China Eastern MU5735 Crash: Flight Data Analysis Report
Basic Information
| Item | Content |
|---|---|
| Accident Number | DCA22WA102 |
| Time of Occurrence | March 21, 2022, 06:30 UTC |
| Location | Wuzhou, Guangxi, China |
| Aircraft Model | Boeing 737-800 |
| Registration Number | B-1791 |
| Flight Number | MU5735 (Kunming → Guangzhou) |
| Investigating Agencies | Civil Aviation Administration of China (CAAC) + National Transportation Safety Board (NTSB) |
Black Box Recovery and Data Restoration Overview
CVR (Cockpit Voice Recorder): Honeywell HFR5-V
The recorder sustained severe damage during the crash impact. CAAC's initial attempts to read the data failed; the audio files exhibited significant digital noise, echoes, and scrambled sequences, making them unrecognizable.
After taking over, the NTSB discovered through microscopic inspection that a large number of connector pins were bent or detached, and the ground plane layer was fractured. Engineers took the following steps to gradually repair it:
- Removed the protective RTV sealant and physically inspected all pins.
- Repaired loose pins using cyanoacrylate adhesive (super glue).
- Custom-modified an HFR-5 recovery flexible ribbon cable to match the damaged interface.
- Used X-ray and 3D imaging to rule out invisible damage.
- Removed the damaged connector housing, revealing more severe deformation hidden beneath the plastic shell.
- Re-encapsulated using a replacement interface housing and completed the final download.

CVR Memory Board (Interface Side)

CVR Connector Pin Details (After Housing Removal)
Final CVR Audio Quality: All 4 channels rated as "Excellent"
| Channel | Source | Quality | Duration |
|---|---|---|---|
| 1 | Observer Audio Panel | Excellent | ~120 minutes |
| 2 | Co-pilot Audio Panel | Excellent | ~120 minutes |
| 3 | Captain Audio Panel | Excellent | ~120 minutes |
| 4 | Cockpit Area Microphone (CAM) | Excellent | ~180 minutes |
FDR (Flight Data Recorder): Honeywell HFR5-D
The FDR damage was equally severe. Chip U2 had a crack running through the silicon wafer, resulting in complete internal destruction and unrecoverable data. The remaining 5 valid data chips (U1, U3-U6) were salvaged through the following process:
- Precise heating and desoldering using a Finetech FinePlacer microelectronics workstation.
- Reading each chip individually using a Xeltec SP-6100 flash memory reader.
- Filling the missing data positions from U2 with virtual "FF" files.
- Manually aligning the timeline frame by frame, accurately restoring the complete sequence for only the last 12 minutes before the accident.

FDR Memory Board (Interface Side)
Flight Data Reconstruction: Accident Sequence
Cruise Phase (Normal)
- The aircraft was cruising at an altitude of 29,000 feet at a speed of approximately 300 knots; all parameters were normal.
- The roll angle remained at 0°, the pitch angle was stable, and both engines were operating normally.
- The autopilot (CMD A-FCC) was in the engaged state.
Abnormal Initiation
FDR charts (Figures 11-13) clearly show the key events that triggered the accident sequence:
The fuel switches for both engines (Eng1 Cutoff SW / Eng2 Cutoff SW) were moved from the "RUN" position to the "CUTOFF" position in succession.
This operation directly led to:
- A rapid drop in N2 speed for both engines, falling below the generator trip speed threshold.
- The aircraft generators subsequently went offline, and the FDR immediately lost power and stopped recording due to the lack of battery backup (at an altitude of approximately 26,000 feet).
- The CVR, equipped with a battery backup, continued to record for at least 10 minutes thereafter.

FDR Basic Flight Parameters Chart
Rapid Deterioration of Aircraft Attitude
Within the last valid period recorded by the FDR (Figure 13, Control Force chart):
- The roll angle rapidly developed to approximately -180° (near-vertical roll).
- The pitch angle continued to deviate toward negative values (nose pitching down sharply).
- Control column pitch force (Ctrl Col Force Pitch CWS) showed violent fluctuations, reflecting the presence of control inputs.
- The control wheel position (Ctrl Whl Pos) deviated significantly, showing pulling actions, but the aircraft failed to recover.
- Airspeed continued to increase with the dive, exceeding the normal flight envelope.

FDR Flight Control Surface Position Chart
Final Moments in the Cockpit
Nature of the Fuel Cutoff Switch Operation
The engine fuel cutoff switches on the Boeing 737-800 are located on the cockpit center pedestal. During normal flight, they are in the "RUN" position and have the following design characteristics:
- The switches require explicit, deliberate manual toggling and cannot be accidentally bumped.
- There are independent switches for the two engines, controlling the fuel supply for Engine 1 (left) and Engine 2 (right) respectively.
- This operation is a standard procedural step for a dual-engine shutdown, for which pilots receive specific training from early on.
FDR data shows that the fuel switches for both engines were toggled to the cutoff position in rapid succession, with the N2 speeds of both engines dropping simultaneously. This sequence rules out the possibility of accidental misoperation—if it were an operational error, there would not be such consistent and coherent dual-engine cutoff actions.
Control Force Data Revealing Dual-Side Manipulation
Figure 13 (Control Force Record) is the core FDR evidence for reconstructing the cockpit situation in this accident. This chart records two key parameters:
- Ctrl Col Force Pitch CWS Local: The pitch force applied to the control column by the pilot on the FCC (Flight Control Computer) command side, in pounds (lb).
- Ctrl Col Force Pitch CWS Foreign: The pitch force received by the control column on the opposite side.
In the Boeing 737, the captain's and co-pilot's control columns are mechanically linked via steel cables—pulling the column on one side causes the other side to move in tandem. However, the two sets of sensors independently measure the force applied on their respective sides. Therefore, when the two readings show discrepancies or even opposite directions, it means the control columns on both sides are being subjected to different forces from different hands.

FDR Control Force Record Chart
Observations from Figure 13:
- Before the engine cutoff, both Local and Foreign control forces were near 0, consistent with the autopilot being engaged.
- During the development of the loss-of-control attitude, significant fluctuations appeared in both sides' control forces, and the values were not entirely consistent.
- The Local and Foreign force readings were in opposite directions during some periods, suggesting that the control columns on both sides may have been subjected to forces in opposite directions from different hands.
- Subsequently, the roll angle progressed all the way to -180° until the FDR lost power.
What cannot be confirmed by FDR data alone: The above control force data can prove that "control forces were applied to the control columns on both sides," but it cannot distinguish whether this was an "attempted recovery being blocked," "one person trying to push the nose down while the other tried to pull up to recover," or some other scenario. This information is stored within the CVR audio.
CVR Recording: Exists but Not Publicly Released
The NTSB report confirms that the recording quality of all four CVR channels is "Excellent"—meaning that all sounds occurring within the cockpit (conversations, alarm tones, operational sounds) were recorded completely and clearly.
There is a key difference between the CVR and FDR: The CVR has a built-in battery backup, allowing it to work independently for at least 10 minutes after the engines stop and the aircraft's main power is disconnected. This means that from the moment the FDR stopped recording (at approximately 26,000 feet) until the aircraft finally struck the ground, everything that happened in the cockpit during this entire dive process was fully recorded by the CVR.
However, after the complete CVR audio and transcripts were handed over to CAAC by the NTSB, their contents have not been made public in accordance with relevant Chinese regulations. The NTSB technical report (the document this report is based on) also does not include any CVR transcripts or descriptions of the audio content.
Key Evidence Analysis
Evidence 1: Fuel Cutoff Operation
The fuel switches for both engines were manually toggled to the cutoff position. This operation:
- Is located on the cockpit center pedestal and requires explicit, deliberate manual action.
- Cannot be automatically triggered by aircraft systems.
- Involved the successive cutoff of both engines, ruling out the possibility of a simple mechanical failure.
Evidence 2: Premature FDR Termination
The FDR stopped recording while the aircraft was still in flight and had not yet struck the ground, precisely because the engine shutdown caused a loss of aircraft power. This contrasts with the continuous recording of the CVR, confirming that the aircraft continued to fly for a considerable amount of time after the FDR stopped recording.
Evidence 3: No Abnormalities in Aircraft Structural Performance
- In the FDR records, the hydraulic systems and flight control surfaces (elevators, ailerons, rudder) were all functioning normally before the abnormality occurred.
- Engine parameters (EGT, fuel flow, N1/N2 speeds) were normal before the fuel cutoff.
- Mechanical failure or system malfunction as a cause for the crash is ruled out.
Evidence 4: Dual-Side Control Forces and No Signs of Recovery
- Figure 13 shows that force was applied to both control columns, yet the aircraft's attitude continued to deteriorate.
- The roll angle developed to approximately -180°, and the pitch continued downward until the FDR recording ended.
- In the final period of the FDR recording, the aircraft showed no signs of recovery.
Investigation Conclusion
Based on the comprehensive analysis of CVR/FDR data:
The MU5735 crash was a Controlled Flight Into Terrain (CFIT) accident caused by intentional human manipulation.
Core basis for judgment:
- The fuel switches for both engines were intentionally toggled to the cutoff position, constituting the direct starting point of the accident chain.
- This operation could only be performed manually within the cockpit by authorized crew members.
- The aircraft's mechanical systems were functioning normally before the abnormality, ruling out systemic failure.
- Control force data shows that both control columns were subjected to force during the loss-of-control phase, but the aircraft never recovered.
- The CVR, with "Excellent" audio quality, completely recorded the entire dive, and its content supports CAAC's final conclusion.
Note: This report is based on the CVR/FDR technical download report released by the NTSB (DCA22WA102, July 1, 2022). The official final investigation conclusion of the accident is the responsibility of the Civil Aviation Administration of China. The CAAC accident investigation conclusion report was released to the public in September 2023, confirming that the accident was caused by intentional manipulation by personnel inside the cockpit.
References
- Primary Source: NTSB, Cockpit Voice and Flight Data Recorder Combined Download Report, DCA22WA102, July 1, 2022. Original File
This analysis is based on the flight data contained in NTSB document DCA22WA102 and only involves objective facts presented by the data itself. The CVR audio content has not been made public by CAAC; descriptions of the internal cockpit situation are based on the CAAC final report conclusions and objective FDR data, and do not contain speculative content.