Skip Ribbon Commands
Skip to main content
Safety Recommendation Details

Safety Recommendation A-00-097
Details
Synopsis: .In this letter, the national transportation safety board recommends that the Federal Aviation Administration (FAA) take action to address the following safety issues: air carrier pilot training in landing techniques and bounced landing recovery, training tools and policies that promote proactive decision-making to go around if an approach is unstabilized, the use of on board computers to determine the required runway length for landing, md-11 handling characteristics and structural integrity requirements, and hard landing inspection requirements. The safety board identified these issues in its investigation of the 1997 accident involving Federal Express flight 14 in Newark, NJ. This letter summarizes the Board's rationale for issuing these recommendations
Recommendation: TO THE FEDERAL AVIATION ADMINISTRATION: Require, on all MD-11's equipped with the flight control computer-908 software, the retrofit of digital flight data recorder systems with all additional parameters required to precisely identify and differentiate between pilot and longitudinal stability augmentation system (LSAS) elevator control activity, including control column force, inertial reference unit pitch rate, LSAS command signals, elevator positions, and automatic ground spoiler command signals.
Original recommendation transmittal letter: PDF
Overall Status: Closed - Unacceptable Action
Mode: Aviation
Location: Newark, NJ, United States
Is Reiterated: No
Is Hazmat: No
Is NPRM: No
Accident #: DCA97MA055
Accident Reports: Crash During Landing Federal Express, Inc., McDonnell Douglas MD-11, N611FE
Report #: AAR-00-02
Accident Date: 7/31/1997
Issue Date: 8/25/2000
Date Closed: 3/24/2006
Addressee(s) and Addressee Status: FAA (Closed - Unacceptable Action)
Keyword(s): Flight Computer, Recorders

Safety Recommendation History
From: NTSB
To: FAA
Date: 3/24/2006
Response: The FAA continues to disagree with the Safety Board on this recommendation and has stated that (1) LSAS is not a unique and novel system that requires recording in accordance with 14 Code of Federal Regulations Section 25.1459(e) and (2) there is no other supporting regulatory basis to require the recording of LSAS parameters. The FAA has suggested that, using a Boeing-developed algorithm, the LSAS parameters can be reliably calculated from other signals already recorded by DFDRs on FCC 908-equipped airplanes. The Safety Board has carefully reviewed data from tests of this algorithm and indicated to the FAA that we do not believe this algorithm is capable of supplying sufficiently accurate data needed for an accident investigation in which LSAS operation may be a factor. Further discussion between our agencies is not likely to resolve this disagreement. Consequently, Safety Recommendation A-00-97 is classified "Closed-Unacceptable Action."

From: FAA
To: NTSB
Date: 9/8/2005
Response: Letter Mail Controlled 9/14/2005 3:02:19 PM MC# 2050429 Marion C. Blakey, Administrator, FAA, 9/8/05 The Board's position is that a requirement for additional flight data recorder (FDR) parameters is supported by 14 CFR 25.1459(e), which states, "Any novel or unique design or operational characteristic of the aircraft shall be evaluated to determine if any dedicated parameters must be recorded on flight recorders in addition to or in place of existing requirements." The FAA has determined that LSAS is not a unique and novel system that requires recording in accordance with 14 CFR 25.1459(e). Such pilot-machine interactions are endemic on modem aircraft. Furthermore, the FAA has determined that there is no other supporting regulatory basis to require the recording of longitudinal stability augmentation system (LSAS) parameters. Therefore, in the absence of voluntary compliance by the operators there is no reasonable way to require the recommended parameters be implemented without changing the FDR operational rules. However, understanding the Board's concerns, the FAA has worked with Boeing to validate a method, using optimal control logic, to derive a pitch rate from the available DFDR data, which analytically derives LSAS commands. Based on Boeing flight test data gathered during a manual flight director approach where there is a high degree of elevator activity during a closely coupled pilot-tracking task, the calculated LSAS input closely resembles the actual LSAS command data. Therefore, until such time as our FDR operational rules are updated requiring additional parameters, such as LSAS commands, the FAA still believes that the capability to analytically derived LSAS commands addresses the intent of this safety recommendation I believe that the FAA has satisfactorily responded to this safety recommendation, and I look forward to your response.

From: NTSB
To: FAA
Date: 1/13/2005
Response: On November 13, 2001, the FAA responded to Safety Recommendation A-00-97 and indicated its belief that the already-recorded elevator position and control column position data can be used to derive the difference between pilot and LSAS elevator control activity. As a result, the FAA did not believe the inertial reference unit pitch rate and automatic ground spoiler command signals were required to identify and differentiate between pilot and LSAS elevator control activity. On July 29, 2002, the Safety Board replied that during the investigation of the FedEx flight 14 accident, Board staff met with stability and control engineers from Boeing to review the FCC-908 software functionality and to perform flight simulator tests with the FCC-908 software upgrade in place. During these meetings, Boeing engineers indicated that LSAS command signals would be required to precisely identify and differentiate between pilot and LSAS control activity after an accident or incident involving an MD-11 with the FCC-908 software upgrade. Boeing also indicated that it had developed Service Bulletin (SB) MD-11-22-20 to enable FCC output of LSAS command signal parameters to the digital flight data acquisition unit, which would in turn enable those LSAS parameters to be recorded by the DFDR. The Board stated its belief that the LSAS command signal parameters, and the additional parameters in Safety Recommendation A-00-97, should be recorded on the MD-11 DFDR to satisfy 14 Code of Federal Regulations Section 25.1459, based on the "novel or unique design" aspects of the FCC-908 software. In July 2002, the Safety Board further stated that it had contacted Boeing and asked for proof that LSAS control activity can be analytically derived. The Boeing staff stated that they had never attempted to derive LSAS control activity, but that they believed it possible and would attempt to do so. Boeing attempted to match the LSAS parameters recorded on the DFDR from a United Parcel Service (UPS) MD-11 equipped to record LSAS parameters with LSAS signals derived from other DFDR signals that would be available from a DFDR not equipped to record LSAS parameters. In July 2002, Boeing still had not achieved a satisfactory match between the derived and measured LSAS data. The Board noted that the flight for which the data were being analyzed was not an accident flight and featured data from a benign and largely steady-state flight. In accidents and incidents that the Board investigates where this data may be needed, the DFDR may record rapid and large pitch control oscillations, making a derivation even more difficult. The Board disagreed with the FAA's position that it is practical or possible to reliably and accurately determine LSAS activity without the inclusion of the recommend data on MD-11 DFDRs and at that time classified Safety Recommendation A-00-97 "Open-Unacceptable Response." At the February 3, 2004, meeting, the FAA stated that Boeing had developed a revised and improved algorithm that the FAA believed was accurate and readily usable. The FAA asked the Safety Board to evaluate this algorithm, and the FAA stated its belief that the algorithm was an acceptable alternative. Shortly after that meeting, the FAA indicated to the Board that (1) it had determined that LSAS is not a unique and novel system that requires recording in accordance with Section 25.1459(e) and (2) there was no other supporting regulatory basis to require the recording of LSAS parameters. The FAA supplied a graph comparing the LSAS signal derived using Boeing's revised algorithm with the actual LSAS signal recorded on the DFDR. The FAA also reported that (1) only 14 of the 195 MD-11s registered in the United States have the LSAS parameters on their DFDR despite Boeing SB MD-11-22-20 and (2) no other operators plan to make this change. The Safety Board reviewed the data comparing the algorithm results with actual recorded data and does not believe the algorithm results are sufficiently accurate for accident investigations that might examine the role of the LSAS in an accident or incident involving large, high-rate control column inputs and pitch axis activity. The Board also continues to believe that the MD-11 LSAS functions are a unique and novel design under Section 25.1459(e). The Board, therefore, concludes that the need for this recommendation remains and that Boeing's algorithm is not an acceptable alternative. The Board asks the FAA to reconsider its decision to not take the recommended action. Pending the recommended action, Safety Recommendation A-00-97 remains classified "Open--Unacceptable Response."

From: FAA
To: NTSB
Date: 5/26/2004
Response: email: A-00-97, the algorithm that was developed is sufficient and we feel a satisfactory alternate action to respond to this recommendation.

From: FAA
To: NTSB
Date: 4/22/2004
Response: email: The FAA has determined that LSAS is not a unique and novel system that requires recording in accordance with ยง 25.1459(e). Such pilot machine interactions are endemic on modern aircraft. Furthermore, the FAA has determined that there is no other supporting regulatory basis to require the recording of LSAS parameters. However, the FAA, being sensitive to the NTSB concerns, has worked with Boeing to validate a method, using optimal control logic to derive a pitch rate from the available DFDR data, that analytically derived LSAS commands (see attachment 1 for details). Based on Boeing flight test data, the calculated LSAS input closely resembles the actual LSAS command data (see attachment 2 for data). The NTSB did meet with Boeing to discuss the latest methods and associated data from the UPS MD-11. The test data supplied to the NTSB in our last response was obtained from a UPS MD-11 equipped with the expanded parameter DFDR and the data was captured during normal revenue service. Based on the discussion at the SWAT meeting, the NTSB's position is that the derivation of the test data would be better evaluated during an unstablized approach with large column inputs. This would require Boeing to conduct a dedicated test. It is unlikely that Boeing would be willing to voluntarily conduct such testing since the MD-11 is out of production and they would need to obtain an aircraft. However, as noted in the SWAT meeting, the NTSB generally does not accept the alternative of calculating or deriving a signal from other data on a DFDR. Regarding the question if the data from the normal landing scenario represent a valid worst case LSAS activity tracking task. This question received mixed reaction from Boeing specialists. Some believe a test more representative of the Newark accident with large control column inputs could yield data skewing errors associated with the DFDR sampling rate limitations of the data used to calculate the LSAS commands. Others believe that during such large column position deflections, although there may be some data skew due to the sample rate limitations, the LSAS commands can be derived with reasonable results. It is also noted that during large control column deflections, LSAS contribution to elevator deflection is reduced due to the LSAS 5 degree limit of elevator authority and that during a normal stabilized approach, LSAS has a much greater affect on elevator position, as a percentage of elevator deflection. The FAA continues to believe this data indicates that the derived LSAS commands track reasonably well during a closed loop tracking task such as in the landing phase, where LSAS is very active. United Parcel Service is currently operating 12 MD-11s and is the only operator that records LSAS commands since they elected to install the expanded parameter DFDR. Gemini Airlines is in the process of installing the expanded DFDR on their 2 MD-11s. Therefore, there are 12 airplanes, soon to be 14, out of 195 MD-11s that have the capability to record LSAS commands. We are willing to facilitate a meeting between the NTSB staff members, FAA staff and Boeing if the Board would like to discuss our methodology for analytically deriving LSAS commands. Although this was not discussed during the February 3, 2004 SWAT meeting, if the NTSB continues to believe that FAA's proposed analytical method of LSAS derivation does not meet the intent of the recommendation, we may consider issuing a SAIB which would recommend the installation of expanded DFDRs that can readily record the LSAS signals.

From: NTSB
To: FAA
Date: 7/29/2002
Response: During the investigation of the accident that prompted this recommendation, Safety Board staff met with stability and control engineers from Boeing to review the FCC-908 software functionality and to perform flight simulator tests with the FCC-908 software upgrade in place. During a meeting on February 7, 2000, Boeing engineers indicated that the LSAS command signals would be required to precisely identify and differentiate between pilot and LSAS control activity after an accident or incident involving an MD-11 with the FCC-908 software upgrade. Boeing also indicated that it had developed Service Bulletin MD-11-22-20 to enable FCC output of LSAS command signal parameters to the digital flight data acquisition unit, which would in turn enable those LSAS parameters to be recorded by the DFDR. The Safety Board continues to believe those LSAS command signal parameters, and the additional parameters in Safety Recommendation A-00-97, should be recorded on the MD-11 DFDR to satisfy 14 CFR Part 25.1459, based on the "novel or unique design" aspects of the FCC-908 software. The Safety Board contacted Boeing and asked for proof that LSAS control activity can be analytically derived. The Boeing staff contacted stated that they had never attempted to derive LSAS control activity, but that they believed it possible, and would attempt to do so. In February 2002, Boeing obtained a 128-word DFDR from a United Parcel Service (UPS) MD-11 that was equipped to record LSAS parameters. Boeing attempted to match the recorded LSAS parameters with LSAS parameters derived from other DFDR signals that would be available from a DFDR not equipped to record LSAS parameters. At this time, Boeing still has not achieved a satisfactory match between the derived and measured LSAS data. The flight for which the data were being analyzed was not an accident flight and featured data from a benign and largely steady-state flight. The Board notes that in a major accident, the DFDR may record rapid and large pitch control oscillations, making a derivation even more difficult. The Safety Board continues to believe that all MD-11s equipped with FCC-908 software should record LSAS parameters, as recommended. The Board disagrees with the FAA that it is practical or possible to reliably and accurately determine LSAS activity without this data and asks the FAA to reconsider its position. Pending completion of the recommended action, Safety Recommendation A-00-97 is classified "Open--Unacceptable Response."

From: FAA
To: NTSB
Date: 11/13/2001
Response: Letter Mail Controlled 11/19/2001 10:56:51 AM MC# 2010909: The FAA agrees that the digital flight data recorder (DFDR) should record parameters that are indicative of the pilot's control inputs. The MD-11 flight data recorders will provide this capability when they are updated and brought into compliance with the FAA's Expanded Parameter Rule to revise 14 CFR 121.344. The final rule becomes effective in August 2001 and requires the addition of the control column position, which is a measurement of the pilot's primary pitch control input. It should be noted that the existing MD-11 DFDR already records the elevator position and will continue to do so after implementation of the final rule. It should also be noted that, although the longitudinal stability augmentation system (LSAS) command signals are not recorded on the DFDR, LSAS commands to the elevator can be analytically derived based on the control column position and elevator position since the LSAS provides a series input to the elevator. With regard to the inertial reference unit pitch rate and automatic ground spoiler command signals, the elevator position and control column position data can be used to derive the difference between pilot and LSAS elevator control activity. Consequently, those signals are not required to identify and differentiate between pilot and LSAS elevator control activity. I believe that the FAA has addressed the full intent of this safety recommendation, and I consider the FAA's action to be completed.

From: NTSB
To: FAA
Date: 6/13/2001
Response: Pending the results of the FAA's review, Safety Recommendation A-00-97 is classified "Open-Acceptable Response."

From: FAA
To: NTSB
Date: 11/13/2000
Response: Letter Mail Controlled 11/15/2000 3:27:59 PM MC# 2001695 The FAA is reviewing the MD-11 longitudinal stability augmentation system function to determine if additional parameters need to be recorded on the flight data recorder for identifying and distinguishing between pilot and longitudinal stability augmentation system elevator control activity. It is anticipated that the review will be completed by May 2001. I will inform the Board of the FAA's course of action to address this safety recommendation as soon as the review is completed.