Airbus Flight Instructor
Airbus · Knowledge Base

Navigation — Position, Accuracy, and Radio Tuning

The first of the FM's three trades: knowing where the aircraft is, knowing how trustworthy that answer is, and keeping the right navaids tuned. This article walks the position pyramid, the accuracy bookkeeping (EPU, RNP, GPS PRIMARY), the position-disagreement alert, the tuning hierarchy, and alignment through polar operations.

 IRS1  IRS2  IRS3 ──(weighted average)──► MIX IRS position (pure inertial)
   └──┬─(each blends GPS)─► GPIRS ×3 ──(best pick)──► GPS position
      │
 VOR/DME/LOC ──(onside receivers)──► radio position
                                        ▼
        FM position = MIX IRS + bias (a correction vector toward GPS or radio)

Per FCOM DSC-22_20-10-20:

Each FMGEC computes its own aircraft position (called the 'FM position') from a MIX IRS position and a computed radio position or GPS position. The FMGES selects the most accurate position, considering the estimated accuracy and integrity of each positioning equipment. GPS/INERTIAL is the basic navigation mode, provided GPS data is valid and successfully tested.

GPS/inertial is not a bonus — it is the staple diet.


1. Three ingredients — and the bias

MIX IRS: the weighted average of the three IRS positions. A drifting unit has its weight automatically reduced; a failed one drops each FMGEC to a single IRS (onside or IRS 3), and every IRS position and velocity is continuously tested and excluded when unfit.

GPIRS — where the "GPS position" really comes from:

Each IRS computes a mixed IRS/GPS position called the GPIRS position. For this, each IRS can independently select their GPS source in order to maximize the availability of GPS data. Of the three GPIRS positions that each FMGEC receives, the FMS selects one GPIRS position based on a figure of merit and priority. The FMS uses the following hierarchy to perform the selection: ‐ Onside GPIRS position ‐ GPIRS 3 ‐ Opposite GPIRS position.

GPS data is blended inside the IRS first; the FMS then picks among three GPIRS candidates — onside, then 3, then opposite (article 18's own-source bias, same family). A GPIRS failing the integrity tests (horizontal integrity limit, automatic bad-satellite exclusion) has the whole GPS mode rejected in favour of radio updating. One exam-grade note: normally ADIRU 1 takes GPS 1 and ADIRU 2 takes GPS 2 (ADIRU 3 follows the ATT HDG selector) — but after one GPS is rejected, all three ADIRUs may quietly share the survivor: the GPS MONITOR page shows the dead receiver's data as dashes while GPS PRIMARY LOST may never appear — the stealth form of a single GPS failure.

Radio position: computed from onside navaids in five recipes — DME/DME, VOR/DME, LOC, DME/DME-LOC, VOR/DME-LOC — with LOC serving as the approach's fast lateral corrector, and the opposite side's receivers borrowable when hardware fails.

And the bias:

Each FMGEC computes a vector from its MIX IRS position to the radio or GPIRS position. This vector is called the 'bias'. Each FMGEC continuously updates its bias, if a radio position, or a GPIRS position is available. If an FMGEC loses its radio/GPIRS position, it memorizes the bias and uses it to compute the FM position, which equals the MIX IRS position plus the bias.

Lose every external source and the bias freezes — the FM position carries the last known correction forward on inertial legs (the essence of IRS-only navigation). At takeoff the FM position is snapped to the runway-threshold database coordinates (plus any takeoff shift entered on PERF TO for an intersection departure) — a correction suppressed when GPS PRIMARY is active, GPS being more trustworthy than the survey. Monitoring speaks at two levels: an FM-versus-single-IRS split beyond a time-scheduled threshold raises the MCDU's CHECK IRS (1/2/3)/FM POSITION; an FWC-detected abnormal IRS drift raises the ECAM's FM/IR POSITION DISAGREE.


2. The navigation-mode ranking

The decreasing order of priority is: ‐ IRS-GPS ‐ IRS-DME/DME ‐ IRS-VOR/DME ‐ IRS only. During approach flown with ILS / GLS or SLS , the system performs a temporary lateral update, via one of the following modes: ‐ IRS-GPS/LOC ‐ IRS-DME/DME-LOC ‐ IRS-VOR/DME-LOC ‐ IRS-LOC.

Four rungs — GPS, then DME pairs, then VOR/DME, then inertia alone — each able to borrow the localizer for a temporary lateral polish on an LS approach (article 11's enhanced LOC capture used FMS trajectory to help the beam; here the beam repays the favour).


3. Accuracy — EPU, RNP, and GPS PRIMARY

EPU (estimated position uncertainty) — two FG standards print two tables; both shown, the L2 column being the newer:

Mode L2 EPU L1 EPU
IRS/GPS metre-class figure of merit (GPS rejected above 0.28 NM) 0.05 NM
IRS/DME/DME converges toward 0.28 NM 0.27–0.37 NM
IRS/VOR/DME 0.1 + 0.05 × DME distance, min 0.28 0.30–0.42 NM
IRS only grows +8/0/+4/0/+2 NM/h in 30/60/30/60-min segments +6/0/+4/0/+2 NM/h in 40/50/40/45-min segments

Note: After an IRS alignment, the EPU is set at 5 NM. ‐ At takeoff, the EPU is set at 0.2 NM.

(The other standard prints 0.2 NM post-alignment — configuration difference, both recorded.) The segmented IRS-only growth models the Schuler oscillation: drift, steady, drift again — not a straight line to the bad.

RNP and HIGH/LOW:

‐ If the EPU does not exceed the RNP, accuracy is HIGH ‐ If the EPU exceeds the RNP, accuracy is LOW. The RNP is displayed in the REQUIRED field of the PROG page. The displayed RNP is (in a decreasing order of priority): ‐ The value that the flight crew entered ‐ The navigation database procedure value ‐ The system's default value

Defaults by area: en-route/oceanic 2, terminal/takeoff 1, GPS approach 0.3, otherwise 0.5 NM — and enter something too generous and the system talks back (PROCEDURE/AREA RNP IS XX.XX). Then the crown:

When in IRS/GPS mode, the GPS PRIMARY status combines two different criteria: ‐ The accuracy criterion previously described (HIGH/LOW accuracy) ‐ An integrity criterion: This is the capability to detect a failure and provide appropriate warning of it.

GPS PRIMARY = accurate and honest — HIGH says "I'm right"; PRIMARY adds "and I'll shout if I stop being right". Losing either criterion raises GPS PRIMARY LOST (clearable on the MCDU, not clearable on the ND, and garnished with a triple click during a non-precision approach). The responsibility transfer is the point to teach: with GPS PRIMARY, accuracy is HIGH by construction and no periodic check is required; lose it and the periodic PROG-page accuracy check becomes the crew's job. (UPGRAD/DOWNGRAD messages are suppressed in IRS/GPS mode for the same reason — the PRIMARY machinery took over the announcements.) And a manually deselected GPS is politely re-offered: within 80 NM of T/D or at the approach-phase switch, the scratchpad reminds "GPS IS DESELECTED".


4. NAV FM/GPS POS DISAGREE — the geometry of the threshold

The lower ECAM displays the 'NAV FM/GPS POS DISAGREE' amber caution, when the GPS PRIMARY function is active, and when either of the FMGEC positions deviates from the GPS positions 1 or 2 by more than: ‐ 0.5' of latitude ‐ For longitude: • 0.5' of longitude, when the aircraft latitude is between 0° and 45° • 0.7' of longitude, when the aircraft latitude is between 45° and 60° • 1' of longitude, when the aircraft latitude is between 60° and 70°. Above 70° of latitude, a longitude difference does not trigger the alarm.

Meridians converge: at high latitude one minute of longitude is fewer metres, so the same physical error needs more minutes — hence 0.5′ → 0.7′ → 1′, and above 70° the longitude term is abandoned altogether (takeoff phase inhibits the alert). The QRH's treatment skeleton (representative operator, by scenario): climb/cruise/descent — check PROG accuracy: HIGH → NAV may continue with ARC/ROSE NAV; LOW → HDG/TRK plus raw data; then POSITION MONITOR: one FM agreeing with its onside GPIRS → fly that side's AP/FD; both disagreeing → deselect GPS and fly raw data (the fault is in GPS, not the FMs). On an ILS/LOC approach: don't use NAV, continue on the beam. On an RNAV(GNSS)/RNP approach: visual references insufficient → go around — an FM-type approach with a doubtful position has no fallback. VOR/NDB: HDG/TRK plus raw data.


5. Tuning — three hands, one etiquette

Three paths: FMGEC automatic, MCDU RAD NAV manual, and the RMPs (the back door when both FMs or both MCDUs are gone — article 11); manual outranks automatic. Architecture: in dual/independent modes each FMGEC tunes its own side (one VOR, one DME, one ILS, one ADF); in single mode the healthy FM tunes both sides; a failed receiver is shared. The automatic VOR priority:

Automatic tuning complies with the following priorities for tuning the VOR: 1. The specified navaid for the approach 2. The navaid that the flight crew should use to compute the current radio position 3. For display purposes: ‐ A specified navaid for the active flight leg ‐ The 'TO' waypoint (TO WPT), if it is a navaid ‐ The 'FROM' waypoint (FROM WPT), if it is a navaid ‐ A waypoint farther along the flight path, if it is a navaid ‐ The navaid closest to the current position of the aircraft.

Steal the position-computation station with a manual tune and the MCDU answers TUNE BBB FFF.FF — give it back. The DME set is a quiet multitasker: one interrogator runs five frequencies at once (four on the earlier standard) — one for display, two for the DME/DME position (fully automatic, geometry-optimised, invisible to the crew), one for the VOR/DME position, one paired with the ILS. ILS auto-tuning has three occasions: takeoff/preflight phase when the departure runway has an ILS (the considerate default for an immediate return), an ILS approach in the plan, and within 300 NM direct of destination; a manual frequency clashing with the automatic request raises RWY/LS MISMATCH (article 11's source). ADF joins when the TO/FROM fix is an NDB or an NDB approach is selected.

The SELECTED NAVAIDS page lists everything the onside FMGEC is using — and offers only deselection (up to 4 navaids, lasting the whole flight) or total inhibition: a station NOTAM'd for calibration gets blacklisted and the radio position stops eating it. Manual-tuning etiquette: enter the ident first; resort to a frequency only on NOT IN DATA BASE — and mind the trap: a frequency entry makes the system attach the nearest same-frequency station's ident, possibly the wrong one (and a mixed-type mismatch may tune only half) — ignore the small-print ident and verify yourself. Identification: a decoded ident on PFD/ND matching the chart absolves you of Morse-listening; mismatch → audio (decoders mis-read occasionally).


6. Alignment — and the poles

The preferred way of initializing the ADIRS is to use the GPS position, because this does not require any flight crew intervention. Less than 1 min after setting the ADIRS to NAV, the INIT page displays the GPS position without displaying the ALIGN IRS prompt.

GPS initialisation is the default: within a minute of NAV selection the INIT page fills itself, no ALIGN IRS prompt shown (entering the company route paints the airport reference and the prompt — not pressing it leaves the GPS position in charge; the manual path exists for GPS-less days). The clock:

A normal alignment takes between 5 and 10 min (depending on the local latitude). At latitudes between 73° and 82° (North or South), the alignment takes about 17 min. The ADIRS cannot be aligned beyond 82° (North or South). A fast alignment takes about 30 s, and it is used for improving the ADIRS initial position accuracy, when time is limited.

Beyond 82° the earth-rate component is too weak for the gyros to find north — no alignment. A flashing IRS IN ALIGN memo flags one of four anomalies: the aircraft was moved (alignment restarts itself), the MCDU-sent position disagrees with the memorised one, or with the GPS one, or the latitude disagrees with the alignment's own solution. The half-failure mode, per a representative operator QRH (ATT mode): select ATT on the affected IR → hold speed, heading and altitude constant for 30 s (a fresh attitude datum) → enter the heading on the IRS MONITOR page → cross-check against the standby compass periodically and re-enter — the ATT-mode heading is a watch that loses time. After landing the system even grades each IR's average drift against the threshold-plus-400-m geometry — maintenance trend data on the IRS page.

Polar rules (above 60°): at 78°N (60°S) the FM switches its position mathematics to earth-centred XYZ coordinates — latitude/longitude is singular at the pole, cartesian isn't, which is why the FM can fly directly over the pole. MAG↔TRUE switching is seamless — no lateral discontinuity in NAV, and HDG/TRK targets convert by local variation (no swerve on switching). The ceiling: when TRUE reference is selected, only CAT I ILS approaches may be flown (article 12's TRUE-north cap) — and selecting TRUE against a MAG-coded approach drops NAV at procedure entry. The reminder chain: CHECK NORTH REF (amber, database-dependent) when the terminal area's coding disagrees with the selection; SELECT TRUE when entering the polar area still on MAG; from 65°N the ND's corner shows GRID TRK (grid track plus XTK, the polar monitoring pair). VOR needles in ARC/ROSE NAV whose station reference disagrees with the cockpit selection appear magenta with a CORR flag — raw data plus local variation; ROSE VOR/ILS stays raw (amber MAG/TRUE side note); CORR appearing at all means check both ends, with a 5° cross-check tolerance. And overflying the pole itself:

When approaching the pole, a CHECK HDG message appears... Disregard the ECAM Procedure. The message is due to the fact that the IRSs do not reach the pole simultaneously; thus the headings do not turn simultaneously from 360 to 180 °.

Three IRSs cannot cross the point at the same instant, so their headings flip 360→180 out of step — a geometric false alarm, and one of the manuals' rare explicit "disregard the ECAM".


7. Operations

The navigation-accuracy cross-check (the craft that returns when GPS PRIMARY is lost): tune a real station on RMP/RAD NAV, read raw bearing/distance on the ND, compare with the PROG page's computed values (article 19 uses the same craft in backup navigation, entering station coordinates on the B/UP PROG page); polar tolerance 5°.

Three layers of GPS rejection, kept distinct: ① FMS-level integrity rejection (automatic, falls to radio updating); ② manual DESELECT (the both-FMs-disagree branch of the QRH procedure); ③ ADIRU-level source rejection (three units sharing one receiver, GPS PRIMARY LOST possibly absent — check the GPS MONITOR page for dashes).

Interfaces: GPS PRIMARY availability prediction and RNP AR dispatch → article 34; GPS PRIMARY LOST during an RNAV(RNP) approach → article 11/31; the IRS failure spectrum → article 19.

[!warning]- Four misconceptions this article corrects (1) "GPS position" is not raw receiver output — GPS blends inside each IRS (GPIRS), and one failed receiver can leave all three ADIRUs sharing the survivor with no GPS PRIMARY LOST shown. (2) IRS-only drift is not linear — the EPU grows in Schuler segments (fast, hold, slower), and takeoff resets it to 0.2 NM. (3) GPS PRIMARY present means no periodic accuracy check is required — its loss transfers that duty to you; HIGH alone is accuracy without the honesty guarantee. (4) CHECK HDG at the pole is a false alarm by geometry — the IRSs flip heading out of step; the FCOM says disregard the ECAM.


Self-test

[!note]- Q1. The FM-position formula — and when does the bias freeze?

FM position = MIX IRS position + bias, the bias being the vector from MIX IRS to the radio/GPIRS position, continuously refreshed. It freezes when all external sources are lost — inertial navigation carries the last known correction forward.

[!note]- Q2. The GPIRS selection hierarchy — and how can one GPS fail without GPS PRIMARY LOST appearing?

Onside GPIRS → GPIRS 3 → opposite GPIRS, by figure of merit and priority. After one receiver is rejected, all three ADIRUs can select the surviving GPS — integrity holds, so no LOST message; the GPS MONITOR page's dashes betray it.

[!note]- Q3. Why does the IRS-only EPU grow in segments — and what is the EPU at takeoff?

The segments model the Schuler oscillation (fast/hold/slower alternation), not linear decay. At takeoff the EPU is set to 0.2 NM by the runway update.

[!note]- Q4. The RNP display's three-level priority and the four default area values?

Crew entry > database procedure value > system default. Defaults: 2 NM en-route/oceanic, 1 NM terminal/takeoff, 0.3 NM GPS approach, 0.5 NM otherwise.

[!note]- Q5. GPS PRIMARY's two criteria — and who checks accuracy after it is lost?

Accuracy (HIGH: EPU ≤ RNP) plus integrity (the ability to detect its own failure and warn). With PRIMARY lost, the crew takes over periodic accuracy checking on the PROG page.

[!note]- Q6. The DISAGREE latitude threshold — and why does longitude stop counting above 70°?

0.5′ of latitude (longitude 0.5′/0.7′/1′ by latitude band). Meridians converge: a minute of longitude shrinks with latitude, so the same distance is ever more minutes — above 70° the longitude comparison is meaningless and is dropped.

[!note]- Q7. The first priority of automatic VOR tuning — and what does TUNE BBB FFF.FF mean?

The specified navaid for the approach. The TUNE message is the FM asking for its position-computation station back after a manual tune displaced it.

[!note]- Q8. The trap in tuning by frequency?

The system attaches the nearest same-frequency station's ident — possibly the wrong station (and a type mismatch may tune only half). Ignore the attached ident and verify against the chart; prefer ident entry.

[!note]- Q9. The three alignment durations and the 82° line — and ATT mode's three steps plus its standing burden?

Normal 5–10 min; 17 min between 73° and 82°; impossible beyond 82°. Fast alignment ~30 s for position refresh. ATT mode: hold speed/heading/altitude 30 s, set the heading on IRS MONITOR, then periodically cross-check and re-enter against the standby compass — the heading drifts like a slow watch.

[!note]- Q10. CHECK HDG at the pole — the response? Can you fly CAT II on TRUE reference? And TRUE selected against a MAG-coded approach?

Disregard the ECAM — the IRSs flip 360→180 out of step. No — TRUE limits you to CAT I ILS. And NAV disengages at procedure entry when the coding and the reference disagree.


Key takeaways

Theme The one thing to remember
Pyramid Inertia is the draft, GPS/radio are the margin notes (bias); notes stop, the draft carries the last correction
Ranking GPS > DME/DME > VOR/DME > inertia; the LOC polishes laterally on approach
GPS PRIMARY Accurate and honest — present: no checks needed; lost: the checking is yours
DISAGREE 0.5′ lat; longitude tolerance widens with latitude and vanishes above 70°
Tuning Manual beats automatic, but return the position station; ident first, frequency last
Alignment 5–10 min / 17 min / never beyond 82°; GPS initialises hands-off
Polar Switch TRUE entering, grid track for monitoring, suspect magenta CORR, ignore CHECK HDG at the pole — and TRUE caps you at CAT I

References

Position computation (MIX IRS, GPIRS hierarchy, radio recipes, bias, takeoff update, monitoring messages) and mode ranking per FCOM DSC-22_20-10-20 (position sections); EPU tables (both FG standards), RNP priorities and defaults, GPS PRIMARY criteria and messages, and the FM/GPS disagreement thresholds per the accuracy section; tuning paths, VOR priorities, DME multitasking, ILS/ADF auto-tuning, deselection and manual-tuning rules per the tuning section; alignment, ATT mode, drift grading and the polar rule set per the alignment/polar sections (all FCOM DSC-22_20-10-20). ATT-mode and FM/GPS-disagree procedures per a representative operator QRH. The pyramid diagram is an integrative synthesis. GLS/SLS wording is quoted as printed; those functions are not installed on the baseline airframe of this series.

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.