Abnormal I: Unreliable Airspeed and the ADR Fault Family

After AF447, "unreliable airspeed" became mandatory recurrent training industry-wide. Its danger is not losing a number but that the false number looks like a true one: three pitot tubes ice identically, three ADRs lie in chorus, and the voting mechanism votes out the only one telling the truth. This article walks the order "how the deception happens → when to suspect → memory items → troubleshoot and isolate → fly impaired (tables/BUSS) → the ADR fault-family ECAM," packing the QRH 23.03 series and the FCTM PR-AEP-NAV chapter into an actionable framework.

1. The deception mechanism — the glory and the shadow of majority rule

The normal fault-accommodation logic. Per FCTM PR-AEP-NAV:

The fault accommodation logics rely on a voting principle: When the data provided by one source diverges from the average value, the systems automatically reject this source and continue to operate normally using the remaining two sources.

One source goes bad and the system digests it silently (only a CAT 3 SINGLE downgrade, or a NAV IAS DISCREPANCY caution). The shadow is here. Per FCTM PR-AEP-NAV:

The systems reject the correct ADR and continue to operate using the two erroneous but consistent ADRs. The flight crew can encounter such a situation when, for example, two or all three pitot tubes are obstructed at the same time, to the same degree, and in the same way.

Majority rule guarantees "consistency," not "truth" — the common root of the two counter-intuitive disciplines "keep one on against a double-false" and "do not reject the outlier." Two bad but inconsistent is another picture. Per FCTM PR-AEP-NAV:

Both the AP and A/THR disconnect. If the disagree lasts for more than 10 s, the PRIMs trigger the NAV ADR DISAGREE“ ECAM caution. The flight controls revert to ALTN 2 law. The high speed protection and low speed protection are lost.

(And it is latched until a PRIM reset on the ground with no hydraulic pressure — AP/A-THR can re-engage if the disagreement was transient, but the law does not return.) The five-case portrait table (the raw material for diagnosis): rain water, drain holes clear — a transient drop, fluctuation, step down then gradual return (self-healing); drain holes also blocked — permanent low IAS; heat failure/icing, drains clear — total pressure leaks to static, IAS drops/fluctuates, A/THR erratic; total blockage incl. drains — total pressure sealed: constant IAS in level flight, but IAS rises in climb and falls in descent (the AP, to hold its target, pitches ever more nose-up in the climb — the AF447-type trap); static ports blocked on the ground — normal on the roll, altitude frozen after lift-off, IAS drops after lift-off, falls climbing and rises descending. The official conclusion. Per FCTM PR-AEP-NAV:

The above table clearly illustrates that no single rule can be given to conclusively identify all possible erroneous airspeed/altitude indications cases.

And a statistic that shifts the decision weight. Per FCTM PR-AEP-NAV:

At high altitude, typically above FL 250, the cases of unreliable speed situation are mostly a temporary phenomenon: They are usually due to contamination of the pitots, by water or ice, in particular meteorological conditions. In-service experience shows that such a contamination typically disappears after few minutes, allowing to recover normal speed indications.

(High up, do not rush to "execute" an ADR — stabilise, wait a few minutes, and many cases fix themselves; the low-altitude blockages are the permanent ones.)

2. When to suspect — the symptom list and two rules of "who to trust"

There are two entries: an ECAM call (the UNREL SPD PROC… APPLY line, e.g. NAV ADR DISAGREE, A.ICE PITOT), or the crew's own suspicion. The FCTM's suspicion list (condensed): a disagreement between three sources and the standby; fluctuating/unexpected changes; basic parameters that do not add up. Per FCTM PR-AEP-NAV:

• The altitude does not increase, whereas there is an important nose-up pitch and high thrust • The IAS increases, whereas there is an important nose-up pitch • The IAS decreases, whereas there is an important nose-down pitch

Odd AP/FD/A-THR behaviour; warnings that contradict the airspeed — hiding two rules. Per FCTM PR-AEP-NAV:

• Rely on the STALL warning. Erroneous airspeed data does not affect the STALL warning, because the STALL warning is based on AOA data

• Depending on the situation, the OVERSPEED warning may be false or justified. When the OVERSPEED VFE warning triggers, the appearance of aircraft buffet is a symptom that the airspeed is indeed excessive.

(The stall warning is believed unconditionally; the overspeed warning takes buffet as its witness.) Other symptoms: barometric altitude inconsistent with RA; aerodynamic noise trending opposite to the IAS (the ear is an independent sensor); the normal gear-extension impossible on approach (the LGCIU airspeed interlock jammed by a false speed).

3. Memory items — save the life first, reason later

The trigger is not the symptom itself but three questions (FCTM): have you lost situation awareness? are the current pitch/thrust inappropriate for the phase? is the flight path unexpected — any doubt about short-term safety, do it:

Disconnect AP/A-THR/FD, then set parameters by window (QRH 23.03A): below the thrust-reduction altitude 15°/TOGA; above it and below FL100 10°/CLB; above FL100 5°/CLB; flaps 0–3 maintain, FULL retract to 3; check speed brakes retracted; gear up; level off at MSA for troubleshooting. The character of these parameters: a coarse "safe climbing" trajectory in any weight, any configuration — but not precise, so the FCTM follows immediately. Per FCTM PR-AEP-NAV:

The flight crew must apply the QRH procedure without delay, because flying with the memory pitch/thrust values for an extended period of time can lead to exceed the aircraft speed limits.

(15°/TOGA at light weight accelerates past VMO — the memory item is a tourniquet, not a bandage.) The bottom line throughout: RESPECT STALL WARNING.

4. Stabilise and troubleshoot — disconnect, heat, and slow the conclusion

The level-off's first action is always the three disconnects. Per FCTM PR-AEP-NAV:

In all cases, the initial actions are to disconnect the automations. This prevents the Flight Guidance to use erroneous data for the computation of the aircraft guidance.

(The "AP pitching ever more nose-up to hold speed in the climb" is the result of not disconnecting.) If the A/THR disconnected itself, mind the Thrust Lock trap. Per FCTM PR-AEP-NAV:

If the A/THR automatically disconnects, the Thrust Lock function activates. The thrust is locked at its level at the moment of the disconnection until the flight crew moves the levers. The thrust may be locked at idle, due to normal A/THR behaviour, or due to the use of erroneous data.

(Do not move the levers and the engines idle while you descend — after disconnecting you must take the levers.) Troubleshoot in three steps: probe/window heat ON (treat the commonest cause first); cross-check all indications — memorising one hardware fact: ADR 3 and the standby share the same probes, so "the standby and ADR 3 agree" is not two independent witnesses but possibly two mouths of one blocked tube; then the chapter's heaviest WARNING. Per FCTM PR-AEP-NAV:

WARNING Do not instinctively reject an outlier ADR.

When one indication differs from the others, the flight crew may be tempted to reject the outlier information. However, they should be aware that two or even all three ADRs can provide identical but erroneous data.

External witnesses for truth: pitch/thrust against the QRH's "resulting speed," GPS ground speed (wind-corrected), RA, aerodynamic noise. A special case — radome collapse (after a bird strike / hail). Per FCTM PR-AEP-NAV:

As a result, the flight crew should only assess the reliability of the air data sources when the aircraft trajectory and configurations are stable.

The isolation decision tree (this group's text): recognise the reliable one → use it, turn off the sick, apply the ECAM; cannot recognise →. Per FCTM PR-AEP-NAV:

• Above FL 250, turn two ADRs off to prevent the flight control laws from using two consistent but unreliable ADR data. The flight crew must keep one ADR on. For flight continuation, the flight crew uses pitch and thrust tables of the QRH • Below FL 250, turn off all ADRs then use the BUSS for the flight continuation.

The depth of "keep one on": with a single ADR left the PRIMs no longer dare use any ADR for protection, but neither can they be fooled by a double-false majority; high up, keeping it also bets on the statistic — it may fix itself in minutes. Set ALT RPTG OFF too. Why is FL250 the BUSS threshold? The BUSS scale is an AOA-derived "flyable band," and high-altitude Mach effects distort the AOA-speed relationship so the green is no longer reliable — so high up, prefer the tables to the BUSS.

5. Table flying — a craft of "attitude sets speed"

The core method (with FPV or GPS altitude available). Per FCTM PR-AEP-NAV:

• If the pitch necessary to maintain level flight is above the pitch target of the table, the aircraft is slow. Then increase the thrust • If the pitch necessary to maintain level flight is below the pitch target of the table, the aircraft is fast. Then decrease the thrust.

(In level flight, attitude is the mirror of speed — at the same altitude and weight, flying slow means pitching high.) With neither FPV nor GPS altitude: set pitch and thrust by table and wait — "expect a significant time to stabilise and important altitude variations during it." Change one parameter at a time; a three-step configuration change: adjust thrust to accel/decel → move flaps one step → return thrust to the table value as pitch reaches the table target. This configuration's key numbers (QRH 23.03D, 200 t as the memory anchor): level cruise — 4° at/below FL250, 3° above; FL300/200 t = 72% N1 (275 kt); descent — clean idle 2° (200 t → 255 kt); final — CONF 3 gear-down 4°/48% N1. The radome-collapse correction. Per QRH 23.03D:

If the failure is due to radome destruction, the drag will increase and therefore N1 must be increased by 3% in cruise and 1.5% in approach. Fuel flow will increase by about 10%.

6. BUSS and ALL ADR OFF — the "flyable band" after total blindness

This is a non-reversible BUSS: turning all three ADRs off triggers it, replacing both speed tapes, not recoverable before landing. The six rules once active. Per FCTM PR-AEP-NAV:

When the BUSS is active: ‐ The AP/FD and A/THR must be disconnected ‐ The F/CTL Laws are in Alternate law ‐ The STALL warning remains operative ‐ Cabin pressure must be controlled manually

CAUTION When flying with the BUSS, do not use the speed brakes. Flying with speed brakes extended affects the relationship between the speed and AOA, and therefore the BUSS may provide erroneous data.

(The whole BUSS scale is built on the clean-configuration AOA-speed table; deploy the speed brakes and the table is void.) A self-check. Per FCTM PR-AEP-NAV:

The BUSS is directly based on the current AOA. Any longitudinal input on the stick will induce an AOA change, and therefore will cause the BUSS to move. If not, the flight crew must disregard the BUSS and use pitch and thrust tables.

The ALL ADR OFF QRH (called by NAV ADR 1+2+3 FAULT): fly the green (retract configuration at the top of the green to bank energy); PFD altitude switches to GPS; TCAS and altitude reporting both dead; cabin pressure manual — MODE SEL MAN, target cabin V/S climb 500 / descent 300 ft/min, target cabin altitude by cruise level (FL410 → 8000 ft… FL250 → 2500 ft); apply the landing-distance procedure. On approach. Per FCOM PRO-ABN-NAV [QRH] ALL ADR OFF:

Before extending the slats/flaps, it is better to fly at the bottom of the speed scale green area, and to be in straight flight.

Note: Do not select CONF FULL so as not to degrade the handling qualities. Since ADR parameters are lost, the SFCCs will set CONF 1+F instead of CONF 1.

The approach target speed is ready-made: the BUSS green triangle = VAPP; at FLAP 2 use gravity gear extension (the normal airspeed interlock is dead) — at the cost of no nose-wheel steering and the gear doors staying open; on final, cabin fully manual (MAN V/S FULL UP), ΔP zero before opening the door.

7. The ADR fault-family ECAM — single, dual, triple, and "argument"

Single NAV ADR 1(2)(3) FAULT: re-source + turn off the sick — ADR 1 fault: AIR DATA SWTG to CAPT ON 3, ADR 1 pb OFF; ADR 2 fault: F/O ON 3; ADR 3 fault: SWTG to NORM. ADR 1's own easter egg (the taproot of the EGPWS article). Per FCOM PRO-ABN-NAV:

Due to the fact that the enhanced functions of the EGPWS or the predictive functions of the GPWS (for aircraft equipped with T2CAS or T3CAS ) are inhibited, the GPWS TERR FAULT light and the GPWS SYS FAULT light come on. If this occurs, the flight crew should switch off the GPWS TERR pb-sw and the GPWS SYS pb-sw.

STATUS bill: CAT 3 SINGLE ONLY; INOP: that ADR (+ GPWS/GPWS TERR with ADR 1). A general order rule (dual fault on the same ADIRU). Per FCOM PRO-ABN-NAV:

In case of simultaneous failure of ADR and IR (same ADIRU) apply ADR FAULT procedure before IR FAULT procedure.

Dual NAV ADR 1+2(1+3)(2+3) FAULT: set the SWTG by combination, both sick pbs OFF; the transponder moves house — 1+3 fault → ATC SYS 2, 2+3 fault → SYS 1. The big bill: F/CTL ALTN LAW (PROT LOST) + MAX SPEED 330/.82 (VMO protection gone, a hand-drawn red line); FOR LDG FLAP LVR 3; a long INOP queue: F/CTL PROT, dual AP, A/THR, CAT 2, TCAS (1+2), the relevant XPDR. Triple fault → §6. The NAV ADR DISAGREE (this group's 1746-1780 text) trigger is precise. Per FCOM PRO-ABN-NAV:

This alert triggers when the following conditions occur: ‐ The PRIMs rejected an ADR, or an ADR is set to OFF ‐ The information received from the two remaining ADRs is different.

The procedure's first step is triage: AIR SPD X CHECK (two PFDs + standby) — speed disagrees → the whole UNREL SPD procedure takes over; speed agrees → the problem is in AOA. The bill: ALTN law and latched, 330/.82, and a frozen rudder-travel limit. Per FCOM PRO-ABN-NAV:

The rudder travel limit value is frozen at the moment when the failure occurs. Therefore, to prevent damage to the aircraft structure, use the rudder with care, when the speed is above 160 kt. At slats’ extension, full rudder travel authority is recovered.

(The high-speed travel limit protects the structure; frozen at a small cruise value it is fine, but frozen at a large low-speed value, full rudder at high speed is a structural risk.) Capability: CAT 3 SINGLE ONLY (this group).

8. The AOA family and two "false-protection" topics

NAV AOA DISAGREE (this group's text): after the third is off/rejected, the remaining two AOA argue — the bill nearly copies ADR DISAGREE: ALTN (PROT LOST) + 330/.82 + frozen rudder, INOP adds CAT 2. Single AOA FAULT (CAPT/F-O/STBY): awareness + INOP only — but its sequel in icing is the prelude to ABNORMAL V ALPHA PROT. ABNORMAL V ALPHA PROT (QRH chapter; aircraft without OEB49): the mechanism is that the PRIM computes protection from a combined value of three AOA. Per FCTM PR-AEP-NAV:

However, the high AOA protection and the display of the V Alpha Prot may be affected by inaccurate AOA measured values in some exceptional conditions: ‐ One or several AOA probes blocked, or ‐ Damage to the probe that causes a small offset in the measurement of the AOA value.

Symptom: the top of the black-and-amber strip (V α prot) climbs abnormally as Mach increases, nearing the current speed — a false AOA makes the aircraft think it is near the stall, and the protection may activate spuriously. Monitoring and the red line. Per FCTM PR-AEP-NAV:

During stabilized flight path with steady heading, if the margin between the V Alpha Prot and the current airspeed decreases below 15 kt, the flight crew must stop increasing both Mach and altitude.

Why is locking Mach and altitude enough? The physics: higher Mach → protection triggers at a lower AOA (V α prot climbs); higher altitude (constant Mach) → level flight needs a larger AOA (again nearing the protection) — move neither knob and the margin stops worsening; to widen the margin, descend. Another discriminator: Abnormal V Alpha Prot does not cause unreliable speed indications (the speed tape is genuine — do not take the wrong procedure). And NAV ALT DISCREPANCY: triggered when the two PFD altitudes differ by more than 500 ft (STD) / 250 ft (QNH), with a CHEK ALT prompt; handling: three-way check against the standby, AIR DATA SWTG the ADR 3 to the faulty side.

Key numbers

Self-test

[!note]- Q1. Three pitots ice identically — how does voting respond? Does ECAM tell you? Your first external witness? Voting keeps two "consistent but wrong," so it does not detect it — possibly no ECAM. First witness: pitch/thrust against the QRH resulting speed, or GPS ground speed / RA / aerodynamic noise.

[!note]- Q2. IAS rising in the climb, AP pitching ever more nose-up — which of the five cases, and the mechanism chain? Total blockage including drains: total pressure sealed, so IAS rises with altitude in a climb; the AP, holding its speed target, pitches up more and more (the AF447 trap).

[!note]- Q3. The three memory-item trigger questions? At FL120 above thrust-red altitude, what values? Why not fly them long? Lost SA? / pitch-thrust inappropriate? / unexpected path? At FL120 above thrust-red: 10°/CLB. Not long, because the memory values can exceed speed limits.

[!note]- Q4. Why can "standby and ADR 3 agree" not convict ADR 1/2? ADR 3 shares probes with the standby — they are not two independent witnesses, so a common-mode blockage fools both.

[!note]- Q5. FL330, cannot identify the bad ADR: how many off, how many on, why? And FL180? FL330: turn two off, keep one on (prevent a double-false majority feeding the flight controls) + tables. FL180: turn all off → BUSS.

[!note]- Q6. V α prot climbs to 12 kt from current speed — climb to FL400 for smooth air? Add Mach to save time? Where is the way out? No and no: increasing altitude or Mach shrinks the margin. Descend (constant Mach) to widen the margin.

Key takeaways

References

• FCOM PRO-ABN-NAV — memory items, ALL ADR OFF, ABN V α PROT, ADR single/dual/DISAGREE, AOA family, ALT DISCREPANCY.
• FCTM PR-AEP-NAV — voting mechanism, five cases, FL250 statistic, symptom list, Thrust Lock, do-not-reject-outlier, radome, table technique, BUSS rules, ABN V α PROT mechanism.
• QRH 23.03D — pitch/thrust tables + radome correction.

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.