RW 4, with a left turn at taxiway D to exit and taxi to the ramp. Handed off to the tower on final, TAC-01 checked in with Wilkes-Barre tower and was cleared to land. TAC-01 touched down on RW 4 at 2248Z in by what witness accounts described as a normal landing, minus speed brakes, of which use was optional for landings. TAC-01 continued its rollout however with no apparent braking action noticed by ground witnesses. In the cockpit of the CT-39, the crew experienced brake failure, and as later evidenced by overhead panel switch positions found in the wreckage, the T-39s hydraulic pump switch was switched to OFF by the crew and the Auxiliary Hydraulic Power Switch was set to ON, with still no braking action apparent [The early T-39 series were not equipped with thrust reversers]. Still rolling at approximately 95 knots, and with over half of the runway consumed already from attempting to stop with both the main as well as the auxiliary braking systems to no avail, the crew pulled the emergency brakes T-handle located on the center console, and effective braking was noted from skid marks on the runway at approximately 750’ remaining and 17’ left of runway centerline. The CT-39s ground path began to correct back to runway centerline, and the aircraft departed the end of RW 4 at approximately 50 knots, rolled through some sod, and became airborne again momentarily as it went over the 125' steep terrain drop off at the end of the runway. Crashing down into the terrain bottom, the CT-39’s fuselage ruptured, the aircraft broke up, and immediately erupted into an intense post-crash fuel fire. The local controller, having witnessed the accident sequence, initiated the crash alarm from the tower as the CT-39 was nearing the end of the runway. At the fire station, located between the General Aviation ramp and the passenger terminal on the field, a medium CFR/ARFF truck with two firefighters responded a little over 30 seconds after the klaxon sounded in their station, and arrived on scene within 2 minutes. However the CT-39 was engulfed in flames, and there were challenges in immediately accessing the wreckage in order to begin firefighting efforts due to the terrain. All 5 persons onboard the jet suffered fatal injuries ranging from blunt force trauma, to thermal burns and smoke inhalation. The aircraft was destroyed.
Probable Cause:
*Brake Failure- Anti-Skid System- Total/All
*Brake Failure- Hydraulic Main/Auxiliary- Partial
*Anti-Skid/Brake Failure Checklist- Incorrect Application- Flight Crew, 62-4496
Secondary Factors:
*Emergency Brake System- Late Initiation- Flight Crew, 62-4496
*Maintenance- Anti-Skid Control Valve- Overhaul Improper- Aircraft Depot
Tertiary Factors:
*Maintenance- Hydraulic Accumulators, Main/Auxiliary- Overhaul Improper- Aircraft Depot
*Maintenance- Recordkeeping- Poor/Faulty- Dedicated Crew Chief 62-4496
*Terrain- Rough/Uneven/Steep
*Runway Arresting System- Unavailable
MikeD says:
This accident brings up four specific areas of discussion related to the causal factors listed above:
1. The CT-39A hydraulic and normal/emergency braking system
2. 62-4496 Flight crew actions on landing rollout
3. 62-4496 Dedicated Crew Chief maintenance recordkeeping deficiencies
4. Maintenance Depot Overhaul Discrepancies
The CT-39A Hydraulic and Normal/Emergency Braking System: The North American T-39A-model series of aircraft utilize a hydraulic system consisting of a 3000 psi main system and a 3000 psi auxiliary accumulator system. The main hydraulic system operates the landing gear, main gear inner doors, speed brake, wheel brakes, and nose wheel steering (NWS) via a hydraulic pump and an accumulator that is pressurized by engine bleed air. The hydraulic pump is electrically actuated whenever system pressure is below 2700 psi, and terminates operation when the system pressure reaches 3000 psi again. The auxiliary accumulator system is a backup system in the event of main system failure and must be manually activated in order to power the speed brake, wheel brakes, and NWS systems only, and provides sufficient system pressure to complete a landing were the main hydraulic system to fail. The auxiliary system pressure should only be used (system switched ON) only as-necessary and required in order to preserve auxiliary accumulator pressure. The ON/OFF switches for both systems are on the center overhead panel, while their respective pressure gauges are located on the bottom-center of the main instrument panel. In normal operation, the Main system is ON, and the Auxiliary system switch is OFF.
The wheel brake system on the T-39 series aircraft is normally run through the main hydraulic system, and can also be run through the auxiliary hydraulic accumulator when that system is activated. Additionally, there is a manual/emergency braking system activated when the emergency brake control T-handle on the center console is pulled out, and this system has its own fluid supply from two small reservoirs located in the nose gear well, with separate brake lines to the wheel brakes. One critical limitation of this emergency system, and a limitation that’s covered by a WARNING in the USAF Technical Order T.O. 1T-39A-1 operations manual, is that upon activation of the emergency braking system, the brakes for both the pilot and co-pilot run through the same brake control valve. Only one pilot should attempt to pump the brakes of the emergency system when activated, since with both pilots pressing the brakes together when this system is in use results in each pedal canceling one-another out, and ends up with no braking action occurring. The use of the emergency braking system therefore, requires some degree of crew coordination in order to prevent this anomaly from occurring.
The antiskid system portion of the wheel brakes is electrically actuated by a switch located just above the hydraulic system switches on the overhead center panel, and hydraulically operated through the main hydraulic system only, and via the left and right antiskid control valves to their respective wheel brakes. In the event of antiskid system power failure or complete electrical failure, normal brakes are still available to the flight crew.
During all ground operations, primarily taxi operations, there is a specific NOTE in T.O. 1T-39A-1 requiring the co-pilot to monitor hydraulic pressure. Ironically, there have been at least two taxiing accidents involving Sabreliner aircraft where hydraulic pressure was lost from the main hydraulic system: one in February 2008 at Fort Lauderdale-Executive (ICAO: KFXE) involving a Sabreliner 80 where the main system failed, the emergency system was improperly operated by the flight crew and the plane collided with a few other aircraft; and one incident in July 2013 involving a Sabreliner 65 at Las Vegas-McCarran (ICAO: KLAS), where the main hydraulic system had failed on final approach, the auxiliary system was activated and an uneventful landing and runway exit was made, but on taxi-in, the brakes and NWS ceased working and the flight crew lost control of the aircraft, with it striking a metal pole.
62-4496 Flight Crew Actions on Landing Rollout: One of the limitations accident investigators faced with the accident of 62-4496, was that there was no Cockpit Voice Recorder (CVR) or Flight Data Recorder (FDR) information available, as these items weren’t required by regulations back then, and thus weren’t installed. So the only verified information regarding flight crew action(s) or inactions(s) can be gleaned by the cockpit switch positions for the main and auxiliary hydraulic systems and antiskid system, by witness testimony, and by tire skid marks at the departure end of RW 4. All else would be conjecture. As will be discussed later on in this synopsis, 62-4496 had a history of brake and antiskid problems which had been known by both maintenance as well as the flight crew, as they recurred here and there. The problem resurfaced during landing at KAVP, and appears to have been the worst occurrence, and ultimately the fatal one. From T.O. 1T-39A-1, the procedure for antiskid/brake failure is described as such:
“Braking action failure may be caused by mechanical malfunction, normal hyraulic system failure, or a malfunction of the antiskid system (T.O 1T-39-888 C/W) [the Tech Order complied with regarding antiskid system installation]. Since diagnosis of particular circumstances may be difficult, the following procedure ensures that, regardless of cause, actions are taken sequentially to provide for the various contingencies:
WARNING: The ANTISKID INOP light indicates only a limited number of failure modes. If abnormal brake operation is encountered at any time, use the following procedure:
1. BRAKES- RELEASE
2. ANTISKID- OFF [items 1 and 2 being boldface/immediate action]
3. Brakes- Apply
If no brakes, proceed to Step 4.
4. Aux Hydraulic- ON
5. Brakes- Apply
If auxiliary system fails, proceed to Step 6.
6. Emergency Brake T-Handle- Pull
7. Brakes- Apply
Apply rapid, intermittent pressure to the top of the brake pedals to pump up the brakes. A few applications will be required to obtain pressure for emergency braking. Do not release the pedals completely, otherwise the pumping action will have to be repeated.
WARNING: When the emergency brake system is used, only one pilot will operate and pump the brakes to ensure effective braking.
NOTE: Do not attempt to taxi on the emergency brake system, other than to clear the runway.”
Based upon the cockpit switch positions found in the wreckage of the overhead panel, the anti-skid switch was in the ON position, the main hydraulic switch was in the OFF/RESET position, and the auxiliary hydraulic switch was in the ON position. The emergency brake T-Handle was pulled. Based on these findings, the boldface critical/immediate action items were not accomplished correctly, specifically with item 2 not being accomplished. As it turns out and will be explained shortly, item 2 was the one item that if accomplished, would’ve allowed the brakes to work again, due to there being a specific malfunction having been found within the anti-skid system of 62-4496, specifically with the right anti-skid valve. The failure to accomplish item 2, prevented items 3 thru 5 from taking effect. Late application of item 6, the emergency brake T-Handle, made application of item 7 late due to the need to pump up the brakes 3 to 5 times before item 7, brake application, could be successfully initiated. And item 6 was additionally dependant on good crew coordination in the heat of the moment, to ensure that both pilot and co-pilot didn’t attempt to pump the brakes concurrently, thus cancelling out the brakes altogether. Whether this occurred or not prior to the runway skid marks at the 750’ remaining point where braking first took effect, is unknown, due to the absence of any CVR/FDR data. As such, braking began far too late to prevent exiting the runway at the departure end with significant speed still, and proceeding down the steep embankment thereafter.
62-4496 Dedicated Crew Chief Maintenance Recordkeeping Deficiencies: Along with the aforementioned privilege of the MAJCOM commander having his own dedicated utility transport jet, he also had his own dedicated crew chief (DCC) for that aircraft. The DCC is the primary enlisted maintenance person responsible for the upkeep and maintenance status of a particular USAF aircraft. The DCC can be a ground crew chief, such as on fighter aircraft, or can be a flying crew chief, such as is found on transport aircraft when they are flying cross-country. The DCC keeps his own tools and small parts supply onboard the jet, and supervises any maintenance done by others on the aircraft, ensuring and being fully knowledgeable of every discrepancy, big or small, for his aircraft. Normally helped by an assistant crew chief (ACC), the DCC takes pride in his aircraft, since on combat aircraft and on some support aircraft, his/her name will be on the aircraft.
The DCC for 62-4496 had been selected by General Creech, and was the DCC who came with the aircraft when it was inherited by General O’Malley. The DCC was very familiar with all the intricacies of 62-4496, and was given a form of carte blanche to do what was needed to keep it in top shape, as well as any priority he needed for parts or equipment. 62-4496 was kept maintained and clean/polished at all times. But there was a significant item that wasn’t being accomplished correctly, that had it been, it could’ve highlighted a trend that had been developing with 62-4496 for a number of months, following an inspection that had been accomplished at an overhaul facility at Andrews AFB (ICAO: KADW) at the beginning of the Calendar Year (CY), which will be discussed more in-depth later in this synopsis.
For any USAF aircraft, the complete set of aircraft forms that every USAF aircraft is required to have, is known as the 781 book. This particular book is divided into sections, denoted by a specific suffix. The 781H, for instance, known as the Aerospace Vehicle Flight Status and Maintenance, is the primary form at the front of the book that shows what the aircraft’s status is: is it up, is it down. Does it have an inspection due, were the daily inspections complied with, and what is the aircraft total time, and individual sortie times filled in below that. Other 781 forms for other purposes also exist within that book.
The DCC for 62-4496 did not maintain a specific form known as a 781A, although he did maintain the other parts of the 781 manual, for reasons unknown. The 781A, known as the Maintenance Discrepancy and Work Document, is a form that is used to record specific problems encountered with the airplane, or any work being done to the plane, as well as how it affects the up/down status of the airplane, what actions were taken to fix/clear the specific discrepancy, and who signed it off. This form is the best to show items that potentially trend, as the normal pilot can thumb back usually a few pages and see what other items have occurred, as well as what was done about them, if anything. What the DCC did maintain was a personal “discrepancy log” of his own for 62-4496 which, while nearly identical to a 781A in terms of what a problem was, what the correction action was, and when it happened; he didn’t keep his in the book with the rest of the 781 forms, therefore no one other than himself really knew what was up with 62-4496 at any given time, nor could they potentially be the “second set of eyes” to see the trending information the DCC was dealing with regarding failures of the braking system over the period of time from when the plane had gone through the depot facility at Andrews AFB, until the week of the fatal crash. Nor could the normal “quality Assurance” (QA) maintenance inspectors take a look at the aircraft and attempt to spot trends, due to the same lack of information in the 781 books. Only the DCC had this information, and he was able to do business this way, directly contrary to USAF regulations at the time, because of the latitude extended to him by both General Creech as well as General O’Malley regarding doing whatever he saw fit in the maintenance of 62-4496.
Some of the specific notes found in the DCCs personal log post-accident, contained the following items, with emphasis on aircraft brake and anti-skid problems that were recurring multiple times over 34 flights following a time-phase inspection conducted in January at the depot at Andrews AFB, are cited below:
- 14 January – Aircraft returned to Langley AFB after Isochronal at Andrews AFB [on 5 January] (notably major work was done on the brake system including change-out of both normal and auxiliary hydraulic accumulators).
- 18 January – Aircraft pulled to the left on landing.
- 30 January – Anti-skid kicked off. (General O’Malley was in the pilot’s seat and had trouble holding the aircraft as he applied power prior to take off.)
- 1 February – Still having problems with anti-skid.
- 15 February – Changed “O” ring on power brake control valve.
- 1 March – Replaced “O” Ring on power brake control valve fitting; it has been leaking again.
- 12 March – Pilot’s brake pedals failed to stop aircraft on touchdown, master caution light illuminated (co-pilot was able to get effective braking action).
Had the DCC been entering these discrepancies into the correct Form 781A, either in addition to, or in lieu of his personal logbook; there’s a good chance that the above trends could’ve been spotted a mechanic working on the aircraft, a QA inspector, a pilot, or nearly any maintenance person who knew how to read the forms. As is, many of these trends, particularly the brake issues, stemmed from problems with shoddy workmanship encountered from the USAF depot where 62-4496 had received a phase inspection in January, 3 months prior to the accident. For reasons unknown, the DCC for 62-4496 either didn’t notice a trending problem, or else didn’t find it significant enough to do anything about it.
Maintenance Depot Overhaul Discrepancies: On 5 January 1985, 62-4496 went to KADW to receive a 6 day time-phased inspection where certain time-phased parts were being replaced as necessary, as well as other inspection items taking place. On 11 January, 62-4496 returned to KLFI, and as short as a few days later, problems began recurring with the braking and anti-skid systems of the aircraft, among other things. Found in the wreckage, many of these parts were recovered, and as they were torn down for inspection by investigators, shoddy workmanship began to be discovered in certain critical parts that are primary causal factors to this accident. The primary causal factor, failure of the main/auxiliary brake system, was traced to the right anti-skid control valve of 62-4496. Teardown analysis of this valve post-accident, revealed that when it was overhauled and rebuilt during it’s inspection at KADW, it was incorrectly reassembled and adjusted such that a pole met an armature in the valve, thus not allowing brake pressure to be achieved within the system. Why this had not manifested itself to the extent it did on the fatal flight, is anyone’s guess; but it did show that there was a trending problem occurring, even though each fix was believed to have fixed the problem, when in reality, those only masked the true culprit. The failure of this anti-skid control valve would’ve still occurred the day of the accident, however had the crew of 62-4496 correctly followed the checklist for anti-skid/brake failure, specifically item 2 regarding turning off the anti-skid switch, that would’ve removed the anti-skid valve from the equation, and allowed hydraulic pressure to build, and pressure would’ve been available when the auxiliary control switch was selected ON by the crew, if not earlier.
Additional factors found regarding shoddy workmanship, but not causal to the accident itself, were the hydraulic accumulators for both the main hydraulic system, as well as the auxiliary system. The main system accumulator had an improperly installed backup O-ring assembly, while the auxiliary system had a damaged piston, determined to have been caused due to damage in the reassembly phase prior to its reinstallation in 62-4496. Neither of these factors were causal to the accident, as when tested, both hydraulic accumulators operated normally. So these are tertiary findings, but still factors nonetheless, even though they were not causal.
MikeDs Final Thoughts: This accident is one of those where although the accident occurred through a series of maintenance discrepancies, there was still a “way out” had the crew executed the emergency checklist appropriately and covered the initial item that would’ve taken care of the rest in one fell swoop: the anti-skid being switched off. Barring that, they still had another trick up their sleeve with the emergency brake handle, but it’s late application didn’t allow time to stop this already fast moving emergency, figuratively and literally speaking. The items in question were critical action items, and were likely so for good reason, what the North American company described in the narrative prior to the procedure as “….actions taken sequentially,” pending they are indeed accomplished “in sequence”.
Once the emergency brakes were activated, it was far too late to stop on the remaining runway. At nearly most other airports, a runway overrun like this would only result in no-damage at best, and maybe serious damage at worse, as the aircraft rolled through the sod. KAVP unfortunately, has the severe terrain drop offs at each end of the runway, a place that the crew of 62-4496 couldn’t avoid going once they found the proverbial square corner they were in following getting the late activation of the emergency brakes. For it’s part, KAVP….due to it’s terrain drop off’s at the end of the runway….became in 2008 one of 45 planned and funded airports to receive the Engineered Material Arresting System (EMAS). Where runway arresting systems have been in use for decades ever since the advent of jets decades ago, at both military as well as joint-use civilian fields for jets both with as well as without tail hooks; these systems haven‘t been prevalent at civilian fields, even the non-tailhook pop-up systems that catch the main landing gear of an aircraft such as the MA-1A. Somewhat understandable as they are expensive to both install, as well as have trained crews to test them and maintain them. EMAS, on the other hand, gives a cheap and easy alternative incorporating collapsible and crushable concrete at the runway overruns for a certain distance that the overrun aircraft can enter and decelerate through as it sinks through the crushable materials. The system works much in the same way that the sand on the runaway truck ramps located at the bottom of highway downhill segments/grades work. Had an EMAS-style system been engineered and available 23 years prior to its installation, CT-39A 62-4496 would’ve been nothing more than a safety hazard footnote occurrence in the history of the USAF, vice the major tragedy it became, and the reminder it gave us that indeed The Runway to Hell is Paved With Good Intentions.
Incidentally, a landing overrun accident of a civil Saberliner 60 occurred 7 November 1992 at Phoenix-Sky Harbor International Airport (KPHX), in nearly identical dynamics as the accident here of CT-39A 62-4496.
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.
USAF CT-39A 62-4496:
Last edited: