Understanding Autoland

[JetPilot500]

Hi all,

I don't fly an airplane with Autoland, but doing some research on it, so hoping someone could help.

At what point does the Autopilot/Autoland stop using Glideslope for guidance? And what happens next? How does it gently find the runway?

Looking for technical information as much as possible.

Thanks!!
Tony

[tubby linton]

Flare mode in the Airbus engages at 50 feet radio altitude.. The previous mode being GS.The transition is based upon current vertical speed. The autoland has cerified limits for glideslope angle, aircraft flap configuration and also the aircraft weight , all,of which can vary vertical speed required for the approach.

[172_driver]

On the 737, G/S is tracked until flare is initiated at 50 ft AGL. I am not sure if it's smart enough to adjust the flare technique depending on conditions. Did one the other day with a tailwind approaching the limit (10 kts) and it was firm !!

For LOC tracking, I've heard (without finding the source, think it was 737 CBT) that the last 50' LOC tracking is abandoned and current ground track is maintained. Don't quote me on that. Quite common to land a wee bit off center line.

[Denti]

The fail operational 737 can continue the autoland with a complete groundstation failure (loss of ILS/GLS signal) below 200ft (alert height). However, it might not be able to completely cover sudden changes in wind direction and therefore a go around is advised. We tried it in the simulator though and even a windshift of 25kts crosswind from one direction to the opposite, basically a 50kts wind shift in less than 50ft, did not lead to a displacement of more than 8m from the centreline.

Haven't tried it in the airbus yet, but a loss of G/S signal above 100ft or a loss of LOC signal above 15ft lead to a red autoland light which requires a go around.

[underfire]

A wander a bit from the thread, but the systems are coupled (FW I understand), so here goes...

Curious on the AB, the windshear detection system, which I understand is coupled with the Autoland, when it senses certain movements, will trigger a GA as well.

Any thoughts on how well, or not well, that wind shear system performs?

[NSEU]

At what point does the Autopilot/Autoland stop using Glideslope for guidance?

It depends on the aircraft. On the 747-400, for example, the glideslope is abandoned at flare height which is between 40 and 60 feet, depending on descent rate. The autopilot aims for a touchdown with a descent rate of 1.5 feet per second (radio altitude), 450 feet past the glideslope transmitter. On the 747-400, localiser is used during rollout, perhaps down to 40kts (retuning of receivers is allowed by airplane logic below this speed)

Like the other aircraft types, there is assistance provided by the inertial reference system. For example, when the G/S is in use, ground signal loss can be tolerated for certain time periods (e.g. 10 seconds). A sensible "inertial path" is maintained. Rapid deflections of ground signals due to, say, other aircraft interfering with the signals, are recognised as such (and the aircraft doesn't go crazy trying to follow the signals).

[Denti]

Rapid deflections of ground signals due to, say, other aircraft interfering with the signals, are recognised as such (and the aircraft doesn't go crazy trying to follow the signals).

Interesting point. Apparently that didn't work out all that well for the 777 from singapore airlines in munich a while back. We tried the same in the 737 simulator and the results were similar, even more annoying was the fact that we could not disconnect the autopilot by overpowering the rudder, that worked only on the yoke or via the normal disconnect buttons. As it is the most intuitive way to disconnect we actually managed to overpower the simulators hydraulics by kicking into the rudder.

[safetypee]

The modes indicated to the crew might not be quite the same as in the autopilot computation; this depends on system design and aircraft type.
The correct approach path is achieved by changing attitude to centre the GS beam; there may be many other damping factors, and in more modern systems the ability to smooth the approach with flight path angle or inertial inputs.

As altitude decreases the angular GS signal becomes more sensitive, and as a reflected beam it suffers hyperbolic distortion – it will be above the ideal glidepath. To counter these problems the autopilot pitch demand will fade-out the GS or use a mixture with an averaged attitude, FPA, or inertial path before reaching the flare altitude.

Autopilot flare modes are usually aircraft type specific, but as a generalisation it is the need to reduce descent rate as altitude reduces by increasing pitch (Ht rate–x ~ Ht, where –x ensures a positive touchdown, the aircraft aims for zero vertical speed just below the runway).
Some systems may vary the flare initiation altitude according to windspeed to help constrain the touchdown footprint.

[Rick777]

The A320 I used to fly tracked the localizer for rollout guidance. If you try to turn off the runway without disengaging the autopilot it will immediately try to correct to centerline as soon as you release the tiller. Don't ask me how I know.

[FE Hoppy]

EJets fade the GS out in 2 seconds from passing 100'. Flare mode engages at 50'.

[Uplinker]

@Rick777 you are correct. The autopilot and tiller commands are summed, so as soon as you release the tiller, the NWS will return to the LOC centre line, unless you have disconnected the A/P.

[JammedStab]

Originally Posted by NSEU

Rapid deflections of ground signals due to, say, other aircraft interfering with the signals, are recognised as such (and the aircraft doesn't go crazy trying to follow the signals).

True. I have seen the needle deflecting back and forth rapidly(not fully) at about 400 feet on an autoland due to a departing aircraft and the aircraft continued a steady flightpath.

[JetPilot500]

Thank you everyone, very helpful info!

[Specaircrew]

Originally Posted by Denti

The fail operational 737 can continue the autoland with a complete groundstation failure (loss of ILS/GLS signal) below 200ft (alert height). However, it might not be able to completely cover sudden changes in wind direction and therefore a go around is advised. We tried it in the simulator though and even a windshift of 25kts crosswind from one direction to the opposite, basically a 50kts wind shift in less than 50ft, did not lead to a displacement of more than 8m from the centreline.

Haven't tried it in the airbus yet, but a loss of G/S signal above 100ft or a loss of LOC signal above 15ft lead to a red autoland light which requires a go around.

Any chance that you could let me know where I can find a written reference for that Denti as I can't find it in the 737 FCOM? As you say the Airbus is quite specific in detailing its triggers for an autoland warning below alert height whereas Boeing seem to be a bit more vague. I'm an SME in the world of ATPL training and EASA have an exam question which asks what you'd do if the glideslope signal was lost at 80ft on a fail operational autoland approach. Clearly the A320 would have no problem as it wouldn't trigger an autoland warning but allegedly ( according to EASA) the correct answer is 'a warning will be triggered and a go around must be initiated'. I suspect tthat this is because the EASA learning objective states: 'Describe typical failures that, if occurring below the alert height, will trigger a warning: all autopilots disengage; loss of ILS signal or components thereof; excessive ILS deviations; radio-altimeter failure. Having already dismissed this as bo**ocks from an Airbus persepective I'd like to do the same for fail operational equipped 737's.