Dealing with aft swept wing. General ideas.

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

Mac790

Well-Known Member
Joined
Jun 22, 2008
Messages
1,529
Location
Poznan, Poland
I'm collecting informations about aft swept wings. Somebody might wonder why I'm thinking about swept wings? Generally there are at least two reasons for it, first transonic speeds, second let's call it "problems" with C.G, in my case it's latter.

I could solve it "easily" with just one of the following approach:
- change design to more "conseravtive" one, and remove sweep completely
- try to move all unnecessary equipment in my case to the front of the fuse, to solve C.G problems
- move wing back, and decrease sweep
- incorporate bigger wash-out, that would be probably the best idea, but, the airfoil that I'm considering has much higher cd at lower (negative) AOA, for example cd at cl 0.1 is 2 times bigger that at cl 0.3.

I believe the key is to use each of above approaches (without first of course) partly, I mean move all possible components forward, move wing a little bit back, decrease sweep a little bit, add moderate wash-out. I was thinking also about possible mixing aerodynamic and geometric twist, but honestly I haven't heard about approach like that, are there any red lights for it?

I could also use more radical approach and add some unnecessary weight (lead) to the front of the fuse :).
Other ideas, boundary layer fence, dog-tooth, I'll need to check out also winglets influence.

Do you guys have other approaches for that, any good papers, there is at least one NACA Langley report about it, back from 30's that deals with tapered and swept wings, but the problem is that they were using thinner airfoil sections at the tip, with approach like that, even rectangular wing stall at the tip, so I wouldn't call it a very useful report, but I'll take a look at it again, maybe I missed something ( I read it long time ago).

To make things harder, the wing I'm thinking about has high AR around 13 at the moment, I want limit this thread only to problems with lift distribution, and wing tip stalls problem, I want exclude problems with pitching moment at the stall for swept high AR wings, and eventually stability issues.

Seb
 
Last edited:

Kristoffon

Well-Known Member
Joined
Sep 4, 2009
Messages
348
Location
Brazil
Re: Acceptable wing sweep.

Have you done a weight estimate for your wing yet? With AR of 13 the sweep will make the wing a lot longer than it already is and therefore heavier due to necessary added structure. Remember AR is determined by the span which is the straight line wingtip to wingtip.

How much sweep are we talking about?

A sketch of your aircraft would help a lot with possible solutions.
 

Mac790

Well-Known Member
Joined
Jun 22, 2008
Messages
1,529
Location
Poznan, Poland
Re: Acceptable wing sweep.

At the moment I'm looking rather for general rules of thumb, ideas, etc, which I could incorporate during first iterations, rather than for final solutions. My current details are based on a conceptual design software rather than on hand calculations, (I wasn't really considering designing an aircraft, so I'm currently more familiar with general rules than with particular equations, I have all necessary literature, and I'm going to change it as soon as time permit).

Currently we are talking about 10-15 deg sweep, I'm looking also at examples in GA category, I'm familiar with Ez family, for example Varieze has 26 deg swept and 3 deg twist.

Despite I have 3D renderings of the airplane based on those software calculations, everything that I can share at the moment is wing, it has 13 deg sweep in current configuration (AR should be in range 12-14, I tried few different ones), it isn't final version, it's doubtful that I'll finish with without any taper for outer section.

Like I said, looking for general ideas, and eventually some papers, ideally with some test (for example wind tunnel results).

I'm familiar with following links ( but those are rather for flying wings)
OTW index
Wing Design Parameters
The Panknin Twist Formula

I looking for a fish-rod, rather than for a fish.

Seb
 

Attachments

  • wing.jpg
    wing.jpg
    20.7 KB · Views: 777
Last edited:

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,312
Location
Grand Junction, Colorado
Re: Acceptable wing sweep.

The spreadsheet you can download here uses the Panknin formula too with some bells and whistles added (such as multiple taper). Scroll down to "Flying Wing Calc" there are imperial and metric versions. However, unless you're designing a flying wing, that formula may not be what you want. If your airfoils have a negative pitching moment you'll get huge numbers for twist at only 13 degrees sweep since the point of the formula is to get the Cm of the wing to zero and sweep also produces a negative moment. Balancing moments on a swept wing gets pretty involved (that's why there are 14 variables in the formula). It may be better to settle for some sweep induced moment and trim it out with the tail and/or CG. Your big challenge is going to be building a spar that's stiff enough without being too much heavier than the unswept version. Consider a deswept spar like they used in the SB-13. I believe there are a couple of articles in OTW about the SB-13 and one of them talks about the spar. In other ways I wouldn't recommend following their example too closely. They had to reflex the elevons 1 degree and add fences to fix a tip stall problem that could have been avoided with washout or some other attention to the stall characteristic of the outer wing in the design (Rutan uses vortilons for exactly this reason but at least they are very low drag unlike fences and reflexed elevons). Even with the deswept spar and the aerodynamic ironmongery it still had a speed limit because of a flutter problem. I haven't heard anything specific but I wouldn't be surprised if it's body freedom flutter because the spar is too heavy.


 

Attachments

  • jwra.JPG
    jwra.JPG
    36.1 KB · Views: 1,033

Topaz

Super Moderator
Staff member
Joined
Jul 29, 2005
Messages
14,943
Location
Orange County, California
There is a NACA paper that modifies Schrenk's Approximation to swept wings, and with a surprising degree of accuracy compared to some of the simpler numerical methods. I have a copy, but I'm away from my office and don't have it here. I'll post a link to it when I get back on Monday.

So I take it the general problem is that your wing is too far forward for the current CG location? If that's the case, then looking at the options you posted, my 'priority list' would be:

1) Try and move the wing aft
2) Try and move heavy items (battery, engine, people) forward
3) *Minor* sweep of the wing (<=10deg.)
4) Trash the overall configuration and start over.
...
...
...
5) Everything else on your list.

Some ideas just don't work out. I have notebooks full of designs that were ultimately discarded because some significant aspect simply wouldn't work. As amateur, beginning designers, we don't have the experience to 'do it right the first time' as often as someone like Orion or one of the other professional engineers.

As others have said, seeing at least a sketch overview of the configuration would be very helpful. It doesn't have to be CAD or pretty, just representative.
 

Mac790

Well-Known Member
Joined
Jun 22, 2008
Messages
1,529
Location
Poznan, Poland
Thanks guys, great comments so far, this exactly what I was looking for :).

Norman,

I've heard about SB-13 and it's problems, this exactly what I want to avoid, the good thing about "my" wing is that the outer sections has only slightly more than 3m in length, so it should be "easier" to design them properly from structural point of view, than those 7.5m for a SB-13. I have also lower AR, and less sweep, I think I'll reduce sweep to around 8-10 degs for the first real iteration, and move wing a little bit aft. Thanks for that attachment with those fences, I completely forgot that Eze's also have them, but if I remember correctly it was used only for planes with modified canard airfoil, it wasn't necessary for airplanes with original GU airfoil, it was something about different downwash of a new canard airfoil, but like I said I'm missing details now.

Topaz,

If you can send those papers I'll be very grateful, of course you don't need to hurry, I don't need them now, I just checking out some possibilities, and I'm pretty excited, this is first in loonngg time configuration that I'm happy about.
So I take it the general problem is that your wing is too far forward for the current CG location?
Honestly, with 13 deg sweep, and fuel there is no real problem, but it's currently based only on software results, I noticed some bugs in it, I'll have to check it out manually, but it will take some time, like I said above I wasn't concerning designing own craft, so I have some serious gaps in some areas (even in some of those basic ones too), at the end I'll be looking at someone like Orion to check it out, but it's future $$, I just want to learn and make as much as possible on my own.
Regarding to the list of changes proposed by you, I have similar feelings, first I'll try to move wings aft, and use smaller sweep, let's say around 8-10 deg, and will see what I'll get.

The general idea for crating this thread, was to get some inputs, I just don't want to put myself in the corner, like you know reading literature for a year, doing calculations for another one, and after that you are realizing to you have flatter issues at 150 mph because of huge sweep, etc.

Thanks again, if someone can recommend, better literature than Raymer for sweep wing I'll be grateful.

Seb
 

Topaz

Super Moderator
Staff member
Joined
Jul 29, 2005
Messages
14,943
Location
Orange County, California
I've got the paper in PDF form, and know right where it is. I'm just not where that computer is, and I don't have a copy on my 'cloud' storage. I'll shoot you a copy on Monday. Won't be a problem at all.
 

Mac790

Well-Known Member
Joined
Jun 22, 2008
Messages
1,529
Location
Poznan, Poland
3) *Minor* sweep of the wing (<=10deg.)

I was wondering where is that 10 deg, coming from, from practice (your calculations), from literature, etc, could you say a little bit more.

I was looking for some info, and I found an interesting discussion at RC page (unfortunately there isn't much about it in those books that I have, actually there is very brief calculation example for aft swept wing, but they only concentrate on the pitch-up problems during high G's, solution to that is dog-tooth), here is link to the most interesting post in it RC Groups - View Single Post - My theory on why swept wing tipstall,
TFLG said:
The problem with using washout for this situation is that the effect is highly non-linear, so that if you try to correct it with linear washout, the tips may be ok and the root may be ok, but the lift in the mid-span will be wrong. Ideally you should have non-linear washout,

this one isn't also very encouraging (in relation to RC again)
stumax said:
I spent months of computational fluid dynamic analysis time analysing wing flow fields, and this was what I learnt....As for swept back wings, don't do it if you're flying subsonic unless it's a scale model.

It's now a little bit clearer why there is almost no aft swept wing in GA sector besides Ez family, unless we are talking about small sweep, like 5-8 deg, which can be found in numerous aircraft, like Tiger Moth, Fly Baby Biplane, and probably in a few more. So I think that for a first iteration I'll reduce sweep to around 5 max 8 degs, move wing aft, and try to play with C.G I have a couple ideas, of course besides using lead ;).

edit found two interesting papers, this time for real airplanes

http://naca.central.cranfield.ac.uk/reports/arc/rm/2908.pdf
http://www.gdt-systems.com/files/SubsonicAirfoilsAndWings.pdf

Seb
 
Last edited:

Topaz

Super Moderator
Staff member
Joined
Jul 29, 2005
Messages
14,943
Location
Orange County, California
There isn't any hard and fast dividing line on sweep. I had to pull a number out of the air and ten degrees is a nice conservative one.

All the reasons about tip stall come down to this: the tip is more highly loaded than the root. The reasons for THAT can vary with sweep angle (at large sweeps), but the lift distribution is the fundamental key.

With all due respect to those RC fans here, the RC world is filled with so much mythology about aerodynamics that I certainly wouldn't use their information as a design resource. Exceptions exist where sound design principles are used (our Norman comes to mind immediately) but by and large the majority of their 'theories' are more eyeball 'engineering' than anything useful.
 

Jay Kempf

Curmudgeon in Training (CIT)
Supporting Member
Joined
Apr 13, 2009
Messages
4,597
Location
Warren, VT USA
There isn't any hard and fast dividing line on sweep. I had to pull a number out of the air and ten degrees is a nice conservative one.

All the reasons about tip stall come down to this: the tip is more highly loaded than the root. The reasons for THAT can vary with sweep angle (at large sweeps), but the lift distribution is the fundamental key.

With all due respect to those RC fans here, the RC world is filled with so much mythology about aerodynamics that I certainly wouldn't use their information as a design resource. Exceptions exist where sound design principles are used (our Norman comes to mind immediately) but by and large the majority of their 'theories' are more eyeball 'engineering' than anything useful.

Cos 10° contributes over 98% of the original value to the calculation, so I think if the function to describe lift distribution is based on a cosine function 10° is a good rule of thumb. The lift distribution of a swept wing is not calculated using the actual length of the spar for instance but the lateral distance between tips so it works. Compressible flow works the same way. Anything approaching 100% of the original number is considered equivalent to the original number. Picking a reasonable tolerance is just how an assumption works in an engineering calc. It can be thought of as garbage in garbage out. Or understanding the accuracy going in allows and understanding of the accuracy going out. Flutter is probably the overwhelming variable as you sweep the spar off of the AC.
 

Mac790

Well-Known Member
Joined
Jun 22, 2008
Messages
1,529
Location
Poznan, Poland
Mac said:
I completely forgot that Eze's also have them, but if I remember correctly it was used only for planes with modified canard airfoil, it wasn't necessary for airplanes with original GU airfoil, it was something about different downwash of a new canard airfoil, but like I said I'm missing details now.
I'm still missing details about it, as I haven't found that particular discussion/report (or whatever it was), but it doesn't have much in common with discussed here conventional airplane. In canard airplanes of course canard have to stall before wing, so in short wing never see those high AOA, it never stalls ( at least in theory) so it's really hard to say is it that 3 deg washout, and fences for this particular wing are good enough to prevent tip stalls, those wings work fine for Ez family, but this same wing might not work in conventional design, in which wing "see" those high AOA.

Topaz said:
All the reasons about tip stall come down to this: the tip is more highly loaded than the root. The reasons for THAT can vary with sweep angle (at large sweeps), but the lift distribution is the fundamental key.
Lift distribution is important for sure, but one more thing must be taken into account, its' spanwise-flow, for the swept forward it flows from tip to root, for aft swept wings it flows from root to the tip. Seems it's more visible during lower speeds or higher AOA, for example I've read report about Pond Racer accident, I remember story about flames from engine coming over a canopy (caused by a spanwise-flow), it vary with a sweep angle, so it's probably higher for bigger sweep, and lower for lower sweep, but it exist, and it really doesn't matters how much sweep we are talking about 5,10,15, 25 etc, degs. The key question here is, when we should start to worry about it?

There is also one more problem which might be caused by spanwise-flow, if stalls for whatever reason starts at the root or in the middle of the span, this spanwise-flow might move it to the tip, for this reason we probably see those fences, dog-tooth, etc.

I still looking for more info, but it seems that the best (safest) idea for aft swept wing is to fly backwards :roll:

It's also interesting to see a difference between F-18 Hornet and F-18 Super Hornet, in the SH they used dog-tooth, probably to improve performance at lower speeds and/or at higher AOA, if you take a closer look you will notice a small drift of boundary layer caused by dog-tooth, probably in this case at high AOA, and I'm not talking here about those created by wing tip and strakes, which are much larger.

Seb
 

Attachments

  • span.jpg
    span.jpg
    71.5 KB · Views: 3,839
  • F18fBankingFrontView.jpg
    F18fBankingFrontView.jpg
    22.9 KB · Views: 512
Last edited:

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,312
Location
Grand Junction, Colorado
A Dog-tooth does pretty much the same job as a vortilon. Granted the wing outboard of the dog-tooth has more camber forward than the inboard part but it's the vortex that makes the big difference. This has been demonstrated by filling in the step to get rid of the discontinuity. Such a modification allows the span-wise flowing attachment line to flow to the tip instead of being kicked up onto the top surface. When the LEV is suppressed the stall delaying effect is also suppressed.

4-19.jpg
 

Topaz

Super Moderator
Staff member
Joined
Jul 29, 2005
Messages
14,943
Location
Orange County, California
Okay, Seb, here's the link to that NACA report on my server: http://bit.ly/g4Dqcw

For anyone that downloads this: This is a rough, approximate method, useful for a quick general determination of the lift distribution of swept wings. Simple enough that you can set it up in an Excel spreadsheet if you do curve-fits to the sweep-effect curves in the paper. I built one that extrapolates between the curves as well for any sweep value from 0°-45°, but I don't recall where that is at the moment. I'll post it if I find it.

The method posted here isn't good enough for a final design, but might help you get 'into the ballpark' for more accurate methods later.

Seb, since this is an empirical adjustment to Schrenk's, taken from wind-tunnel data, it probably already accounts for any span-wise flow situation. Obviously it won't handle fences, vortilons, or other flow-control devices, but it's a good first-cut for a 'clean' swept wing.
 

Mac790

Well-Known Member
Joined
Jun 22, 2008
Messages
1,529
Location
Poznan, Poland
Norman,

Thanks for clarifying it a little bit more, seems that my description was a little bit brief, just in case for some of us 2 more sketches do describe it a little bit more. I believe that Orion would classified them as a fixes, rather than part of design process, it's really hard to find modern plane with a dog-tooth.

I had some fun in relation to that today, first I found in one book following sentence:
Unusual change of sweep usually shows that a designer had problem with the C.G, after it was to late to hang the wing anywhere else
next I found Orion comment in relation to dog-tooth (I'm backing myself that he was talking about straight wing, but I believe he would has similar opinion about dog-tooth in relation to swept wings).
Orion said:
...Moving to your wing, like it or not, cuffs are either fixes or an up front admission that you are unable to design a proper wing. Even if a few technical papers say otherwise, looking at industry history, all cuff use is to fix a problem and that's simply how they're perceived.
So it seems that thinking about dog-tooth or similar tool, in relation to that swept wing, it's like fixing a fix in this situation :gig:.

Topaz,

Thanks for those papers, I'll take a closer look at them ( I just read them briefly), should be useful in the beginning of design proses, maybe even later too.

Seb
 

Attachments

  • dog-tooth.jpg
    dog-tooth.jpg
    45.3 KB · Views: 6,282
  • dog2.jpg
    dog2.jpg
    63.4 KB · Views: 553
Last edited:
Top