Aerobatic Airfoils Design Considerations? 3

[cludwig]

Hello all,
I've got B.S. in Aerospace Engineering and I am trying to design my own personal homebuilt aircraft. It will be two place aerobatic, designed for about 8G solo, 6G dual, fully inverted systems, etc, etc. Wing will have no twist, 1 degree or less dihedral, zero sweep about the quarter chord, and a taper ratio of 0.5. (or 2 using the other convention)
I read all there is to read about aerobatic airfoils. Most everything is propretary. It was a quick read. I have airfoil coordinates for a Sukhoi SU-26M and an airfoil similar to those used by CAP and Extra. I can also get data for the one design. But all these airfoils follow a similar principle in design and I am not sure if it is what I need.
Can anyone give me a place to start my design from? My main consideration is for an airfoil that breaks in the stall and reattaches readily. I also want a fairly high CLmax so my wing isn't the size of a barn door.
Also, what software can I use to reliably predict stall and post stall behavior of an airfoil? Mind you, I am a homebuilder, not Boeing. I can't sell my Jeep to pay for a software license.


Any help is immensly appreciated!!
Chris.

 

[MikeHalloran]

Screw software. Build an airfoil test rig on your jeep, build some test foil sections, and measure what they do. A jeep can't go fast enough to measure high speed drag, but it can sure go fast enough to measure stall characteristics on any airfoil you'd want.





Mike Halloran
Pembroke Pines, FL, USA

 

[rb1957]

surely you don't need to generate your own data, with all the uncertainities that come with that !

there are many aerobatic planes from the US and Europe which will use US and European airfoils, for which you should be able to get the data much more easily that Russian designs. Personally I'd start with Pitts (yes, i know thats a bi-plane).

 

[btrueblood]

I'd also look at trainer aircraft for the US military (starting from the PT-19), most of these had "soft" stall and good recovery characteristics as part of their design definitions.

 

[cludwig]

I can't copy a Pitts airfoil. The top (and forward) wing is a laminar airfoil. The lower (and rearword) wing is not a liminar flow wing. The laminar wing stalls at a slightly lower angle of attack than the other. When the laminar wing stalls (forward wing), the nose of the plane drops. This happens regardless of whether the plane is right side up or upside down. That is how Mr. Pitts made his planes safe in the stall right side up and upside down. I can't apply biplane theory to monoplanes in an aerobatic case.

Military trainers are no help to me. They all fly right side up. I'm wanting a fully symmetrical wing. Also, they can use a rectangular wing or geometric twist to encourage a root stall and thus allowing gentle stall charactoristics.

I can't use geometric twist because the wing needs to be fully symmetrical. I also can't use a rectangular wing since I am aiming for a design loading of 8G. Rectangular wings carry too much weight and inertia on the tips, not good when pulling 8G or rolling at 360 degree/sec. I prefer a 1/2 taper ratio to optimize the structure better. A 1/2 taper ratio wing with no twist typically stalls at about mid half span and separation bubble grows both rootward and tipward. This blankets the ailerons pretty quickly.

I've finally got headed down the right path though. The first airfoil design software I tried used an empirical method to predict stall. It wasn't good and every stall looked the same.

I switched to XFOIL and although the interface is enough to do my head in, it is much more reliable. I can finally see the real diference between these airfoils. There seem to be 3 schools of thought in aerobatic airfoil design:

School 1: Medium-high CLmax (~1.6), low drag throughout. Really hangs on around a corner and avoids burning off energy.

School 2: Massively high CLmax (~2.0), low drag at high angles of attack, but the downside is high drag at more normal angles of attack. Stall is VERY sharp. These airfoils have a great ability to pull high G at low speeds but the burn off speed quicker: DR-107, DR-109, CAP family.

School 3: Conservative throughout. Lower CLmax (~1.35) with a mushy stall and very little break. This is used on the SU-26M and seems to be typical of the Rusioan planes. Higher drag, especially when pulling G. Looses energy quickly. Soft stall.

I am thinking of combining 2 of the 3 schools. Put a conservative airfoil with an early stall on the root and a low drag, high lift airfoil on the tip. I can get a aerodynamic washout of about 5 degrees to force a root stall first and keep the ailerons working futher into the stall. That will make a safe plane during an approach to landing stall. The low drag high lift tips will help pull around a corner at higher G. When the tips finally stall, it will be sharp but only after the preceding root stall. Thus I am hoping for a soft root stall followed by a sharper tip stall such as when needed for snaps and spins.

And when I am all done playing with the theory, I will do one of two things: Strap a model of a wing to my jeep and head for the open road to check the stall break or just find a plane that stalls exactly as I want and copy the airfoil/wing design.

Thanks Mike for the fool proof strategy.

Feel free to drop me a note if anyone has more info for me,
Chris.

 

[dgapilot]

5 degree twist seems like a lot. It will help when right side up but will hurt when inverted. Most of the airplanes hI've worked on have 1.5 to 2 degrees of washout.

 

[cludwig]

The twist is not geometric, it is aerodynamic. The wing is identical right side up as upside down, it is perfectly straight. Root arifoils are selected that stall prior to the tip airfoils. 5 degrees difference in critical angle of attack is about the biggests difference you can get using this method. I may choose less (2-3), it depends which airfoil I use on the tip.

Sorry for the lack of clarity.
Chris.

 

[btrueblood]

Are you really planning to pull 8G negative?

 

[cludwig]

Hell No! Negative G hurts!! I don't like much over 3 or 4 G negeative. Designing for -6G would be enough from an operational standpoint.

 

[rb1957]

-6G would be ultimate load then (-4G limit) ?

 

[btrueblood]

So, if you don't need +8G and -8G lift, why design with a symmetric airfoil?

 

[CarbonWerkes]

Sorry to jump on this thread late.

Cludwig, Im curious as to why you would want to reinvent the wheel here. Having flown unlimited aerobatics in an EA300, and like most acro pilots- having worked up to that from C152s through Decathalons and the Pitts stuff, I can tell you that the stall on the Extra/Edge is tame and very easy to resolve relative to the other aircraft mentioned. There is not much warning of impending stall, but that is typical of sym wings. So why not just use a profile from a Cap232 or Ea300, and use some of the very well established construction methods for the Giles, etc? Forget about CFD sims. Very few can even deal with post-stall, and that is without the contributions of prop wash, roll, yaw, etc.
Hope this helps a little-
Best,
Rob

 

[MattAero]

Carbonwerkes is right about reinventing the wheel and forgetting about CFD. I have forty years of aerodynamic experience and can assure you that CFD and even 3D wind tunnel investigations will not predict the stall and post stall characteristics that you will see in the aircraft at flight Reynolds number conditions. Hence, research what has been already done and pick the design that best suits your requirements.

 

[cludwig]

to btrueblood: The symmetric airfoil is so you can fly clean inverted figures. For an example, go fly a 2 point hesitation roll in a Citabra. Then fly it in an Extra. You will see the difference. An unsymmetrical airfoil requires much, much higher alpha to generate the lift required for level flight. That much higher alpha requires much higher control displacement and forces. The manuever is just harder to fly well with a non-symmetric airfoil. Though to be fair, the difference in that example will have more to due with the wing placement of the citabra (it is too far from the thrust line).

As to Carbonwerks and Mattaero: You're both right. If I do go ahead with building my own plane, it will have a wing profile nearly identical to the Extra. After all, it works and well.

For now though, this is all a self-education exercise. I don't expect so much to actually start building, only to learn from the design process.

Thanks for all the posts,
Chris.

 

[TractorPilot]

Hey Chris,
Did you happen to come across airfoil data for the Extra wing? I am also designing a aerobatic plane for competition acro. The airfoil for the Extra seems to be quite similar to the Roncz airfoil on the Edge 540, but I have been unable to find any info on either... I've drawn up an airfoil that looks similar, and I've ran it through X-Foil and Javafoil. I'm pretty sure it will work, but I still would like to compare to an Extra airfoil.

J.

 

[cludwig]

TractorPilot,
Please click on the link below and call or email me. I'd like to discuss this with you offline.

Regards,
Chris.

http://directory.utk.edu/search.jsp?query=Chris+ludwig

 

[cludwig]

These airfoils of the type on the Edge and Extra were first proposed by Dr. Eppler. (at least that is the oldest reference to the design that I can find, correct me if I am wrong, I'd love to see read more on the subject) Eppler proposed an airfoil with an elliptic nose and flat plate top and bottom.

You can model this and generate coordinates by making a piecewise function of an ellipse and two straight lines. The ellipse has it's center on the chord at the point of the desired max thickness. Ellipse minor axis gives the desired thickness while half the ellipse major axis gives the desired point of max thickness for the airfoil. Next locate two tangental lines from the ellipse to the trailing edge. This can be done in excel relatively easily. I suggest using loop functions to find the correct slope for the lines.

Eppler wanted airfoils that would have low drag when pulling high G and relatively high high drag on a downline.

The low drag at high G was achieved by moving the max thickness point forward of the typical 25% (or later)chord area. Most of this family of airfoils operates in the range of 14 to 17% chord for the location of max thickness. This forward thickness keeps the airfoil flying at much higher angles of attack and generating higher CL values. You can expect Clmax values in the range of 1.6 to 1.8 and even 2.0 for these airfoils depending on RE.

The high drag coefficient (think slightly higher, it isn't much higher at all) on a down line (AOA=0) comes partly from the fattish wing and also from the forward transition to turbulent flow. At zero angle of attack, the transition points are up in the 15 to 20% chord ranges. Profile drag from the wing is higher at zero AOA and lowest up in the 5 to 7 AOA range.

Eppler designed his airfoil e472 based on these principles although that airfoil deviates slightly from the idea of perfectly flate top and bottom. Eppler refered to this family of airfoils as "Extra" airfoils.

Roncz probably followed Eppler's lead and designed the Edge airfoil in a similar manner as Eppler's airfoils. (I believe he also wrote a version of Eppler's code with a better correction for drag with separation bubbles) Both the Edge and Extra have the flat plate sides. I can only guess whether Roncz deviated from the ideal elliptical nose as I haven't seen coordinates from that airfoil.

The Extra airfoils are 12% thick at the tip and 15% thick at the root (from their website). They appear to have a max thickness point in the range of 14 to 15% chord, and also a finite thickness trailing edge (about 5 to 6 mm). The ailerons have a sharper trailing edge (less thick, beveled to a point) than the rest of the wing.

Still, all this matters very little on a plane with 80% span ailerons. What good is 2-D analysis on idealized airfoils if you are gonna chop the back 30% off and install an aileron with big gaps in it's place. The typical aerobatic aileron is symmetrical or almost nearly so. At full difflection, the nose pops up into the flow on the opposite side of the wing. The gap will be relatively small in at full diflection and the flow will first cross the gap, then accelerate out around the curvature of the nose before being turned to follow the aileron surface. It makes for the lowest radius turn possible and improves the chances that the flow stays attached. Also, the ailerons are slightly thicker than the wing at the cutout. This encourages the flow to reattach after jumping the gap (helps when there is little or no aileron diflection).

My point: You can copy an Edge or Extra airfoil design precisely, but if you screw up the aileron design, it won't roll very well at all due to separation over the ailerons.

Well, now you all know what I know, or what I think I know. Where does that leave me? What sort of great aircraft am I designing? None. I'm gonna buy an old cheap aerobatic plane (Pitts S1-S maybe) and leave the expensive toys to the boys with excess cash.

I can't afford all this fun.

 

[rb1957]

I admit that i haven't heard on "Edge and Extra" airfoils before ... they sound like a very simple construction, presumably with the loss of some aerodynamic performance. Biulding a tapered wing "S"houldn't be all that difficult ... you'd make tool ribs at teh two ends and tool stringers could define the spanwise taper, build the intermediate ribs between the tapering tool stringers, replace tool strginers with the real ones, skin it over, "lick of paint".

btw, if the two flat panels of the airfoil meet at the trailing edge, I'd suggest that they are tangent to the ellipse making the fwd part of the airfoil, rather than meeting at the diameter (and causing a shape change of the sirfoil contour.

 

[cludwig]

You're spot on about the flat sides not meeting the ellipse at the point of max thickness. They meet slightly further aft at the point where the slope of the ellipse matches the slope of the line. That is what I meant by tangent point. The flat plates will either meet in a point at the trailing edge or meet with a finite thickness trailing edge.

Historical note: Eppler's software prefered sharp trailing edges so he typically designed his airfoils that way. (to ensure experiemental results were as close as possible to the analytical results - hence e472 has a sharp trailing edge) XFOIL is more accurate when there is a finite thickness at the trailing edge so most airfoils designed with XFOIL have a finite thickness trailing edge. (Most airfoils built have some small finite thickness at the trailing edge so this is useful)

 

[MadProf]

Hi Cludwig and Fellows,
My idea may not be a definitive solution to the problem as it throws up several other challenges to be met.
However, if one were to start with a biplane configuration having both wings symmetrical, probably laminar flow and neither wing set forward of the other, consider this:
A deployment mechanism embedded in the wing roots designed to expand the chord section of one wing pair; or the other; or neither; but not both simultaneously;
Slotted and slatted wing substructure rendered active when the mechanism expands the chord section at the wing root - inactive when contracted, such that the wing assumes its normal dimensions, shape and flow characteristics when contracted, but a low stall, lifting multi-element configuration when expanded;
Slots and slats in the lower wing oriented to produce positive lift when the aircraft is right-side-up and those in the upper wing oriented to produce positive lift when upside-down;
Wingtips remain fixed, such that deployment increases wing taper and the mean chord section is increased by not quite half of the increase at the wing root. Thus, the slots and slats are themselves tapered when they're active.
Pros:-
1. Lift is generated by the expanded wing and permits level flight at zero pitch at a given airspeed - even climb at higher speed.
2. Stall speed is reduced significantly.
Cons:-
1. Complex aerofoil section requires probable application of fairly advanced composite materials technology to retain robustness and wing flexion characteristics for aerobatics.
2. Challenges in designing and constructing a reliable and safe deployment mechanism that is easy to operate.
3. Greater difficulty and expense in construction on both counts.
4. Possible airworthiness certification difficulties with air safety bureaucracies.

I guess you might actually have to sell more than just the jeep to get this one working, but if you did, you'd have one hell of an aerobat....
Regards, Kerry (Mad Prof).