AIRCRAFT ACCIDENT REPORT EASTERN AIR LINES, INC. McDONNELL -DOUGLAS DC-9-31, N8967E AKRON-CANTON REGIONAL AIRPORT, NORTH CANTON, OHIO

Transcription

1 L._ L ARCRAFT ACCDENT REPORT EASTERN AR LNES, NC. McDONNELL -DOUGLAS DC-9-31, N8967E AKRON-CANTON REGONAL ARPORT, NORTH CANTON, OHO NOVEMBER 27, 1973 ADOPTED: NOVEMBER 5, 1974 b 5 NATONAL TRANSPORTATON SAFETY BOARD Washington, D.C REPORT NUMBER: NTSB-AAR L

2 1. Report No. 2.Governent Accession No. NTSB-AAR Title and Subtitle Aircraft Accident Re ort - Eastern Air Lines, McDonnell. Douglas DC-9-3!, N8967E, Aron-Canton Regional Airport, North Canton, Ohio, Noveber 27, Author(s) TECHNCAL REPORT DOCUMENTATON PAGE 3.Recipient's Catalog No. 5.Report Date Noveber 5, Perforing Organization Code 8.Perforing Organization Report No. 9. Perforing Organization Nae and Address 1.Wor Unit No. National Transportation Safety Board Bureau of Aviation Safety Washington, D. C Sponsoring Agency Nae and Address NATONAL TRANSPORTATON SAFETY BOARD Washington, D. C A 11.Contract or Grant No. 13.Type of Report and Period Covered Aircraft Accident Report Noveber 27, Sponsoring Agency Code 15.Suppleentary Notes This report does not contain Safety Recoendations. 16.Abstract An Eastern Air Lines McDonnell-Douglas DC-9-31 crashed at Aron-Canton Regional Airport, North Canton, Ohio, on Noveber 27, 1973, at 2129 e.s.t. The aircraft ran off the end of runway 1 after copleting a precision approach and landing, traversed 11 feet of unpaved ground, and plunged over a 38-foot ebanent. The aircraft was daaged substantially by the ipact, but there was no fire. The 21 passengers and 5 crewebers sustained various injuries. The National Transportation Safety Board deterines that the probable cause of the accident was the captain's decision to coplete the landing at an excessive airspeed and at a distance too far down a wet runway to perit the safe stopping of the aircraft. Factors which contributed to the accident were: (1) Lac of airspeed awareness during the final. portion of the approach, (2) an erroneous indication of the speed coand indicator, and (3) hydroplaning. 17. Key Words Aircraft Accident, LS approach, tailwind, wet runway, hydroplaning, overshoot, speed coand indicator. 15.Security Classification (of this report) 2.Security Classification (of this page) 18.Distribution Stateent This report is available to the public through the National Technical nforation Service, Springfield, Virginia No. of Pages UNCLASSFED UNCLASSFED 5 NTSB For (Rev. 9/74) ii 22.Price

3 ... TABLE OF CONTENTS r Synopsis... nvestigation... History of the Flight... njuries to Persons... Daage to Aircraft... Other Daage Crew nforation... Aircraft nforation... Meteorological nforation... Aids to Navigation... Counications... Aerodroe and Ground Facilities... Flight Recorders... Aircraft Wrecage... Medical and Pathological nfortion... Fire... Survival Aspects... Tests and Research... Other nforation... Analysis and Conclusions... Analysis... Conclusions... (a) Findings... (b) Probable Cause... Page Appendices : Appendix A nvestigation... Appendix B Crew nforation... Appendic C Aircraft nforation... Appendix D Flight Data Recorder Graphs... Appendix E Cocpit Voice Recorder Transcript... Appendix F Airport Diagra... Appendix G. Approach Plate for Aron-Canton Regional Airport iii t

4 . c

5 File No Adopted: Noveber 5, 1974 NATOJXL TRANSPORTATON SAFETY BOARD WASHNGTON, D. C ARCRAFT ACCDENT REPORT EASTERN AR LNES, NC. McDONNEZL-DOUGLAS DC-9-31, N8967E AKRON-CANTON REGONAL ARPORT, NORTH CANTON, OHO NOVEMBER 27, 1973 SYNOPSS An Eastern Air Lines McDonnell-Douglas DC-9-31 crashed at Aron- Canton Regional Airport, North Canton, Ohio, on Noveber 27, 1973, at 2129 e.s.t. The aircraft ran off the end of runway 1 after copleting a precision approach and landing, traversed 11 feet of unpaved ground, and plunged over a 38-foot ebanent. The aircraft was daaged substantially by the ipact, but there was no fire. The 21 passengers and 5 crewebers sustained various injuries. The aircraft landed about 2,2 to 2,6 feet beyond the threshold of runway 1. The weather at the tie consisted of low ceilings, light rain showers, fog, and l-ile visibility. The National Transportation Safety Board deterines that the probable cause of the accident was the captain's decision to coplete the landing at an excessive airspeed and at a distance too far down a wet runway to perit the safe stopping of the aircraft. Factors which contributed to the accident were: (1) Lac of airspeed awareness during the final portion of the approach, (2) an erroneous indication of the speed coand indicator, and (3) hydroplaning. 1.1 History of the Flight 1. NVESTGATON Eastern Air Lines, nc., Flight 3 (JL4 3), a McDonnell-Douglas DC-9-31 (N8967E), was a scheduled passenger flight fro Miai, Florid to Aron-Canton Regional Airport, North Canton, Ohio, via Pittsburgh, Pennsylvania. n Pittsburgh, the captain received the 29 L/ Ohio Valley States weather reports which included the following Aron-Canton observation: Scattered clouds at1 feet, easured ceiling-3 feetovercast, visibility- 1% iles, thunderstor, light rain showers, fog, wind-3 at 5 n. - 1/ All tieshereinare Eastern standard tie, based on 24-hour cloc.

6 - 2- EA 3 departed Pittsburgh at 211 with 21 passengers and a crew of 5. Their coputed taeoff data was based on a gross weight of about 83, lbs., and according to the flightcrew, the aircraft perfored in accordance with their calculated data and airspeeds. The flight was cleared to, and flew at 6, feet. 2/ - EA 3 contacted Aron-Canton approach control at The approach controller stated, ''You can have your choice, either the localizer bac course one nine approach or the LS one approach, landing straight. Wind is one six zero degrees variable both sides at eight to one two, altieter two niner seven four. Weather is indefinite ceiling two hundred, sy obscured, visibility one and one-half iles, light rain showers and fog, and the runway one visual range is ore than six thousand feet." The captain chose the instruent landing syste (LS) front course approach to runway 1 in order to use the electronic glidepath inforation. The captain and first officer stated that the weather ade the use of the bac course approach inadvisable, Although the visibility was above the published inius, they noted that the reported ceiling was below the iniu descent altitude (MDA) for the bac course procedure. They also new that the runway was wet, and that they would be landing downwind. Although neither pilot coputed the exact value of the tailwind coponent before landing, they new it did not exceed the axiu allowable of 1 n. While the flight was being vectored toward the LS final approach course, the captain authorized a flight attendant to observe the approach fro the cocpit jupseat. Before clearing the flight for the approach, the approach'controller infored the, "Eastern three hundred, a copany DC-9 just too off. Said when he landed (on runway 19), there was soe water on the runway, but the braing action was pretty good." The flight acnowledged this essage. The tower controller cleared the flight to land, and reported the wind as, "One six zero degrees at niner." (Based on this wind, the tailwind coponent was 8 n.). The first officer positioned the flaps at the - 2/ All altitudes and elevations are ean sea level, unless othe.rwise indicated, - 3/ Aircraft landing distance deterinations are based upon an approach airspeed which is 13 percent of the stall airspeed (1.3 Vs).

7 - 3- a captain's coand, and both the captain and the first officer stated that the flaps were extended to the 5' position before the aircraft passed the outer arer (OM), The checlists were accoplished in accordance with copany procedures. The captain stated that the approach was flown within "noral para- - eters." The approach lights were in view at 4 feet, and the runway lights could be seen at 3 feet, He was satisfied with the approach except for the speed coand indicator, which was reading slow. A/ According to the captain, this occurred about the tie the first officer called that the runway was in sight, and the airspeed was 125 KAS. He ade rears to the first officer about the speed coand syste's slow indication and a loading proble, and then added thrust oentarily until "the aircraft was in good position for the landing." Before this conversation, the first officer had told the captain that they were *(A big fast." e first officer stated that his airspeed indicator at that tie indicated 13 to 135 KAS. He reebered the captain's rears, and he verified that the captain's speed coand pointer was reading slow; however, he did not recall the indications on his indicator at any tie during the approach. He confired that the captain added thrust, but he stated that the sequence of events occurred before he told the captain that the runway was in sight and while they were still in instruent flight conditions. About 24 seconds before touchdown, the first officer called "Three hundred feet...above the glide slope" and, shortly thereafter, "Minius... the airport's on your left." There were no other altitude calls during the approach, and no callouts of airspeed deviations were ade during the final part of the approach. Neither creweber recalled the indicated airspeed just before touchdown. * The flight data recorder airspeed trace disclosed that the indicated airspeed began increasing 35 seconds after the OM was passed. During the final inute of the approach, it increased steadily fro 132 KAS, and attained a axiu of 142 KAS when the first officer called '"inius." At touchdown, the airspeed trace was about 139 KAS. According to the crew, the aircraft touched down about 1, to 1,5 feet beyond the runway threshold and on the centeine; it did not float before the touchdown. The spoilers deployed autoatically at touchdown. Noral reverse thrust and braes were applied iediately after touchdown, and reverse thrust was aintained until the aircraft slowed to between 6 and 7 KAS. At that tie, deceleration seeed to stop. Reverse thrust was increased to axiu continuous thrust (MCT), - 4/ The speed coand indicator oves vertically over a scale. The pointer is centered when the coputed optiu speed for a aneuver and indicated airspeed correspond. The pointer oves below center when the indicated airspeed is below the coputed optiu speed and above center when the speed is higher.

8 -4- and engine copressor stalls were heard. The captain said that the braes felt noral, but that the aircraft's response to the was not. The first officer turned the antisid syste off in response to the captain's coand; he then "got on" the braes with the captain.. The captain stated that these actions too place about the point where the runway lights were "just starting to go orange." The first officer stated that the airspeed was about 7 to 8 KAS. They noted that the airspeed was less than 6 KAS when the aircraft left the runway. After the aircraft left the paveent, it traversed about 11 feet of level ground, plunged over a 38-foot ebanent, and stopped in a level attitude in a field below. The tie was Soe passengers noticed the lac of deceleration during the landing roll. One stated that the aircraft was beyond the terinal building before it landed. (See Appendix F.) n addition, four witnesses at the airport saw the flight land. Three said that the aircraft touched down at, or beyond, the intersection of runways 1/19 and 5/23. The fourth said that the aircraft landed before it reached that intersection. The accident occurred during the hours of darness and at latitude 4 54'58''N and longitude '32W. 1.2 njuries to Persons njuries - Crew Passengers Others Fatal Nonfatal 5 21 None 1.3 Daage to Aircraft The aircraft was daaged substantially by the ipact. 1.4 Other Daage None 1.5 Crew nforation The crewebers were qualified and certificated for the flight. (See Appendix B.) 1.6 Aircraft nforation N8967E was certificated and aintained according to Federal Aviation Regulations. (See Appendix C.)

9 -5- The aircraft's weight and center of gravity at the tie of the accident were 8, lbs. and 23.8 percent ean aerodynaic chord (MAC), respectively. Both were within specified liits. f the runway is wet and E the tailwind coponent is 1 n, the axiu allowable aircraft gross weight for a landing on runway 1 is 92,9 lbs. c 1.7 Meteorological nforation The following are selected surface weather observations at the Aron- Canton Regional Airport on Noveber 27, Record Special, indefinite 2 feet, sy obscured, visibility- 1% iles, light rain showers, fog, teperature-58of., dew pint-54~f., ~ind-15~ at 1 n, altieter setting in. Thunderstor ended 22, oved northeast, pea wind-31 at 7 n. 27. Pressure falling rapidly Special, indefinite 2 feet, sy obscured, visibility-1% iles, light rain showers, fog, teperature-59'f., dew point- 54OF., wind-16 at 1 n, altieter setting in. The official precipitation records for the airport indicated the following rainfall aounts: - Tie nches LL - A Beechcraft King Air, N711MC, landed at the Aron-Canton Regional Airport about 5 inutes before EA 3. Before the King Air landed, the approach controller ased the pilot to chec the braing action, because he had received two reports of poor braing. A twin Cessna had reported soe hydroplaning, and an Allegheny Aiines DC-9 had reported poor braing because of water on the runway. The approach controller also stated that two aircraft had landed since those reports and that they reported no difficulties. The CVR confired that this transission was available in EA 3's cocpit and that intracocpit conversation occurred during that transission. The captain stated that he did not recall hearing the transission. The first officer reebered the reference to the aircraft's call sign, but did not reeber the contents of the essage. The King Air's flightcrew did not give a braing action report to the tower, but subsequently stated that they had experienced light rain and a tailwind on the approach, and that they saw the runway at the

10 -6- iddle arer (MM). The pilot stated that the runway was wet with puddles of water on it. He used reverse thrust only to stop the aircraft and, therefore, was not aware of hydroplaning. Two DC-9's landed at the airport before EA 3: Allegheny Aiines Flight 915 (AL 915) at 228 and Eastern Aiines Flight 573 (EA 573) at 2&8. The pilot of A, 915, a DC-9-31, stated that the weather was near inius with light to oderate rain and a wind of 13 at 5 n. He stated that the approach to runway 1 was noral. After touchdown, "Maxiu reversing and braing was used and considerable hydroplaning was experienced." The pilot reported to the tower that the braing action was llpoor. The pilot of EA 573, a DC-9-14, ade two unsuccessful LS approaches to runway 1. The first approach was abandoned because of the reportedly poor braing action, a high Vref speed of 133 KAS, and a tailwind, which resulted in a descent rate of "about 1,5 feet per inute." On the second attept the flight touched down. At touchdown, the crew was infored that the wind was 16' at 1 n. During a postflight interview, the captain stated, "Since we touched down rather long, 1,5 to 2, feet down the runway, coupled with the last wind report, elected to abort the landing even though the ground spoilers had actuated." A bac course approach to runway 19 was then ade and the captain had no difficulty stopping his aircraft after landing. 1.8 Aids to Navigation The LS at Aron-Canton Regional Airport provides a front'course approach to runway 1 and a bac course approach to runway 19. The inbound heading for the front course approach is 6O. There is a low frequency hoer located at the OM. The OM and MM are located 3.7 ni and.7 ni, respectively, fro the runway threshold. The glidepath angle is 2.9'. The iniu crossing altitudes at the OM and MM are 2,416 feet and 1,452 feet, respectively, The published inius for this approach are runway visual range (RVR) 2,4 feet, or + ile. The decision height OH) is 1,413 feet (2 feet above the runway touchdown zone. ) The published iniu for the LS bac course approach is 3/4 ile visibility. The MDA is 1,56 feet, or 343 feet above the touchdown zone of runway 19. After the accident, the pertinent airport equipent, NAVADS, and the LS were inspected and flightcheced; they operated within prescribed paraeters.

11 Counications Not applicable. 1.1 Aerodroe and Ground Facilities L The Aron-Canton Regional Airport was certificated under 14 CFR 139 for scheduled air carrier operations on April 14, t is served by three runways: Runway 1/19> runway 5/23, and runway 14/32. Field elevation is 1,228 feet. Runway 1 is 6,398 feet long and 15 feet wide and is paved with ungrooved asphalt. The usable runway beyond the glidepath intersection is 5,458 feet. The runway is equipped with high-intensity runway lights, a high-intensity approach light syste with sequence flashers, and a transissoeter. The final 2,2 feet of runway 1 is indicated by highintensity aber runway lights. The distance fro the south end of runway 1 to its intersection with runway 14/32 is about 1,2 feet, and to its intersection with runway 5/23, about 2,25 feet. During an inspection of runway 1 on the orning after the accident, rubber deposits were found in the touchdown zones at both ends of the runway; however, no other significant runway surface containation was noted. Over the ajor portion of the runway, the tire tracs of EA 3 could not be distinguished fro those of other aircraft. The accident aircraft's tire tracs becae distinguishable over the last 2 feet of the runway. Within this 2 feet, an area about 3 feet long contained blac rubber ars; that area was very close to the end of the runway Flight Recorders c The aircraft was equipped with a Fairchild Model F-5424 flight data recorder (FDR), serial No The flight recorder and foil recording ediu were undaaged, and all paraeters had been recorded. The last 5 inutes of the recorder traces were read. The altitude inforation was based upon a baroetric pressure of in. to convert pressure altitude to ean sea level. No corrections were ade to any other paraeters. (See Appendix D.) The aircraft was equipped with a Fairchild A-1 cocpit voice recorder (CVR), serial No The recorder and tape were not daaged. The final 11 inutes were transcribed for this report. (See Appendix E.) Both recorders were installed in the aft section of the aircraft.

12 Aircraft Wrecage The aircraft stopped on a agnetic heading of 36' and about 38 feet beyond the end of runway 1. The wings were intact and reained attached to the fuselage. There was inor spillage of Jet-A fuel. The leading edge slats were found extended fully; the trailing edge flaps on the right wing were found extended fully, while those on the left wing were in an interediate position. The flap selector handle was found in the 5' detent. The spoilers were down, intact, and attached to the wing. Weeds and grass were trapped between the spoilers and the wing structure. The speed brae lever was in the retracted position. The fuselage forward of station 996 was intact. Both engine asseblies separated fro the aft fuselage section, and the pylons reained attached to the engines. The entire aft fuselage section and epennage, including the auxiliary power unit, separated fro the ain fuselage at the pressure bulhead. The ain landing gear was extended and loced, and had folded aft at its attachents. The nose gear assebly was in the ext d osition and attached to its support structure. The nose gear and * 2 structure were torn fro the fuselage and rotated aft and upward into the electrical and electronic copartent. Continuity of the rudder, elevator, and horizontal stabilizer controls was destroyed by separtition of the epennage fro the fuselage. The ailerons, spoilers, and flaps were not ovable as a result of ground ipact. The ovable reference arers (bugs) of the captain's and first officer's airspeed indicators were set at 116 and 115 n, respectively. The upper left altieter was set at in. and indicated 27 feet. The lower left altieter was set at in. and indicated 1,2 feet. The right altieter was set at in. and indicated 3 feet. All four ain gear tires contained patches of reverted rubber. The tire pressure for tires No. 1 through 4 were 135 lbs., 137 lbs., 125 lbs., and 15 lbs., respectively. Brae wear, tire tread depth (except within the patches of reverted rubber), and inflation pressures were within prescribed tolerances Medical and Pathological nforation There were no fatalities. Of the 16 seriously injured persons, 7 sustained vertebral fractures. Other injuries included contusions, lacerations, and sprains. Most of the passengers and crewebers were treated at the accident site before being taen to the hospital.

13 Fire There was no fire Survival Aspects This was a survivable accident; the cocpit and cabin aint ained their structural integrity; the tiedown chain- reained intact. Because there was no fire, there was aple tie for evacuation and rescue. High vertical ipact forces accounted for the separation of the tail section, the daage to the cabin interior, and the injuries sustained by passengers and cr eber s. The flight attendants responded effectively to the eergency. They used their copany-issued flashlights, an electronic egaphone, and other eergency equipent to direct the rescue operations. Passengers reported, "a second or two of confusion," followed by an ordey evacuation. The collapse of the overhead racs onto the bacrests of the seats during the final ipact caused head injuries and interfered with the use of the overwing exits. The eergency lights functioned propey, except that the purser was unable to reove the quic-release light unit at the ain entry door for use as a hand-held flashlight. All exits were opened except the captain's sliding window; the captain exited through the rear of the aircraft. The inflatable slide at the galley door operated norally and was used by several people to exit and reenter the aircraft to assist with the rescue. Because of the proxiity of the ain entry door to the ground, the flight attendant detached the girt bar; therefore, the slide did not deploy when the door was opened. The slide at the tail exit was torn off with the structure that separated fro the aircraft. The first officer exited the aircraft through his sliding window. The captain and a flight attendant entered the cabin via the cocpit door and helped supervise the evacuation. Several passengers were iobilized and had to be carried out on bacboards. All occupants reportedly were reoved fro the site within 2 inutes after the accident Tests and Research Aircraft Coponents and Systes The brae syste coponents and the daaged antisid control box were exained. Two of the four circuit boards tested satisfactorily; however, two boards had cracs and broen transistors which prevented testing. The four transducers, hydraulic fuses, braes, and brae servo valves also tested satisfactorily.

14 - 1 - The coponents of the pitot syste tested satisfactorily. The three pftot heads were clean with no evidence of distortion. The pitot and static drains were clean. Both airspeed indicators were tested, and both operated within liits. All ajor coponents of the speed conrand syste were tested and were found to function propey Runway Coefficient of Friction At the Safety Board's request, the National Aeronautics and Space Adinistration (NASA) conducted slipperiness and drainage tests on runway 1 using the NASA diagonal braed vehicle (DBV). Three test zones, A, B, and C, were established along the runway, and covered with water to depths ranging fro dap to.2 in. The following stopping distance ratios (SDR) - 5/ were established. - Test Zone Feet fro Threshold Rubber Deposits SDR A i/ 1,865 to 2,58 None to spots 1.31 B 2,58 to 3,58 None to spots 1.6 B to C 3,58 to 4,4 Spots to ediu 1.61 C 4,4 to 5,4 Mediu to heavy 1.8 A wet runway slipperiness reference for civil aircraft operations ay be deterined fro the Federal Aviation Regulations for aircraft landing certification (14 CFR ) and aircraft landing oderations (14 CFR - L ). This reference slipperiness level is equivalent to an SKR of a According to a NASA Langley Research Center study, runway water depth during a landing deterines the type of hydroplaning phenoena that could occur. Tests indicate that an aircraft is susceptible to both viscous hydroplaning and reverted rubber hydroplaning (when the wheels are loced) when water depths range fro dap to.5 in. All three types of hydroplaning (dynaic, viscous, and reverted rubber z/) ay - 5/ SDR is the ratio of the wet runway stopping distance to the dry runway stopping distance for an aircraft of the sae weight, speed, and configuration. - 6/ Test Zone A included an asphalt patch at the intersections of runways 1 and 5/ / Dynaic hydroplaning is the result of the hydrodynaic lift forces developed by a tire oving across a fluid-covered surface. Viscous hydroplaning or sidding is the result of the reduction of the friction coefficient caused by the lubrication properties of a thin fluid between the tire and the runway. Reverted rubber hydroplaning occurs fro an increase in tire print pressure as a result of prolonged viscous or dynaic hydroplaning to the point that the tire elts and reverts to its original unvulcanized state.

15 occur with.5 to.1 in. water on the runway. The speed of the airplane ust exceed the tire dynaic hydroplaning speed (approxiately 9 ties the square root of the tire pressure) for dynaic hydroplaning to occur Perforance Data At the Safety Board's request, the McDonne11-Douglas Aircraft Corporation furnished stopping distances for the DC-9-31 aircraft, The following paraeters reained constant for all coputations: Field elevation-1,228 feet, teperature-15o C, aircraft landing weight-8, lbs., slats extended, and 5' flaps. An 8-n tailwind coponent was applied to all coputations. Ground spoilers were assued to be deployed autoatically and fully extended 1 second after touchdown. When braes were included in the deceleration dat braing began 1 second after touchdown, and full braes were applied 2 seconds later. The variable conditions and the results are set forth below. Condition 1 assued that the runway was dry. The aircraft crossed the runway threshold at an altitude of 5 feet, at 1.3 Vs (115 KAS), and landed at 1.25 Vs (11 KAS). Forward thrust reained at idle throughout the landing roll. Based upon these criteri the ground stopping distance was 1,818 feet. A landing speed of 139 KAS increased the stopping distance to 2,964 feet. Condition 2 was set up to investigate the effect of 8 percent N1-8/ reverse thrust on the ground stopping distance on a wet and dry runway. An SDR of 1.71 was applied to approxiate the wet runway braing conditions. t was assued that the thrust levers were pulled into the reverse detents 3 seconds after touchdown and that 8 percent Nl was attained 8 seconds after touchdown. The following data were coputed: Conditions Stopping Distance (feet) Dry - Wet (1) 8 percent N1 reverse thrust and braing 2, 54 3,678 (2) 8 percent N1 reverse thrust and no braing, 5,697 5, Other nforation Eastern Air Lines Copany DC-9 Flight Manual Procedures Eastern Air Lines requires the pilot not flying the aircraft to ae the following calls on all approaches: - 8/ A easureent of thrust expressed in ters of a percentage of the rotational speed of the N1 (low pressure) copressor.

16 At 1, feet above field elevation, call out altitude, airspeed, rate of descent. On instruent approaches, he will also call out the result of flag scan. At 5 feet above field elevation, call out altitude, airspeed, rate of descent. On instruent approaches, he will also call out the result of flag scan. Call out significant deviations fro prograed airspeed and desired descent rate. For instruent approaches, call out DH or iniu descent altitude. Call out "runway in sight--right/left or straight ahead." According to Eastern Air Lines DC-9 training procedures, once the runway is called "in sight" only airspeed deviations should be called to the pi1ot''s attention during the reainder of the approach. The copany recoended that these speeds be called out as variances fro the Vref or "bug" speed, for exaple, "bug plus" or "bug inus" the aount of variation. The Eastern Air Lines DC-9 flight anual states, in part, "After touchdown, apply reverse thrust soothly but quicly, using 8 percent N1 as desired axiu; however, axiu continuous thrust (MCT) ay be used if conditions require axiu stopping effort. As speed decreases to approxiately 7 nots, reduce power to prevent surging. dle reverse thrust should be used until forward thrust is required to taxi." With regard to airspeed control procedures during landing approaches, the flight anual states in part: "The speed 1.3 Vs, reference speed (Vref), is used to deterine FAR landing distance and is used as target speed on final approach." The flight anual also discusses factors which affect the Vref speed, such as headwind coponent and gust factors, and how these should be anaged. The discussion closes with the following stateent: "Over the threshold, only the gust factor should be aintained above Vref. 1 8 Eastern Air Line's airspeed control procedures require that a target speed of Vref + 5 KAS flown on the final approach. The additional 5 KAS is to copensate for unnown, or undeterined, wind effect and is to be bled off slowly in order to cross the runway threshold at Vref Speed Cond Syste Speed coand deviation pointers are located on the right side of the captain's and first officer's flight director indicator. They indicate deviation fro optiu speed for the low-speed odes of flight and

17 aid the pilot in aintaining a safe argin above the stall speed. The fast-slow pointers receive electrical signals fro a speed coand coputer, which receives inforation fro an acceleroeter, vertical gyro, an angle of attac transducer, the right flap transducer, slat relays, the flight director controller, throttle switches, and the air/ground oleo relay. With an aircraft weight of 8, lbs., landing gear extended, and flaps and slats fully extended, the pointer would center at an indicated airspeed of Vref + 3, or 118 ns. At lesser flap settings or higher gross weights, the pointer would center at correspondingly higher indicated airspeeds. A alfunction of a slat relay would cause the instruent to center at a higher indicated airspeed without a warning flag (about 2 KAS at this aircraft's configuration). Eastern Air Line's training curriculu for the speed coand syste was, for the ost part, liited to the presentation afforded the pilot during the taeoff, the go-around aneuver, and engine-out aneuvers. With regard to the final approach, the copany taught that the.airspeed indicator was the priary instruent for airspeed controi that scheduled airspeeds should be aintained, and that the speed coand was a bacup reference instruent. t did not teach the effect which variances of gross weight or center of gravity locations would have on the speed coand syste's fast-slow presentation. There was no presentation or possible alfunctions which could alter the speed coand reading, without a display of the warning flag. The only alfunction deonstrated in the flight siulator resulted in the display of warning flags. The captain's and first officer's depositions corroborated that this was the extent of their training on the syste Flight Attendant Stations The flight attendant's presence in the cocpit was not precluded by the provisions of 14 CFR , and based on the nuber of passenger seats on the aircraft, only two flight attendants were required in the passenger cabin. The third flight attendant was assigned to the flight to assist in a eal service between Miai and Pittsburgh, and as such, was designated as the "extra" attendant. According to the copany's flight attendant anual, the extra attendant shares the duties of the other flight attendants and will occupy the forward jupseat with the senior flight attendant on taeoff and landing. A copany flight attendant supervisor stated, however, that the anual does not expressly preclude a flight attendant fro occupying the cocpit jupseat during a taeoff or landing, and that the captain had the authority to authorize a flight attendant to sit in a seat other than that to which she was assigned.

18 APALYSS AND CONCLUSONS 2.1 Analysis When the flightcrew was offered a choice between a front course LS approach to runway 1 or the bac course approach to runway 19, they had received adequate weather inforation to ae the selection. Based on the reports they had received at Pittsburgh and en route, they new about the ceiling, the rain, and the surface winds at Aron-Canton. Therefore, the captain new that his choice of the front course approach not only would require landing on a wet runway, but also would subject his aircraft to the effects of a tailwind that approached the aircraft's axiu allowable coponent. Before clearing EA 3 for the approach, the controller told the flight that a copany DC-9 had reported that there was water on the runway and that the braing action on runway 19 was "pretty good". Before that report, a twin Cessna and an Allegheny Aiines DC-9 had reported poor braing and hydroplaning, but.these reports were 1 hour old. There was no further corroboration of the Allegheny DC-9's report, thus, the inforation given to EA 3 was the latest data available to the controller. The captain's and first officer's recollections of the indicated airspeeds during the ey oents of the descent vary. The captain stated that his reaction to the slow indication on the speed connnand indicator occurred just after passing the MM, whereas, the first officer placed this event between the OM and M. Correlation of the CVR and FDR data disclosed that the speed coand indication was entioned just after the OM was passed and after the first officer called that they were fast. According to the first officer, the indicated airspeed at that oent was 132 KAS. The FDR trace, at that tie, indicated 132 KAS. Therefore, the Safety Board concludes that the captain noted the slow reading of the speed coand syste between the OM and M. The target airspeed for the approach was Vref + 5 KAS (12 KAS), and the runway threshold was to be crossed at Vref. The FDR disclosed that the lowest recorded indicated airspeed on the approach was 132 KAS, and that occurred about 15 seconds after the OM was passed. The airspeed reained at 132 KAS for about 16 seconds, and then began to increase steadily until it reached 142 KAS, about 5 seconds before touchdown. The landing speed was 139 KAS.. During the approach, the crew discussed the high indicated airspeeds and slow indications on the speed connnand indicator. Since there was no change in aircraft configuration and since the descent rate reained substantially unchanged, the airspeed increase confired the captain's stateent that he added thrust in response to the slow speed coand reading. He stated that the additional thrust was aintained until the landing was

19 assured. This assessent could not have been ade until he had the runway in sight. Thus, it was apparent that the additional thrust was applied about 3 seconds after OM passage and was aintained until after the first officer called the runway in sight, 1 inute 22 seconds later. During that interval, the airspeed increased about 1 KAS. The copany's training curriculu on the speed coand syste did not include a discussion of possible alfunctions that could cause an erroneous speed coand indication without causing the warning flag to be displayed. However, they did teach that the airspeed indicator was the priary instruent for control of airspeed and that the target speed, based upon the estiated landing weight, was to be flown on final approach. The captain new that the speed coand syste would portray the optiu speed for the aneuver being perfored, and that it was based on aircraft configuration and gross weight. He also new that his Vref speed of 115 KAS was predicated on an 8,-lb. landing weight. The slow indication occurred at about 13 to 135 KAS, and it could not have been valid unless the aircraft's gross weight was about 1, to 15, lbs. Although the captain entioned "a loading'proble" in connection with the speed coand reading, he also stated that he had not noted any significant variations between their coputed taeoff data and aircraft perforance at Pittsburgh; consequently, an error of this agnitude should have ade the captain suspicious of the operation of the speed coand rather than aircraft loading. nstead, he chose to react to the speed coand reading without requesting a crosschec of his instruents with those of the first officer. The coponents of the speed coand syste functioned norally when tested; therefore, the reason for its slow reading reains undeterined. The evidence indicated that, once both pilots had the runway in sight, the prescribed copany airspeed control procedures were not followed. This is substantiated by the fact that neither pilot recalled noting the airspeed during the final oents of the approach, although it was about 25 KAS above Vref. The lac of airspeed awareness also explains the captain's stateent that the approach and landing appeared noral to hi--so noral that he never considered rejecting it. The autoatic actuation of the ground spoilers identified the touchdown point on the CVR. Fro that point, the tape terinated in 26 seconds. Therefore, the ground distance traversed by U 3 was coputed using the FDR airspeed trace for those 26 seconds. The coputation was based on average indicated airspeed selected at the idpoint of 1-second intervals. The 8-n tailwind coponent was then added to the indicated airspeed to deterine a ground velocity. The coputed ground distance traversed by the aircraft was 4,53 feet. The-aircraft stopped 6,777 feet beyond the threshold of runway 1. Based on these figures, EA 3 touched down 2,275 feet beyond the threshold of runway 1.

20 The distance fro the OM to the runway threshold was 22,481 feet, and the elapsed tie fro arer passage to touchdown was 1 inute 4.5 seconds. During that tie, the flight traversed a ground distance of about 25,13 feet, which indicates that the aircraft landed about 2,648 feet beyond the threshold. This estiate was based on average true airspeeds selected over 1-second intervals. An 8-n tailwind coponent was added. These coputations support the stateents of the passenger and the witnesses who said that the aircraft landed at or beyond the intersection of runways 1 and 5/23. The Safety Board concludes that the aircraft landed about 2,2 to 2,6 feet beyond the threshold of runway 1, which left 3,8 to 4,2 feet of usable runway in which to stop the aircraft. The NASA tests disclosed that at water depths ranging fro dap to.2 in., the average SDR over the last 3, feet of the runway was about Although the exact aount of water on the runway could not be deterined, an approxiate SDR at the tie of landing can be derived fro the available evidence. The average deceleration rate of EA 3 was about 6 feet/second 2. Over the last 4 feet of the runway, where evidence of viscous and reverted rubber hydroplaning was found, this rate decreased to 3 feet/ second 2. This rate was achieved by applying braes shortly after touchdown, using 8 percent N1 reverse thrust over the first portion of the landing roll, and full reverse thrust over the latter portion. Deceleration rates were coputed fro the anufacturer's wet runway stopping distances. The following table represents a coparison between these rates and EA 3's perforance as noted above. Condition ueceleration rate (f t./sec 2) Braes + 8 percent N1 reverse thrust Braes + forward idle thrust No braes+8 percentn1reverse thrust EA OO Based on these rates, and stateents by the flightcrew describing the braing conditions, the Safety Board concludes that the runway conditions at the tie of EA 3's landing equaled, or probably exceeded, an SDR of f the aircraft had been flown according to recoended copany procedures, it would have landed about 1, feet beyond the runway threshold. Applyirg an SDR of 1.71 and an 8-n tailwind, the wet stopping distance without the use of reverse thrust would have been 3,19 feet, for a total distance of 4,19 feet. Thus, if the recoended approach speeds had been adhered even with the long landing, the aircraft ight have been

21 stopped between 5,39 and 5,79 feet beyond the long landing was not the priary causal the threshold. area. Therefore, Based on an SDR of 1.71, the 147-n landing velocity (139 KAS + &n tailwind) increased the wet stopping distance without reverse thrust fro 3,19 feet to 5,68 feet. Therefore, based on the estiated touchdown points, it was ipossible to stop the aircraft on the runway without the use of reverse thrust. Copany procedures authorize the use of MCT reverse thrust "if conditions require axiu stopping effort." The fact that the captain did not apply full reverse thrust until well into the landing roll further substantiates that the landing appeared noral to hi. According to the anufacturer, 8 percent Nl reverse thrust would reduce the landing roll to 3,678 feet. Therefore, if the aircraft touched down at 2,2 to 2,6 feet fro the runway threshold, it ight have been possible to stop with 12 to 52 feet of runway reaining, provided the reverse thrust was applied as set forth in the stated paraeters, and provided it operated at its axiu effectiveness throughout the application. The copressor stalls that occurred when the captain applied MCT reverse thrust coproised the effectiveness of the reverse thrust and ay have rendered it totally ineffective, as evidenced by the fact that during the tie it was increased to MCT, the deceleration rate deteriorated to less than the rate coputed for 8 percent N1 reverse thrust only. t is apparent that the anner in which the landing was accoplished placed the aircraft in a position in which stopping was dependent upon the added effects of reverse thrust. Although it was established that the deceleration rate of the aircraft was less than that expected on a runway with an SDR of 1.71, it could not be deterined whether this was the result of dynaic hydroplaning, the interaction of actual SDR and the antisid syste, or a cobination of both. To the extent that conditions for dynaic hydroplaning existed, the fact that the landing was ade at a velocity 4 n above the theoretical tire dynaic hydroplaning speed, resulted in a considerable increase in the duration of the aircraft susceptibility to this type of hydroplaning. There is no doubt that viscous and reverted rubber hydroplaning occurred during the latter portion of the landing roll. The evidence disclosed that the decrease in the deceleration rate during the final portion of the landing roll was practically siultaneous with the first officer's acnowledgeent of the captain's coand to turn off the antisid syste, and with the beginning of the tire ars on the runway. Since both pilots stated that they were on the braes at the tie, the deactivation of the antisid syste probably produced a loced wheel sid, and the deterioration of braing perforance.

22 Considering the runway conditions, the tailwind, the aircraft's excessive landing velocity, and the touchdown point, the captain should have questioned the feasibility of bringing the aircraft to a safe stop within the confines of the runway. He had control over every facet of the approach and landing, and the decision to go around could still be ade even after the wheels contacted the runway. Therefore, the Safety Board concludes that although hydroplaning and its effects contributed to the accident, the captain's decision to coplete the landing under the existing adverse conditions was the priary factor in this accident. With regard to the flight attendant, it appears that her presence in the cocpit had no effect on the anner in which the crew executed the approach. However, the Safety Board stresses that the operator as well as the pilot-in-coand should be fully cognizant of their respective responsibilities for assuring that persons aditted to the flight dec have assigned functions to perfor and that they are authorized by Federal regulations and copany procedures. 2.2 Conclusions a. Findings The flighterew was certificated and qualified to conduct the flight. The aircraft was certificated and aintained in accordance with FU rules and copany procedures. The captain was aware of the tailwind and wet runway when he decided to ae a front course approach to runway 1. + The flight acnowledged reception of a report stating that there was soe water on the runway, and that a copany DC-9 had reported that braing action was "pretty good" on runway 19. The Safety Board was unable to deterine the cause of the erroneous indication on the speed coand indicator. The indicated airspeeds on the approach were 17 to 27 KAS above the coputed Vref. The aircraft landed about 2,2 to 2,6 feet beyond the runway threshold, leaving 3,8 to 4,2 feet of paved surface on which to stop. The stopping distance required under the existing runway conditions without the use of reverse thrust was 5,68 feet.

23 Hydroplaning occurred during the landing roll. b. Probable Cause The National Transportation Safety Board deterines that the probable cause of the accident was the captain's decision to coplete the landing at an excessive airspeed and at a distance too far down a wet runway to perit the safe stopping of the aircraft. Factors which contributed to the accident were: (1) Lac of airspeed awareness during the final portion of the approach, (2) the erroneous indication of the speed coand indicator, and (3) hydroplaning. BY THE NATONAL TRANSPORTATON SllFETY BOARD JOHN H. REED Chairan FRANCS H. McADAMS Meber LOUS M. Meber TMYER SABEL A. BURGESS Meber WLLAM R. WEY Meber Noveber 5, 1974

24

25 c 1. nvestigation APPENDX A NVESTGATON & HEARNG The Safety Board was notified of the accident at 233 e.s.t., on Noveber 27, An investigation tea went iediately to the scene. Wor groups were established for operations, air traffic control, witnesses, weather, huan factors, structures, powerplants, systes, aintenance records, cocpit voice recorder, and flight data recorder. Parties to the investigation included: Eastern Air Lines, nc., Federal Aviation Adinistration, McDonnell-Douglas Corp., Air Line Pilots Association, and nternational Association of Machinists. 2. Hearing. There was no public hearing. Depositions were taen on July 31, a

26 APPENDX B CREM NFORMATON Captain Willia H. Hill Captain Willia H. Hill, 4, was eployed by EAL on February 18, The captain held Aiine Transport Pilot Certificate No with an airplane ultiengine land rating, and coercial privileges in airplane single engine land. He was type rated in Convair 24/34/44, Locheed Electr and DC-9 aircraft. He had a First-class edical certificate dated Septeber 4, 1973, and was required to wear glasses for reading. The captain had checed out on the DC-9 aircraft on July 1, He had 1,881 flight-hours, 736 of which were in DC-9 aircraft. During the previous 9-day, 3-day, and 24-hour periods, he had flown 135 hours, 41 hours, and 2 hours 42 inutes, respectively. His last two proficiency checs were on Septeber 15, 1973, and April 17, His last line chec was May 9, 1973, and he had copleted recurrent ground training on Septeber 24, The captain had been off duty about 53 hours before reporting for the flight. At the tie of the accident, he had been on duty 4 hours, 15 inutes, of which 2 hours, 42 inutes were flying tie. First Officer Andrew R. McQuigg First Officer Andrew R. McQuigg, 28, was eployed by EAL July 29, He had a Coercial Pilot License No with an instruent rating. He had airplane ultiengine land, single engine land, and glider aircraft ratings. His First-class edical certificate was issued on June 15, 1973, with no liitations. He had upgraded to the E-9 aircraft on October 31, He had accuulated 7, flight-hours, 23 of which were in the E-9 aircraft. During the previous 9-day, 3-day, and 24-hour periods, he had flown 85 hours, 26 hours, and 2 hours 42 inutes, respectively. His last proficiency chec and ground school training were copleted on October 31, 1973; these copleted his DC-9 upgrading curriculu. He had been off duty about 111 hours before reporting for the flight. At the tie of the accident, his on-duty and flight ties were the sae as the captain s. The flight attendants were qualified according to FAA and copany regulations, requireents, and procedures.

27 APPENDX C ARCRAFT NFORMATON N8967E, a DC-9-31, was anufactured in August 1968, and registered to Eastern Air Lines, nc., on August 23, A standard airworthiness certificate was issued for the aircraft on August 13, The aircraft had accuulated 15,615 flight-hours at the tie of the accident. Aircraft and coponent records showed that all inspections and overhauls had been perfored within prescribed tie liits and that the aircraft had been aintained according to copany procedures and FAA regulations. All applicable airworthiness directives had been coplied with as of Noveber 27, The aircraft was equipped with two Pratt and Whitney JT-8D-7 engines. The No. 1 engine, serial No , had 14,8 hours since overhaul, and the No. 2 engine, serial No , had 13,927 hours since overhaul. 5

28

29

30

31 APPENDX E TRANSCRPT OF ADDTONAL! PERTNENT CO"UNCA!EONS FROM COCKPT VOCE RECORDER DOUGUS MODEL E-9-31, N89673, EASTEBN ARLNES FLGHT 3, AKRON-CANTON REONAL ARPORT, NORTH CANTON, OHO NOVElvBER 27, 1973 Cocpit area icrophone voice or sound Radio transissions fro N8967E Voice identified as Captain Voice identified as First Officer Voice identified as Voice/source unidentified Aron-Canton Approach Control King Air 7llMie Chaie Unintelligible word or words Nonpertinent word Brea in continuity Questionable text Editorial insertion Pause USTerrh S-a-R Note: Ties expressed in t - of final ipact as 2l29:OO.O c.d.t. tie using the tie

32 APPENDX E zo 9 CJ u $ c -P (H a c cd zo M M 4 d -P cn 54 dh d d cu t

33 APPENDX E. M. b- v- +J= ria cdo 3- c c +J - H v 5 cd +J Q d E 1 C. 8 c a 2 c $ x. X E: E. E: +J fi-i 3-3 c 3 R 3 c. d x cd cn v) 2

34 APPENDX E * 9 h aj 2 a Ti F4. -P M r i 4 H ri (d n C. H * tr, aj s h M

35 F: c, c, 3 - APPENDX E d n $3 yc

36 APPEXDX E L. 2 vi M!- n.. cn cu W V GJ d h" s 4J z 5 1 u 5 sr.d!= E H g d x H c- +1,. s Q) E u * * d h

37 b APPENDX E *.... f- A P E! - c, M

38 APPENDX E * C. h 4J * $, bo 2 4J a 3 M c E +-' -. d 4: A U H

39 APPENDX E a d 2? * B r: e,.p x rn %- e, X 2 n 5 so W UJ d (u d C. cn L W L a arn d! 4J a - H rn w > > - HE: g

40 APPENDX E n 4 n 5 W ho F: a F: cd u. -P % d v3 v3 * i-l. ho

41 r-, n n c.. 8

42 '" '" M F: $4 - * d -P E CH d *ri E d cd R - al 2 5 P * M * W +, M

43 APPENDX E 2 M d a3.\ d L. a3 P U $ 3 c i C. H 2 a3 : n h a3 8 U.ri

44 APPENDX E aj W con n O M -PC!.d W Hw x r;l h s -P Ei c? rn h C. 2 4 rn -P rn 3 a r-i -lj N B -. * * e. 2 M.r( * n n a H

45 APPENDX E F: -P c: E: (do Ln $ Q) Pl a i w B -P 5 cd h 2 v1. F P r: B E:.??- Y.. M cuu 3

46 APPENDX E L * * * E H 2 E Q) +-.e F: E CJ c. F: M % 3 -tj F B -!J a i. n a d 2 d G.d & a 2 Y a rn 2. $4 u a 9 u u H U 4 2 u

47 APPENDX E A t- t n n a 2 *rn

48 APPENDX E 'a cd P a vi? B 4J f d. n - 5 R * 3 * * cu

49 APPENDX E e i 4 u B u cn 4 3 -P a cd E P 4-1 o % % o cd *

50 APPENDX E E B 1 " u. cu d -P Ld *.. o\ co '! co z2 NV 3 -=! 3.. co $3 cu cs (u f

51 APPENDX E & rn L * 4J -P 6. ' -P v a 9 v z.ri rn bo v 2 a 3 o v n 2.. \o \o L n ;?; co 33 cu f-.-. # cuu

52 APPENDX E cu cu n a).. 2 q co cu v 3 -E!. -P a e, -4. c F! M c v cd - E * cd v n sr, % 2 %- M 3 M -P H cn cu o 3.A E. a c s v % 2. M - R - co f- ;' c\!.. n M M.. L" c. n.. co 33 cu co '?! cu a) cd o -P a il

53 APPENDX E cu F1 -ts d B L c

54 - 5 - APPENDX F MAY appesen Approach Chal AKRON, OHO a AKRON-CANTON REG'L APT. lev 1228' 4" 55'N 81' 27'W NOTE: Custos on prior request. VOT 11.6 i HlRL HlALS SFL HlRL VAS-1 MRL MlRL RElL VAS-L ale in Feet D O ZOO 3 t r i : i 1 ADDTONAL RUNWAY NFORMATON USEABLE LENGTHS -LANDNG 8EYOND- LGHTNG Threshold Glide Slope TAKE. OFF RVR 5458' MlRL AlN FllFD WTHN AKE-OfiF AlN CELNG-VSBLTY lal 5 M,n OFFAllN P,rsn 3 FLED Mln WllHlNl lvrhlnc TAKE-Ofi Htn.ZHr3&4En Othcr SCHEDULED. c_ ialltla tnon-s All Carrnurr ls&proach Approad AR CARRER TAKE - OFF ALTERNATE "LLUSTRATON ONLY - NOT TO BE USED FOR NAVGATONAL PURPOSES"

55 APPENDX G MAY AKRON, OHO AKRON.CANTON lower Apt. lev 1228' AKRON-CANTON REG'L Var 4" W LS Rwv 1 GS NDB Ry-y 1 ApproachtR) Departure (R) Ground LOC 19.5 CAK :<' 87"-6* MSA 7'.186" "-186' Gpperen Approach Chad 3S". 135". 315' 31' 1 27'.1445' 1538' (NDB) 25 NoPT LOM MM PULL UP: (Miniu altitude to coence turn 1613') clibing RGHT turn to 3' direct ACO VOR and hold EAST, RGHT turns, 272" inbound, or as directed. STRAGHT-N LANDNG RWY 1 DH 413'(2'/1 DH 463'(257 nl54'(327') CRCLE.O.LAND "LLUSTRATON ONLY - NOT TO BE USED FOR NAVGATONAL PURPOSES" U, S, GOVERNMENT PRNTNG OFFCE: h4 3-1