I really like the idea of low aspect ratio.
The problem of low aspect ratio
But they have one major problem. Bad characteristics at low
speed. Well, they can fly slowly, but they need large angles of
attack to obtain this low speed. This leads to high landing
What is the main reason? Span efficiency! If you reaction is
"Huh?", let me explain. The vortexes created at the tip of a wing
influence the airflow over the wing. The smaller the span area that
is influenced the better the span efficiency.
The created vortexes of low aspect ratios are huge at low speed.
These huge vortexes influence a certain area of the very short
span. This leads to bad span efficiency.
I hope to show a way to resolve this problem.
Many of you have already seen winglets. Some of you have seen
experiments using "feathers" (I haven't found any official name, so
I call them feathers and will do so for the rest of the text).
How do I hope they work?
So, under the wing there is a positive pressure and above the
wing there is a negative pressure. The positive pressure will of
course leak towards the negative pressure. This happens along the
tip of the wing. Wingtips have a certain form and according to that
form a certain length of space where the air can leak to the other
side of the wing. The shorter this space the more force the leaking
air will have (I don't mean that high aspect ratios (= long wings
with short airfoils) have larger vortexes then low aspect ratios) .
Imagine you push a bucket of water over that was standing on a
table. At the table edge each centimeter or inch will get a certain
amount of water over it. If you pushed the bucket towards the short
side of the table each centimeter or inch will have to swallow more
water than if you pushed it to the long end of the table.
Now, how do you increase the wingtip edge length? You could do
that in two ways. The first way is to increase length of the tip
chord. But that would be difficult in construction. The wingtip
would get more lift than the root chord. This could leads to extra
forces on the spar at the root. The second way to increase the
length of the wingtip edge is to deform the wingtip. By making
featherlike extensions on the tip you made that edge longer. There
are more centimeters or inches along the wingtip.
I hope that this will reduce the force of the leaking air. If it
does it will also reduce the force of the vortexes. And that would
make that the airplane has a better span efficiency.
A second thing about the feathers that could help reduce the
vortexes is the use of several leaks instead of one leak. Let me
explain. An unfeathered wingtip has one major leak, the entire
wingtip. A feathered wingtip can have more leaks. I don't mean that
the leak power is greater, but that between each feather you will
have a leak. If you only have one leak, this certainly will lead to
a vortex. But two leaks next to each other disturb each other in
creating a vortex. Just try it out in water. Making quick circles
with one finger in the water leads to a vortex. Now try doing this
with two fingers. Just try to create two vortexes next to each
other. It will be more difficult. So when you make experiment, make
sure to use several feathers (definitely more than 2).
Scale model testing
There is one thing that also needs to be tested by models. Do
the feathers need to be installed on the top of the wingtip (as
seen in the drawings) or must they be placed on the bottom? I
cannot imagine what the difference will be, but I am sure there
will be a difference.
One mistake you may not make is the next. Do not turn one of
these feathers into an aileron (= control area which creates a roll
maneuver). The spaces between the ailerons will lead to leaks at
unwanted places. That is why I drew a kind of V-tail. I wanted to
cooperate all control areas into this tail. Kind of a combination
off elevons (= combination of elevators and ailerons) and a
vertical tail. I didn't have an opportunity to test it. But I leave
that fun to those who want to try out this idea into some
Feel free to use this idea in your models. Please report me any
good or bad flight results.
Some people mentioned to me that there already is such a design.
It is called WingGrid. The glider has rectangular
"feathers". An article tells that the results were promising.
I was glad to see that my thought about these vortex-reducing
feathers actually works (imagine a man dancing while singing "it
works, it works, it woooorks").
Notes from viewers
I got this remark from Serge Krauss: "like your
bird-wing (feathered) idea, but believe that it would change the
nature of the low-aspect-ratio plane. The advantage of high CL from
high angles of attack comes from the vortices rotating the air
stream back over the wing to re-attach the flow over a great part
of its (otherwise stalled) area. It may be that in the case of ARUP
shapes, with round trailing edges, the vortices actually meet at
the aft center wing. So if you interrupt these large vortices, you
may raise the CL and lower the CD at lower angles, but prevent the
wing from reaching those very high CL's at high angles of attack by
reintroducing the stall."
I replied: "I like to react on your comment on the
Birdwing-idea. OK, I see now that the stall speed will increase due
to a lower CL max. I even did find on the new site of Winggrid that
the test plane had a higher stall speed. So, it proofs that you are
right. Anyway... I did not doubt your remark. You know your
One thing you might not know: an American is working on testing
the idea with models. He did build a small glider and found it
"remarkably stable". I forwarded your remark and he replied that he
still is interested in the idea. Just love the guy. :)
He hopes to develop the idea into a powered ultralight. He hopes to
get an ultralight with a larger speed range than the current
So, the wing will stall earlier. But the low aspect ratio could
be built lighter than an airplane with the same wing area and a
conventional AR (I think). Wing loading gets lower. Lift needed to
take-off is less. Could a higher speed (higher than the
pure-very-low-aspect-ratio-plane-stall speed) and a lower Cl result
in the same stall speed as the conventional airplane? If the stall
speed can be kept the same, one would benefit from the other pros
of the concept.
Other pro's are:
- Ease in
construction (all part (ribs for instance) are bigger and easier to
construction is better crash survivable
- The hanger
can be smaller
- ... And the
looks are better. :)
"His reaction was: "Sounds like a good plan to me! I hope to
learn the outcome of the experiments. That the Horton "wingless"
seemed to work well with its "tip" plates indicates that you may be
on a good path."
In the August 2001 edition of the magazine Kitplanes I found a
article by Howard Levy about the Marsden Skylark.
The Skylark has two winglets on each wingtip. At the first flight
the Skylark had no winglets, quickly one was added, soon a second
was added. The first winglet is placed at about 1/3 of the
tipchord, the second is canted at 45° at is placed at the rear of
the tipchord. David Marsden (of Edmonton, Alberta, Canada) mentions
that the use of the second winglet showed a further improvement of
further about 50% over the single winglet. The vortex gets split
into three parts.
This shows again that maybe my Birdwing-idea can be used in
planes with more normal aspect ratio.
If somebody can get me into contact with David Marsden I can
learn more about his winglets and ask his opinion about my
Birdwing. All extra info will be placed in the site.
On 26 May 2004 I got a mail from John G. Tarsikes,
Jr. He talks about my Birdwing. I guess he suggest several
good things, like ...why not start building it accually. Here is
"Well, I saw the Birdwing and thought you were sneaking in and
looking at my old designs! I have recently renewed my interest in a
design I worked up about twenty years ago, so I was "Googling" and
hit on your site.
I have always been intrigued by low aspect ratio designs in that
they afford such a low weight to strength ratio and low production
cost. You could reasonably construct a proof of concept aircraft
for about $400 USD plus engine. What are you waiting for?
Points to consider:
- Do not worry
about the high angle of attack on approach. While this may be
disconcerting, but is really not a problem.
- Keep the
proof of concept a tractor. The wing will perform longer and
respond quicker with the prop moving air over the airfoil. This
could be a life saver on those early flights while a learning curve
is being developed.
- Select a
nice fat auto-stable airfoil. Basically you are building a "control
wing" aircraft without a fuselage. To go up, push in the throttle.
The flare or extreme slow flight is the only phase where serious
pitch inputs will be required. WIG will be a factor in the landing
- Keep it a
tail-dragger. Design it to sit on the ground a degree or two above
the MAX angle of attack to get the most useful performance. It
should land slow enough and short enough at that angle to land into
the wind even across a runway. The crosswind component will be to
high to land any other way, so a trike is only useful to any degree
on a pusher, and you should not tackle a pusher on a concept
aircraft as radical as this anyway. Too many radical concepts
applied at once are a sure formula for disaster. Keep it SIMPLE,
then add new stuff when the data is in on the least radical
configuration. C/G limits will put the pilot in a fine position in
the airframe to have unrestricted visibility even in the flare.
Forget a pilot pod. Just glass the LE (Ed.: LE = "Leading Edge" or
in other words the front of the wing).
My goal was to keep the span down enough to make the airframe
street legal as a trailer. Tow it home and stuff it in the garage.
The only tear down would be to fold the rudder-vators inward. Given
these dimensions I come up with a 12 G airframe weighing in
complete, less engine, prop and fuel just below 80 pounds.
The control mixing circuits are the most daunting construction
items. Everything else is soooo easy, it should only take eight to
ten hours to construct the bare airframe.
I use bolted 6061 T6 angle for all my designs now. It is the
easiest, quickest and cheapest way to go. My first flyer with
this technique spanned 24 feet and was 19 feet long. The completed
fuselage was 36 pounds. I quit crunching number when I could not
find a cluster not capable of ten Gs + and -. I had not set out to
design an ultralight, but the aircraft ended up weighing only 246
Many thanks for this letter, John. It kind of sparked the old