StarDestroyer.Net BBS

Tribble wrote:Appearing to be willfully ignorant. He has the Internet, and he is posting on this forum. He clearly has enough command of the English language to be posting coherent (if ridiculous) responses. Would it kill him to use Google and/or WIki once in a while before saying something? "How do planes fly?" or "How do planes take off / land?" or "What is the safety record of air travel?" are not hard research topics - in fact, the basics of flight were taught in primary school, at least where I live. He has been asked multiple times to do a little research once in awhile, yet he constantly refuses.

There's a lot of bad information out there on the internet as well as good data - would Archinist be able to distinguish between the two?

Sorry to wake this thread up again, but this Guardian piece appeared relevant, and interesting.

Crash: How computers are setting us up for disaster

It tells the story of the Air France Flight 447, which crashed in 2009. In it, as this writer tells it, the airplane DID more or less stand up on its tail, and then slide backwards into the ocean--at 125 mph.
It's mainly talking about what it's calling "The paradox of automation", where the better the automation is, the less skilled the human operators tend to be, so when something goes badly wrong and the automation dumps control back in the human's hands, the human is shocked, confused, desperately trying to work out what's going on, and woefully out of practice.

In this crash (although I'd suggest reading the linked piece, the writer puts it better that my summarising here), an iced-over sensor meant the autopilot had to switch off, so the pilot had to take over. However, the same iced-over sensor meant the Fly-By-Wire had to switch off too, giving the pilot control without the computer automatically smoothing out and re-interpreting the pilot's control movements to what was actually safe. The pilot was jerky, and then he made the mistake of pulling up. The plane climbed steeply, warning the pilot constantly of an imminent stall, while the pilot kept pulling up (a theory put forward is the crew interpreted the warning as the Fly-By-Wire announcing it was correcting the problem-that, instinctively, they didn't believe it was possible to stall the plane due to the automation). The plane climbed high, losing airspeed, until finally it no longer had enough and just fell back down tail-first.

So I hope that's put your mind at ease, Archinist.

I must take exception to what that writer wrote. I have read the actual accident report, both in the original French and the English translation and that is NOT what happened!

The airplane did NOT "stand up on its tail". It did enter a deeply stalled condition, but "deep stall" doesn't mean "stand up on the tail". Stalls, as I have said, occur between 17 and 22 degrees pitch up from direction of travel (number range given because the exact stall angle varies from one design to another and I don't know the exact figure for the Airbus).

The airplane did NOT "tailslide" into the water. It was more of a "falling leaf" stall, which, well, it resembles a falling leaf fluttering to the ground. When the plane hit the water the fuselage was almost level in relation to the ocean surface as near as anyone can tell.

Not one but TWO airspeed sensors had iced over - if only one was out of commission the airplane wouldn't have turned things over to the humans because 2 of the 3 would be working. The end result is that no one on board the airplane, not the humans and not the computer, had any way to know for sure what the plane's airspeed was.

At night, in a storm, that sort of instrument failure is almost always fatal. The computer's not working. The humans don't have enough information. They will most likely lose control, as happened here, and crash. The only way to survive that situation is to somehow figure out what you're airspeed is and/or configure the airplane for a particular airspeed and hope to god things haven't gone too far.

The fly-by-wire in an Airbus NEVER turns off - if it did, you wouldn't be able to control the airplane because there is NO direct connection between the controls in the cockpit and the flight control surfaces. If the plane loses all power then what is essentially a small windmill deploys to provide power for the controls, as happened in the two prior loss-of-power episodes I mentioned involving Airbuses.

In a prior thread I cover the Air France Flight 447 crash in some detail so if you're interested please read it.

There are actual issues with increasing automation and human pilots, but Air France is not an example of that. The linked article in the above post is very misleading. I'd include more detail here, but I have to get ready for work. Maybe when I get back.

Fly-by-wire is pretty much mandatory for planes above a certain size anyway. They're simply large enough that trying to use a manually linked control system would be unmangeable for anybody of normal strength, and the systems are reliable enough that it's unusual when a plane *is* built with manual controls; the A-10 Thunderbolt, for example, has one such, which has actually saved its bacon a time or two since its role takes it close enough to ground combat that there's a decent chance of electronics and such getting knocked out.

B-52 has manual flight controls and it's still one of the heaviest planes ever. The reason is they are power assisted, this began began in the early 1930s era. The A-10 has this, near everything since the era of MiG-17 has had this for jets. The issue with the A-10 was the hydraulic lines being shot out would become probable with the level of AA fire it would otherwise withstand, as well as the requirement for low cost in an era when computers were seriously expensive. A-10 program did begin around 1965, as a propeller driven aircraft, after all. Cables being smaller, and able to be nicked and not break, simply survive better. But milspec data cables also withstand that pretty well.

The odds of minimal triple redundant fly by wire systems being directly shot out of action are pretty damn low, you are all but certain to loose all hydraulics or engines first. The best solution for a major aircraft is really to go to all electric flight controls, where your power lines can also serve as added data cables, ensuring only one conductor has to survive to a given surface (the ground can be the air frame) but this really only became feasible in the 1990s.

Meanwhile though fly by wire has now highly evolved and can make a plane flyable in situations in which no human pilot could fly the plane, simply because the human pilot cannot independently control paired surfaces in a normal aircraft. Computer has no problem doing that given suitable programming. This has led to among other things massively increased (as much as tripled) angles of attack in the F/A-18E and F-35 compared to legacy planes like the F-16, as a direct practical measure. This also makes them able to survive some pretty extreme damage or loss of surfaces, or say, a control surface being jammed. Its no small part of why swing wing aircraft are gone.

By the way - THIS is an example of the sort of problems we're starting to see with pilots no longer used to flying an airplane (or possibly not adequately trained) relying far too much on automation.

Again - Flight 447 is not an example of that, despite what both the linked article and the Vanity Fair article referenced in that article claims. Yes, the pilots on Flight 447 could have rectified the situation IF THEY HAD HAD ACCURATE INFORMATION. At night, over the ocean, in a storm, in clouds, with failing instruments they did not and could not have solid information on their situation. I don't know about the The Guardian's Tim Harford but William Langewiesche should have known better. Of the two articles, the Vanity Fair one is better but it's also an example of Monday-morning quarterbacking from the safety of an armchair. Yes, there were problematic elements in Flight 447 but the problem of instrument failure in IFR conditions is a tough one that can kill even really good pilots. I'm not convinced the pilots were the key failure here, although there were definite issues with the pilots on that flight. He could have picked a better examples in my opinion.

On a slightly related note:

Broomstick, what's a typical vertical speed for making touchdown? Does it vary depending on aircraft size?

In my limited messing around with flight sims, it seems like the general procedure for landing is as simple as reduce throttle and deploy flaps, pitch up a bit to stay airborne, and fly down at a gentle angle until you hit the ground . . . err, that is, until you touch down. With something big and clumsy like an airliner, I've seen that it's non-trivial to line up on the runway and to make sure that you touch down after the runway starts but while you've still got room to stop.

Of course, I know that general market flight sims are games, not real experience or even real training, and I wouldn't presume to think that I can fly because I have a few hours behind the stick in X-Plane.

Zeropoint wrote:Broomstick, what's a typical vertical speed for making touchdown? Does it vary depending on aircraft size?

The "standard rate of descent", which is just what it says, the usual rate of descent on landing, is about 5-6 vertical miles per hour, or 8-9 kph. That is held up as a standard for all aircraft. Most of them can handle significantly more vertical speed than that if necessary, and helicopters, well, they can really slow down the vertical descent if they want to do so.

In my limited messing around with flight sims, it seems like the general procedure for landing is as simple as reduce throttle and deploy flaps, pitch up a bit to stay airborne, and fly down at a gentle angle until you hit the ground . . . err, that is, until you touch down.

Yes, that's basically it. The devil, of course, is in the details. Usual angle of approach is about 3 degrees although, again, most aircraft can handle significantly more than that if needed or desired. Practice is required to really do it smoothly.

I'm not aware of any airport with scheduled passenger service using a posted/recommended descent angle greater than 4-5 degrees (usually to either avoid obstacles or get into odd spaces, such as in mountainous areas). Anything greater tends to scare the passengers.

The emergency power-off descent in the Piper Arrow, a single-engine retractable plane I was training in briefly, has something like a 2:1 glide ratio without engine power. That means if your engine fails you're coming in for a landing at a 45 degree angle. That can cause pilots to freak out the first time they see it from the cockpit. It is the correct procedure, though, and you can certainly land them using that approach angle. Don't particularly like the view on that one, personally, although I did them just fine. A lot of ultralights have a 30 degree (or more) power-off approach angle, again, even a lot of pilots are uncomfortable with that one but you have to come in that steep to keep up adequate airspeed so your wings don't stall.

Carrier landings are another exception to the general rules, but that's a very specialized segment of aviation.

With something big and clumsy like an airliner, I've seen that it's non-trivial to line up on the runway and to make sure that you touch down after the runway starts but while you've still got room to stop.

The issue there is inertia - airliners are massive constructs and really do illustrate the Newtonian concept of tending to remain headed in one direction until forced to change.

Of course, I know that general market flight sims are games, not real experience or even real training, and I wouldn't presume to think that I can fly because I have a few hours behind the stick in X-Plane.

Some "general market flight sims" are actually pretty good - Microsoft's, for example, has long been used to supplement formal training and under very specific and controlled circumstances the FAA will even allow it to be used and logged in training. The key is for the program to use real physics and traits - which a lot of gamers find either boring or frustrating as hell because real life isn't a video game. If you can get a set up using a stick and rudder pedals even better. I know quite a few pilots who have used flight sims like that to rehearse a flight prior to actually making it. Most non-pilots don't, because they discover that even in a small, primitive airplane long segments of the trip can actually be boring. If you approach their use seriously and carefully you actually can learn quite a bit from them. Or, if you're careless or take short cuts, develop some really bad habits.

Broomstick wrote:The emergency power-off descent in the Piper Arrow, a single-engine retractable plane I was training in briefly, has something like a 2:1 glide ratio without engine power. That means if your engine fails you're coming in for a landing at a 45 degree angle.

2:1 glide is a 22 degree descent angle. The glide ratio on light aircraft is usually about 10:1, so descending that steeply would purely be to build up airspeed for a flare. In fact given that the recommended engine-out landing speed on a Piper Arrow is 79 MPH, which is about the same as the optimal glide speed. While there doesn't seem to be an explicit recommended approach angle for emergency landings, it seems very unlikely that you'd do a such steep descent; you'd likely overspeed, even with the gear and flaps down, and have a nasty crash if you screwed up the flare.

45 degrees (1:1) would pretty much be a splat for any aircraft; the only aircraft with glide ratios that bad are spacecraft during re-entry.

You are correct - but it sure seems alarming.

And yes, the recommended engine out speed is going to be the best glide speed - think about it, you want the maximum efficiency when you no longer have engine power.

The is no recommended glide angle for engine-out approaches because you don't want to concentrate on the glide angle, you need to pay attention to your airspeed and you adjust pitch to get that correct. Whatever pitch angle is required to hit that speed is what you use.

Some ultralights wind up with ridiculous engine-out approach angles because they're so full of drag.