Do pilots really fly for a length of time not knowing they are inverted?

No, he isn’t.

I see nothing wrong with the statement “in a properly flown banked turn, the direction you feel as ‘down’ is directly towards the bottom of the airplane”. My objection was to the idea that this also must be true when the airplane is flying straight and level.

I agree that “straight and level” is commonly taken to imply “upright and unbanked”. But inverted flight can also be both straight and level.

In short, saying “This isn’t really right” does not require that every part of the statement in question be shown to be incorrect - just some part of it.

Actually the human brain does exactly the same thing but at a much shorter time scale. The human inner ear has “gyros” (the semi-circular canals) which DO provide a short-term gravity signal and a “down sensor” (the otoliths)*.

The issue is that the brain treats the gyro “very short term signal” and resets it roughly every second (to be accurate the “gyro signal” converges exponentially towards the “otolith’s down signal” with a time constant of 0.5 to 2s depending of studies). It does so because the inner ear gyros are much less accurate than aircraft gyros (well they’re lightweight and not really high-tech, right?), so the gyro signal can’t be relied upon for very long. Furthermore evolution has found that this worked very well in ecological situations where accelerations are generally brief (unlike in aircrafts) which means that the “down signal” is quite reliable.

And so humans get disoriented by a few seconds of linear acceleration because their “gravity perception” system is not tuned to the time scale of aircraft motion.

*yes, semi-circular canals rotation signals are integrated into a gravity signal at high frequencies. The matter has been somewhat disputed in the 80’s and 90’s but definitely demonstrated in the last decade.

IMO …

I’ve always understood the semicircular canals as integrating accelerometers. As such, from an unaccelerated reference frame they can be used to compensate a local “down” signal for brief perturbations from true “down” caused by local accelerations.

The problem is they saturate quickly and so as you say they’re washed out over a second-ish.

The net effect is that given a long-term stable “down” signal that’s always close to true down, they’re good at measuring transients to indirectly derive an accurate orientation vector vs true down. Which works great for the design case of a terrestrial animal that runs lots & jumps some.

There really is no human system which provides a stable platform ignoring any and all accelerations.

Said another way, gyros are long term stable in global down with long term drift compensators to maintain global down against local perturbations whereas humans are short term stable in local down with short-term compensators to derive the global/local delta and hence global down. Briefly.

All IMO and I hope to learn more.

The canals measure transient angular velocity which is used to track the “true down” in head coordinates (when the head pitches or rolls or rotates relative to gravity in general) by velocity to position integration of rotation signals.

During transient accelerations in the absence of head rotation, there is a transient perturbation of the “local” down (measured by the otoliths). But the brain’s “true down” signal, which depends on the canals (at high frequencies) doesn’t change. The brain compares both and correctly interpret the perturbation of the “local” down as a linear acceleration.

So the canals/otolith system extracts both the “true down” and the linear accelerations. This works only at high frequencies (since, indeed, the canals and the velocity to position integration are inaccurate for anything else than transient rotations).

Ah yes. Mechanically, the gyros are really stable in global down. The canals/otolith system are different; mechanically they are not stable at all relative to global down (they just rotate with the head). It is the brain that builds a global down signal which is (short term) stable.

In the end, the two systems are analogous but the underlying mechanics are completely different.

Thanks. It’s cool to dig a couple more layers deep into how this stuff works.

I don’t see why anyone would think this. Inverted flight is 180 degrees off from being level. There may be cases where a symmetrical object can be considered level in multiple orientations, but an airplane isn’t one of them.

I think level here means “neither ascending nor descending”.

I always understood “level” in “straight and level flight” to include “wings level” or “laterally level” (as well as constant altitude), but I’ll grant that there may be a bit of ambiguity here.

Thank you all (especially LSL) for your explanations of the artificial horizon!

OP here. * spits out mud, toads, pieces of bog, silt *

Looking at this, now: Sensory illusions in aviation - Wikipedia

I really wish I was in a plane with these things being explained. I understand, I think, “the leans,” (understand as in some sort of real-world acquaintance). The money quote about “up,” is here, from Wiki “Graveyard spiral”:The pilot mistakenly believes he or she is flying with the wings level, but with a descent indicated on the altimeter and vertical speed indicator. This usually leads to the pilot “pulling up” or attempting to climb by pulling back on the control yoke. In a banking turn, however, the plane is at an angle and will be describing a large circle in the sky. Pulling back on the control yoke has the effect of tightening that circle and causing the plane to lose altitude at an increasing rate, like water swirling in a drain or funnel. An increasing component of the lift being generated by the wings is directed sideways by the bank angle, not only pushing the plane “upward” into the turn, but reducing the amount of lift which is holding the plane up. At that point the aircraft is describing a descending circle or spiral, with a flight path that again resembles being in a funnel. In the ever-tightening, descending spiral the aircraft eventually exits the base of the clouds and/or impacts the ground.[1]

But I’m having trouble “feeling” it, I think for the following reason, which I put here because I don’t think I’m that different than any other confused non flyer: I visualize “turning left” in the sky like a kid vrooming his toy car in one direction, then, you know, turns it. (Yaw only, a handy diagram now tells me.) Whereas, for some reason, I know that “going up” is not translation only along one axis, and involves a roll.

Given that–and here I actually tried, as Mrs. Bloom watched, to be a plane in a slow roll-angle, and hold that position while trying/thinking about pitching up, while letting the physics process in my brain. Surprisingly (heh) I didn’t “feel it,” intellectually, the “up” causing the down. Then tried to do it, simpler but less impressive for my wife, with palm and outstretched thumb. Still no joy, which annoys me, because I’ve been on a bicycle and you’d think one more dimension and some pressure constraints wouldn’t be that hard to add.

I’ll go hit some YouTube, see what comes up.

When you ride a bike and go around a corner you lean the bike over. Surely you’ve seen and done that.

How come the bike doesn’t fall over the rest of the way? It surely would if it was standing still.

The reason it doesn’t is simple. The centrifugal/centripetal force from the turn exactly offsets the effect of gravity on the leaning bike and rider. If the rider poured water out of a pitcher in the turn the water would fall straight along the plane of the bike and land by where the tires meet the road. It would *not *fall straight down and land directly beneath the pitcher.

When you are riding around the corner you have no sensation of trying to fall over to the inside of the turn. Which sensation you’d surely have if you just sat on the stationary bike and somebody leaned it over on it’s side.

The airplane is no different from that when just considering two horizontal dimensions. In a smooth turn “down” feels like it’s straight towards the floor. Just like it feels when you’re going straight. IOW, you can’t tell the difference between the two situations; they feel the same because in terms of vector forces they are the same in direction though slightly different in magnitude.

The key thing that makes an airplane different from a bike is the freedom in the third dimension. Unlike a bike, as soon as it’s turning, the airplane begins to descend and pick up speed absent any control inputs to the contrary.

And even if the pilot is lousy at flying instruments he’ll probably see the altimeter start to unwind. At a bare minimum even with no instruments at all he’ll soon hear the increased wind noise from increasing speed.

Not being able to tell a turn from straight he’ll assume he’s going straight. And the cure for an unwanted descent is to pull back to reduce the descent angle or maybe even to climb a bit.

Which just makes the problem worse. Very quickly he’s confused, panicked, and shortly thereafter committed.

What a perfect description. Thank you.

Hmm… I thought in making a turn, a pilot banks and pulls back to get the turn going and uses a bit of counter rudder to keep the nose up?

Speaking of what ‘down’ means, does anyone remember a GQ thread from loooong ago that discussed it? ISTR good arguments for down being directly towards the center of the earth and for it being towards the relative ground. Or something like that, if I remembered I could find the thread. Anyone?

No, banking gets the lift vector angled to the inside of the turn so you’ve got the lift vector now with a lateral component. In doing that it loses some vertical component which would cause the aircraft to descend so you pull back on the controls a bit to increase the lift vector and restore the vertical component thus preventing a descent, this also introduces a pitching moment which helps with the turn. The rudder’s use depends a lot on the aircraft but in general it can be said to be used to keep the aircraft flying “true”, that is to keep the tail directly following the nose, it is not normally used to hold the nose of the aircraft in any particular spot unless you are flying aerobatics.

In large aircraft the rudder is incredibly powerful and its manual use is pretty much restricted to keeping straight during landing and taking off and compensating for asymmetric thrust during an engine failure. Rudder input during turns is handled by the yaw damper.

You know, I think I understand, I know I do, I think (:)) about my physiology and the physics of the graveyard spiral. It’s back to the physics – mentioned in practically every post by pilots, here, for example:

It’s the physics of that part that I don’t intuit. Remember the kid turning the toy cars left and right…? I don’t understand why the lift lessens as a matter of aviation reality.
ETA: re-reading Richard immediately above, I visualized the vector addition of a lateral component, so it loses a vertical, aaand…I lost it. I somehow am comfortable with vector graphics, if that helps get me a step closer.

Consider if you banked the wings 90º, the lift vector is now entirely horizontal, so no vertical lift component at all which means nothing holding you up. If you have the wings level the lift vector is entirely vertical so all of it is working to hold you up. In between level and 90º bank the lift vector has some vertical component and some horizontal component, but the vertical component is always less than when the wings were level so you need to increase the lift vector somehow to restore that vertical component.

Fundamental concept: The vertical component of the lift vector must equal the weight of the aircraft to maintain level flight.

OK, to put the nail in the coffin of how airplanes do not fly: wings level, you want to turn left: turn the rudder right, apply force to that side (how? search me–you catch a jet stream at right angle to the rudder, maybe, or a light a little rocket engine) and you can turn left without banking, and…without descending.

And that’s not how an airplane flys. What does “pull back on the controls” mean?

The lift vector is perpendicular to the wings. Suppose you’re flying along, straight and level, and your wings are generating 3,000 pounds of lift (and your plane must weigh 3,000 pounds). If you now start banking, raising one wing and lowering the other, you still have 3,000 pounds of lift but it’s not vertical anymore. If your bank angle is 30°, the lift vector has a vertical component of 3,000 * cos 30°, and a horizontal component of 3,000 * sin 30°. The horizontal component makes the plane turn. The vertical component, however, is less than the weight of the plane so you’ll start to lose altitude. To fly a level turn, you pull back on the stick and add power. You want to increase the amount of lift so that the vertical component will exactly offset gravity. According to my figuring 3464 * cos 30° = 3,000. When you’ve turned as much as you want to, you roll the wings back to level, release the back pressure on the stick, and ease off on the throttle.

It’s just that easy.

Does anybody else want to tackle adverse yaw?

Same thing in a car going around a curve – that’s why tight curves like freeway turn ramps are banked. Also race tracks – more banked, because they are going at a much faster speed.

<sidenote> I once asked a highway engineer about the angle of the banking: If we’re taking a ramp from one freeway at 65 mi/hr, merging onto another freeway also with traffic at 65 mi/hr, why is the ramp curve banked for 35 mi/hr, forcing traffic to first slow down, and then speed back up again?

His reply was: This is Minnesota, stupid: if it’s a January blizzard with glare ice on the road, and you’re creeping along at 10 mi/hr, if we built it banked for 65 mi/hr, you’ld slide right down to the bottom and into the ditch. Plus there are lots of vehicles (like school buses & loaded semi trucks that don’t travel at the speed limit.