Figure: The recovered acme screw, from NTSB Report, Figure 12c.

Eddie Sez:

First the facts:

  • The MD-80 was designed without a fail safe mechanism on the horizontal stabilizer.

  • Alaska Airlines extended lubrication, inspection, and replacement intervals on the horizontal stabilizer.

  • A lead mechanic at Alaska Airlines' Oakland maintenance facility reported maintenance short comings to the FAA fifteen months prior to the crash, and was placed on paid leave. As it turns out, over two years prior, he had ordered the horizontal stabilizer jack screw on this aircraft be replaced. His recommendation was overruled by the next shift.

  • The aircraft was put back into service, its next overhaul would be in two and a half years.

  • The pilots at first experienced a jammed stabilizer. They accomplished the appropriate checklists and ended up with an airplane that was barely controllable. The horizontal stabilizer jack screw was bare of any lubrication and the threads had been ground off. The stabilizer would not budge.

  • There were no procedures on what to do next and the pilots thought a successful landing was in doubt. They elected to use primary and secondary motors to free the stabilizer. This moved the stabilizer free and dislodged a retaining nut on the stabilizer. The stabilizer moved violently enough to completely dislocate the jack screw. The airplane entered an uncontrollable dive and all were killed.

The pilots fought the airplane all the way down, never once giving up. I cannot fault them for anything they did. I will relay a personal story in hopes that it may suggest a lesson for the future.

  • Fifteen years prior to this mishap I was one of several Boeing 707 captains flying for an Air Force squadron in Hawaii. One of our airplanes shuttered violently during a stabilizer trim preflight check and I refused to fly it. The next crew did fly it. Of course nothing happened but I pushed the matter up the chain of command and the airplane was taken out of service. They found the jackscrew threads had been ground away. They replaced the jack screw. I had no idea about jackscrews having their threads ground smooth, but I knew the airplane shouldn't have done that with stabilizer movement.

  • During certification the manufacturer flight tests various maneuvers to prove they can be done. They do this in sterile environments, good weather, and no passengers on board. If your trouble shooting suggests you try a maneuver, pull a circuit breaker, or try something not in the book, you have become the test pilot in less than ideal conditions flying with passengers on board.

I do not begrudge these pilots anything, they were given an impossible situation by a poor aircraft design and what was, at that time, a criminal maintenance program. But I think they could have survived a landing with the jammed stabilizer on the very long and wide runways at LAX or even an hour east at Edwards. Trying to break the stabilizer free with an untested procedure sealed their fates.

What follows are quotes from the sources listed below, as well as my comments in blue.


Accident Report


Narrative

[NTSB Aircraft Accident Report, AAR-02/01, section numbers noted]

Figure: Installation of jackscrew assembly, from NSTB Report, Figure 3.

[§1.6.1] Longitudinal control for the MD-80 (and for all DC-9, MD-90, and 717 series airplanes), that is, control of the airplane's pitch movements, is provided by the horizontal stabilizer and the elevators. The horizontal stabilizer is mounted on top of the 18-foot-high vertical stabilizer; they are connected by two hinges at the aft spar of the horizontal stabilizer and with a single jackscrew assembly at the front spar of the stabilizer in a T-tail configuration (see figure 3).

The primary trim motor is activated by the pilots' control wheel switches, the alternate by the autopilot. Mechanical trim wheels either side of the cockpit can be used to manually move the primary trim system.

The jackscrew was designed for a service life of 30,000 hours and not subject to periodic inspections, until 1966 after several assemblies were discovered to have excessive wear.

[§1.6.3.4] On September 26, 1997, the airplane entered a C check (which was completed on October 1, 1997) at the OAK maintenance facility, which included an end play check. The airplane had accumulated 17,699 flight hours. The records indicated that the elevators and horizontal stabilizer, including the jackscrew assembly, were lubricated.

[§1.6.3.4] September 27th (Saturday) — nonroutine work card (MIG-4) was generated following an initial end play check on the accident airplane by a day-shift mechanic and an inspector. The MIG-4 noted the following discrepancy: "Horizontal Stab—acme screw and nut has maximum allowable end play limit (.040 in.)." The "planned action" box, which was filled out by the day-shift lead mechanic and inspector, stated, "Replace nut and perform E.O. 8-55-10-01."

[§1.6.3.4] September 30th (Tuesday) — the graveyard-shift lead mechanic marked through the line on the MIG-4 that called for replacement of the jackscrew assembly and wrote, "re-evaluate test per WC [work card] 24627000." A different mechanic and inspector made an entry in the MIG-4's "corrective action" section, which stated, "Rechecked acme screw and nut end play per WC 24627000. Found end play to be within limits .033 for step 11 and .001 for step 12. Rechecked five times with same result."

Figure: The September 27, 1997, Non-routine Work Card (MIG-4), from NTSB Report, Figure 11.

[§1.1] FDR information from the accident flight indicated that during taxi for takeoff, when the FDR started recording, the horizontal stabilizer was at the 7° airplane-nose-up position, which was the takeoff pitch trim setting. About 1337, the accident airplane departed PVR as flight 261.

[§1.1] FDR data indicated that during the initial portion of the climb, the horizontal stabilizer moved at the primary trim motor rate of 1/3° per second from 7° to 2° airplane nose up. According to FDR data, the autopilot was engaged at 1340:12, as the airplane climbed through an altitude of approximately 6,200 feet. Thereafter, the FDR recorded horizontal stabilizer movement at the alternate trim motor rate of 1/10° per second from 2° airplane nose up to 0.4° airplane nose down. At 1349:51, as the airplane continued to climb through approximately 23,400 feet at 331 knots indicated airspeed (KIAS), the CVR recorded the horizontal stabilizer move from 0.25° to 0.4° airplane nose down. This was the last horizontal stabilizer movement recorded until the airplane's initial dive about 2 hours and 20 minutes later. At 1353:12, when the airplane was climbing through 28,557 feet at 296 KIAS, the autopilot disengaged.

[§1.12.2] The acme screw was found cracked but attached to the support assembly. Metallic filaments were found wrapped around the central part of the acme screw.

[§2.2.2] FDR data indicated that the horizontal stabilizer's last movement during the climbout was to 0.4o airplane nose down at 1349:51 as the airplane was climbing through 23,400 feet at 331 knots indicated airspeed (KIAS). After this, no horizontal stabilizer movement was recorded on the FDR until the airplane's initial dive 2 hours and 20 minutes later. This cessation of horizontal stabilizer movement is not consistent with a typical MD-80 climb profile, which normally would require additional stabilizer movements to maintain trim. Thus, as the accident airplane continued to climb above 23,400 feet without any horizontal stabilizer movement, the autopilot would have attempted to achieve trim by continuing to add elevator input to compensate for the lack of movement of the horizontal stabilizer. At 1353:12, when the airplane was at approximately 28,557 feet and 296 KIAS, the autopilot was disconnected, most likely by the flight crew in response to the trim annunciator warning light, which would have illuminated to indicate the airplane's out-of-trim condition.

It appears these metallic filaments came from the acme nut used to retain the acme screw below the lower stop and may have caused the initial jammed stabilizer.

[§1.1] FDR information and airplane performance calculations indicated that, during the next 7 minutes, the airplane continued to climb at a much slower rate. During this part of the ascent, the elevators were deflected between -1° and -3°, and the airplane was flown manually using up to as much as 50 pounds of control column pulling force. After reaching level flight, the airplane was flown for about 24 minutes using approximately 30 pounds of pulling force at approximately 31,050 feet and 280 KIAS. The airspeed was then increased to 301 KIAS, and the airplane was flown for almost another 1 hour 22 minutes using about 10 pounds of pulling force. At 1546:59, the autopilot was re-engaged.

[§1.1] At 1607:54, a mechanic at Alaska Airlines' LAX maintenance facility contacted the flight crew on the company radio frequency and asked, "are you [the] guys with the uh, horizontal [stabilizer] situation?" The captain replied, "affirmative," and the mechanic, referring to the stabilizer's primary trim system, asked, "did you try the suitcase handles and the pickle switches?" At 1608:03, the captain replied, "yea we tried everything together." At 1608:08, the captain added, "we've run just about everything if you've got any hidden circuit breakers we'd love to know about 'em." The mechanic stated that he would "look at the uh circuit breaker uh guide just as a double check." The LAX mechanic then asked the flight crew about the status of the alternate trim system, and, at 1608:35, the captain replied that "it appears to be jammed . . . the whole thing, it [the AC load meter] spikes out when we use the primary, we get AC [electrical] load that tells me the motor's trying to run but the brake won't move it. When we use the alternate, nothing happens."

During this time there was also considerable pressure put on the captain by Alaska Airlines personnel to continue to SFO, their destination, but he stuck to his guns and insisted on getting the airplane on the ground sooner and on a longer runway with less crosswind.

[§1.1] At 1608:50, the LAX mechanic asked, "you say you get a spike on the meter up there in the cockpit when you uh try to move it with the primary right?" According to the CVR transcript, at 1608:59, the captain addressed the first officer before responding to the mechanic, stating, "I'm gonna click it off you got it." One second later, the first officer replied, "ok." At 1609:01, the captain reiterated to the LAX mechanic that the spike occurred "when we do the primary trim but there's no appreciable uh change in the uh electrical uh when we do the alternate." The LAX mechanic replied that he would see them when they arrived at the LAX maintenance facility.

Figure: Flight path, from NSTB Report, Figure 1.

[§1.1] At 1609:13, the captain stated, "lets do that." At 1609:14.8, the CVR recorded the sound of a click and, at the same time, the captain stating, "this'll click it off." According to FDR data, the autopilot was disengaged at 1609:16. At the same time, the CVR recorded the sound of a clunk, followed by two faint thumps in short succession at 1609:16. The CVR recorded a sound similar to the horizontal stabilizer-in-motion tone at 1609:17. At 1609:19, the CVR again recorded a sound similar to the horizontal stabilizer-in-motion tone, followed by the captain's comment, "you got it?" (FDR data indicated that during the 3 to 4 seconds after the autopilot was disengaged, the horizontal stabilizer moved from 0.4° to a recorded position of 2.5° airplane nose down, and the airplane began to pitch nose down, starting a dive that lasted about 80 seconds as the airplane went from about 31,050 to between 23,000 and 24,000 feet. Figure 1 shows the accident airplane's flightpath, starting about 1609 (about the time of the initial dive) and ending about 1620 (about the time of the second and final dive). Figure 2 shows the radar altitude data and selected ATC transmissions for about the same time period.

[§2.2.3] Consequently, although operation of the primary trim motor as part of troubleshooting attempts earlier in the flight did not release the jam, the torque created by the primary trim motor when the captain activated the primary trim system at 1609:16 apparently provided enough force to overcome the jam between the acme nut and screw. Over the next 3 to 4 seconds, the increasing angle-of-attack of the horizontal stabilizer and the increased elevator deflections would have in turn increased the tension loads on the acme screw, contributing to the motion of the acme screw upward through the acme nut.

[§2.2.3] Therefore, the Safety Board concludes that the accident airplane's initial dive from 31,050 feet began when the jam between the acme screw and nut was overcome as a result of the operation of the primary trim motor. Release of the jam allowed the acme screw to pull up through the acme nut, causing the horizontal stabilizer leading edge to move upward, thus causing the airplane to pitch rapidly downward.

The airplane could have been landed safely but the odds were getting worse as their troubleshooting continued and exacerbated the damage. The crew was dealing with a situation not found in any of their checklists or taught in any of their training. But they had been trained to land the airplane with a jammed stabilizer. Sometimes it is better to accept the airplane as it is and leave the trouble shooting for on the ground.

[§1.1] At 1609:26, the captain stated, "it got worse," and, 5 seconds later, he stated "you're stalled." One second later, the CVR recorded a sound similar to airframe vibration getting louder. At 1609:33, the captain stated, "no no you gotta release it ya gotta release it." This statement was followed by the sound of a click 1 second later. At 1609:52, the captain stated, "help me back help me back." Two seconds later, the first officer responded, "ok."

[§1.1] One second later, at 1609:55, the captain contacted the Los Angeles Air Route Traffic Control Center (ARTCC) and stated, "Center Alaska two sixty one we are uh in a dive here." At 1610:01.6, the captain added, "and I've lost control, vertical pitch." At 1610:01.9, the CVR recorded the sound of the overspeed warning (which continued for the next 33 seconds). At 1610:05, the controller asked flight 261 to repeat the transmission, and, at 1610:06, the captain responded, "yea we're out of twenty six thousand feet, we are in a vertical dive . . . not a dive yet . . . but uh we've lost vertical control of our airplane." At 1610:20, the captain stated, "just help me."

[§1.1] At 1610:28.2, the captain informed the Los Angeles ARTCC controller, "We're at twenty three seven request uh." At 1610:33, the captain added, "yea we got it back under control here." One second later, the first officer transmitted, "no we don't." At 1610:45, the first officer stated, "let's take the speedbrakes off." One second later, the captain responded, "no no leave them there. it seems to be helping." At 1610:55, the captain stated, "ok it really wants to pitch down." At 1611:06.6, the captain stated that they were at "twenty four thousand feet, kinda stabilized." Three seconds later he added, "we're slowing here, and uh, we're gonna uh do a little troubleshooting, can you gimme a block [altitude] between uh, twenty and twenty five?" FDR data indicated that, by 1611:13, the airplane's airspeed had decreased to 262 KIAS, and the airplane was maintaining an altitude of approximately 24,400 feet with a pitch angle of 4.4°. At 1611:21, the controller assigned flight 261 a block altitude of between FL 200 and 250. Airplane performance calculations indicated that between about 130 and 140 pounds of pulling force was required to recover from the dive.

[§1.1] At 1611:43, the first officer stated, "whatever we did is no good, don't do that again." One second later, the captain responded, "yea, no it went down it went to full nose down." Four seconds later, the first officer asked, "uh it's a lot worse than it was?" At 1611:50, the captain replied, "yea yea we're in much worse shape now," adding, at 1611:59, "I think it's at the stop, full stop . . . and I'm thinking . . . can it go any worse . . . but it probably can . . . but when we slowed down, lets slow it lets get down to two hundred knots and see what happens."

[§1.1] At 1612:33, the captain told LAX maintenance, "we did both the pickle switch and the suitcase handles and it ran away full nose trim down." At 1612:42, the captain added, "and now we're in a pinch so we're holding uh we're worse than we were." At 1613:04, the captain indicated to LAX maintenance that he was reluctant to try troubleshooting the trim system again because the trim might "go in the other direction." At 1613:10, the LAX mechanic responded, "ok well your discretion uh if you want to try it, that's ok with me if not that's fine. um we'll see you at the gate." At 1613:22, the captain stated, "I went tab down . . . right, and it should have come back instead it went the other way." At 1613:32, the captain asked the first officer, "you wanna try it or not?" The first officer replied, "uhh no. boy I don't know." Airplane performance calculations indicated that about 120 pounds of pulling force was being applied to the pilots' control columns at this point.

[§1.1] At 1614:54, the Los Angeles ARTCC controller instructed the flight crew to contact another ARTCC controller on frequency 126.52, which the flight crew acknowledged. At 1615:19, the first officer contacted another ARTCC controller on 126.52 and stated, we're with you we're at twenty two five, we have a jammed stabilizer and we're maintaining altitude with difficulty. uh but uh we can maintain altitude we think . . . our intention is to land at Los Angeles." The controller cleared the airplane direct to LAX and then asked, "you want lower [altitude] now or what do you want to do sir?" At 1615:56, the captain replied, "I need to get down about ten, change my configuration, make sure I can control the jet and I'd like to do that out here over the bay if I may."

[§1.1] After the radio transmission, the captain told a flight attendant that he needed "everything picked up" and "everybody strapped down." At 1617:04, the captain added, "I'm gonna unload the airplane and see if we can . . . we can regain control of it that way." At 1617:09, the flight attendant stated, "ok we had like a big bang back there," and, the captain replied, "yea I heard it." At 1617:15, the captain stated, "I think the stab trim thing is broke." At 1617:21, the captain again told the flight attendant to make sure the passengers were "strapped in now," adding 3 seconds later, "cause I'm gonna I'm going to release the back pressure and see if I can get it . . . back."

[§1.1] At 1617:54, the captain stated, "gimme slats extend, and, at 1617:56.6, a sound similar to slat/flap handle movement was recorded by the CVR. At 1617:58, the captain added, "I'm test flyin now." At 1618:05, the captain commanded an 11° flap deployment, and, at 1618:07, a sound similar to slat/flap handle movement was recorded. At 1618:17, the captain stated, "its pretty stable right here . . . see but we got to get down to a hundred an[d] eighty [knots]." At 1618:26, the captain stated, "OK . . . bring bring the flaps and slats back up for me," and, at 1618:36.8, sounds similar to slat/flap handle movement were recorded. At 1618:47, the captain stated, "what I wanna do . . . is get the nose up . . . and then let the nose fall through and see if we can stab it when it's unloaded."

[§1.1] At 1618:56, the first officer responded, "you mean use this again? I don't think we should . . . . if it can fly." At 1619:01, the captain replied, "it's on the stop now, it's on the stop." At 1619:04, the first officer replied, "well not according to that it's not." At this time, FDR data indicated a horizontal stabilizer angle of 2.5° airplane nose down. Three seconds later, the first officer added, "the trim might be, and then it might be uh, if something's popped back there . . . it might be mechanical damage too." At 1619:14, the first officer stated, "I think if it's controllable, we oughta just try to land it." Two seconds later, the captain replied, "you think so? ok lets head for LA."

Throughout the crew resource management was very good, as this exchange illustrates. Unfortunately, the last change in slat/flaps may induced more stress on the horizontal stabilizer than it could withstand.

Figure: Acme nut, from NSTB Report, Figure 16.

[§1.12.2] The upper mechanical stop was found attached to the splines of the acme screw, and the clamp bolt for this stop was in place. The lower mechanical stop and its clamp bolt were recovered as a separate unit from the acme screw.

[§1.16.2.5] The acme nut was recovered attached to vertical stabilizer wreckage in its normal position but separate from the acme screw and the remainder of the jackscrew assembly. On-scene visual inspection of the thread area on the inside diameter of the acme nut showed a relatively smooth, flat surface, with only small ridges of the acme thread remaining.

It appears the upper portion of the stabilizer exerted so much force on the acme screw that the acme screw stripped itself out of the acme nut and the stabilizer flipped into the relative wind, causing a total loss of control.

[§2.2.4] At 1619:29, the captain ordered redeployment of the slats and flaps. At 1619:35, the flaps were transitioning from 7° to 11°. At 1619:36.6, the CVR recorded the sound of an "extremely loud noise." Immediately thereafter, the airplane began its final dive. At the time of this pitchover, radar detected several small primary returns, consistent with parts of the vertical stabilizer's tip fairing being torn from the airplane as the fairing brackets broke, which would have allowed the horizontal stabilizer to move well beyond the 3.6° airplane-nose-down position it was being held at by the brackets. Therefore, the extremely loud noise recorded on the CVR at 1619:36.6 was likely made as the fairing brackets failed and caused the loss of the tip fairing and structural deformation of the tail under flight loads, resulting in local aerodynamic disturbances.

[§1.1] About 5 seconds later, the CVR recorded the sound of a series of at least four distinct "thumps." At 1619:24, the first officer asked, "you feel that?" and the captain replied, "yea." At 1619:29, the captain stated, "ok gimme sl---." At 1619:32.8, the CVR recorded the sound of two clicks similar to the sound of slat/flap movement. At 1619:36.6, the CVR recorded the sound of an "extremely loud noise" and the sound of background noise increasing, which continued until the end of the recording. At the same time, the CVR also recorded sounds similar to loose articles moving around the cockpit. FDR data indicated that at 1619:36.6, the flaps were extending and the slats were moving to the mid position. The next few seconds of FDR data indicated a maximum airplane-nose-down pitch rate of nearly 25° per second. The FDR recorded a significant decrease in vertical acceleration values (negative Gs),26 a nose-down pitch angle, and a significant decrease in lateral acceleration values. By 1619:40, the airplane was rolling left wing down, and the rudder was deflected 3° to the right.

Figure: Radar altitude, from NSTB Report, Figure 2.

[§1.1] FDR data indicated that, by 1619:42, the airplane had reached its maximum valid recorded airplane-nose-down pitch angle of -70°. At this time, the roll angle was passing through -76° left wing down. At 1619:43, the first officer stated, "mayday," but did not make a radio transmission. Six seconds later, the captain stated, "push and roll, push and roll." FDR data indicated that, by 1619:45, the pitch angle had increased to -28°, and the airplane had rolled to -180° (inverted). Further, the airplane had descended to 16,420 feet, and the indicated airspeed had decreased to 208 knots.

[§1.1] At 1619:54, the captain stated, "ok, we are inverted. . . and now we gotta get it." FDR data indicated that at this time, the left aileron moved to more than 16° (to command right wing down), then, during the next 6 seconds, it moved in the opposite direction to -13° (to command left wing down). At 1619:57, the rudder returned to the near 0° position, the flaps were retracted, and the airplane was rolling through -150° with an airplane-nose-down pitch angle of -9°. After 1619:57, the airplane remained near inverted and its pitch oscillated in the nose-down position.

[§1.1] At 1620:04, the captain stated, "push push push. . . push the blue side up." At 1620:16, the captain stated, "ok now lets kick rudder. . . left rudder left rudder." Two seconds later, the first officer replied, "I can't reach it." At 1620:20, the captain replied, "ok right rudder. . . right rudder." At 1620:38, the captain stated, "gotta get it over again. . . at least upside down we're flyin." At 1620:49, the CVR recorded sounds similar to engine compressor stalls and engine spooldown. At 1620:54, the captain commanded deployment of the speedbrakes, and, about 1 second later, the first officer replied, "got it." At 1620:56.2, the captain stated, "ah here we go." The FDR recording ended at 1620:56.3, and the CVR recording ended at 1620:57.1.

[§1.1] The airplane impacted the Pacific Ocean near Port Hueneme, California. Pieces of the airplane wreckage were found floating on and beneath the surface of the ocean. The main wreckage was found at 34° 03.5' north latitude and 119° 20.8' west longitude.


Probable Cause

[NTSB Aircraft Accident Report, AAR-02/01, §3.1

[NTSB Aircraft Accident Report, AAR-02/01, §3.2


See Also:

Abnormal Procedures & Techniques / Maintenance Malpractice


References

May Day: Cutting Corners, Cineflix, Episode 5, Season 1, 15 October 2003 (Alaska Airlines 261)

NTSB Aircraft Accident Report, AAR-02/01, Loss of Control and Impact with Pacific Ocean Alaska Airlines Flight 261, McDonnell Douglas MD-83, N963AS, About 2.7 Miles North of Anacap Island, California, January 31, 2000