Facet Opel

(Moved from "Affecting Thrust, Propeller". See photo there for an image of this aircraft.)

Originally posted by orion
Regarding the Australian Facet Opel, I too had a problem finding anything until about a year ago, by sheer coincidence, I ran into someone who actually knew the designer. Several others locally to him, apparently had followed the development so that in the future they too could build the little plane. The story I got was that the designer did not feel comfortable releasing the data until he flew the bugs out. Unfortunately due to his untimely death, that never happened.

Perhaps you might know one thing that I can't see in the available photos. I looks like the aircraft has elevons extending from outboard of the vertical stabilizers to near the tips, rather than separate elevators and ailerons. Do you know which configuration was in use? Some of the stuff in Nickel & Wohlfart's book suggest that separate surfaces might be desirable in a flying plank, and I was wondering what the Opel's designer was using.

The other piece of information I'd like to know was the approximate aspect ratio. Looks mighty small, although he appears to be making use of that to increase the chord and maybe get a little more pitch damping out of the wing. I'm completely guessing on that latter - any thoughts?

Virtually all planck designs I've run across (which in actuality is only a few) have used only elevons and a control mixing device. I've seen a few sketches of two sets of surfaces but never seen it in practice. I think what most shoot for is sufficient wing area and chord length so as to dampen some of the sensitivity and dynamics, and somewhat compromise the top end performance. The compromise though is relatively slight as long as the airplane is maintained as a point design.

As far as the physical characteristics of the Opel are concerned, I never saw any drawings, only the few published photos. It however is a low aspect ratio wing but I think it works fine due to its relatively low loading and substantial chord. It also looks like the control surfaces are a good proportion of the chord but the perspective makes that hard to judge. I also don't recall any photos with sufficient detail to show how many control surfaces it had. But I think you're right - some of the damping probably does come from the long chord and probably a bit from the constantly reflexed surfaces.

Interesting. If I recall, I picked up the idea of using separate elevators and ailerons from Marske, who uses that configuration in all his later designs. I believe his claim is that if you design the wing to stall near the root first, the separated flow blankets the elevators and prevents them from developing enough nose-up pitching moment to force a stall over the entire wing. Another 'natural' stall-limiting device, and one that potentially only limits the pitch control authority from the point of initial stall and not earlier. I couldn't find anything in Nickel and Wohlfart's book that ruled this out for a 'plank (although they favor full-span elevons for efficiency's sake), so I thought I'd try it. But I'm going on memory again here.

I suspect you've seen the same two or three photos of the Facel Opel that I have, and that there probably aren't any more on the web. They don't show a lot of detail.

Thanks for your thoughts on chord and area as it relates to damping dynamic behavior. I'd hoped that was the case, but I hadn't figured out a way to quantify it yet, beyond some very 'ballpark' methods that I was using with regards to my control surface effectiveness. Too 'ballpark' to risk my life upon, I'm afraid. I certainly agree that a little extra area or chord was worth the slight drag penalty if it meant the airplane was easier to fly.

And yes, all this talk has be thinking back on my earlier design. Tempting, if I could be more confident of the flying qualities for a newbie pilot such as myself.

One other subtopic regarding damping dynamic characteristics has to do with mass distribution. This isn't discussed much because as was pointed out in the previous thread, most of the flying wing efforts seem to be concentrated on gliders, not powered aircraft. However here the Facet Opal (I checked the spelling) is a pretty good example of a potential approach - notice that the cockpit pod sits relatively forward of the wing. If we examine the photos, we can see that the mass center of the pilot is actually at, or just in front, of the wing's leading edge.

The nose and the pilot are therefore balanced by the mass distribution of the wing, as well as by the location of the engine, drivetrain and of course, prop (the spinning prop also delivers a bit of damping). You don't benefit from this in a glider since there the effort is to place the pilot as close to the CG as possible so as to minimize the effect of differering pilot weights.

But on a powered craft you can distribute the masses in such a way that your dynamic factors become more tolerable due to the inertial charactiristics of the displaced masses. Of course you can't get carried away here since the aero controls are sized for the particular airframe and wing, and without a conventional tail, have substantially less effectiveness, but some potential can be gained by careful placement of the aircraft major components.

I can't find any pictures of it on the web, but I know I've seen some before! Or atleast I think I know what your talking about.

At one time I had about half a dozen pictures of the Opal but now all I have left is the one below - I'm not sure where all the others disappeared to.

Wait a minute.. this site might have some pics?

http://www.twitt.org/baxter.html

Originally posted by Topaz
It looks like the aircraft has elevons extending from outboard of the vertical stabilizers to near the tips, rather than separate elevators and ailerons. Do you know which configuration was in use?

A constant chord plank can use either system because the longitudinal distance from the CG to the trailing edge is constant along the span. Forward swept planks like Marske and Fauvel have the pitch control at the center and ailerons at the tip because the tip doesn't provide any moment arm.

Marske doesn't use any washout because the slight forward sweep has the same effect, the root stals first

Okay, this is just great stuff. Orion, I want to kind of 'parrot' back some of what you were saying regarding mass distribution in my own words, to make sure I'm really grasping it.

If I'm understanding you correctly, having the large masses a little farther away from the CG than 'absolutely possible' could be a beneficial thing. Essentially you're 'lengthening the pendulum' and slowing down the period of the phugoids to something more comfortable from the pilot's point of view, and the added pitch-axis inertia would help damp their formation in the first place. Obviously this could be overdone, such that the extra inertia of the system works against you once it's rotating about the pitch axis, and the control surfaces (and static stability) couldn't generate enough opposing moment to get you back to straight and level in the first place.

Is that about right? I had been quite concerned abou the 'overdoing' it part, and overwhelming the weak control authority and damping in my 'plank design. Perhaps I was putting too much weight on that effect. Nickel & Wohlfart also mention a high-frequency, medium-amplitude pitch ocillation in turbulence connected with several flying planks and low-sweep flying wings, notably the SB 13 sailplane. They call it "pecking". I would imagine 'lengthening the pendulum' might help reduce or prevent that.

I went back to my notes and texts and did a little more research on pitch control placement for flying planks. The results were:

• Full-span elevons have the least effect on the induced drag with control-surface deflection, as one might expect.
• Center-mounted elevators (with outboard ailerons) affect the induced drag the most, and may promote tip stall because of the change in the lift distribution with 'nose up' deflections.
• Outboard elevators (with either spoilers or inboard ailerons) are less effective as pitch control devices, because of tip losses (end plates can reduce the effect)
• Full-span elevons worsen the landing problem I alluded to in earlier posts: When they are deflected for 'nose-up', they immediately reduce section Cl over the entire wing, before eventually increasing the angle of attack to the point that the overall wing CL goes up and you begin to climb. If you make a sharp nose-up control motion at low altitude (such as during flare), this can potentially cause the aircraft to settle suddenly and perhaps impact the ground. Limited-span pitch controls reduce this effect *somewhat*, but it's pretty much inherent in flying planks to one degree or another.
• The increase in induced drag for center-mounted elevators can be reduced by increasing the local wing chord over the span of the elevator itself, along with some small adjustments to the wing planform near the tips. Nickel & Wohlfart give a theoretical basis for this and a method to determine the required increase in chord. Done properly, they claim that this 'fix' works "independantly of the CL value and the deflection angle" of the elevator. From what I can see, both Fauvel and Marske are using this technique - or at least they both are extending the elevator chord on their later designs.

Combining this with Marske's claim about center-elevators having a 'self-limiting' pitch authority near the stall led me to go for center elevators on my own design. When I ran lift distributions on my airplane, the initial stall was just outboard of the ends of the elevator, but still more than a quarter-span away from the tips. I was hoping this would still give me fairly tame stall characteristics from the point of view of the planform (my airfoil was quite benign all along). I wasn't getting initial separation over the elevators as a result of the planform, but I was hoping for some there due to my small 'fuselage pod' that would help a little bit - if only by providing some 'shake' to the pitch controls near the stall.

I'd be very interested in your thoughts about this material. I'm taking Nickel & Wohlfart pretty much on faith. They have an obsession with induced drag (being soaring enthusiasts), but since my chosen engine was quite small (55hp), I figured following in their footsteps a bit couldn't hurt my climb rate any.

Norm - Any further insights on Marske's design rationale regarding planform and control surface configuration selection? He's so darned secretive and the person representing him publically either doesn't have access to the information or is unwilling to share it.

Regarding the mass distribution, I think you're right on.

I have encountered the issue of pitch oscillation before but not to any level of detail - I've never heard it called "pecking" although that's prabably a pretty good description. The few instances I've come across regarding this behavior seem to be more connected with gliders, not powered airplanes. The gliders fly slower so any gust induced motion would be worse than in an airpalne that has a measurably greater cruise speed.

The control surface configuration is probably one of the areas that you could get a good handle on using some level of scale modeling. No, you wont get the quantatative data but the qualatative info just might be sufficient to what you're after. And as you indicated, for a powered airplane, the variations of induced drag are probably not as critical as control harmony and control effectiveness.

As far as the lift loss due to elevon deflection on flare is concerned, the way I do first cut work on this is to consider the wing planform as two seperate lifting planes. The division is made at the point where the camber line dips below the chord plane. Although very approximate, it does allow you to handle the lift vectors seperately and do the trim calculations almost as for any other airplane, just a very close coupled one. As such, the flare vectors can be easily calculated and accounted for when you're looking to define your maneuvering envelope.

Sorry this is so quick - I'm heaed out of town in a few minutes so no time for details.

Originally posted by Topaz


Norm - Any further insights on Marske's design rationale regarding planform and control surface configuration selection? He's so darned secretive and the person representing him publically either doesn't have access to the information or is unwilling to share it.

I've got Marske's little booklet from 1970. I don't know that he has published anything since then. Over the last few years I've seen 4 or 5 letters from him on the Nurflügel list but looking them up may be difficult since it's a fairly old mailing list that's changed servers several times, each time losing the archives. Also Mat Redsel owns the name "Jim Marske" and uses it in the header of his e-mail instead of his own name. The first two Marske wings, the MX-1 and Pioneer, were constant chord. In it's original form the MX-1 had tip fins. One of the modifications to that ship was to remove the fins and rudders and replace them with small split flap type drag rudders on the outboard end of the elevon's. Yaw control was not what he expected with these drag rudders. At some point he met Al Backstrom and during their conversation he asked why Backstrom's elevon's were inset instead of extending clear to the tips like on the Pioneer. The answer was that the tip vortices would make that part useless. Marske realized that his little drag rudders were right in the disturbed region and asap shaped some styrofoam into trailing edge blocks and shortened his elevons to move the draggers inboard. This didn't have a noticeable affect on performance but his yaw rate went up substantially. He also tried some different roll control devices on the Pioneer but a Hershey bar just has too much built in resistance and that's why the Pioneer had tapered wings. He never said a word about the fact that, for a given root chord and wing area, a tapered wing has a higher aspect ratio and lower span^2 loading than a constant chord wing. Back then he said that the reason for sweeping it forward was to get the pilot out in front so he could see and he has some drawings showing where the CG is on various planforms. The fact that this allows you to avoid building a washout jig is just a happy accident.

While looking around to see if I could find pictures of the early gliders and some information about the roll spoilers he used on later versions of the Pioneer-1A I found this page that has a lot of the text frome the booklet plus vintage color pictures. The booklet just has drawings and B/W pictures
http://www.continuo.com/videowebpage...2/interest.htm

Originally posted by orion
Regarding the mass distribution, I think you're right on.

Very good. Thanks for the feedback!

Originally posted by orion
The control surface configuration is probably one of the areas that you could get a good handle on using some level of scale modeling. ... And as you indicated, for a powered airplane, the variations of induced drag are probably not as critical as control harmony and control effectiveness.

I agree completely, although the fact that my budget favors smaller engines means I need to pay it some attention. I haven't had a chance to do any work with the holiday weekend, but I want to run some examples on the proceedure from "Tailless Aircraft..." regarding extending the chord on a centrally-mounted partial-span elevator as a means of reducing the induced-drag penalty and keeping the flight-behavior as benign as possible. The authors made an interesting notation regarding configurations with endplates, whereby the 'optimized' chord length could be equal to that of the neighboring wing, or even shorter, due to the influence of the endplates on the lift distribution. I'd like to see how that works out.

Originally posted by orion
As far as the lift loss due to elevon deflection on flare is concerned, the way I do first cut work on this is to consider the wing planform as two seperate lifting planes. ... As such, the flare vectors can be easily calculated and accounted for when you're looking to define your maneuvering envelope.

Hmmmm... There's some food for thought. How would you calculate the lift coefficient for each separate 'plane', given that they in fact are part of the same wing and their flows influence each other greatly? I have memory of an early wind-tunnel lab I took where we took pressure measurements along the upper and lower surfaces of an airfoil, reduced the values to pressure coefficients and from there derived the overall lift coefficient for that section under those conditions. Is that process applicable to the lift of partial-chord segments? I could see using a CFD program to calculate the pressure coefficients for a finite-span wing, then applying those values the lab process I mentioned above. Actually, couldn't you then derive lift, drag, and moment values for the entire wing, and from that the sum vectors for that flight condition? Am I anywhere near what you were talking about???

Originally posted by orion
Sorry this is so quick - I'm heaed out of town in a few minutes so no time for details.

No worries. I hope you had a safe and happy trip.

Originally posted by Norman
I've got Marske's little booklet from 1970. ... While looking around to see if I could find pictures of the early gliders and some information about the roll spoilers he used on later versions of the Pioneer-1A I found this page that has a lot of the text frome the booklet plus vintage color pictures.

Very interesting reading! I'd been on the parent site before, but I hadn't seen that portion.

My previous flying-plank design was quite similar to his first airplane, with constant-chord wings, endplate vertical stabilizers, and centerline landing gear. Obviously the span was shorter, as I was intent on a sportplane instead of a sailplane.

Thanks, Norman.

Regarding the calculation of lift vectors, as I indicated, this is only an approximation since as you indicated, the flow from one section of the airfoil does influence the other. The approximation however really looks only at the camber line. Essentially, although I haven't tried this yet, you could use an airfoils section program where you can model the camber line itself, then have the program calculate a lift-curve plot for the two segments of the camber profile, and use that data for the first cut.

If however you have access to any form of CFD analysis, that of course will deliver much more useful, complete and accurate data, making my approximation look somewhat crude.

Ah, okay, got it. Interesting idea! I do have X-foil and know how to use it fairly well. Would you put an arbitrary thickness distribution on each camber segment, or use something closer to the parent airfoil?

As for CFD, I was taking a look at Drela's AVL routine, but the most recent version is only available as Fortran 77 source code, and requires the user to recompile for their own system. I'm not averse to doing the compile (my computer days go back far enough for that), but I haven't had a Fortran compiler of any kind in years, let alone one for Windows XP. Bit of a pain, but I don't have thousands of dollars to spend on CFD software.

Any suggestions as to reasonably-priced or shareware alternatives? The advantage to AVL in my eyes was that it was fairly complete, including some pre- and post-processing. I don't mind something with a crude interface (I'm comfortable with X-foil). I also don't need the capability to model multiple separate airfoils (lacking a second wing or a tail) but I'd like to be able to model winglets on a wing of arbitrary planform, sweep, and twist, including control surfaces.

If I do (now or whenever) decide to restart a flying-wing project, this seems like it would be the way to go, unless I can decode the rest of the methods in Nickel & Wohlfart's book sufficiently to make that work. Some of the math is still beyond me, but I think perhaps I understand the theory well enough to make it work with CFD doing the crunching.