I promised I would make a drawing to show the normal forces on the airplane with roll. Notice that just the gravity vector goes away. The tail determines the AOA of the wing so keeps working on roll. Luckily most airplanes pitch to the belly so the normal force acceleration toward the canopy is never noticed. If you have no rudder deflection or a really pure airplane with rudder deflection the airplane will go toward the canopy at 1g acceleration. To stop it you must dump the elevator trim to 0 degrees

Here the diagram really is, maybe.

I found an EXCELENT diagram of the forces... in a [:-]1938[:-] Model Airplane News magazine.

Now to get it scanned.... The article its in is dealing with how to build a model that MUST self-level, as its a free flight model. But the description and diagram of the forces involved is very complete.

Can't really ignore gravity in knife edge... its a force acting on the model... and compensating for it causes other forces to come into play (using rudder and trying to fly on the fuselage side area... and correcting for roll with ailerons... and....)

.
"pitch to the belly".. what causes this "normally"
In a steady knife-edge configuration, the wing is still lifting, right? Unless the alpha is changed (using the elevator/mix) to eliminate that.
KE on my Kadet generally requires much down elevator to get the flight path straight due to pitching to the canopy (semi-symmetrical wing). Mac's hand-eye coordination is much better than mine, and can do a KE all day long with his Kadet.
Pushing the nose down in a longitudinal roll is common with a semi-symmetrical wing.
And the only way to get a roll out of a normally dihedralled RE airplane, which goes "towards the canopy" when rolled.

Don't mean to ignore gravity, if you are trying to do knife edge it is a big force. But it is in the Y axis of the airplane, along the wing. Assume you control the roll to 90 degrees with no rudder input.

Paul, one time when I was at the Byron Expo I saw a J3 knife edged, I think it was flown by Don Lowe but won't swear to it in a court of law.

Say FHHuber, what is your first name? I can't wait to see the article. I love reading the old magazines. At Purdue in the stacks were bound volumes of Air Trails (I think that was the magazine). Anyway I would read them instead of studying. Grades showed that too!

Years ago, when the YF-22 was being tested, I read a Lockheed-Martin article describing a test pilot's experience when rolling the plane at extreme alpha and low airspeed. The pilots were tempted to exit the maneuver prematurely, thinking that the plane was not responding correctly. It looked to them like they were way off heading during the roll. (Remember, no one had rolled this kind of airplane at these airspeeds and angles of attack before.) The quote that stuck in my mind was "The airplane rolls around the velocity vector". I've created two brief QuickTime movies to illustrate the phenomenon.

In the first movie, the airplane is in normal cruise, and the velocity vector is in line with the fuselage, as we'd expect, and it's obvious that the plane "rolls around the velocity vector".
http://homepage.mac.com/mikejames/rc...alpha_roll.mov

In the second movie, the airplane is in level flight, but at a high angle of attack. Now, when the airplane rolls around the velocity vector, you'll notice that the nose deviates to nearly 90 degrees off heading, although this actually correct for the configuration, as Ben illustrated.
http://homepage.mac.com/mikejames/rc...alpha_roll.mov

Both these movies, and some others, are on my "Movies" page, at http://homepage.mac.com/mikejames/rc.../rcmovies.html

Unfortunately... the Apple download site for Quicktime to play on a Windows based PC has a major spyware affixed firmly to the player. (and the same spyware is affixed to all the mirror sites for Quicktime...)

Since I hate spyware of all forms, this makes Quicktime about useless for giving access to video files. Maybe AVI? That is a basic video type that I believe both Windows and MAC support without having to download a player.

I am impressed.

As your figure shows for normal rolls and low angles of attack with no elevator or rudder input the airplane initial roll acceleration is about the body axis but as the roll reaches steady state conditions (which is fairly quickly) the angle of attack is stabilized and the roll transfers to the velocity vector (of course with no elevator or rudder input the airplane comes down due to gravity)

When we do our rolls and input rudder and elevator we are forcing the airplane to roll about the body axis, which is hopefully at a low angle of attack to look really good.

A typical light aerobatic ship that will do high angle of attack (70 degrees or so) and hovering maneuvers can be seen to roll about the body axis in the maneuvers. In that case for a roll to the right that goes from 0 to 90 to 180 to 270 to 360 the angle of attack goes from 70 to 0 to -70 back to 0 and the sideslip angle goes from 0 to 70 to 0 to -70. Also the elevator is constantly moving to keep the airplane from transferring to a velocity vector roll.

In the second movie, pilot "disorientation" is no doubt created by the cockpit motion due to the displacement off the roll axis, giving him a different cue than he might expect because of the odd vertical g's and lateral swaying he no doubt experiences.
Probably loses a lot of airspeed also, so it might not be too popular if someone in the vicinity has evil intentions towards your longevity.
The off-axis nose motion is somewhat similar to the inertial-coupling that did in the short-finned F-100s, which pitched 90° to the line of flight and broke up when rolling. Extending the vertical on the F-100 fixed that.
And created the birth of ventral fins on other planes of that era.
I've watched contrails from the F-22 and others doing amazing feats of manuvering over the years at EAFB.
The prolonged " Harrier" for one... plane pitches up and stops any forward motion, then descends nose-high with no yaw for 1000's of feet. The engine contrail comes out of the exhaust pipe and then goes straight up as the plane sinks.
The purpose of this manuver isn't obvious, nor is there any observable benefit to losing all airspeed in a combat situation, but it sure interesting to see!

BTE, Jim, what did you use to do the simulations?
It could be interesting to do that with some of the questions that arise about our toys.