The previous lesson built the machine and its databases; this one operates it. What the FMS provides in flight, what its numbers are made of and when they lie, how it fixes its own position and what happens as the inputs fall away, where its data must yield to raw data, and the CDU/MCDU through which all of it is worked, ending with the pre-flight set-up flow every FMS shares whatever its make.
What the FMS provides
In operation the FMS supplies typical data in five families:
lateral and vertical navigation guidance — the steering LNAV and VNAV fly.
present position.
time predictions — estimates for every waypoint and the destination.
fuel predictions — fuel remaining at each point ahead.
altitude and flight-level predictions — where the profile will have the aircraft at each point.
The rest of the lesson is about what those numbers rest on.
Computed from what
Every output blends three ingredients: inputted data, database data and measured data. The crew enters weights, winds and the route; the databases supply the world and the aircraft model; the sensors measure what is actually happening; and the FMS projects the three forward into its predictions. The blend is also the vulnerability: the measured ingredients correct themselves continuously, but the inputted ones are only as good as the entry, and the FMS computes confident, precise-looking numbers from whatever it was given. Consequences of incorrect entries scale with what the entry feeds: a wrong weight poisons the take-off speeds and every fuel figure; a wrong wind bends the time and fuel predictions; a wrong route point steers the aircraft, in LNAV, somewhere nobody cleared it to go.
Fuel predictions, honest and otherwise
In standard operations the fuel calculation is the performance database at work: the engine and drag models, the entered weights, the cost index speeds, and the forecast winds and temperatures along the route combine into fuel-remaining figures at every point ahead. The accuracy therefore leans on the entered zero-fuel weight and the fuel load, the quality of the wind and temperature forecast, and the performance factor that matches the model to this airframe.
The model has a blind spot: it assumes the aeroplane it describes. In abnormal configurations, engine out, gear down, flaps or spoilers extended, the calculations lose accuracy if the FMS cannot detect the failure: a dragging airframe burns more than the clean model predicts, and the predictions drift optimistic, precisely when fuel matters most. The discipline is knowing which failures the FMS sees and which it does not, and distrusting the fuel predictions in any configuration the model was never told about.
Tuning its own radios
To fix its position the FMS does not wait for the crew: it has dedicated radio-navigation receivers that it tunes automatically, picking stations from the navigation database, best-placed for a position fix, tuning them and using the results, all without a knob being touched. The crew’s receivers stay free for the aids the crew wants displayed.
The FMS position
The typical position inputs are the course’s roll call: GPS, the IRS of module 5, DME ranges and VOR bearings from the auto-tuned receivers, the LOC beam on approach, and the runway threshold (RWY THR), the known point the position can be pinned to as the take-off roll begins. From these the system creates its own FMS position fix, and every FMS calculation is based on that FMS position. Depending on the type of system, the fix is built one of two ways: from a single source, the most accurate data available at the time, the system promoting and demoting sources as their quality changes; or from multiple sources combined, a position derived from the inputs together, the blend the inertial lesson met as the FMC position.
As inputs fall away, the fix degrades gracefully but measurably, and the syllabus singles out one loss: GPS, whose absence bears directly on the capability of performing RNP/PBN approaches (performance-based navigation: approaches flown on the FMS position alone, permitted only while the system can demonstrate the required navigation performance, the position accuracy the procedure demands). Lose GPS and the demonstrable accuracy sags; sag far enough and the approach capability is gone, exactly the downgrade logic the autoland lesson taught for its own category.
FMS data against raw data
The FMS’s picture and the raw picture are not the same thing:
Following FMS data — following the computed position against the database’s picture of the world.
Following raw data — following the actual received signal: the localizer needle, the VOR bearing, the NDB pointer (the non-directional beacon, an old radio aid flown by a simple bearing pointer).
For an instrument approach procedure (IAP) built on a raw aid, LOC, VOR or NDB, there may be limitations on using FMS data as the primary source: the procedure is legally defined by the beam or bearing itself, the FMS position can be wrong exactly when its inputs are poorest, and the database’s coding of the procedure may differ from the published track. The working rule the syllabus wants: the FMS may fly the approach, but the raw data is the reference, displayed, monitored, and obeyed when the two disagree.
The CDU/MCDU
Everything above is worked through the CDU/MCDU, whose purpose is to be the crew’s interface into the FMS: the keyboard for every entry and the pages for every readout, the same unit the inertial systems and ACARS already borrowed.
The typical layout has four working parts:
the screen — showing one page of data at a time.
the line select keys — flanking it, each key owning the line beside it, for copying data in and out.
the menu select keys — jumping straight to the major pages.
the alphanumerical keys — for typing.
Two conventions make the unit workable. The scratchpad is the working line at the bottom of the screen: typed characters land there first, and a line select key then copies the scratchpad’s content into the chosen field (and system messages surface there too). And the fields themselves declare their importance by symbology:
rectangular boxes mark compulsory information the FMS cannot function without.
dashed lines mark optional information that refines the computation but does not gate it.