Rfd-richard's Final Design

[Richard Miller]

Dave and Rotaryphile, principally;

You two impress me as mature, sensible, and informed individuals, which is kind of rare, as you well know, so I would like to invite you to look over my shoulder and comment on Richard's Final Glider Design. Others are of course wellcome.

Wing: Span = 33.3-ft. Area = 100 sq.ft. Taper ratio about 3:2 and 4 or so degrees of sweepforward. A little dihedral in all likelihood.

Stabilizer: A lifting surface of 40-sq.ft. AR=4. Marked sweep; almost fantail. Mounted only a foot or two behind the wing leading edge, in the downwash.

TLC = Tandem Lift Centers. Lateral maneuver by tilting the aft surface. Pitch maneuver by weight shift.

The only thing new about this is my realization that using a lifting stabilizer to turn a full-scale design is neither the cumbersome nor impractical idea I imagined. Once I had my head straight on that all the rest was clear.

Here, as I see it, are the advantages of this configuration:

1) All the flying surface is lifting surface.
2) The wing tends to be flat to the ground in [skidding] turns.
3) The total volume of two surfaces, and hence the mass, is less than that of a single surface. Ceteris paribus.
4) There must be some advantage to having the stabilizer in the wing downwash.

That'll do for a start.

And yes, it is the bird configuration. You struggle with glider design your whole life and then end with the simplest, most efficient (?) and obvious solution possible.

Comments and critique please.

-Richard

[Dave Robelen]

Hello Richard,
Boy, that is the ultimate "slowflyer". In a serious vein, I would not be all that comfortable combining weight shift for pitch control with a tilting stab for lateral control. It sounds like the effectiveness of the lateral control will be highly dependent on the specific CG during a manuever. I suppose my concern is based on the idea that you could lose the effective lateral control in the cases where you shift the CG forward to drop the nose and pick up speed. Or in other words, the amount of lateral control would be closely linked with the airspeed.
Back in the "good old days" when I was employed at NASA we used dynamically scaled models to check issues like this with new configurations. My reccomendation would be to build a scale model and get some "stick time" with your feet on the ground while you finalize the design and look for any dangerous tendencies.
My $.02 worth, Dave Robelen

[Richard Miller]

Dave,

I should have made it clear that this - AVP/TLC - is not your common iron-plate glider. Not meant to climb high, go far, go fast. I am intent on just hanging out at low altitudes, like the birds do.
Thus no great demands on speed range n' like that.

I am, as I speak - well, I was doing it a little while ago - designing the dynamically scaled model and I would never dream of going further without getting a sufficient amount of stick time on that, and having others share it to eliminate bias.

Dynamically scaled means, in this case, something on the order of 68-72" span, wing loading to be figured out. I'd like this one to be close enough to, say, a Gentle Lady to be able to make some meaningful comparisons of servo loads and turning response, stuff like that.

I DO want and DO appreciate you two-cents worth. I did not solicit it just to provide extra work for you.:)

More as there's more to write about.

-Richard

[Dave Robelen]

Hello Richard,
I understood you were asking for a meaningful opinion. There is a second concern when you design for a large tail that is up loaded. As a rule this setup will lead to a small decalage which could make dive recovery rough even with a rearward CG shift. What I am basing my concerns on is not the smooth, trimmed flight under good conditions, but the effects of a gust disturbance that could change the attitude of a lightly loaded machine abruptly.
What about the possibility of the same configuration, but with the tail assembly gimbal mounted in the yaw/pitch axis? Some degree of aerodynamic control could offer a substantial margin of safety.
Some of these design issues could be addressed with the classic small all balsa glider before moving to a full dynamic scaled model. I am more than a bit envious considering that I have wanted to tackle a project in full scale myself for a bit now. I wish you the best and will be glad to continue the dialogue.
Regars, Dave Robelen

[Dave Robelen]

Hello Richard,
Another thought. When I speak of a "dynamically scaled model" I am referencing the Froude system. Thus model linear dimensions are some scale=N, the model areas would be full scale/ N squared, and the model weight full scale/ N cubed.
Most likely you already knew this info, but someone else might benefit by running it by.
Regards, Dave Robelen

[Richard Miller]

Dave,

You hit right on perhaps my greatest concern. It's not quite enough to say that I only intend to fly this glider in benign conditions and don't contemplate upsets. You have to have some margin of safety.

A tube [d.=2?"] would be integral with the horizontal surface. It would be supported by two bearings, one aft at the center of lift of the surface, also a fuselage/pod hardpoint, and forward just above the small of the pilot's back. Tube and surface rotate/tilt as a unit.

From horns, right and left, cables would run through pullies at the pilot's hips and forward to a pivoting bar. Pull right/push left and the left side of the stab goes down. I can hardly say how pleased, after jumping through all kinds of hoops - swash plates and gears and whatnot - to fall on something so very very simple.

The tube+stabilizer tilts up and down - greater and lesser decalage - at a the surface leading edge. Fore and aft movement of the control bar would thus effect decalage changes. Kind of like a V-tail mixer. Have I made it all clear?

Were the travel of such a large surface even a few degrees I believe it would have a very strong effect. Quod erat demonstrandum.

What I had in mind with dynamics was nothing more than multiplying the linear difference by 3 when I dealt with weight and wing loading. I know there's a lot more to dynamic similitude than that, but nothing that I can address as I see it.

And this model will be LIGHT. I intend to use struts and see no need for typical, as Gentle Lady-type surface structure for a glider that will be hand launched in sea breazes and never bear much real stress. I'm aiming for a very light loading for the full-scale version, needless to say.

-Richard

[Dave Robelen]

Hello Richard,
It is reassuring that you have provided some degree of aerodynamic control for those moments that we wish would never come. The very fact that you are planning operations in conditions of lift suggests to potential presence of turbulence. Your ideas and messages make interesting reading and I wish you a long life.
When I spoke of the dynamic scaling, it was the model weight that I was most concerned with. For example if you are dealing with a 1/4 scale model of a 200 pd. flying machine, the model would weigh 3.125 pounds for dynamic similitude. This can be significant when testing control responses and handling, because it establishes the flying speed and angle of attack for trim that the full scale machine will see. This of course ignores the effects of Reynolds no., but with a reasonably large model this is generally not a major issue.
I will admit to a bit of a puzzle as to why you favor placing the tail in the wings downwash. Certainly there will be pitching moments imposed on the tail that will vary with the AOA, and this may be favorable, but the flowfield behind the wing has reduced dynamic pressure and potential turbulence that reduce the effectiveness of the tail. Care to share?
Enough for the moment, Dave

[Richard Miller]

Morning Dave,

Well, you did it again. It was that tail in the downwash that was most on my mind as I contemplated this correspondence.

I thought I was getting something of possible value, although I admit I didn't know what it was, nor why. Tails in the downwash is kind of a normal thing in, say, indoor design, and I thought it had to be GOOD. Maybe it would help with AOA, with effective decalage.

HELP!

Meanwhile, airfoils: I always thought Carl Goldberg had a sensible approach to the matter and have, some time since, adopted a bread-and-butter section: flat bottom, about 9% thick at the 37% chord point. Moderate roundness to the l.e. Pretty much "Gentle Lady" section. That for the wing.

After that I founder a bit. The stab area, 40%, is something I just pulled out of the air. It's negotiable. The position, not far behind the tt.e. of the wing is pretty solid. Short to keep the whole glider from getting too long; long to get the effectiveness of the moment arm. The pilot's CG comes into this as well.

The downwash issue is further compromised by the 14-degree forward sweep of the wing t.e. and the 19-degree backward sweep of the stab l.e. :( As soon as the stab starts to tilt things get even more problematical.

So I puzzle over it now, but with the assurance that building and flying a model in, say 1/5 or 1/4 scale, the variables will emerge and the compromises become evident.

[I recollect my morning walks to the hill with a design I called the Manta, this back in the 70s. On my way I'd be telling myself that I'd never understand what was going on. On my way back it would be quite clear.]

You will be informed of size and weight and loading in due course, and can comment. And of course I misspoke myself when I said 3X; what I meant, of course, was the cube of the linear dimension.

-Richard

[Dave Robelen]

Hello Richard,
If it were myself, I would use a symetrical airfoil on the stab. Say about 6 %. The approach to the wing sounds quite reasonable, but why not use a tapered thickness for a bit more strength at the same weight. Someting like 12% flat bottom at the root blending to 9% at the tip. At the Reynolds number of a full scale glider (even a slow one) the flow should stay attached very nicely and may even boost the C sub l a bit. It seems pretty likely that some of the control issues will need to wait for a model to really see whats going to work. I would vote for adjustable dihedral on the model as well. Turning with a tilting tail is a rather weak way of rapidly lifting a wing from a gust condition.
I stiil see some value in a smaller all balsa glider with simple single surface wing airfoil to get a quick look at both stability and control effectiveness. This would be a matter of presetting the control on a free flight model through a matrix, and then the CG the same. Might be fun, and it could save some work on the larger glider.
It would be very helpful to get a measure of the control forces in the two axis with the R/C glider. Intuitively, I can see considerably more force required for the pitch inputs than the tail rotation.
Cheers, Dave Robelen

PS Calling this machine the RFD sounds a bit fatalistic considering you plan to ride in it.

[Rotaryphile]

Am a bit rushed right now, but I agree with Dave's comments. Would very strongly recommend building a small scale model, even a hand launched glider, before committing yourself further. I have found simple profile fuselage hand launched gliders an excellent research tool for checking out unusual configurations. I am highly unenthusiastic about weight shifting for pitch control, since the pitch moment that can be so developed is very small relaltive to what can be obtained from aerodynamic pitch control, and aerodynamic trim control will still be required, in all likelihood. Once the main bugs are worked out of the hand launched glider, a larger RC glider would be the way to go, in my view.

I can see no advantage, other than structural, in having the horizontal stab in the wing's downwash. Use of a high mounted stab, i.e., a T tail usually enables about a 20% reduction in stab area for the same pitch stability, and a corresponding reduction of drag. The low aspect ratio stab will tend to develop rather high induced drag whenever it is developing sufficient lift to make carrying all that wetted surface worthwhile.

Yaw control by tail tilting is a good way to reduce drag, particularly at near-stall airspeed, but it tends to generate rather weak yaw moments at higher speed.

[Richard Miller]

Dave & Rotaryfile,

A small, simple glider is under construction. The span is 16.5", the rest in proportion. The tentative design of an RC version, span to be in the 68-72" range, is on the boards.

Airfoils - I see these as second-order matters; refinement stage, and will consider the matter seriously then. I am always aware of where that 1/2-panel centroid is going to fall and the consequent mass moments/inertia. [The wing actually has double taper and an outer diffusor panel.]

Speaking of which, it occurred to me to day to plot the effective lateral force derived from stab tilt against the force required to tilt the stabilizer. Since area varies to the second power, and the second variable, muscle input, to the 4th [if I have the math correct], large is not good.

So I am rethinking size. But how much force would one need?
And, in reserve, is the idea of using articulated trailing-edge slats/ailerons, working differentially, to induce the stab to tilt. My sense is that it would go a lot quicker that way.

And I'm rethinking position. Above the wing!? I think of the Genesis. I'll play around with that.

C.G. shift. I have in mind a one-speed glider and not really much need to shift. Longitudinal trim in turning? Maybe, with the mass between the lift centers, it can be made to take care of itself.

The other implication of that "final" didn't occur to me. It is now Richard's Ultimate Glider/Design...until or unless that bites me in the butt.

I really do appreciate your comments, suggestions, analyses, and if I skip any in the short run it's because I'm busy trying to solve others.

-Richard

[Dave Robelen]

Hello Richard,
I will be looking forward to the results from the smaller test model. It should be illuminating. Way to go on the name change.
Regards, Dave

[Richard Miller]

Dave & Rotaryphile,

This is just kind of musing on my part. I think about the design, and think some more, and presently I think, why didn't I think of that before. As...

One degree of tilt beginning at 10-degrees is going to be a lot more effective than the 1st degree, and one beginning at 20-degrees more effective still. How much is given up as against how much gained at some designate angle, and is there a sweet spot?

So, being pretty innumerate, which is a kind of blessing in a way [you don't tend to get sidetracked by abstractions as easily] I drew a spread of resultants - 22, 24, 26, 28-degrees - and broke them down into their components, just using a protractor and a cm rule. I like 27-degrees: you've only given up some 10% of the vertical component and have a 46% lateral vector.
*************************************

Longitudinal axis. At its aft end, the station at which lateral forces are going to be effective. A small vector representing that force. The appropriate questions:

1) How strong is it? Strong enough?
2) How quickly can it be deployed?
3) How far above the longitudinal axis is the designer forced, by other considerations, to place it, the penalty being the adverse roll that's induced when, as it almost always the case, it is above?
[When you build hlg and fool around, try an under rudder, you learn about these things. All praise to model building and the knowledge it gives us.]

I've backed off to a 32% horizontal surface. The lateral component of that [46%] is equal to about 15% of that surface area acting laterally. Is that close to, for example, the vertical surface of a "Gentle Lady"? Have I got just about the same lateral force?

On the plus side, that component will have less vertical separation than the - what, 3 or so inches in the case of the "Gentle Lady." On the minus side, the moment arm is considerable shorter.

How quickly can it be deployed? I intend to use a standard "Gentle Lady" servo and compare turning response as a function of that, then try a succession of smaller servos to get a feeling for power in/power out.

By all means, Dave, variable dihedral. There will be one best angle. I shy away from tapering airfoil thickness in the model only from laziness. That's my attitude now. Maybe I'll change.

About enough, eh?

-Richard

[Dave Robelen]

Hello Richard,
At first I thought I was confused with your terminology, but now I am not so sure. Will your "bird" have a deployable vertical surface? any at all? Perhaps you are reffering to the movement of the horizontal tail in roll, yes?
Regards, Dave

[Dave Robelen]

Hello Richard,
You are causing me to think more than usual, quite an excersize these days. In your discussion about the available side force from a rotating tail, I got the impression that you are expecting more side force with increasing angle of rotation. If in fact the CG is placed so that the tail is carrying lift, and you make the necessary adjustments to maintain a level altitude in a turn, I would expect the turning force to weaken at increasing angles as the vertical component of the tail is reduced. Sort of analogous to a banked airplane in a turn and the increased wing loading thing. Am I missing something here?
Regards, Dave

[Richard Miller]

Good Morning Dave,

Maybe I should go all the way back to the beginning of this little passion of mine. That time of the morning when the eagle and the hawk drop off their perches and go huntin'. They know the convection is good up to 3, 400 ft. and they more than likely know where the good stuff is. They just work the lift in a lazy kind of way until they find breakfast.

That's what I'm after, and I stress again that it is a benign kind of soaring as I conceive it. Maybe a little forward stick to get a little extra speed when it's time to land, which is before things get rough, and nothing like cross controls in the landing pattern.

So you do that, fly, then land and have lunch, then get in a real glider and fly to 30,000 ft, or go 500 miles, or whatever strikes your fancy.

What I began with was the consciousness of the bird; not its structure. And it's time to confess that I contemplate, somewhere down the line, using supplementary electric power to provide the same assurance that the bird has, knowing it can flap. The power is just adequate to maintain equilibrium flight, plus a slight reserve to [slowly] gain altitude. One rises and falls with changes of thrust, maneuvers with simple bar movements. Basic attitude need never change. I only jest a little when I say I contemplate the kind of flying machine you could put your grandmother in, give her a little instruction, send her off.

No vertical surface, but a lateral control force by means of the tilt of the rear surface. All the wetted area = lifting area. Pretty much like the bird. Very high efficiency. Very light wing loading. Very low sink rate. Very little power/thrust requirement for sustianing flight.

-Richard

[paulriseborough]

Richard,

Tails are better off positioned in as little downwash as practical for the following reasons:

1) A pitch up results in an increase in wing angle of attack, followed shortly thereafter by an increase in down wash. The increased downwash reduces the angle of attack at the tail which reduces its corrective effect.

2)If the tail is producing lift then its induced drag is increased.

A related but separate efect is that a short tail arm provides less pitch damping. Not a problem for RC unless you're flying pattern, but nasty for fullsize given pilots/padssengers that suffer from motion sickness!!

Paul

[Richard Miller]

Paul,

I'm still trying to get this matter of downwash clear in my mind and I appreciate your comments.

Let us assume - not quite settled at this point - a special case: that we're dealing with a one-speed, one-angle of attack, one-attitude glider intended to be flown in light morning lift. Ideally no changes in the relationship between the wing and the horizontal in the intended flight regime.

Any speed changes, as forcing your way down against the thermals, would be achieved by weight shift. [A little girdle for the pilot with studs on the hips. Walk the studs fore and aft on side plates with slots. Real easy. Real quick.]

Would this in any way modify your point of view?

-Richard

[Dave Robelen]

Hello Richard,
I will jump in here. Every time you wiggle or shift weight to change airspeed, you are changing the AOA and therefore the downwash. My difficulty in placing the tail in the direct wake of the wing is that the down wash not only has low energy, but it is also turbulent and may well have spanwise angularities that will monkey with the effect of the tilting tail.
Cheers, Dave

[paulriseborough]

Richard,

You will be using weight shift for control, but it will not be the primary stability mechanism. Stability will largely come from the fact that the aerodynamic centre is behind the centre of mass.

Placing the tail closer to the wing and in increasing downwash means that you need a larger tail area for stability than a simple moment arm calculation would indicate. When the tail gets very close to the wing it becomes more advantageous to move to a flying wing configuration to avoid the possibility of unpredictable trim changes as the tail moves in and out of the wings wake.

As Dave says, there will be always some change in angle of attack however small so unless you have the reflexes and flight instrumentation of a bird, a statically and dynamically stable vehicle is a good idea.

Sounds like a fun project

Paul

[Richard Miller]

Paul,

Thanks for taking the time to comment. Note that with the lifting horizontal surface - now about 31, down from near 40% of the wing area - there are two centers of lift, with the weight of pilot neatly nestled between them. Sounds good when you put it that way, doesn't it.

For over 40 years I have built models, free flight, all my hlg, with this sort of set up, and all have had very satisfactory longitudinal stability/trim. Thus I have a strong intuition that it will work on a full-scale aircraft, especially one with such inherent pendulum stability, a pilot twice - ?? - the weight of the aircraft, some distance below the wing.

And I do want to use that aft lifting surface for maneuvering, and am convinced it can be done. It is efficiency that drives me.
Nice clean sufaces, no breaks nor gaps nor slots, pure lift vertically directed, except for diverting the vectors to turn. It is necessary to make some sacrifices in other regards - it always is - and have some legitimate questions raised.

So I do appreciate having them raised, and having the opportunity to answer them.

-Richard