Automatic Landings

[Keith.Williams.]

A question for those operating aircraft with full autoland capability.

Do you use it on a routine basis, or simply when the conditions require it?

[Intruder]

Usually, the latter.

We have to exercise the system every 15 days in each airplane to keep the airplane current for Cat 3 landings, but we prefer to keep up our manual skills when possible.

[av8boy]

...or simply when the conditions require it? Like, when you're approaching the runway for landing, for instance?

OK. It was a joke. I don't think I got enough sleep last night...:(

Dave

[Keith.Williams.]

The logic of practicing manual landings is clear enough.

But is there not a similar argument for practicing autolands? Is it really wise to employ your least practiced procedures under what are arguably the most difficult conditions? Most pilots readily accept the requirement to practice manual flying, but very few appear to consider the need to practice automatic flight. Is it really that easy that regular practice is not required?

[Bealzebub]

When the conditions require it, but a minimum of 3 autolands within a 6 month period.

The misconception here Keith is that there really isn't that much to practice. Apart from the standard call outs and monitoring, an automatic approach and landing( by definition) is automatic and usually very routine. This is why the visibility limits are set low for these approaches.

The complications with such approaches and landings (or go arounds) come from various level of equipment failures, lighting failures, engine failures, electrical failures, wind and crosswind variations, etc etc. Of course these things can only be really practiced in the simulator. It is in the simulator that much of this is done. Indeed most airlines allow a credit of 2 autolands for each simulator proficiency check. As there is often a requirement to only do three in 6 months, this only leaves a need to do one in the aircraft in the same period. Many crews will probably average about one automatic landing a month either practice or actual but this is very flexible.

In summary, for ordinary line flying there isn't a great deal to practice. Hope this sheds a bit more light on the subject.

[van der vart]

auto land - completely automatic? the system will take cross wind, windshear into consideration? even the braking?

[bafanguy]

As one writer said, it's a good idea to practise autolands due to the different set of callouts that go with it. The autopilot, autothrottles do slightly different things also. Plus, at least in the USA, a certain number of autolands must be documented in the ship's log for the particular airplane to hold its CAT3 capability. But, we were reluctant to do as many as we should and now and then you'd see a note on the flight plan requesting an autoland to prevent the airplane from going out of currency, requiring MTC testing to put it back on status again. Well, pilots just like to make landings !! But we should be more attentive to the need for autolands.

[spork]

A couple of related questions: 1) If you do an autoland, at precisely what point do you take back manual control? and 2) Would you be fully aware of all that was going on with an autoland in order to efficiently take back manual control if something wasn't quite right?

[411A]

van der vart,


Yes, if the aircraft is so equipped.
The Lockheed TriStar did it best....with a
little help from the Trident folks.:)

[gatman]

We frequently get aircraft saying they will carry out a practice autoland. We, of course, reply no problem but no protection.

How does this affect the autoland as the ILS isn't protected to CAT 3 limits. Does it still count if you get deflections on final and have to take over or does it only count if it does the touchdown?

I've only experienced a proper CAT 3 from the jumpseat once.
Oh S*** would be a good expletive I think. Much respect.

[Keith.Williams.]

I think that bafanguy has probably put his finger on it. "Pilots like to do landings". Or more generally "straight and level is boring so let the autopilot do it, but landings are more fun so do them manually". How often do you manitain your straight/level and navigation skills by doing 3 or 4 hours of manual cruise flight?

This subsconcious attitude is probably a signifiacnt factor in the apparently natural asumption that it is sensible to do as few autolands as possible.

The "Oh S........" feeling described by gatman probably subsides with each autoland you do, but real confidence and competence are likely to come only with frequent practice. This is likely to be as true of autolands as it is of manual flight.

There is clearly a balance to be struck between practicing each type of landing. But the comments in this string suggest that little or no thought is being given to maintaining autoland familiarity. Is it really sensible to do autolands so infrequently that tech log reminders are needed to maintain aircraft CAT3 capability?

[Daysleeper]

KW
The thing about computers, is they dont actually need practice to do things consistently. So there is little point in doing more than needed to prove the system still works.

[Chimbu chuckles]

Where I work we are certified down to Cat3b, no decision...which means cloudbase 0'/75m RVR. The 75 meters RVR requirement only exists so you can then taxi the thing to the gate or they can come and find you in the event of an emergency.

I've yet to do one in anger but they are not hard...we only have one extra call on a full 'for real' autoland...the hard part being taxiing in such conditions.

Certainly the the lack of LVP procedures in force may create problems with other vehicles on the ground bending the localiser signal but from what I hear it rarely causes more than a little wing waggle...although 'departure stage right' has been recorded.

What does take considerable affort on the part of the PNF is monitoring and calling out various failures in a concise manner. Not a lot requires a missed approach...more usually a minor upwards adjustment of the minima, from say 0' to 50R...and the aircraft isn't going to chuck a huge wobbly even if the entire ground based ILS signal fails, it'll merely continue on as if nothing has happened based purely on the IRS derived ground track and ROD. As a result arms don't need to fly and panic does not need to rein...just a calm appraisal and decision.

If an engine fails below 1500R the rudders are mechanically locked, on the 767, and the three auto pilots work to control the rudder and keep the aircraft straight, all the pilot needs to do is bump up the thrust to control speed manually, the auto throttle being disconected asap following failure. That state will exist right through an assy go around until any mode is selected on the MCP that knocks out GA, like LNAV/VNAV/HDG SEL etc..at that point you better be bloody certain your feet are ready or the aircraft will roll over and kill you.

Lotsa fun!!

Chuck.

[spork]

The thing about computers, is they dont actually need practice to do things consistently.Ha ha ha ha ha ha ha ha ha ha ha ha ha ha ha ha ha ha ha ha ha... Oooh nooo - have wet myself... :uhoh:

[Keith.Williams.]

Thank you for that valuable contribution Spork.

Daysleeper, I think that if you read the thread you will find that it has never actually been suggested that computers need practice. But rather that the pilots may (or may not) need more practice in interacting with the computers.

If you think about it, no matter how often you use the autoland system you can never be sure that it will work the next time you try it. So the regulation requiring periodic use for the aircraft to retain autoland capability (quoted by Bafanguy) is rather curious.











;) ;)

[spork]

Pardon my levity, but my career is IT based. The day you knowledgeable and capable guys get wholly replaced by a computer is the same day I stop getting on planes.

I am sincerely interested in this thread, but no one has yet addressed my questions posted earlier.

[moo]

right then spork - as stated before it depends on what autoland capability your aircraft has. I work in Engineering on 747-400 and 777 aircraft and they have full autoland CAT3B capability i.e. they have no decision height and, provided the crosswind isn't above 10-15 knots (straight across) they will land right down to the ground with no assistance from the pilots. In normal operation, one autopilot controls the aircraft. During an autoland, all three autopilots engage automatically somewhere far out (probably on localiser or glideslope capture) as do the radio altimeters (below 2500ft I think) The flight crew are responsible for setting the autobrakes, flaps, autospoilers and speed settings they want for the approach (the autothrottle needs to be told) oh, and thrust reversers do not deploy automatically on landing either. Once this is done, the aircraft will do pretty much everything else - the flight director system will keep the flight crew fully informed of the approach - flare, rollout etc. and the glideslope and localiser 'needles' are shown on both primary flight displays and navigation displays. The radio altimeter call outs allow the pilots to know the aircraft altitude without having to look constantly at the altitude tape. At any point along the approach, the Captain can abort the approach by disconnecting the autopilot and by pressing two buttons forward of the throttles full power is applied automatically and a go around commenced.

Provided the flight plan entered into the Flight Management Computer has an active approach and runway selected for the destination airport, the ILS frequencies don't even need to be tuned, the aircraft will fly the approach, capture the localiser and glideslope and flare, land and rollout and stop on the runway - you have to taxi the damn thing!! :)

its all pretty amazing really :ok:

[spork]

Fascinating stuff. Thanks for that comprehensive answer. Although I can hazard a guess, some bits of jargon I don't understand: full autoland CAT3B, decision height, autobrakes, and rollout.

From the way you put it, I assume that some planes require a higher degree of input than others. One other area of interest for me: on landing there's always ferocious braking despite the runway being apparently more than adequate. Is that a crew decision, or is there no option, ie: brakes full-on or not on? I haven't seen thrust-reverser use for a long time, but I don't fly as much as I used to.

Despite having flown a lot, landing still spooks me to some degree. Knowing the physics of mass, momentum, etc just makes me wonder how it can be done so accurately every time. The only time I've ever really worried was into Heathrow quite a few years ago on a 737 which due to crosswinds was shifting around a lot when close to the ground. The closer we got the worse it seemed.

[flyer4life]

Interesting thread... I practiced a CAT3B approach only today in the sim!

In my company, we're qualified to carry out real 75m visibility autolands if we've done three actual or practice autolands in the last six months. "Practice" means using the autoland procedures but in good weather. Of those three, two can be done in the sim. In addition to the crew qualifications, the aircraft and the runway must also be certified and capable.

As for the brake question, on the A320 we have three settings for autobrake: low, medium and maximum. Our procedures are to use medium. It gives pretty firm braking but there's very little chance of the crew missing the fact that the autobrakes may have failed. Another reason for firm braking is to reduce runway occupancy times at busy airports. Also, some braking systems have better wear characteristics at the high temperatures caused by firm braking. But I wouldn't call it "ferocious"... that's reserved for maximum braking which is used for aborted take-offs.

As for thrust reversers, many airlines use idle reverse ie. reverser doors open but the engine isn't spooled up. It saves engine wear and reduces noise.

flyer.