PPRuNe Forums - View Single Post - Turkish Airlines cargo 747 crashes in Kyrgyzstan
Old 3rd Apr 2017, 18:47
  #351 (permalink)  
EMIT
 
Join Date: Aug 2007
Location: The Netherlands
Age: 67
Posts: 288
Likes: 0
Received 0 Likes on 0 Posts
Hello Check Airman,

It is all a matter of careful wording, either in posts or in (reports about) reports.
In official Boeing documentation, there is a description of how the autoflight system monitors ILS signal for interferences, and how it will act to pass through such a phenomenon if it is short lived or if it persists. This description mentions that the system will maintain an inertial path, but does not attatch numbers to that description.

That description leaves room for interpretation and you can bet your 2 cents that the way the system works will be logical for the way an aircraft flies: if the ILS glideslope happens to be 3 degrees and, once established on the slope, the signal is lost for a short while, then the inertial path will be 3 degrees. If the slope was for 2.5 degrees, then the inertial path will be 2.5 degrees.

In this particular case, the aircraft started to intercept the 9 degree false glideslope - it never established a steady situation, the signal irregularity happened during slope acquisition. In that case, the most logical "best guess" inertial path would be 3 degrees, the "standard ILS slope" until a good quality signal would appear again. In this case, the inertial period lasted so long, that the proper warnings were given by the system, an AUTOPILOT Eicas Caution and AMBER LINE drawn through the pitch FMA. The proper reaction to those conditions is the immediate start of a Go Around.

You must realize that there is a lot of clever algorithm stuff in autopilot systems: for instance, most ILS glideslope intercepts are made from below the slope, in level flight - so you have a period where the ILS deviation would suggest to fly UP to the slope middle line. In reality, the A/P will maintain level flight until it hits the center of the slope.
A false 9 degree slope may exhibit a reversal of signal: fly DOWN when below the middle line, fly UP when above it. Whether or not the signal is reversed depends on the type of antenna array that is employed for the G/S signal. The A/P logic must filter and smooth all the signals and aircraft control commands so that no abrupt and illogical moves are made. The user manual descriptions will never be detailed enough to describe exactly how all that is accomplished.

What the user manuals do very well though is the bottom line - if things are not going in the way they should be going, GO AROUND! I mean, 1 mile to go to the threshold and being 2.000 ft above threshold elevation is a pretty simple indication that the situation is not as it should be.
EMIT is offline