In cruise, the lavatory smoke message appears, the alarm sounds, and two crew members open the door a few moments later: no one, nothing hot, a sweet smell that is already fading. The alarm resets itself and does not return. The question remains, which the procedure does not resolve: what do you write? The detector did exactly its job, and it proved nothing at all. Both statements are true at the same time, and it is the optical mechanism that reconciles them.
What the detector measures, and what it does not measure
A lavatory detector is not a chemical smoke sensor. On transport aircraft, detection is predominantly photoelectric: a light source illuminates a small labyrinth chamber, and a photoreceptor is deliberately placed off-axis, in the shadow. If the chamber is clean, it receives nothing; as soon as particles cross the beam, they reflect a fraction toward it, and the alarm triggers beyond a factory-set threshold.
It all comes down to this: the measured quantity is the amount of light reflected by particles. Not a temperature, not a carbon monoxide concentration, not a chemical signature. The sensor does not know where the particles come from, what they contain, or who produced them. It detects that there are enough, and it reports it.
The choice of photodiffusion is not arbitrary: the labyrinth allows air to enter and blocks daylight. The competing principle, the ionization chamber, suffers from sensitivity to airflow and requires managing a radioactive source in a removable cabin module.
Le détecteur ne fait pas la différence entre fumée et vapeur. Il réagit simplement quand assez de particules bloquent sa lumière. C’est comme un radar qui verrait des objets sans savoir s’ils sont dangereux.
Sur un Airbus A320, un passager utilise une cigarette électronique dans un lavatoire de 2m² avec une température de 22°C. Le détecteur déclenche l’alarme pendant 25 secondes avant que la vapeur ne se dissipe. Aucun signe de combustion n’est observé.
Why vapor triggers the alarm as well as smoke
The efficiency with which a particle reflects light depends on the ratio of its diameter to the wavelength. Far below, it reflects almost nothing; far above, it reflects a lot, but at equal mass, these particles are too few to populate the chamber. Relative to the suspended mass, the response peaks near the micrometer scale, typical of slow-burning combustion aerosols.
The aerosol from an e-cigarette falls into the same range. These are droplets of propylene glycol, glycerin, and water, formed by condensation as they exit the atomizer, typically around the micrometer scale and growing as they cool. From the photoreceptor’s perspective, there is nothing to distinguish them: both clouds speak the same language.
The balance of power is even unfavorable. A vaping cloud is produced in one or two exhalations, while a smoldering fire takes minutes to fill a volume: the threshold is often crossed faster and more decisively with an harmless aerosol than with a real fire.
The temporal signature, the only usable clue in flight
What separates the two clouds is not their optical nature, but their behavior over time. Propylene glycol and glycerin droplets are volatile: mixed with the drier, warmer cabin air, they evaporate, while continuous extraction renews the volume. The profile is characteristic: abrupt rise, threshold crossing, then return below the threshold, typically within a few dozen seconds, without intervention.
A smoldering fire does the opposite. Paper or a cigarette butt burning in a trash bin produces a non-volatile aerosol with a sustained source: opening the door dilutes the cloud without removing what produces it. The alarm persists or returns, with indicators that the aerosol never provides—acrid odor that lingers, hot spot, warm trap on the back of the hand.
Hence the only honest discriminant available to the crew in flight: behavior after ventilation, not the alarm itself. An alarm that resets on its own and does not return is consistent with a volatile aerosol. An alarm that persists or returns indicates combustion, until proven otherwise.
The trap. “Consistent with an aerosol” is not “it was an e-cigarette.” A smoldering fire goes through a discrete phase where smoke production is still low and heating localized: a brief alarm that resets may be the first sign. No reading of the temporal profile dispenses with the physical inspection required by the operator’s documentation, trash bin included. The only irreparable error is treating the start of a fire as a misconduct.
Why some vaping does not trigger anything
The symmetrical error is just as easy to make. The detector does not monitor the room: it monitors the air passing through its chamber. The space is swept by a flow entering under the door and exiting through the extraction; the sensor is placed on this path, and what does not follow it never reaches it.
Exhaling downward, into a damp towel, into the trash bin, or directly into the extraction grate often keeps the load below the detection threshold. Add that the droplets evaporate along the way, and the absence of an alarm ceases to be meaningful.
Both ends of the system event are thus silent. The alarm does not establish the cause; the silence does not establish the absence of fault. A case built on either will not last ten minutes with someone who understands the sensor’s principle.
Tampering with the detector, however, leaves traces
There is, however, a material and lasting fact: the attempt to prevent detection. Paper jammed in the grate, adhesive on the vents, a bag placed over it, unclipped cover. This action has no innocent explanation and changes the nature of the matter: it is no longer about misconduct but about the neutralization, in flight, of a safety device required for aircraft certification.
It still must be seen: this trace disappears with the next pass, and the reset often occurs without checking the grate. The order of actions determines what will remain in the record.
The right reflex. Before any reset, two things and only two: note the exact time on the cockpit clock, and inspect the detector’s grate and cover. These are the two elements that disappear the fastest. The time becomes approximate as soon as it is written from memory at the station, and the grate is cleared by the next passenger. The rest of the report can be written later; these two points, no.
Other causes of alarms, and why they matter
Vaping is not the only thing that fills a measurement chamber. The table below is not for assigning blame: it sets the on-site checks, the absence of which will reopen the case later.
| Origine | Shape of the rise | What is found in the compartment | After ventilation |
|---|---|---|---|
| Smouldering combustion (paper, tobacco, waste bin) | Slow, continuous, sustained | Persistent acrid smell, ash, hot spot, warm flap | Persists or returns |
| E-cigarette aerosol | Abrupt, within seconds | Fleeting sweet smell, condensation on cold surfaces, sometimes the device | Clears by itself, does not return |
| Cosmetic aerosol (hairspray, deodorant, powder) | Abrupt, within seconds | Perfumed smell, dry deposit, bottle among the passenger’s belongings | Clears by itself, does not return |
| Water vapour and high humidity | Gradual, linked to use of the washbasin | Fogged mirror, wet surfaces, no smell | Clears with ventilation |
| Sensor contamination or moisture | No correlation with occupancy of the compartment | Nothing to see, no physical clue | Recurs, always at the same location |
The last line is the one most often forgotten, and it is the most useful. A lavatory that alarms on multiple legs, empty room, and locked door, raises a maintenance question, not a behavior one: dust deposited in the labyrinth raises the sensor’s baseline. Knowing that this station already alarmed the previous week avoids blaming someone for a material defect.
From the alarm to the report: what holds and what does not
A system event is not testimony. It establishes that an optical threshold was crossed at a given time in a given location. It does not establish the substance, the exact location of emission, the perpetrator, or the intent. Everything added to this raw fact is produced by people in the following minutes and cannot be reconstructed later.
| Evidence collected | What it establishes | Its limit |
|---|---|---|
| Detector alarm | An optical threshold crossing, timed, in an identified compartment | Says neither the substance, nor the perpetrator, nor even that anyone was present |
| Direct observation by a crew member | The observed fact, the compartment, the time and the person coming out of it | Only valid if written up the same day; reworded later, it can be challenged |
| Smell noted in the compartment | An indication of the nature of the cloud | Cannot be preserved, passed on or verified by anyone else |
| Device handed over or found on board | The presence of the equipment on the flight | Does not prove use, still less where and when it was used |
| Grille blocked or cover removed | A deliberate action on an aircraft safety device | Only identifies a perpetrator if occupancy of the compartment is otherwise established |
| Consistent passenger statements | Corroboration of the crew’s observation | Without contact details taken on board, they vanish at the jet bridge |
The writing rule that follows is simple: record what was observed, do not qualify. Legal qualification will be done on the ground, with or without the crew. The factual narrative, however, belongs only to them: no one will redo it, and a report that concludes in place of the judge loses its strength along with its credibility.
The channels are not confused either. The event notification, which feeds flight safety, serves prevention and protects the person reporting. The safety report and the form for handing over to authorities serve an administrative or judicial follow-up. Writing one in place of the other weakens both: safety loses data, the file loses its piece.
What to check in your own documentation
The first point to establish concerns the system installed on the fleet: does the alarm lock until manual action, or does it reset automatically as soon as the chamber clears? On a locked system, “it stopped on its own” is impossible and indicates something else; on an automatic reset system, it is the nominal case of a volatile aerosol. The same report is not read the same way depending on the answer.

Next comes what the documentation requires before opening a trash bin, with which extinguisher, and who is authorized to reset the detector. Then a question rarely asked in time: depending on the cabin system architecture, the event may leave a time-stamped trace consultable by maintenance. It is the only element of the file that does not depend on anyone’s memory—and these memories overlap.
Check what the operator says about transporting these devices: in the cabin, under the owner’s custody, never in the hold, and no recharging on board. The reason is the lithium battery and the risk of thermal runaway, not tobacco—it is therefore a distinct basis for intervention from the smoking ban. Finally, check if your report form has a section dedicated to detector tampering; otherwise, this aggravating fact is diluted in a free field that no one will review.
There remains an acknowledged gray area, better to know than to discover. The qualification of the act and its consequences depend on the state of registration, the state of landing, and, on international routes, a conventional regime that states have not all updated in the same way: the same behavior does not produce the same effect on either side of a border. What the crew masters is the quality of their observation—and knowing how to state the difference between a system event and an observation is better, in preparation for a supervisory role, than reciting a procedure. The exchanges on the FL330 professional forum and the cabin supervisory role preparation modules are useful places to compare the drafting of these reports.
The applicable regulatory text, the authority’s instructions, and the operator’s approved documentation prevail over this article, which only has an explanatory value of the mechanism.

Leave a Reply
You must be logged in to post a comment.