Hacker News Reader

Comment by fsloth | Hacker News Reader

There's nothing wrong with unstable airframes. Military craft have required computer assisted flight for decades.

What's problematic with 737 MAX is that the stability assist was obviously installed as an after thought, and even worse, it's dependent on instrumentation that does not have several layers of fault tolerance as would be the case if it was actually built with the care that planes actually require to be as dependable as they are.

So the instability is not a problem. That boeing tried to get around it with a cheap patch of jury rigged software that is depending on input from non-fault-safe instrumentation, is.

View on HN

Inherent instability in a commercial aircraft seems like a very bad idea. Safety has to the top priority. The military has very different priorities.

And would that affect its capacity to glide?

Not in this case, since the instability is caused by the lift generated by the engines.

IIRC, the instability is caused by the engine nacelles (the metal shroud around the turbofan), and therefore is present whether the engines are active or not.

Sure, I'm not saying it's a good idea, just that I don't see any inherent problems with it - given that fly-by-wire systems have been used for decades I would consider them well understood.

As an unrelated side note, I have always said “jerry rigged” and wanted to know which of us have been saying it wrong. Turns out it’s both! https://www.merriam-webster.com/words-at-play/jerry-built-vs...

That's a huge stretch. The fix was to work around a certification requirement that the control forces shouldn't decrease as the aircraft approaches the stall. This is a long way away from the aircraft being inherently unstable.

"Fehrm said Boeing must have added this system on the MAX because when the angle of attack is high this model is less stable compared to prior 737 variants. That’s because the MAX has bigger, heavier engines that are also cantilevered further forward on the wing to provide more ground clearance. That changes the center of gravity."

Fehrm, a former jet-fighter pilot and an aeronautical engineer: https://www.seattletimes.com/business/boeing-aerospace/faa-e...

It seems like you think that contradicts what I said. The system by design adds a little trim in very specific circumstances. That's not the same as being needed because the aircraft is aerodynamically unstable. That is a term which has an actual definition.

That said there are airliners out there which are unstable in the longitudinal axis and require a yaw damper to prevent Dutch roll.

Thanks, you are quite right.

My intention was not to claim 737 MAX had the same aerodynamic nature as, say, F-16, but in the scope of the discussion point out that fixing instabilities with flight software can be made to work reliably.

The main argument was that it was inherently a bad idea to require computer assist to maintain stability and decades of military planes demonstrate that you can engineers such systems to be more or less fault safe.

Agreed. I think unfortunately most of the people outraged by this decision are completely unaware of the complexity of systems on board a modern airliner. There are other systems like the Mach trim system which adjust the trim in response to Mach speed. There are yaw dampers on aircraft with good reputations which are critical for safety of flight (e.g the concensus is the 757 will depart controlled flight without an operative yaw damper). Even on the 737 flying with an in-op yaw damper has been known to make light chop vomit inducing.

There was a sibling comment somewhere in this post that stated that airbus doesn't introduce a layer between the pilot and the aircraft. This couldn't be further from the truth (assuming normal control laws). Applying full back stick on an airbus will hold the aircraft right on the edge of a stall. Applying full left or right on the stick will neatly roll the aircraft to 60 degrees of bank. Releasing the stick at this point will roll the aircraft back to a perfect 30 banked turn in the same direction. This isn't normal behaviour by any stretch.

If anything I think Boeing are suffering from not going to fly by wire control earlier. That would have made this whole subsystem irrelevant. You'd just apply the desired behaviour in the response curves in the control software.

A combat aircraft can trade reliability for superiority, you increase one kind of risk to lower another. This trade-off simply does not exist in civilian aviation.

is that the case? The Osprey certainly tried and failed that trade off, at least in the beginning.

https://en.wikipedia.org/wiki/Accidents_and_incidents_involv...

> There's nothing wrong with unstable airframes.

For military purposes, I agree. There is an end objective to fighter jet development. To own the sky at all costs.

With civilian aircraft, there is too much to lose. You can't have an unstable air frame for civilian commercial mass purposes.

Helicopters are inherently unstable at low speeds, yet there are civilian models.

And if they carried 500 people we’d think that was a bad idea.

Helicopters scare the hell out of me because they resemble nothing that's natural.

I guess you've never seen a maple seed then.

https://duckduckgo.com/?q=maple+seed&t=ffab&iax=images&ia=im...

Maple tree seeds?

(to other repliers)

maple tree seeds resemble auto-rotation in a helicoptor, not powered flight.

It's a nice mental model as to why you won't fall out of the sky necessarily with a loss of power in a heli, though.

Why does that matter? There are plenty of natural things that are scary as hell, and plenty of unnatural things that aren't.

The reason you don't see it in nature is that macroscopic organisms never evolved a continuously rotating joint. It's a hard problem for evolution to solve; how would blood vessels cross that boundary?

But of course machines have no such constraints, and the wheel and axle is hardly a new (or scary) invention.

Your shoulder comes close, on a ball-and-socket model rather than wheel-and-axle.

It doesn't continuously rotate, and it's nowhere close to allowing so.

It will allow you to rotate your arm in the same direction indefinitely.

It does not. Once you go through one rotation you've ripped all the muscles and ligaments off, ripped all your skin, and can no longer rotate it any further. Oh, and your entire arm will die in short order because you've ripped all the blood vessels delivering it oxygen.

Is it possible we're talking about two different things? Making a circle with your arm isn't the same as rotating it, and just making a circle isn't particularly useful for flight or locomotion. Imagine attaching a wing to your arm and making a circle -- half the time isn't going to be attacking the wind the wrong way. This wouldn't work to, e.g., fly a helicopter. You need real rotation.

ATP synthesase and the bacterial flagellum are the two counterexamples both molecular in scale.

My home is filled with things that resemble absolutely nothing natural. It doesn’t scare me in the slightest.

What? Of course there is somethig wrong with unstable airframes. This is not a figther jet where it may make sense to maximize maneuverability, it is a bloody airliner, where stability and reliability should have been the design goals.

I disagree that "instability is not a problem". In military aircraft that are unstable it's common for everyone in the plane to be in an ejector seat or to have a parachute. Can you provide statistics on how many (or what proportion of) military planes with inherent instabilities routinely carry passengers without parachutes?

I suspect military planes have ejection seats not because of built-in instability, but because they have a tendency to get shot at.

It doesn't matter why they have them, the fact that they are there changes the overall risk equation for the aircraft. Many military planes have bailout procedures because everyone has ejection seats or parachutes. Civilian planes universally do not. What is an acceptable risk in one circumstance is not necessarily acceptable in another.

It does matter why they have them, if you're using their having them as evidence that something else is a problem.

Swap out "instability" for "drab paint jobs" and see:

> I disagree that "drab paint jobs are not a problem". In military aircraft that have drab paint jobs it's common for everyone in the plane to be in an ejector seat or to have a parachute. Can you provide statistics on how many (or what proportion of) military planes with inherent drab paint jobs routinely carry passengers without parachutes?

Yes we can make an argument incoherent by swapping words in and out.

In this case the argument is not that a is evidence of b, it is that a ameliorates b. And that the risk of b is therefore priced differently where a is also present. So where a is not present, we must price the risk differently.

How that assertion relates to the historical record may be of interest.

Risk is assessed on a case by case basis. When people ask "is the risk of this system acceptable given its mission role" the answer depends on all of the characteristics of the system. In the case of military aircraft that means that both the fact that the airframe may be dynamical unstable in flight and that every crew member has an ejection seat are taken into account. And whether or not the plane is expected to be used in a combat environment will play a huge role in the acceptable levels of risk as well.

To blithely say "oh well, military planes do X so X must be fine for civilian aviation as well" is either intellectually dishonest or fundamentally naive.

It strikes me as unlikely that the military is taking the attitude "it's OK if our planes fall out of the sky during normal operation, because we have ejection seats".

Therefore, as with civilian planes there is usually no "plan B" (exception https://en.wikipedia.org/wiki/Cirrus_Airframe_Parachute_Syst... ), acceptance criteria for unstable civilian aircrafts must be more strict than for military aircrafts.

E.g. higher redundancy of HW & SW, stricter testing/checks/validation of flight envelopes, more feedback to the pilots about why a computer is performing certain actions, etc... than military aircrafts using the same systems.

> Many military planes have bailout procedures because everyone has ejection seats or parachutes.

Do any modern military non-fighter jets have bail out capabilities? It was my understanding that you couldn't even put on a parachute in flight to jump out. You either had an ejection seat, or no capability at all.

The B-1, B-2, and B-52 bombers all have ejection seats. Some other combat aircraft, like the AC-130, have bailout procedures and the crew usually all has parachutes. Almost all other planes are either fighter jets with ejection seats or transport aircraft which are substantially similar if not identical to civilian aircraft in terms of dynamical stability.

I meant using computer assist to maintaim stability in vehicles as such is well known and widely used feature. The problem is not the computer assist, the problem is that it was obviously implemented not rigorously enough and depends input from non-fault-tolerant sensors (since making it fault tolerant apparently means Boeing would have needed to do some expensive qualification work due to the changes to the sensors).

You can eject from most of those mil aircraft

Eject gonna cost extra with Ryanair

Genuine Ryanair advert: https://youtu.be/Idd32nyf1pc

Reply on news.ycombinator.com