Centaurus
27th Aug 2017, 15:28
The Autumn 1990 issue of BASI Journal No 7 included an article STAY IN TRIM AND OUT OF TROUBLE. Reproduced below:
At New Yorks La Guardia airport on 21 September 1989, a Boeing 737-400 with 63 persons on board crashed into the East River and broke into three sections.
The captain had abandoned the take-off attempt when the aircraft veered to the left but he was unable to prevent its leaving the runway and crashing into the adjacent river. Subsequent investigation revealed that the rudder trim was deflected substantially to the left prior to the commencement of the take-off roll.
On the Boeing 737-300 and 400 two hydraulic systems supply the rudder power control unit which provides directional control to the aircraft (each system can operate independently of the other). Sensing is provided by a feel and censoring unit utilising the mechanical action of springs, cams and rollers. Trim is accomplished by operating the trim knob (switch) located on the aft electronics panel on the centre pedestal, which is immediately forward of the jump seat position.
Rudder trim application will result in uneven rudder pedal positioning. This is because the trim biases the position of the feel and centering unit which feeds back through the control cables and repositions the cables. Reports indicate that inadvertent trim knob operation can occur when a person seated in the jump seat rests their feet on the pedestal against the knob, or when objects are placed on the aft electronic panel.
Full trim deflection will result in a 10.6 cm displacement of the pedals. Since this differential will also result in nose wheel steering tending to turn the aircraft, any significant rudder trim application on the ground should be apparent to the flight crew during the taxi and take-off phase. It should be noted that on this aircraft type, pedal displacement caused by trim can easily be compensated for and directional control maintained with the pedals and trim displacement will not reduce available rudder travel.
As a result of the September accident, a bulletin was issued recommending that flight crew verify that the rudder trim position indicator reads zero and that the pedals are not displaced before take-off. It was also recommended that on completion of the before take-off flight control check, both pilots should ensure that the control wheels and pedals are neutral.
While other factors may have contributed to the La Guardia accident, it is clear that an unrecognised out of trim condition in ANY aircraft can result in disaster.
End of BASI article.
In the Executive Summary to the NTSB accident report investigators stated:
On September 20, 1989, USAir, Inc. flight 5050 was departing New York City's LaGuardia Airport, Flushing, New York, for Charlotte Douglas International Airport, Charlotte, North Carolina. As the first officer began the take off on runway 31, he felt the airplane drift left. The captain noticed the left drift also and used the nosewheel tiller to help steer.
As the takeoff run progressed, the aircrew heard a "bang" and a continual rumbling noise. The captain then took over and rejected the takeoff but did not stop the airplane before running off the end of the runway into Bowery Bay. Instrument flight conditions prevailed at the time and the runway was wet. The National Transportation Safety Board determines that the probable cause of this accident was the captain's failure to exercise his command authority in a timely manner to reject the takeoff or take sufficient control to continue the takeoff, which was initiated with a mis-trimmed rudder.
Also causal was the captain's failure to detect the mis-trimmed rudder before the takeoff was attempted. The safety issues discussed in this report were the design and location of the rudder trim control on the Boeing 737-400, air crew coordination and communication during takeoffs, crew pairing, and crash survivability.
..
NTSB Investigation
The Safety Board collected about 90 reports of rudder trim anomalies for the Boeing -737 300/400 aircraft. The majority of these reports were received after the accident and were from pilots who had heard of or read about the accident in various publications. Boeing knew only of six anomalies and the FAA Maintenance Discrepancy Reports showed none. Many reports described the inadvertent setting of rudder trim by the foot of the jump seat occupant behind the captains seat. Their shoe sole pushed the trim knob counter clockwise and set left trim. Pages 28 and 29 of the NTSB Accident Report covers this in more detail. The NTSB report is well worth reading in its entirety especially as the Boeing 737-300/400 series aircraft is still used by Australian and New Zealand registered operators.
See: http://libraryonline.erau.edu/online-full-text/ntsb/aircraft-accident-reports/AAR90-03.pdf
..
At the time of the accident in September 1989 I was flying Boeing 737-200 aircraft in UK. The 737-200 had a manual rudder trim and I was yet to fly the 737-300. Naturally we had heard about the accident to the USAir 737 but only from media reports. In December 1989 I converted to the Boeing 737-300 using the Aer Lingus simulator in Dublin. Initially the first two simulator sessions were completed without drama despite the check captain being a rather curt type. The third session started off normally until it came to the first engine failure after take-off. This required a SID change of heading. I was in the left seat as PF and during the initial climb on one engine used my right hand to tweak some rudder trim. The rudder trim indicator is out of sight of the PF unless he turns his head around and down to see it. If flying on instruments, it is normal for the PF to either feel for the trim knob or ask the PM to set it for him. Within a few seconds the aircraft seemed to be wobbling all over the place as I tried to maintain the turn at 1000 feet agl.
The instructor was getting testy and I was getting tense at his barking. I asked the first officer to centralise the rudder trim and when this happened my flying smoothed out. I was still on instruments at the time. On the acceleration phase I again tweaked the rudder trim for two or three seconds and sure enough the aircraft started to yaw and roll. On the previous session I had no trouble with asymmetric handling but this was different. The instructor lost his cool and began berating me for sloppy flying.
By now I had enough of this clown and requested he stop the simulator for a coffee break while I cooled down. Reluctantly he did so and while he went his way and the first officer and I went ours, I asked the first officer why he took such a long time to zero the rudder trim when I had asked him earlier. He replied that the rudder trim was full scale left, even though I had only tweaked it for a couple of seconds. I immediately thought of the USAIR 737-400 accident where the crew were certain the rudder trim was zero before the take off, yet somewhere along the way it had gone to uncommanded full scale left.
The first officer and I quickly returned to the simulator and I tweaked the rudder trim left for five seconds and released the knob. To our astonishment the rudder trim indicator kept running until it reached the full left stop. We repeated the exercise with the same results. We then timed how long it took to move from neutral to full left trim. The answer was 29 seconds; about the same time I had encountered the directional control difficulties after tweaking the trim knob.
By now the check captain had returned to the cockpit and asked what was going on. We demonstrated the rudder trim event for his benefit. He still remained convinced my flying was at fault and refused my offer for him to fly the simulator while I failed an engine on him. He did say, however, that he would advise the simulator technicians. I told him we should advise the British CAA especially in view of the USAIR accident. He refused to agree to contacting the CAA so I did so the next day. They were most interested but I heard nothing more.
Several weeks passed and I was flying a British registered B737-300 on charter to an Australian operator. We had flown to Tullamarine (Melbourne) Airport and were doing a turn-around one evening. The inbound leg had been normal.
We carried an LAME and after the doors had closed he asked if he could occupy the jump seat. He was looking around the cockpit as we did our push back checks and remarked did you blokes have a rudder trim problem on the last leg? We looked at the rudder trim and it had moved uncommanded to half left trim. We had no idea when it could have happened but it must have been sometime during the turn around. I immediately wrote it up in the servicing document. The flight was delayed while the LAME did some trouble shooting but everything seemed to work OK.
To this day even in the simulator I make a careful check of the rudder trim as the last item before each take off. Habit, I suppose.
At New Yorks La Guardia airport on 21 September 1989, a Boeing 737-400 with 63 persons on board crashed into the East River and broke into three sections.
The captain had abandoned the take-off attempt when the aircraft veered to the left but he was unable to prevent its leaving the runway and crashing into the adjacent river. Subsequent investigation revealed that the rudder trim was deflected substantially to the left prior to the commencement of the take-off roll.
On the Boeing 737-300 and 400 two hydraulic systems supply the rudder power control unit which provides directional control to the aircraft (each system can operate independently of the other). Sensing is provided by a feel and censoring unit utilising the mechanical action of springs, cams and rollers. Trim is accomplished by operating the trim knob (switch) located on the aft electronics panel on the centre pedestal, which is immediately forward of the jump seat position.
Rudder trim application will result in uneven rudder pedal positioning. This is because the trim biases the position of the feel and centering unit which feeds back through the control cables and repositions the cables. Reports indicate that inadvertent trim knob operation can occur when a person seated in the jump seat rests their feet on the pedestal against the knob, or when objects are placed on the aft electronic panel.
Full trim deflection will result in a 10.6 cm displacement of the pedals. Since this differential will also result in nose wheel steering tending to turn the aircraft, any significant rudder trim application on the ground should be apparent to the flight crew during the taxi and take-off phase. It should be noted that on this aircraft type, pedal displacement caused by trim can easily be compensated for and directional control maintained with the pedals and trim displacement will not reduce available rudder travel.
As a result of the September accident, a bulletin was issued recommending that flight crew verify that the rudder trim position indicator reads zero and that the pedals are not displaced before take-off. It was also recommended that on completion of the before take-off flight control check, both pilots should ensure that the control wheels and pedals are neutral.
While other factors may have contributed to the La Guardia accident, it is clear that an unrecognised out of trim condition in ANY aircraft can result in disaster.
End of BASI article.
In the Executive Summary to the NTSB accident report investigators stated:
On September 20, 1989, USAir, Inc. flight 5050 was departing New York City's LaGuardia Airport, Flushing, New York, for Charlotte Douglas International Airport, Charlotte, North Carolina. As the first officer began the take off on runway 31, he felt the airplane drift left. The captain noticed the left drift also and used the nosewheel tiller to help steer.
As the takeoff run progressed, the aircrew heard a "bang" and a continual rumbling noise. The captain then took over and rejected the takeoff but did not stop the airplane before running off the end of the runway into Bowery Bay. Instrument flight conditions prevailed at the time and the runway was wet. The National Transportation Safety Board determines that the probable cause of this accident was the captain's failure to exercise his command authority in a timely manner to reject the takeoff or take sufficient control to continue the takeoff, which was initiated with a mis-trimmed rudder.
Also causal was the captain's failure to detect the mis-trimmed rudder before the takeoff was attempted. The safety issues discussed in this report were the design and location of the rudder trim control on the Boeing 737-400, air crew coordination and communication during takeoffs, crew pairing, and crash survivability.
..
NTSB Investigation
The Safety Board collected about 90 reports of rudder trim anomalies for the Boeing -737 300/400 aircraft. The majority of these reports were received after the accident and were from pilots who had heard of or read about the accident in various publications. Boeing knew only of six anomalies and the FAA Maintenance Discrepancy Reports showed none. Many reports described the inadvertent setting of rudder trim by the foot of the jump seat occupant behind the captains seat. Their shoe sole pushed the trim knob counter clockwise and set left trim. Pages 28 and 29 of the NTSB Accident Report covers this in more detail. The NTSB report is well worth reading in its entirety especially as the Boeing 737-300/400 series aircraft is still used by Australian and New Zealand registered operators.
See: http://libraryonline.erau.edu/online-full-text/ntsb/aircraft-accident-reports/AAR90-03.pdf
..
At the time of the accident in September 1989 I was flying Boeing 737-200 aircraft in UK. The 737-200 had a manual rudder trim and I was yet to fly the 737-300. Naturally we had heard about the accident to the USAir 737 but only from media reports. In December 1989 I converted to the Boeing 737-300 using the Aer Lingus simulator in Dublin. Initially the first two simulator sessions were completed without drama despite the check captain being a rather curt type. The third session started off normally until it came to the first engine failure after take-off. This required a SID change of heading. I was in the left seat as PF and during the initial climb on one engine used my right hand to tweak some rudder trim. The rudder trim indicator is out of sight of the PF unless he turns his head around and down to see it. If flying on instruments, it is normal for the PF to either feel for the trim knob or ask the PM to set it for him. Within a few seconds the aircraft seemed to be wobbling all over the place as I tried to maintain the turn at 1000 feet agl.
The instructor was getting testy and I was getting tense at his barking. I asked the first officer to centralise the rudder trim and when this happened my flying smoothed out. I was still on instruments at the time. On the acceleration phase I again tweaked the rudder trim for two or three seconds and sure enough the aircraft started to yaw and roll. On the previous session I had no trouble with asymmetric handling but this was different. The instructor lost his cool and began berating me for sloppy flying.
By now I had enough of this clown and requested he stop the simulator for a coffee break while I cooled down. Reluctantly he did so and while he went his way and the first officer and I went ours, I asked the first officer why he took such a long time to zero the rudder trim when I had asked him earlier. He replied that the rudder trim was full scale left, even though I had only tweaked it for a couple of seconds. I immediately thought of the USAIR 737-400 accident where the crew were certain the rudder trim was zero before the take off, yet somewhere along the way it had gone to uncommanded full scale left.
The first officer and I quickly returned to the simulator and I tweaked the rudder trim left for five seconds and released the knob. To our astonishment the rudder trim indicator kept running until it reached the full left stop. We repeated the exercise with the same results. We then timed how long it took to move from neutral to full left trim. The answer was 29 seconds; about the same time I had encountered the directional control difficulties after tweaking the trim knob.
By now the check captain had returned to the cockpit and asked what was going on. We demonstrated the rudder trim event for his benefit. He still remained convinced my flying was at fault and refused my offer for him to fly the simulator while I failed an engine on him. He did say, however, that he would advise the simulator technicians. I told him we should advise the British CAA especially in view of the USAIR accident. He refused to agree to contacting the CAA so I did so the next day. They were most interested but I heard nothing more.
Several weeks passed and I was flying a British registered B737-300 on charter to an Australian operator. We had flown to Tullamarine (Melbourne) Airport and were doing a turn-around one evening. The inbound leg had been normal.
We carried an LAME and after the doors had closed he asked if he could occupy the jump seat. He was looking around the cockpit as we did our push back checks and remarked did you blokes have a rudder trim problem on the last leg? We looked at the rudder trim and it had moved uncommanded to half left trim. We had no idea when it could have happened but it must have been sometime during the turn around. I immediately wrote it up in the servicing document. The flight was delayed while the LAME did some trouble shooting but everything seemed to work OK.
To this day even in the simulator I make a careful check of the rudder trim as the last item before each take off. Habit, I suppose.