Quantcast

"Spar-less" wing construction

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

billyvray

Well-Known Member
Joined
Aug 17, 2005
Messages
773
Location
Newnan, GA
I'm learning about composite construction and design and have a question.
I wonder why there aren't more aircraft (lightweight simple ones) using a composite wing construction in which the skin is laid up in such a fashion to carry the "spar" loads. This would mean a single hotwire cut for the main wing and then lay up the skins, using thinner sections around the spar location. This would eliminate fabricating a spar separately.
I know the Quickie 1 was built this way.
I understand if landing gear and other devices involving stress concentrations were mounted in the wings it would change things. But, if you have just a wing, would this not be a good way to build a composite wing more quickly?

Just something to ponder....


Billyvray
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
6,976
Location
Saline Michigan
You certainly can beef up the skins enough to carry the primary bending loads. This is not skipping the caps, but is distributing them along the skin. The thing that you then have to make sure of is that you can distribute the load into these wide thin caps, which will usually mean quite a bit of BID or BIAX cloth. Remember, if the objective is to save weight, the figure of merit to be minimized is total fiberglass. If all you want to save is time, go buy an airplane.

You still need a shear web tying the top and bottom together. The foam has some strength this way, and can serve in some of the wing. If you analyze a beam at any point along its length (assuming a cantilever wing), you find that the shear web carries all of the load applied outboard of the place where you are looking, and the the shear web has to extend with one cap (or skin) and compress with the other one. If the element starts out square, the shear makes it a parrallelogram, and the bending distortion makes it a trapazoid. Superimpose the two, and that is how the shear web deforms. You will generally be way beyond the ability of 2 lb foam at the root, although you can usually allow the foam to do it at some point along the wing, tails, etc.

Having built my horizontal and vertical tail on foam cores, I can tell you that putting in the shear web is not really a big deal. You hotwire the full form with low area for the caps, and then hotwire where the shear web is laid up. You jig and brace the foam that has the spar in it, layup the shear web, then fit and bond the previously removed parts. The foam cores are now much stiffer. Then you build up the spar caps, then glass the whole thing. All steps are straightforward, it really goes quickly, and this is important, you can build an accurate laminar flow foil with little potential for messing it up. Think about having a bunch of UNI tape or cloth plus a ply or three of BID between the skin plies and the foam, and how much luck you would have replicating the desired skin shape...

Billski
 
Last edited:

billyvray

Well-Known Member
Joined
Aug 17, 2005
Messages
773
Location
Newnan, GA
Yeah I see your point, in that even a simple shear web could make a huge difference in loading capacity, enabling you to more easily account for the upper and lower cap loading.

Off topic, but do you have any pics or info on what you are building?
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
6,976
Location
Saline Michigan
Pics. Hmmm. Thanks for your interest. I have promised to convert my film photos to PDF files and start taking pics with a digital camera. Then, I will be able to do this sort of thing.

Right now, it largely looks like a lot of parts. The basic tub, with firewall, center section of the main spar, and a couple of aft bulkheads are in it, and the horizontal tail is complete, but everything else is just parts. Wingskins, turtle deck, forward deck, gullwing doors and roof, all resting in their molds. Outer panel spars, stock for the ribs and drag spars, etc, sitting in the rack.

I am building the center section drag spar right now. Layout and alignment of the wings follows, which mean install of drag spar, the seatback bulkhead, the instrument panel bulkhead, and commencing wing assembly. There will be some pictures once it gets that far. And it will start to look like an airplane then. In the meanwhile, there is my job, the homeowner's exercise program, snowboarding (it is winter after all), and the wifey poo. All in due time.

Billski
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
Stressed skin construction is sort of the Holy Grail of the composites industry since its application could result in very efficient structures that encompass a fraction of the part count of a more conventional approach. But the concept is not necessarily constrained to composites. Back in the fifties and sixties Ted Smith designed some very innovative structural concepts with aluminum, one result being a nearly monocoque wing. A good example of that type of approach was on the Piper Aerostar where the internal wing structure was pretty much a series of webs reinforcing a thick skin. There was a stub spar at the root for the wing's attachment but the remainder of the loads were carried by the skin.

Composites are ideally suited for this and as long as you consider issues like shear and panel stability, there's no reason why this could not be done, even without an internal foam mass. Unfortunately, the vast majority of the kit companies using composites have not gotten sufficiently beyond metal thinking, resulting in aircraft that have significantly heavier components that are at times a major pain to assemble.

Two of my current projects are based on a stressed skin type of structure and as soon as I get permission from the customers I'll post some pictures. One set of pictures that I can post however is this project from many years back, which was a set of graphite wings I did for a Pitts that raced at Reno. The pictures are of the one piece top wing prior to close-off and the bottom panel, one with tip off and one assembled with the new Frise aileron. Of note is the top wing - what you see is all that's there. The only components on the inside are the fore and aft shear webs and root and tip ribs. The internal structure is there primarily for assembly reinforcement and support for the hard points. Other than that the wing panels are totally hollow.

With a cantilever wing this would require a bit more analysis and design but the potential is there to significantly deviate from the spar rib structure you get with a metal type of approach.
 

Attachments

Corsair82pilot

Well-Known Member
Joined
Jun 18, 2006
Messages
83
Location
MD
You need some structure to keep the top and bottom wing skins or spar caps apart. That is usually a spar web. If you build a foam wing covered with glass or carbon, that works well, too. The only problem is the weight. That much foam still weighs a lot (depending on the size of the wing). The lighest foam you would use is still about 2 lbs/cu ft. Do the math. A single spar web weighs a whole lot less.
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
A couple of folks asked over email so I thought I'd add this - the new wings for the Pitts, installed, were about 40% stronger, at least five times as stiff (although I did not calculate the stiffness difference at the time so that's sort of a guess) and weighed in at seven pounds lighter than the original wood and fabric. The stiffness was apparently dramatically better over the original - the test pilot indicated that the roll response was equivalent to a four aileron equipped airframe (this one had ailerons only on the bottom wing).

Regarding Corsair's note, the interesting thing about this type of composite design is that done right, you actually do not need an internal structure at all, even on a cantilever wing. About the time I did this I was a structural/design engineer for HITCO Aerospace Structures - while there we developed a true monocoque wing (it was a flying wing with a separate body pod for payload) - it had no internal structure at all except for root attachment to the body. It used only cored wing skins on top and cored skins on the bottom. The shear loads were taken up by the aft web (which we needed for the control surfaces) and the leading edge, which of course transitions from the top skin to the bottom anyway and thus also acts as a shear web.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
6,976
Location
Saline Michigan
Some topics related to these last few comments:

The outside skin of a composite airplane has to do many things. It has to carry the loads generated by flight, by landing, by ground handling, rain and snow and hail, bird and bug strikes, be stiff enough to avoid flutter at Vne + few%, being handled during construction, and it has to survive fly-ins. Yeah, the thoughtless folks using it as a writing desk or setting a squalling two year old on it. In little airplanes, I believe that the birds and bees and hands result in more outside skin thickness than the flight loads do.

So, once we have a skin that is thicker than it would otherwise need to be, why not let it do other duty, like carry loads to the leading and trailing edge, where we can have the shear webs that let the skins work as caps. We can also let the skins serve to carry the bending loads. If we had to beef up the skins to get to this, I suspect that a nice simple little spar might have been a lighter way to get here, but being as we are already carrying around pretty beefy skins, we might as well use them for flight loads. One requirement of that is that the in-plane bending stiffness of the skin structures has to be adequate, but with a reasonable core thickness and inner skin layup, that should be fairly easily obtained. I would have to design the Pitts racing wing in each of these ways to decide which way would actually be the best. I suspect that Orion and HITCO probably did that little study. Scale the wing up in size to fighter or airliner levels, and the spar and skin with a few ribs might end up being the way to go...

Also in regards to building a composite wing for the Pitts that is lighter than Curtis Pitt's original beefy spar, many ribs, bracing, and fabric covered design, I would hope so! While it certainly is possible to build an equivalent wing that is heavier in composites, I would hope that a composites engineer and shop decent in their crafts would build lighter and stiffer simultaneously. This is just recognizing the limitations of the original material set vs the new set.

Another way to build a foil with minimum fuss is to build the cored skins, and then build a corrugated shear web. The corrugations would be on the scale of the wing thickness. The webs are not vertical, and you can make the mold out of galvanized steel on a brake. You just make it a little deeper than you need it, leave room forward and aft. When you assemble the top to the webs to the bottom (one assembly step), it sets its own thickness. If you want spar caps in it, you could add them too, but I would prefer to put the bending resistence in the skins, where it is further out and you need less of it... Our friends at AVT (http://www.avtcomposites.com/) showed me that one, which they did for race car foils.

On the topic of foam filled versus hollow, there are break points between hollow being lighter and foam cored being lighter. A bunch of years ago I heard the story about Burt Rutan's Catbird's canard. In an attempt to address this very argument, Burt and company built TWO canards for the Catbird. Recall that the Catbird is a three-surface airplane, with a wing, aft tail with pitch controls, and a canard. They built a conventional foam cored surface, built in a spar, skinned it, finished in the conventional ways. They built a hollow canard too, with a mold set, made a pretty light spar. The lightest way to build it hollow worked out to having outer plies, a core, and an inner plies plus cored ribs at the fuselage and tips. Well, all those inner plies and the adhesive to bridge the small but real gaps between spar and skin and tabbed joints adds up. In this case, the solid cored canard was significantly lighter than the "hollow" one, and so the foam cored one went on the airplane and made history. Scale comes into play. The canard for the Catbird is pretty small. So are all of the foam core foils on the Long EZ/COZY/Velocity/etc, and a lot of other small fast airplanes. So the cores don't add up to much weight there. And as evidence, when I first went to OSH, I walked around the Catbird ( I was awed to be able to just walk around this famous airplane), they had sawn the canard off of the starboard side completely. Why is another story, but you can see the edge of the foam core canard - they just sanded it flush and epoxied over the sawn edge.

Now at the size of Tony's Corsair (even at 82%, it is BIG), I suspect that flight loads may have taken his skins thicker than the minimum to survive the non-flight loads, so his skins might weigh less as a percentage of total airplane weight than mine do. We could very well find that the internal foam would be heavier than cored skins, ribs, adhesives, etc because his wings are so big too.

My airplane (wing area of 97 ft^2, aspect ratio of 8, and it is tapered 3:2) would have been lighter using solid cores, but I had to put 60 gallons of fuel someplace. Yes, I did the analyses, considered strake tanks, tip tanks, fuselage tanks, etc. The wetted area of the airplane would go up in most of the schemes, fuel in the fuselage had all sorts of CG impact, and I decided that I did not want the fuel in the same container as me following a forced landing.

Anyway, you have to consider the tradeoffs and figure out where your airplane stands relative to the break points. Also, please, please remember to optimize the airplane first, then the individual structures and systems. Doing it the other way around makes for poor airplanes.

Billski
 
Last edited:

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
Billski - Interesting you mention the internal corrugated webs - it's something I've been pushing for in several programs but the folks running the shows have always been rather mired in the metal concepts so up 'till recently I've only played around with it on my own.

However, now I have the sole responsibility for several new developments so I get to do what I want (within reason of course). The structure, as evidenced by the attached pictures, is rather beefy but it is designed as a stressed skin concept. By eliminating the typical cored configuration of the skins, the fabrication and assembly are much easier to the point that even with a few minor hiccups along the way the entire stab was assembled within a few days - the elevator took even less. The webs are in a Warren Truss arrangement, which provides significantly superior stiffness over that possible with just vertical webs. The parts are for an unlimited category racing aircraft (designed to 14 Gs) so that's the reason for the substantial gages in the laminates.

There is no spar - the entire laminate is made up of bi-directional prepreg. I did a similar structure for the wing but there I did have to go to a box type spar just simply due to several issues I ran into when doing the spar carry-through and its effects on the fuselage. And yes, the box spar does use uni in the upper and lower caps. But the rest of the wing interior is this type of corrugation.

One drawback to this approach is a bit more tooling investment but that is more than made up for in the time and cost savings on assembly. Also, for swept surfaces like the horizontal stab, the laminate is sort of a pain to smoothly get into the valleys. The company that did my lamination worked at applying the corrugation in one piece per 0-90 deg. layer and two pieces in the 45 deg. layers. That was a lot of work but they did a great job of it. For the wing though we came up with a simpler approach where the laminate is applied in strips.
 

Attachments

PTAirco

Well-Known Member
Joined
Sep 20, 2003
Messages
3,548
Location
Corona CA
And going round in circles: that wandering, corrugated web looks like something that would do well in a metal wing too. If you could form it accurately enough and used a fairly thick skin, you would not need ribs.
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
That's actually true and I've looked at that some time ago, along with bonding possibilities - the problem though comes in in the finish. For the thin skins we use for light aircraft the corrugation essentially prints through so you end up with a somewhat faceted finish. There is a possibility in making the assembly using thicker skins (.040" to .060", depending on several details) but then you have somewhat of a weight penalty and the thicker the skin the more you have to roll it to at least get it close to the final shape prior to bonding, which would be done in a female mold. With today's CNC sheet metal brakes I would foresee that this might be attractive even to tapered planforms. I just haven't taken it far enough to see how well it might work in practice. In theory though it sounds great.
 

Corsair82pilot

Well-Known Member
Joined
Jun 18, 2006
Messages
83
Location
MD
I like it. I feel like an idiot. Do you mind if I try the "internal corrugated webs" on my Corsair82?
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
6,976
Location
Saline Michigan
Hey Tony,

How So? You designed a big airplane using conventional methods. It might be a little conservative, but in my mind, conservative is not a bad thig. Could it be a little lighter? Maybe, but that would have entailed risk too, and in the wing on a one-off airplane, most of us have to be risk-averse.

This type of design, we could call it many things, including "distributed spar" or "whole wing spar" or "stressed skin wing", has some places where it just is not practical. You will note that none of the structures mentioned had to pass bending moments to another structure. In each case, the bending moments were all carried within stressed skin structure itself. And in each case only the lift is reacted out of the structure. That works fine for one piece stabilizers and wings, as well as for biplanes. It works OK for wet wing fuel tanks too. But it is not so good if you have to build the wing in several sections and then attach them together, passing bending moments from piece to piece. Or if you do not have the room to carry the whole wing through the fuselage. Or if you need to interrupt the distributed web in order to stow the landing gear, and so have to work the loads around the gear bay. Or, for that matter, find a place to install control bellcranks and the like. In these places, we have to figure out how to get the load from being distributed to being concentrated in one or two elements. The lightest and most reliable way may just be to use the old reliable spar and skin design with a minimum of ribs or a solid foam core.

Even the venerable Vari-EZ/Long-EZ and its many derivitives reflects this. The spar is built seperate from the skin, and bolts to the center spar, while the foam serves not as shear web but as one continuous wing rib.

A parallel development occurred in design of tactical aircraft fuselages. The stressed skin model was approached starting just before WWII, and continued into the 60's, but a funny thing happened. They ended up so packed with engines and fuel and electronics and other systems, all of which require removable panels for maintenance, that most of the fuselage ended up being a space frame with the removable panels forming most of the outside. So much for stressed skin.

Now, if your bird approaches something resembling production status, and you can build the wing in one piece, and the landing gear stowage method cooperates, and you can fit in the various flap actuation and aileron control systems, you could build your wing this way. I suspect that you may be able to use the corrugated spar for things like the horizontal stabilizer, flaps, and control surfaces. I also suspect that these might actually be best just with hotwired cores instead...

Naw, not an idiot. Just an guy living a world where the theory keeps progressing...

Billski
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
I like it. I feel like an idiot. Do you mind if I try the "internal corrugated webs" on my Corsair82?
By all means - I don't claim any ownership of the concept. Essentially this is just a combination of several ideas, the two primary ones being the stressed skin configuration combined with a modification of the multi-cell wing structure. The latter is simply the application of the Warren Truss to what is conventionally modeled as cells with parallel vertical webs. The angled truss walls provide the assembly with a bit more stability and torsional stiffness and of course are easier to manufacture since they come out of a mold much easier than if I tried to use the vertical wall layout.

One of the major benefits of this approach that I see is the elimination of sandwich cored parts. Most programs within our industry that I've seen do a pretty good job of building sandwiched components but the process can be lengthy since doing the entire layup in a single step, especially on a large part, can be a challenge. Most manufacturers using conventional layup techniques generally have to do a sizable sandwich structure in three distinct steps, which can get rather man-hour intensive.

Vacuum infusion makes the process a bit more efficient since it can allow the whole thing to be laid up in one shot but there is still the potential of getting dry spots, especially on parts that have expansive areas of core.

This approach works well with glass and graphite but of course does require tooling that's well laid out and lofted. Recent practical experience also shows that it is important to incorporate positioning buttons or similar devices into the tools since bonding under vacuum can actually distort the webs by flattening them out of shape - I bonded the elevator corrugation to a skin that was in its own mold. The front part of the corrugation was pinned in place but the aft portion was not, the result being that the pressure slightly distorted the aft-most trapezoid.

If I can take credit for anything it'll be for finally getting this idea out in the open and into practical application within our industry.

As far as its use on your project, Billski has some excellent points, namely in that the use of this type of assembly does require consideration up front since you do have to consider all the ramifications especially as they apply to things like mounting, attachment, internal bays, etc. I too had to make some changes when it came to the details of the wing and although my other projects will use this to greater benefit, practical matters sometimes need to take precedence. But if you can benefit from this approach, you're welcome to try. And of course if you have any questions, you can get a hold of me through here or my other contact information.
 
Last edited:

STDJantar2

Member
Joined
Jan 31, 2008
Messages
24
Hi,

This is my first post. I´m in the process of design and building a 15 meter sailplane. The sparless construction is an attractive idea, however I have some questions to the group:

1 - would the skin be built using a sandwich of bid and uni layers? or would just the BID layers suffice? How one can calculate the BID layer contribution to the bending moment resistance?

2 - how about the fuselage to wing attach points. This is an important feature in any sailplane since all have to be engineered to be taken apart and fitted into a trailer after a out landing.

3 - Would it be stiff enough to resist aeroelastic forces?

Wladimir in Brasilia - Brazil
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
Our initial application of the idea allowed us to baseline the concept using only Bid material however, as the design progressed, we discovered that for the wing we actually did need to build up a box spar that utilizes a combination of Bid and Uni fabrics. In our instance though, the design assumes a continuous carry-through structure (one piece wing).

The most accurate way to calculate the stress distribution within a composite wing is to use finite element modeling and analysis. In order for the program to provide you with the best data, you will need to have a good material property database so that the program has accurate physical constants for the number crunching. You'll also need a FEA program that has the capability to model layered elements and use orthotropic materials. It also helps if the program has the ability to output stresses in the material orientations.

The analysis can be done by hand but doing so is rather time intensive and is very prone to the introduction of errors. As such, the manual method will need to be somewhat conservative, meaning that you'll most likely end up with a heavier structure than what you possibly intended. However to do this in this manner also requires a good material database since you'll need to know the characteristics of the particular material you're working with.

The hard points are critical in the design and this will require a bit of specialized knowledge in order to make sure the added structure attaches with minimum impact on the rest of the laminate. The attachment can be done either through the addition of metal components or through the use of molded spar extensions that simply slide and pin into a fuselage mounted receptacle. The latter is common on a number of European gliders and is also used by Lancair in the attachment of their wings.

And yes, designed properly, the structure can be sufficiently stiff to meet your needs. Actually, the multiple cell configuration is substantially stiffer than a more conventional arrangement.
 

LBarron

Well-Known Member
Joined
Oct 11, 2006
Messages
122
Location
Irving, TX
Having built my horizontal and vertical tail on foam cores, I can tell you that putting in the shear web is not really a big deal. You hotwire the full form with low area for the caps, and then hotwire where the shear web is laid up. You jig and brace the foam that has the spar in it, layup the shear web, then fit and bond the previously removed parts. The foam cores are now much stiffer. Then you build up the spar caps, then glass the whole thing.
Billski
Billski,

Regarding the vertical stab, with your type of construction is the VS then attached to the fuselage by a simple butt joint, with the joint area faired and glassed? Without a spar attaching to a section deeper into the fuselage, is the load carried mainly by the outer layer of fiberglass?

Leland
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
6,976
Location
Saline Michigan
I have heard of folks just tabbing the v-stab to the rest of the airplane, but I looked at the moments and the shear stresses in the tapes and I needed something more substantial.

The horizontal and vertical tails are somewhat offset from each other to minimize inteference drag. The horizontal tail will be bonded into the top of the fuselage tub with a fillet and tapes from the fuselage skin. The horizontal stabilizer has a channel spar somewhat aft in the stab. The vertical stabilizer has a forward spar built onto the core that extends past the root end, with the flat side aft. While the vertical tail has a vertical hinge line, since the surface is tapered, the spar is slightly swept, and the horizontal stabilizer spar has a matching angle on its web. The spars from each will be bonded and bolted together, and then an aft spar will be laid up in the trailing edge of the vertical stabilizer and fuselage. Then to finish it off, the turtledeck will be bonded and tabbed to the tub, and more fillets and tapes will tie them all together.

Result: The H-stab will penetrate the fuselage and be glassed in with fillets and tapes. The V-stab will have one spar bolt and bond to the H-stab spar, and another spar tie it to the fuselage tub. And then more fillets and tapes will tie it all to the fuselage.

Billski
 
Last edited:
Top