FlarePathAir-accident investigation intelligence
Investigation brief · generated 2026-10-09 · source: NTSB public docket and accident database

EMBRAER S A ERJ 170-200 LR — Montgomery, AL, USA, 2022-12-31

NTSB number
DCA23LA109
Operator
Envoy Air Inc.
Registration
N264NN
Highest injury
Fatal
Regulation
Part 121
Damage
Minor
Cohort
Part 121 airline
Recorder data
Tier 1 — decoded recorder tables
Every sentence below is taken directly from NTSB records — the accident database, the docket's recorder reports and the Board's findings — with its source noted. Nothing is paraphrased or inferred.

NTSB summary

After an uneventful flight, Envoy Air flight 3408 (dba American Eagle) taxied with both engines operating to its assigned gate and parked at Montgomery Regional Airport, Montgomery, Alabama. The No. 1 (left) engine was still operating while they waited on ground power due to an inoperative auxiliary power unit (a minimum equipment list item). Shortly after arriving at the gate, the captain saw a “FAIL” symbol on his engine display, felt the airplane shake violently, and noticed that the left engine had shut down. The flight crew subsequently determined that one of the ramp agents had been ingested into the No. 1 engine. Ramp agents from Piedmont Airlines were responsible for supporting airplanes after arrival at the airport. Before the accident airplane arrived, the ramp agents discussed each person’s role and the safety precautions that were to be taken, which included maintaining a position away from the airplane until the engines had spooled down and the red beacon lights were turned off (except when the lead ramp agent had to approach the airplane to chock the nosewheel). Security video showed at least one of the red beacon lights was still on when the accident ramp agent walked directly in front of the left engine before being ingested into the engine. This investigation evaluated the accident ramp agent’s training and American Eagle’s procedures, her medical conditions and toxicology results, and her judgment. In addition, the investigation considered the result of the company’s drug and alcohol use policy and the benefit of classifying ramp personnel as a safety-sensitive position regarding federal drug and alcohol testing requirements. Ramp Training and Operator Procedures The accident ramp agent completed new-hire ramp training in December 2021 and recurrent training in September 2022. The training included a discussion about jetblast and ingestion zones and procedures for safely approaching an engine. The American Eagle Ground Operations Manual, which provided guidance to ground personnel supporting American Eagle regional flights, also specified jetblast and ingestion zones, engine spool down, procedures for approaching an airplane, and the importance of waiting until the airplane’s red beacon lights were turned off before doing so. The manual also stated that a ramp agent should never pass under the fuselage to move from one side of the airplane to the other. According to the lead ramp agent on the day of the accident, ramp personnel were required to stay 15 ft away from an engine, consistent with guidance in the American Eagle Ground Operations Manual. The accident ramp agent should have been familiar with the company’s ramp safety policy because her position required her to continuously work around jet-powered airplanes. The 15-ft guidance provided a buffer that was almost two times the airplane manufacturer’s boundary for the inlet hazard area (8.3 ft). In addition, the lead ramp agent ensured that the accident ramp agent participated in the safety huddle and understood her duties while providing ground support for the Envoy Air airplane. Notably, he tried to warn her, while she was at the back of the airplane, to move away from the operating engine. Thus, the accident ramp agent’s training and the operator’s procedures were not factors in this accident. Medical Conditions and Toxicology Results The accident ramp agent had relapsing-remitting multiple sclerosis, which can adversely affect cognition, including executive functioning, information processing, and working memory. Although no cognitive impairment was documented at the ramp agent’s most recent neurology visits, she had a substantial risk of such impairment because of her condition. This investigation was unable to determine whether she was experiencing a multiple sclerosis flare at the time of the accident. The accident ramp agent also had diabetes that was treated with medications, including insulin. People with medication-treated diabetes are at risk for both abnormally high (from diabetes) and abnormally low (from medication effects) blood glucose. The ramp agent’s postmortem urine glucose result was normal; thus, she was not likely experiencing a major metabolic disturbance from extreme high blood sugar at the time of the accident. The investigation was unable to determine, from the available evidence, whether the ramp agent was experiencing milder high blood sugar effects (such as fatigue) or low blood sugar effects (such as diminished concentration). The ramp agent’s toxicology testing detected delta-9-THC and its metabolites, indicating that she had used a cannabis product. Cannabis has the potential to cause cognitive and psychomotor impairment and can worsen cognitive impairment in individuals with multiple sclerosis. However, a person’s blood concentration of delta-9-THC does not directly predict that person’s impairment. Details of the ramp agent’s cannabis use, including dose, route, timing, and specific effects, could not be determined from the available toxicological evidence. The ramp agent’s toxicology testing also detected carboxy-delta-8-THC, a non-psychoactive metabolite of delta-8-THC (which was not detected). This result indicated that the ramp agent had likely used a product containing chemically synthesized delta-8-THC given that very little delta-8-THC occurs naturally in the cannabis plant. However, because no delta-8-THC or psychoactive metabolite of delta-9-THC was detected in the ramp agent’s blood, she was not likely impaired by delta8-THC effects at the time of the accident. Ramp Agent’s Judgment Multiple cues were available to the accident ramp agent to indicate that the left engine was running. These cues included the airplane’s red beacon lights, which were on and visible; the sound of the operating engine, which would have been discernible, even with the ear protection that the ramp agents were wearing; and the accident ramp agent’s encounter with the left engine’s jetblast (as shown on security video) when placing the safety cone at the tail of the airplane. Therefore, the accident ramp agent was likely aware that the left engine was still operating. However, the accident ramp agent’s actions on the day of the accident were not consistent with those that would be expected for someone in that position. No task on the day of the accident or on any day that accident ramp agent worked would have placed her in front of the left wing and engine. The accident ramp agent’s decision to walk in front of the engine, despite operating engine cues, was also inconsistent with her training and the briefings that she received before the airplane landed. The accident ramp agent’s records showed no previous instances in which she was within the engine ingestion zone while an engine was operating, and, as previously stated, the lead ramp agent reported that he tried to alert her that the left engine was still operating. The accident ramp agent’s behavior at the time of the accident demonstrated that her judgment was deficient. Given her identified medical and toxicological cognitive risk factors, the ramp agent’s deficient judgment was likely due to cognitive impairment. The extent to which individual cognitive risk factors, such as the ramp agent’s multiple sclerosis and cannabis use, contributed to this impairment could not be determined based on the available evidence. Company Policy on Drug and Alcohol Use Piedmont Airlines had a company drug and alcohol policy that prohibited employees from reporting to work in a condition that would impair satisfactory work performance due to drugs or alcohol. The ramp agent acknowledged this policy electronically on November 10, 2021. However, toxicology results showed that the accident ramp agent had used a cannabis product at some point before her work shift. Thus, the company policy did not deter the accident ramp agent’s from working in a potentially impaired condition. The Department of Transportation does not consider ramp personnel positions to be safety sensitive, so the company was not required to provide mandatory drug and alcohol training and perform required drug and alcohol testing, including random testing on its ramp personnel. If these positions had been classified as safety sensitive, the accident ramp agent would have been subject to federally required drug testing, and the company might have been able to detect the accident ramp agent’s use of cannabis and take appropriate action in response, including removing her from safety-sensitive functions.
Verbatim, NTSB aviation accident database (narratives.narr_accf).

Probable cause

The ramp agent’s cognitive impairment, which resulted in her (1) inconsistent behavior with trained procedures and prelanding briefings, (2) presence on the left side of the airplane while the left engine was still operating, and (3) subsequent ingestion into the engine.
Verbatim, NTSB determination.

Flight data recorder

Embraer 170-200 · N264NN — 11,388 samples, 62 parameters, 15:35:10.0 to 15:40:21.9.

ParameterMaxatMinatFinalat
Ground speed kts119.015:35:10.70.015:39:53.70.015:40:21.7
Vertical acceleration g1.515:35:13.10.815:35:13.61.015:40:21.9
Engine 1 N1 %RPM69.015:35:23.00.015:39:53.00.015:40:21.0
Engine 2 N1 %RPM70.915:35:22.30.015:38:53.30.015:40:21.3
Computed from the docket's tabular attachment; times as recorded (local or SRN as published).

Sequence of events (FDR factual report)

TimeEventPage
15:38:36 CSTThe inboard and outboard parking brake pressure statuses changed from not applied to applied. At this time, the statuses for both engine 1 and engine 2 control modes were Idle gas generator speed (Idle NG).5
15:38:45 CSTThe engine 2 control mode changed to shutdown while the engine 1 control mode remained Idle NG.5
15:39:12 CSTThe forward baggage door locked sensor changed from locked to not locked.5
15:39:15 CSTThe forward baggage door closed sensor and the forward baggage door latched sensor 1 transitioned from closed to not closed and latched to not latched, respectively.5
15:39:15.7 CSTThe master warning forward cargo door open enabled transitioned from false to true and the master warning forward cargo door open displayed changed from not for all to active/display.5
15:39:17.7 CSTThe master warning forward cargo door open displayed transitioned to active/acknowledge/displayed (Active/Ack/Display).5
15:39:52.7 CSTThe master warning forward cargo door open enabled transitioned back to false and the master warning forward cargo door open displayed changed to not for all.5
15:39:54 CSTThe engine 1 control mode changed to shutdown.5
Extracted by claude-sonnet-5; 8 of 8 stated events verified against the raw recorder data.

Recorder documents in the docket

GroupTitlePagesType
fdrFLIGHT DATA RECORDER - SPECIALISTA��S FACTUAL REPORT10pdf
fdrFLIGHT DATA RECORDER - ATTACHMENT 1csv
cvrCOCKPIT VOICE RECORDER SPECIALAIST'S FACTUAL REPORT6pdf

Docket: data.ntsb.gov/Docket/?NTSBNumber=DCA23LA109. This brief is an analytical aid assembled from public records; it is not an investigative finding.