EFIS Attitude Disagreement on the Ground

Establish ground truth first. Three independent references agree that the aircraft is level: the surveyed stand, the digital inclinometer (0.1° nose-up) and the standby attitude indicator (0.0°). The captain's PFD showing 2.5° nose-up is therefore wrong by about 2.4°.

The distraction: the two ADCs agreeing on altitude and airspeed is irrelevant to attitude. Attitude data comes from the IRS, not the ADC. An engineer who chases the ADC because 'the two sides disagree' is looking in the wrong system.

Prime suspects, in order:

1. IRS 1 attitude output (position data being good does not prove the attitude output is good).
2. The captain's PFD/display unit or its wiring.

Do not decide by majority vote. The standby is a self-contained gyro and the inclinometer is a physical reference; these are stronger evidence than 'two out of three boxes say zero'.

Discriminating test: cross-select the captain's PFD to IRS 2 and, if the panel allows, the first officer's PFD to IRS 1.

• If the 2.5° error follows IRS 1 (now appearing on the first officer's display), the IRS is faulty.
• If the captain's display still reads 2.5° while fed from IRS 2, the display or its data path is faulty.

What realignment can and cannot fix. A full alignment uses the earth's rotation and gravity to establish attitude and heading. It corrects errors caused by:

• movement or vibration during a previous alignment,
• an alignment carried out with a poor position entry,
• a fast/quick alignment that did not complete correctly.

Why it is unlikely to help here:

• An alignment error from motion is normally a few tenths of a degree, not 2.4° on a stationary aircraft.
• Both units were aligned at the same place at the same time; IRS 2 aligned correctly, so the procedure and the aircraft state were acceptable.
• Position data from IRS 1 is already correct, which argues against a failed alignment.

Correct approach: realigning as a diagnostic step is acceptable only if the aircraft is genuinely stationary and level, the correct position is entered, and the result is re-checked against the inclinometer. If the 2.4° bias persists after a valid alignment, the unit has failed its attitude accuracy requirement and must be tested to the fault-isolation manual and replaced. Never 'adjust' a horizon to match another display - the error would then be hidden and could be present in flight.

Before release:

1. Complete the source-select test to positively identify whether the fault is IRS 1 or the captain's display/data path.
2. If IRS 1 is implicated, run the manufacturer's built-in test and the fault-isolation procedure; replace the unit if it fails the attitude accuracy check.
3. If the display is implicated, check the ARINC wiring, connectors and the display's own test page, then replace the display.
4. After replacement, re-align and verify all three attitude sources against the inclinometer on the level stand - agreement within the manual's tolerance, typically a fraction of a degree.
5. Record the before/after readings and the reference used.

Why no deferral:

• Attitude is primary flight information. A 2.4° pitch bias is small in cruise but significant in the flare, where the pitch attitude window is only a few degrees; it also corrupts flight-director and autopilot commands.
• The error is a known, quantified defect in a primary display, not an intermittent snag. An MEL cannot be used to dispatch an aircraft with a display that is knowingly giving false primary data.
• The other pilot's display and the standby are not acceptable substitutes for a required primary instrument under most approval regimes; the standby is a backup, not an operative source.
• If the true cause were a failed IRS, the same unit supplies heading and position - the defect could be broader than the pitch error alone.