Leak Detection

The bleed ducts carry 200 °C air through places no human sense can patrol — the unpressurised belly, the pylons, the wing leading edges (article 01). So the hot ducting gets a nervous system: loops of overheat-sensing elements laid alongside every run. This article covers the chemistry that makes a wire "feel" heat, the loop geometry and its AND logic, and the three isolation playbooks the system executes before the crew has finished reading the alert. The procedures themselves follow in article 08.

      LEFT sub-system (both BMCs watch)  ‖ crossbleed valve = the boundary ‖  RIGHT sub-system
  ┌─ pylon 1: single loop (204 °C)       ‖                    ┌─ pylon 2: single loop (204 °C)
  ├─ left wing + fuselage: loops A+B (124 °C) — AND           ├─ right wing + fuselage: loops A+B (124 °C) — AND
  └─ APU duct (tailcone→belly): loops A+B (124 °C) — AND (analogue to BMC 1 only)

1. Mission: ambient heat, not duct temperature

Per AMM 36-22-00:

The leak detection system serves to detect any ambient overheat in the vicinity of the hot air ducts which run through the engine pylons, the wings and the fuselage. The system is made up of sensing elements connected in series which form the detection loops. The purpose of this monitoring system is to prevent any damage to the structures and components which could result from duct leak or rupture.

Note the phrase ambient overheat. These loops do not measure the air inside the duct (that is the outlet-temperature sensor's job, article 05); they measure whether the air around the duct has been cooked by an escape. The alert says LEAK because in this neighbourhood, external heat is a leak. The FCOM's crew-level line carries the same wording — per FCOM DSC-36-10-60: The air leakage detection loops detect any ambient overheat in the vicinity of the hot air ducts in the fuselage, pylons and wings.

This active layer pairs with the passive one from article 01: loops are the nerves (alarm and isolate), blow-out panels and relief doors are the skeleton (survive the pressure if nothing catches). Together they are the complete answer to "what if a duct bursts".

2. Geometry: split at the valve, doubled where it matters

Per AMM 36-22-00:

The leak detection system is divided into two sub-systems, left and right. The separation is made at level of the crossbleed valve. Each sub-system: - operates independently, - is made up of overheat sensing elements connected in series which form detection loops, - is monitored by two computers identified Bleed Monitoring Computers (BMC).

Why cut the map at the crossbleed valve? Because the end-game of every leak alert is isolate the sick half of the network — close that side's bleed and close the crossbleed. Detection zones must match isolation boundaries, or alarms and valves would talk past each other.

A dual loop detection system is installed in the wings and the mid fuselage and APU. The two loops (A and B) are routed in parallel along the air ducts. An "AND" logic ensures the interconnection of the loops in the BMCs to prevent spurious warnings. In each engine pylon, a single detection loop ensures the leak detection. The loop is installed near the pylon ventilation duct.

And the AND logic's full meaning, in the FCOM's one must-know sentence:

A wing or APU leak signal is activated when the two loops detect a leak, or when one loop detects the leak and the other is inoperative.

[!warning]- Not a pure AND — a disabled loop hands over its vote Both loops healthy: two votes required (spurious-warning protection). One loop failed: the survivor's single vote counts double — per AMM 36-22-00, if one of the loops (A or B) is inoperative, the overheat detection is ensured by the remaining loop. Detection capability never degrades; what a single loop failure costs is the false-alarm protection. Which is why a single dead loop is silent (a STATUS-page maintenance message at most) while a double failure raises the DET FAULT alert — and why some operators' MELs treat wing-loop DET FAULT as a no-go while giving the APU loop several dispatch paths (article 11).

Why do pylons get away with a single loop? The pylon threshold sits high (204 °C, section 3) in a small, hot, engine-adjacent zone that also lives under the ATA-26 fire-detection loops; the wing and fuselage loops guard long corridors at a low threshold where one false vote would cost half the network plus wing anti-ice — those need corroboration.

On certain airframe configurations a conditioned-service-air system for cargo is present; its ducting is watched by elements wired into the left-wing loops, and a leak there additionally closes that system's isolation valve — a configuration note, not a new architecture.

3. The eutectic-salt element: a fuse that conducts instead of melting open

Per AMM 36-22-00:

The overheat sensing elements have a solid nickel center conductor embedded in a ceramic insulation of porous aluminum oxide. An inconel tube contains these components and is sealed on both ends. Eutectic salt fills the space between the tube, the ceramic insulation and the center conductor.

When the overheat sensing element is exposed to hot gases (or any other heat source) whose temperature reaches the eutectic point, the impedance between the nickel conductor and the inconel outer tube immediately decreases. The conductor is no longer insulated but continuity is established between the tube and the conductor. Therefore, a ground signal is emitted.

The physics is elegant: below its eutectic point the salt is a solid insulator; at the eutectic point it melts as a whole into an ionic conductor — no gradual zone, a chemical switch that closes at temperature. The BMC keeps a monitoring voltage on the loop; wherever the salt melts, the core wire grounds to the tube and the impedance collapses. Cool the spot and the salt refreezes, insulation returns — the signal is self-resetting, which is the physical root of the FAULT-light behaviour in the leak procedures (light follows the hot spot, not the switch — article 08).

Two trigger grades, tailored to the neighbourhood's normal body heat:

The overheat signals are triggered for temperatures of: - 124 deg.C plus or minus 7 deg.C for the overheat sensing elements located in the wings and fuselage - 204 deg.C plus or minus 5 deg.C for the overheat sensing elements installed in the engine pylons.

A pylon sits beside a running engine and is legitimately hot, so its alarm point moves up to 204 °C; wing boxes and fuselage bays are normally cool, so 124 °C there already proves hot air has escaped. These two numbers surface verbatim in the alert trigger conditions of article 08.

Series wiring buys total coverage at a price. Per AMM 36-20-00: The sensing elements are connected in series and monitor all points along the length of the ducts. and An increase in temperature along a short piece of one element is sufficient to cause an alarm signal. The whole run is guarded with no dead spots — but the alarm cannot say where along the run. ECAM reports L WING LEAK, never "third rib from the root". Hence every leak procedure isolates a whole side rather than hunting the hole.

4. Three isolation playbooks — who else gets locked out

The FCOM's action matrix is the system-side script behind every alert of article 08. Per FCOM DSC-36-10-60:

‐ In case of wing leak signal • The engine bleed valve and HP valve on the affected wing are automatically closed. • The associated ENGINE BLEED FAULT light on the AIR panel are illuminated • The X-bleed valve automatically closes (except when manually selected open). • If the APU bleed valve is opened, and if the leak affects the left wing, the APU bleed valve automatically closes.

‐ In case of pylon leak signal • The bleed valves and the HP valves on the affected side are automatically closed • The FAULT light associated with the related engine is illuminated on the AIR panel • The X-bleed valve automatically closes (except during an engine start or manually selected open). • If the APU bleed valve is opened, and if the leak affects pylon 1, the APU bleed valve automatically closes (except during engine start).

‐ In case of APU leak signal • The APU bleed valve automatically closes • The FAULT light illuminates on the APU BLEED pushbutton on the AIR panel. • The X-bleed valve automatically closes (except when manually selected open). • If the APU bleed valve is closed, the engine 1 bleed valve and engine 1 HP valve automatically close.

Three readings unlock the table. First, the left-sidedness again: the APU's delivery corridor is the left half, so left-wing and pylon-1 fires cut the APU's supply; right-side events only need the crossbleed wall. Second, the exemptions: engine starts get grace (the interface unit's start signal is the source), and a manually selected open crossbleed is respected — the automation will not override your selector, which is exactly why a dedicated alert exists to tell you to close it yourself (AIR BLEED LEAK, article 08).

[!warning]- The strangest line: APU valve closed, alert still live → close ENG 1 If the APU loop fires while the APU bleed valve is already closed, the system closes engine 1's bleed and HP valves. Follow the deduction: the APU duct is hot, but the APU cannot be feeding it — so the heat is arriving backwards from the crossbleed duct; the 0.029-psi check valve of article 04 has failed open; and the air pushing through it belongs to the left half's supplier, engine 1. ECAM names the finished deduction APU LEAK FED BY ENG — the system plays detective and arrests the supplier. Procedure and dispatch consequences in articles 08 and 11.

[!warning]- The APU loop has NO engine-start exemption of its own Wing and pylon closures spare the crossbleed and APU valve during a start — but per AMM 36-22-00: The APU bleed valve closes if a leak was detected during engine start phase. The exemption logic protects the start's air supply; an APU-loop alarm means the supply line itself is burning, and protecting it would mean feeding air to the fire. The start loses.

And the closing NOTE that shapes crew expectations — per AMM 36-22-00:

NOTE: As long as the fault continues, the valves cannot be opened even if the system is reset.

While the hot spot exists, the isolation is latched. No reset, no pushbutton cycling, will reopen a leak-closed valve — which is why the leak procedures of article 08 never contain an OFF-THEN-ON attempt, in pointed contrast to the regulation faults of article 09.

The full annunciation set, for completeness (AMM 36-22-00): the amber FAULT on the matching pushbutton (ENG BLEED 1 for left loops, ENG BLEED 2 for right, APU BLEED for the APU loop — The FAULT legends of the pushbutton switches remain on as long as the fault continues.), MASTER CAUT, single chime, the E/WD alert and the BLEED page auto-display — tier 4 of the taxonomy in article 05.

5. When the nerves lie: heat without a leak

The elements answer to temperature, not to causes. On a baking-hot ramp the wing leading edge can genuinely reach the 124 °C band with nothing leaking, and the SOP carries a specific inoculation. Per FCOM PRO-NOR-SOP-21:

On ground, hot weather conditions may cause overheating to be detected around the bleed ducts in the wings, resulting in AIR L(R) WING LEAK alert. Such warnings may be avoided during transit by keeping the slats in CONF 1, when the OAT is above 30 °C.

Extending the slats opens the leading edge and lets the oven ventilate. A perfect closing lesson for the chapter's sensors: before acting on any detector, ask what it actually measures — this one measures heat near a duct, wherever that heat came from.

Self-test

[!note]- Q1. What do the loops actually measure, and why is that the right thing to measure here?

Ambient overheat in the vicinity of the ducts — not duct-internal temperature. The ducting runs through unpressurised, unattended zones where escaped 200 °C air is undetectable by crew senses; external heat in those bays is the signature of a leak.

[!note]- Q2. Where is the left/right boundary and why there? Which areas get double loops and which a single one?

At the crossbleed valve, because detection zones must coincide with the isolation boundary the alerts will act on. Wings, mid-fuselage and the APU duct carry parallel loops A+B under AND logic; each pylon has a single loop near its ventilation duct.

[!note]- Q3. State the AND logic exactly, including the degraded case, and what a single loop failure does and does not cost.

A wing or APU leak signal requires both loops to detect — or one loop detecting while the other is inoperative. A single failed loop costs only the spurious-warning protection; detection capability is preserved by the survivor. It is silent apart from a maintenance-level status entry; only a double failure raises DET FAULT.

[!note]- Q4. Explain how a eutectic-salt element produces a signal and why the alarm self-resets. Give both trigger temperatures.

At the eutectic point the salt melts into an ionic conductor, collapsing the impedance between nickel core and inconel tube — a ground signal. When the spot cools the salt refreezes and insulation returns. 124 ± 7 °C for wing/fuselage elements; 204 ± 5 °C for pylon elements.

[!note]- Q5. Reconstruct the APU LEAK FED BY ENG deduction chain.

APU loop alarms + APU valve already closed → heat cannot be the APU's own supply → it is flowing backwards from the crossbleed duct → the APU check valve has failed open → the source feeding it is the left half's supplier → close engine 1's bleed and HP valves.

[!note]- Q6. Why do wing/pylon closures carry an engine-start exemption but the APU loop does not?

The exemption exists to protect the start's air supply. An APU-loop alarm indicts the supply duct itself — keeping it open would feed the fire, so the APU valve closes even mid-start.

[!note]- Q7. OAT 35 °C on a turnaround — what protects you from a phantom WING LEAK, and what principle does it teach?

Keep the slats in CONF 1 during transit to ventilate the leading edge. The principle: the sensor reports heat near the duct regardless of cause — know what your detector measures before you act on it.

Key takeaways

References

System purpose, series sensing elements, left/right split at the crossbleed valve, dual-loop A+B with AND logic, single pylon loops, element construction (nickel core, porous aluminium-oxide ceramic, sealed inconel tube, eutectic salt) with the impedance-collapse mechanism and both trigger values, single-loop-inoperative behaviour, the full annunciation and automatic-closure lists with their exemptions, the APU start-phase closure note and the cannot-reopen latch per AMM 36-22-00 (Description and Operation). Crew-level loop description, the activation sentence and the three-case action matrix per FCOM DSC-36-10-60. Series-coverage sentences per AMM 36-20-00. Hot-weather slats note per FCOM PRO-NOR-SOP-21. The playbook table and the detective framing of the FED BY ENG case are integrative syntheses of the referenced text. Dispatch consequences reflect some operators' MEL practice and are developed in article 11.

Independent study material, not an Airbus publication and not endorsed by the manufacturer. Always defer to the current operator FCOM, FCTM, and QRH for operational use.