A Klasjet Boeing 737-500, registration LY-KLJ performing flight KLJ-125 from Madrid,SP (Spain) to Kaunas (Lithuania) with 57 passengers and 8 crew, was climbing out of Madrid's runway 14L when the crew stopped the climb at about 5000 feet due to problems with their navigation systems. The crew decided to return to Madrid, the aircraft entered a climb at more than 3000 fpm followed by a descent with more than 2500 fpm, misaligned with their assigned runway 18 aborting the approach early, positioned for a second approach but again were left of both runways 18 and went around. Due to weather the crew subsequently decided to divert to Getafe Airforce Base,SP (Spain) where the aircraft managed a safe landing about 40 minutes after departure.
Klasjet reported the passengers were never in danger, the crew put flight safety as first priority and therefore decided to divert to Getafe.
The aircraft was carrying the basketball team of Zalgiris, who had won their match against the Euroleague team of Real Madrid and had earned their qualification for the Euroleague Playoff (which the club stated was "unthinkable"), and were returning to their home in Kaunas.
The aircraft is still on the ground in Getafe a week (168 hours) after landing.
On Apr 12th 2019 Spain's CIAIAC reported that during preflight preparations a fault with one of the aircraft's navigation systems had been spotted. An engineer and the crew reviewed the minimum equipment list (MEL) verifying that the aircraft was permitted to be dispatched with this equipment inoperative and released the aircraft under the related MEL requirements. During climb out a number of faults occurred also in the first officer's navigation systems prompting the crew to return to Madrid without the automatic navigation equipment available. Following two landing attempts which ended in go arounds due to adverse weather conditions the crew declared emergency and diverted the aircraft to Getafe Airforce Base for a safe landing. Spain's CIAIAC have opened an investigation into the occurrence.
Spain's CIAIAC released their final report concluding the probable cause of the incident was:
The investigation has determined that the incident was caused by the problems the crew had operating the aircraft in instrument flight conditions after losing both of the aircrafts automatic flight systems.
The CIAIAC summarized the sequence of events:
During the pre-flight inspection, they [the crew] identified a fault in the captains automatic flight system, so both pilots, with help from a company engineer, reviewed the Minimum Equipment List (MEL) and verified that it allowed dispatching the aircraft with this equipment inoperative.
They decided to proceed with the flight and after receiving the relevant clearance, took off from runway 14L at 14:15:26.
During the climb, several faults occurred with the first officers automatic flight system, which eventually became inoperative at 14:17:32, so the crew decided to return to the departure airport while flying without assistance from the automatic flight systems.
Spains air traffic control manager, ENAIRE, reported that shortly after takeoff, the crew declared an emergency but did not report the exact nature of their problem.
At 14:20, the operations supervisor informed the approach controller that he was transferring him an aircraft that had just taken off from Madrid-Barajas whose crew had declared an emergency and wanted to return to the airport.
Based on information provided by the controller, he cleared the runway 18R localizer by diverting two aircraft, AC/1 and AC/2, which were on approach to this runway. He also instructed them to adjust their speeds to maintain the required separation as much as possible, both to each other and to those that were already positioned at the localizer.
The traffic that declared an emergency did not lower its speed properly and crossed the two localizers at 250 kt behind a third aircraft, different from the two mentioned above, A/C 3.
The controller then corrected the approach vector he had initially provided so that the aircraft could intercept the localizer for runway 18L (180 heading).
The crew ended up going around at 14:23:10.
The controller asked if they had problems with the speed, since he did not know the nature of their emergency, but the crew again requested vectors to land.
Because of its position, the aircraft had to climb to maintain the minimum altitude, but the controller saw that it was not climbing.
He could not turn it toward Casas de Uceda because there were aircraft at the runway 18R localizer and it would have had to continue toward heading 360 to try to intercept it from behind.
Twice he informed the crew that they were below minimums and instructed them to climb, but they did not carry out this instruction, since the aircraft was at 4,400 ft and entering an area where the minimum was 6,700 ft.
It turned to 220 at 220 kt and positioned itself behind A/C 2, which was at the other localizer.
It managed to intercept the localizer 11 NM out, at an altitude of 4,300 ft and a speed of 170 kt.
At 8 NM out, it had not acquired visual contact with the runway. Its speed was 150 kt.
It remained at the localizer, and the controller transferred the aircraft to a colleague who was on the final approach sequence (TWR 118.680) just before the aircraft reached a distance of 4 NM from the DME.
At a distance of 2.5 NM DME, it changed localizers and the controllers realized that it had gone around again. It was 14:33:58.
When the aircraft declared the emergency, the airport activated the local alert and the airport firefighters were standing by to respond to any situation during the landing on runway 18L.
After two failed landing attempts, the aircraft was diverted by air traffic control to the Getafe Air Base (LEGT) in Madrid, where it landed on runway 23 at 14:53.
There were no injuries and the aircraft was not damaged.
At Madrid-Barajas, the runway from which the aircraft had taken off was checked at 14:29. No foreign debris was found.
The CIAIAC analysed:
When analyzing this event, the first thing to note is the fact that having both autopilots become inoperative did not prevent proceeding with the flight in instrument conditions, since the instruments required to carry out a flight of these characteristics were available to the crew at all times; namely, they had the artificial horizon, altimeter, variometer, anemometer, compass, turn and bank indicator and the engine instruments (intake pressure and engine pressure ratio).
The crewmembers had instrument flight ratings and, based on the information gathered, had considerable experience, both in general and on the type.
In the case of the captain, he was a type rating instructor (TRI(A)), meaning he was not only very familiar with the airplane, its systems and its operation, but he had to be able to explain these concepts, that is, to convey them during training to other crewmembers in an operational setting, and to other pilots in general in a training setting.
If the first officer was the pilot flying (PF), it was because the captain noticed that the air speed cursor flag on his own side was inoperative, and since it is the captain who has to know all the aircraft systems well, he might have thought that the flight computer was affected. They therefore engaged the computer on the other side to do the flight.
An analysis of the data recorded on the FDR does not show any attempt to engage the autopilot on the captains side (AP A). The flight director on that side (F/D A) was off from the start of the flight, although it should have been on, since the F/D is independent from the autopilot. The crew had set up the cockpit to have both the AP and F/D on the first officers side be the master, meaning they were turned on first so that for both systems, the ones on the right side provided the guidance. The AP were inoperative, but if both F/D had been on, the indications provided would have been from the right side.
However, the data recorded in the FDR show that the F/D display on the left side was off, meaning the captain did not have guidance from the F/D on his side.
The data recorded in the FDR clearly indicate that they had considerable problems maintaining the basic flight parameters (altitude, speed, heading, etc.). This is most obvious when analyzing the turns.
The bad weather conditions complicated the operation.
There was turbulence and significant cloud cover between 2,000 and 4,000 ft in and around the Madrid-Barajas Airport. This prevented the crew from making a safe visual approach, and they probably did not have the runway in sight.
After the second failed landing attempt, the aircraft was diverted to the Madrid-Getafe Air Base, probably because ATS thought that a third attempt would entail delays and safety risks for other traffic arriving at the Madrid-Barajas Airport, but not because the weather conditions were clearly better at the Madrid-Getafe Airport. The clouds were just as low, though perhaps not as dense on the surface, which helped with the landing.
Also worth noting is the communication made at 14:46 between approach (APP) and the tower at the Getafe Air Base (TWR GE), in which APP reported that the crew did not speak English well and were unable to clearly describe the emergency they were experiencing.
Their English proficiency was rated as 4, which should have been sufficient to correctly communicate the nature of the problem.
The problems maintaining the basic flight parameters due to the absence of automatic control systems indicates that neither the decision making nor the cooperation between the crew were adequate.
All of these aspects, which could be the subject of a safety recommendation, were already noted by the Operator in its report on the event; as a result, no additional recommendations are necessary.
The Operator also noted the need to track the defects associated with the autopilot system.
The CIAIAC described the maintenance following the occurrence:
A review of the aircrafts technical logbook (TLB) since the day of the incident showed that the maintenance tasks on it were performed by six different maintenance technicians, identified in the TLB as follows:
The maintenance done and the entries in the TLB are described below:
On 5 April 2019 at the Barajas Airport, before the incident flight, the daily check, done by technician LT.ES.66.404 is signed in TLB #02387. The same TLB shows a fault of the air speed cursor as a result of the failed test conducted by the pilots during the pre-flight inspection (see 1.6.5, List of pre-flight procedures for the pilots), which displayed a flag in the EADI. Because of this, the crew decided to defer the Air Speed Cursor Flag, as per MEL 34-1-04.
After the incident, with the aircraft parked at the Getafe Air Base, the crew noted the following faults in technical log page (TLP) 02388:
I. Auto pilot A and B Inop.
II. Airspeed Cursor Flag (Left side)
III. Time to time Speed trim and mach trim fail.
As a result of these entries, technician 503 did the operational check of the autopilot as per AMM 22-11-33, with the result being Autopilot A = Fail and Autopilot B = OK.
- On 6 April 2019, technician 503 deferred Autopilot A as per MEL Section 22-01A (TLP02394).
- On 9 April 2019, technician LT.ES.66.404 logged the replacement of the hydraulic pump module assembly in the TLB, as per AMM Section 29-15-95 (entry TLM02390).
- On 15 April 2019, the same technician, 503, replaced the A flight control computer (AFCC A), the A digital air data computer (DADC A) and the accessory unit, which were annotated in the logbook as entry TLP02388.
He also logged a wiring inspection, finding and repairing an insulator in poor condition (entry TLP02395), and he expanded the HIL for the A autopilot.
There is no record of any entries involving the results of the checks after replacing the FCC or the DADC. According to the AMM, a test is required after these changes (AMM Section 22-11-33).
Had the test been done and yielded a satisfactory result, the deferred item could have been closed out; instead, the deferred item was extended. Neither the document for extending the deferred item nor the entry in the TLB for said extension could be found.
That same day, technician LT.ES.66.404 logged in TLB 02392 the performance of JIC B71-00-03L (entry TLP02396).
The airplane flew from the Getafe Air Base to the Vilnius Airport (Lithuania). No entry was made regarding all of these previous faults.
- On 24 April 2019, due to the presence of Autopilot A inop in the HIL, #24, entry TLP 02398 was made in the job sheet (Job Sheet #02398) made by a different technician, 382, as documented in the TLB. However, the job sheet was signed by technicians 480 and 185 as the inspectors.
The entries make reference to the following actions:
- Troubleshooting performed that involved checking the entire cabling installation in the area of FCC A. The built-in test equipment (BITE) was also used to run a test of the DFCS, specifically, the sensor values. The test showed that the hydraulic pressure switch for actuator A on the elevator was in the wrong position. As a result, both the actuator and the pressure switch were replaced. The hydraulic pressure switch on the autopilot for the elevator was also checked and the system was tested, with satisfactory results.
- Following these actions, deferred HIL entry #24 Autopilot A inop, was closed out.
- On 24 April 2019, a test flight was conducted (entry TLP02400) by the same captain from the incident flight, who noted the following:
- DURING CLIMB APR. FL240 LOST OF EADI ON LH (BECOMES BLANK). AFTER 4 s POWER RESTORES AUTOMATICALLY ON EADI, EHSI. AUTOPILOT, FLIGHT DIR, AUTOTHROTLE DISCONNECTS AUTOMATICALLY. YAW DAMPR STAYS IN OFF.
- Subsequently, another operator, technician 127, did troubleshooting as per AMM 34-22-00, which gave a TEST FAIL SG, as a result of which the #1 symbol generator was replaced.
- The troubleshooting was continued and relays C801, R3, R41 and R349 were replaced as per AMM 24-21-00. The AC generation system test was then performed again, giving a satisfactory result.
As per BOEINGs recommendation, the inertial reference units (IRU) were replaced.
- On 20 April 2019, a test flight was carried out, with no faults being identified (entry TL). (editorial note: the test flight was actually on Apr 30th 2019)
Metars Madrid Barajas Airport:
LEMD 051400Z 23007G21KT 9999 BKN018 08/04 Q0998 R88/29//95 NOSIG=
COR LEMD 051330Z 24008KT 200V280 9999 SCT025 BKN040 07/04 Q0998 R88/29//95 NOSIG=
LEMD 051300Z 24013KT 200V280 8000 SCT018TCU BKN030 07/04 Q0999 RESHRA NOSIG=
LEMD 051250Z 23012KT 210V280 6000 -SHRA SCT020TCU BKN025 07/05 Q0999 RESHRA NOSIG=
COR LEMD 051238Z 22016KT 4000 +SHRA SCT020TCU BKN025 06/05 Q0999 TEMPO NSW=
LEMD 051230Z 23017KT 6000 2500S SHRA SCT020TCU BKN025 07/04 Q0999 NOSIG=
LEMD 051200Z 16014G24KT 110V190 9999 -RA BKN020 BKN040 09/06 Q0999 NOSIG=
COR LEMD 051130Z 16012KT 120V200 9999 BKN020 BKN040 08/05 Q0999 R88/CLRD// NOSIG=
LEMD 051100Z 16010KT 120V190 9999 SCT015 BKN070 08/05 Q0999 R88/29//95 NOSIG=
LEMD 051030Z 17010KT 130V190 9999 SCT012 BKN030 07/04 Q0999 R88/29//95 NOSIG=
COR LEMD 051000Z 17009KT 070V190 9999 SCT012 BKN030 07/05 Q1000 R88/29//95 NOSIG=
LEGT 051400Z 20011KT 200V270 9999 FEW017 SCT020 BKN024 09/04 Q0998=
LEGT 051300Z 25010KT 220V280 9999 -RA SCT012 BKN024 BKN033 07/05 Q0999=
LEGT 051243Z 26013G23KT 230V290 9999 -RA FEW012 SCT015 BKN021 07/04 Q0999=
LEGT 051208Z 22015G25KT 170V250 2500 +RA FEW007 BKN014 BKN017 06/03 Q0999=
LEGT 051200Z 19012G22KT 7000 RA SCT012 BKN017 BKN035 08/06 Q0999=
LEGT 051100Z 18012KT 150V210 9999 SCT014 BKN016 BKN025 09/05 Q0999=
LEGT 051000Z 17010KT 140V200 9999 SCT011 BKN015 BKN027 08/05 Q1000=