I've been working on improving the flying characteristics of my flying plank design, especially as regards to the 'sinking' problem on landing if the pilot abruptly pulls back on the stick during the flare. I had a theory that I tried out and I thought I'd share the results.
The basic problem is that when you pull 'up' elevator on a flying plank, you are, in effect, 'putting the flaps up', as the elevator moves up for nose-up trim. The Cl in the elevator area immediately goes down, and you can actually sink a few feet before the increased positive moment pitches the airplane up and the increased AOA causes an overall lift increase on the wing. If you're close to the ground when that happens, you can touch down prematurely and a bit more firmly than you intended. One of the 'downsides' to the 'plank, I'm afraid.
One way (suggested by a couple of flying-wing references) to mitigate the problem is to shorten the span of the elevators and simply deflect them more. I don't like that solution because it really messes with the overall lift distribution on the wing at high CL values when you really want it to be as near-elliptical as possible (climb, especially).
I had a thought that if I decreased the chord of the elevator instead of shortening the span, there might be a 'sweet spot' where the reduction in lift was less than the change in moment. My theory went that since you were reducing the area of the surface but increasing the effective arm of the control surface, one might keep the control effectiveness of the elevator without reducing the lift quite so much.
Well, no, as it turns out. I did a series of runs in X-foil in viscous mode and found that, apparently, any elevator chord much smaller than 25% results in the control surface getting 'buried' in the separated flow on the back part of the wing and losing effectiveness at higher angles of attack. In fact, it could result in control reversal where you get reduced AOA as a result of pulling more aft stick. Blech.
Below is a graph of my results. The runs were done at a Mach number of 0.05, an angle-of-attack of 15 degrees, and a Reynolds number of 2,500,000. The flap was hinged at the upper surface, as anything else made the problem worse. The listed 'flap' (elevator) deflections are all in degrees 'up', rather than down. The airfoil is my own reflexed modification of an existing 'foil. It hasn't been optimized for my airplane yet, so the final Cm values may differ from what you see here. I did runs for elevator chords of 15%, 20%, and 25%. Notice what starts to happen at an elevator deflection of about 10 degrees.
Just to make sure this effect wasn't peculiar to my airfoil, I ran it on a couple of 'pre-existing' reflexed airfoils, and then on a couple of 'conventional' airfoils as a cross-check. The 'normal' airfoils showed less of the problem, but it was still there. The other reflexed airfoils had pretty much the same result as mine.
Equally interesting was that there really wasn't much of a difference in Cm between the various flap chord percentages in the range when they were all linear. I hadn't expected that.
Oh well, back to shortening the elevator span.
The basic problem is that when you pull 'up' elevator on a flying plank, you are, in effect, 'putting the flaps up', as the elevator moves up for nose-up trim. The Cl in the elevator area immediately goes down, and you can actually sink a few feet before the increased positive moment pitches the airplane up and the increased AOA causes an overall lift increase on the wing. If you're close to the ground when that happens, you can touch down prematurely and a bit more firmly than you intended. One of the 'downsides' to the 'plank, I'm afraid.
One way (suggested by a couple of flying-wing references) to mitigate the problem is to shorten the span of the elevators and simply deflect them more. I don't like that solution because it really messes with the overall lift distribution on the wing at high CL values when you really want it to be as near-elliptical as possible (climb, especially).
I had a thought that if I decreased the chord of the elevator instead of shortening the span, there might be a 'sweet spot' where the reduction in lift was less than the change in moment. My theory went that since you were reducing the area of the surface but increasing the effective arm of the control surface, one might keep the control effectiveness of the elevator without reducing the lift quite so much.
Well, no, as it turns out. I did a series of runs in X-foil in viscous mode and found that, apparently, any elevator chord much smaller than 25% results in the control surface getting 'buried' in the separated flow on the back part of the wing and losing effectiveness at higher angles of attack. In fact, it could result in control reversal where you get reduced AOA as a result of pulling more aft stick. Blech.
Below is a graph of my results. The runs were done at a Mach number of 0.05, an angle-of-attack of 15 degrees, and a Reynolds number of 2,500,000. The flap was hinged at the upper surface, as anything else made the problem worse. The listed 'flap' (elevator) deflections are all in degrees 'up', rather than down. The airfoil is my own reflexed modification of an existing 'foil. It hasn't been optimized for my airplane yet, so the final Cm values may differ from what you see here. I did runs for elevator chords of 15%, 20%, and 25%. Notice what starts to happen at an elevator deflection of about 10 degrees.
Just to make sure this effect wasn't peculiar to my airfoil, I ran it on a couple of 'pre-existing' reflexed airfoils, and then on a couple of 'conventional' airfoils as a cross-check. The 'normal' airfoils showed less of the problem, but it was still there. The other reflexed airfoils had pretty much the same result as mine.
Equally interesting was that there really wasn't much of a difference in Cm between the various flap chord percentages in the range when they were all linear. I hadn't expected that.
Oh well, back to shortening the elevator span.
Attachments
Last edited:
