52 NOVEMBER 2009

Jim Koepnick
NEAL WILLFORD, EAA 169108

Canard design considerations

G
rowing up, my annual vacation was
attending the EAA convention with
my dad and brother. I consider myself
fortunate, but perhaps the biggest drawback of
frequent attendance is that the memories started
blending together—unless something profoundly
unique stood out.
As a 12-year-old attending Oshkosh ’75,
two things did just that. The first was the hot,
dry weather. The second was a unique little
homebuilt parked on its nose. It was Burt Rutan’s
prototype VariEze. It had a canard or tail-first
configuration, was powered by a pusher engine,
and featured a moldless composite-construction
method pioneered by Rutan. Plans for a refined
version of the VariEze were made available the
following year, and it became extremely popular
with homebuilders. Over the next 10 years . . .

EAA Sport Aviation 53

. . . the Rutan Aircraft Factory (RAF) developed several more by the National Advisory Commit-
tee for Aeronautics during the 1940s
canard designs, including the popular Long-EZ. Rutan stopped
found that aircraft with a stick fixed
selling plans in 1985 to focus efforts on his other company, static margin of at least 8 percent
Scaled Composites. Though plans for Rutan’s canards are no MAC had good flying qualities. Any-
longer for sale, there are a few Rutan-inspired designs still thing less than 3 percent MAC was
considered dangerous.
available to the homebuilder. An aircraft’s neutral point is influ-
Decades before Rutan, the Wright When properly designed and enced by a variety of factors, includ-
brothers’ 1903 Flyer was the first built, the canard configuration can ing the size and aspect ratio of the
successful application of the canard provide stall departure resistance. wing, canard or horizontal tail, the
configuration for powered flight. This can be a challenge and should distance between those surfaces, the
According to a 1986 article by Seth not be taken lightly by the aircraft fuselage shape, and power effects.
Anderson in a NASA publication, designer. With that caveat, let’s take A stable aircraft also requires that
the Wrights selected it based on a look at some of the design consid- the pilot be able to trim at a desired
intuition and Wilbur’s concern that erations for canard aircraft. airspeed, or at least hold it in trim
Otto Lilienthal’s fatal accident was All naturally stable aircraft are con- with a reasonably low force. For most
associated with his glider’s aft-tail sidered stable in pitch as long as the conditions this requires some amount
configuration. The early Wright center of gravity (CG) is located ahead of load on an auxiliary flying surface
aircraft were not stable in pitch, of the aircraft’s neutral point. The (either a horizontal tail or canard).
though, and in 1909 Orville urged neutral point is defined as the loca- An exception to this is a flying wing,
Wilbur to adopt an aft-mounted tail tion along the wing’s mean aerody- which can be trimmed through the
to improve stability. Orville did add namic chord (MAC) where a change careful selection of airfoil(s), elevator
an aft tail to the canard-equipped in angle of attack does not result in a deflection, and/or a combination of
Wright Model A in 1909, and by change in pitching moment. A pilot wing sweep and twist. However, for
1910 the brothers had dropped flying with the CG at this location our discussion we are going to ignore
the canard altogether. By that time would find that the aircraft would this option.
most other aircraft designers had not tend to return to the trimmed air- At first glance, you might think
also been migrating to this configu- speed if disturbed in pitch, either by a that in trimmed flight a canard sur-
ration. In the following decades, gust or control movement. Instead, it face with the same area and location
canard aircraft were designed and would tend to stay at the new angle ahead of a wing would carry a simi-
built by a variety of designers, but of attack. This would require constant lar load as a horizontal tail located
not with the level of success Rutan pilot attention and be annoying and the same distance behind the wing.
had with his designs. Canard: A Rev- potentially deadly. Consequently, the This is not the case, as we will see
olution in Flight, by Andy Lennon, aft center of gravity limit is set ahead in a moment. An isolated wing has
provides an interesting historical of this location to ensure that the air- a neutral point at its aerodynamic
overview of canard aircraft for those craft has what aerodynamicists call a center, which is the location where
wanting a closer look. “positive static margin.” Flight testing the wing’s lift is applied and also the
pitching moment is constant. The
aerodynamic center is at 25 percent
Jim Koepnick

MAC in theory and doesn’t vary too

far from this point in reality. Adding
a horizontal tail behind the wing
moves the neutral point aft of the 25
percent position, with the exact loca-
tion depending on the wing’s down-
wash characteristics as well as the
size and location of the horizontal
tail. Since stability requirements dic-
tate that the CG needs to be located
ahead of the neutral point, the end
result is that the CG is generally close
to the 25 percent MAC position for
a conventionally configured aircraft.
The situation reverses when a sur-
face is placed ahead of the wing, with
the neutral point now moving ahead
of the 25 percent MAC. Estimating
the amount it moves forward is more
complicated because the wing is par-
tially in the canard’s downwash, and
Due to its unique center-of-gravity configuration, the Long-EZ parks with its nose it’s best determined through wind
wheel retracted. tunnel testing or using a computer

54 NOVEMBER 2009
aerodynamic vortex lattice or panel
method program. Assuming that the
canard aircraft’s neutral point loca-
tion is known, placing the CG ahead
of it for stability will result in a CG
well forward of the 25 percent MAC.
It’s pretty straightforward to deter-
mine the load sharing between the
wing and horizontal tail or canard
required to trim an aircraft once the
CG location is known. Figures 1 and
2 show how the load sharing require-
ments differ between the two con-
figurations. Each figure compares the
percentage of canard or horizontal tail
area (in terms of the combined area
of the wing plus the auxiliary surface)
to the percentage of total lift that the
surface needs to keep the aircraft in
trim. The estimated trends in both fig-
ures were generated using a computer
vortex lattice program and include the
following assumptions:
ate a moment that tends to twist
the wing’s leading edge down and
trailing edge up. This nose-down
moment is reacted by the horizontal
tail and reduces the amount of lift it
provides. Most airfoils used on light
aircraft have a CM between 0 and
-0.1 (flaps up), which becomes more
negative with flaps down.
Depending on the horizontal tail
size and wing pitching moment coef-
ficient, Figure 1 indicates that a con-
ventional aircraft can have a slight
lifting load in flight. For those cases,
the CG is behind the wing’s 25 per-
cent MAC, and consequently the
horizontal tail has to provide a lifting
load to trim the aircraft. Moving the
CG forward increases the static mar-
gin and has the effect of shifting the
curves down in the figure, and it may
result in a zero or slight downward
load on the horizontal tail.
One of my initial assumptions was
that the wing and either horizontal
or canard have equal aspect ratios.
This is normally not the case for
aft-tail configurations, but they were
kept equal for this study to keep the
number of variables to a minimum.
Using a lower aspect ratio for the
horizontal tail results in the same
general trends, but with the curves
shifted down some.

• CG located 10 percent MAC

ahead of the neutral point.

• Wing and horizontal tail/canard

have an aspect ratio of 7.5.

• Tail arm equals 2.5 times the

wing MAC.

• No fuselage or power effects.

Figure 1 shows the trends for

a horizontal tail configuration, and
several observations can be made.
The first is that the percentage of
lift the horizontal tail provides goes Figure 1. Approximate aft horizontal tail lift requirements.
up with increasing tail size. This is
because increasing tail area moves
the neutral point farther aft from the
wing 25 percent MAC. And since I
kept the CG a constant 10 percent
MAC ahead of the neutral point, this
means that the CG moves farther
away, too. The farther the CG is from
the wing’s 25 percent MAC, the great-
er the load the auxiliary surface needs
to provide. Though the trends are for
a large range in tail area percentages,
most “conventional” aircraft have
horizontal tail areas in the 15 percent
to 20 percent range. Consequently
we will focus on that range for the
other observations.
The second observation is that
the horizontal tail load is affected by
the magnitude of the wing’s pitch-
ing moment. Most airfoils have a
negative pitching moment coeffi-
cient (CM), meaning that they cre- Figure 2. Approximate canard lift requirements.

EAA Sport Aviation 55

While the tiny tailwheel pant/rudder of the Quickie in its original form was perfectly adequate in flight, it resulted in demanding
ground handling characteristics. A taller vertical fin and a conventional rudder surface made the Quickie easier to fly and taxi.

The last observation from Figure moment coefficient becomes more to scale effects, and aerodynamicists
1 is that the percentage of lift (either negative due to deploying flaps. This found that they can quantify those
up or down) on the horizontal tail is is one of the reasons flaps are rarely effects by examining an airfoil at dif-
quite a bit less than its percentage of used on canard designs. The Beech ferent Reynolds numbers. Reynolds
total lifting surface area. The implica- Starship was one of the few that did number is the ratio of the inertia
tion is that the horizontal tail’s lift use wing flaps, and it incorporated a forces of the air flowing around an
coefficient requirements are lower forward-swinging canard to increase airfoil to the viscous forces of the
and historically have been met by distance from the wing to the canard. air. The inertia forces depend on the
using a symmetrical airfoil equipped Other designs have used speed brakes air’s density and speed, whereas the
with an elevator. located on the belly of the fuselage to viscous forces are a measure of the
Figure 2 shows the estimated allow steeper approaches for landing. air’s “stickiness.” At sea level, stan-
load-sharing trends for a canard The final observation is that unlike dard-day conditions, the following
configuration. The major difference the aft horizontal configuration, formula can be used to estimate the
from Figure 1 is that the canard the percentage of canard loading is Reynolds number of an airfoil:
always has a positive lifting load. always higher than its percentage of
As I mentioned ear- total lifting surface Reynolds number ≈ 9360 x airspeed x chord
lier, this is because area.
the canard aircraft’s The canard’s higher lift The canard’s …where the airspeed is in miles
neutral point (and requirements mean that higher lift require- per hour and the wing chord is in
resulting CG loca- ments mean that feet. For example, the Reynolds num-
tion for stability) is careful attention must be careful attention ber for a wing on a light aircraft can
always ahead of the must be paid to range from 2 million at stall speed
wing’s 25 percent paid to the the design or selec- to 7 million or more at top speed.
MAC. Focusing on design or tion of the airfoil. However, a canard’s Reynolds num-
the same 15 percent Designers often use ber on a homebuilt can be around
to 20 percent area selection of a high aspect ratio 500,000 or so. Normally an airfoil’s
range we did earlier, planform for the maximum lift coefficient goes down
you can see that the the airfoil. canard, because with decreasing Reynolds numbers.
canard would need it stalls at a lower Rutan initially used the same airfoil
to provide roughly 20 percent to angle of attack than a comparable (NASA’s GAW-1) for both the wing
30 percent of the total lift to keep lower-aspect-ratio surface, and you and canard on the prototype VariEze.
the aircraft in trim. Moving the CG want the canard surface to stall at a Initial flight testing showed that the
farther forward would result in shift- lower angle of attack than the wing. aircraft had a high stall speed, so he
ing the curves up some, indicat- The consequence is that the canard’s installed a new canard with an airfoil
ing an increase in canard loading. chord length tends to be short com- specifically designed to operate at
The canard lift requirements are fur- pared to the average wing chord. low Reynolds numbers. This airfoil
ther increased if the wing’s pitching An airfoil’s characteristics are subject had a rather unwieldy designation of

56 NOVEMBER 2009
GU25-5(11)8 and is shown in Figure
3. Subsequent flight testing demon-
strated an 8-knot decrease in mini-
mum flight speed with the canard
airfoil change. More details about
the development of the VariEze and Figure 3. Canard airfoil used on the VariEze and Long-EZ.
Long-EZ can be found in Burt Rutan’s
article “Tale of Three EZs,” published
in the February 1980 edition of EAA
Sport Aviation.
It is also important that the
canard airfoil has gentle stall char-
acteristics. NASA wind tunnel test-
ing on a powered, full-scale VariEze
(“Wind-tunnel investigation of a
full scale canard configured general
aviation airplane,” NASA TP-2382,
Lon Yip, 1985.) included a con-
figuration with the GU25-5(11)8
canard airfoil being replaced with
a 12 percent thick symmetrical
airfoil. The data showed that the
“stock” VariEze had good post-
canard stall characteristics. Replac-
ing the original canard airfoil with
a symmetrical one showed that the
aircraft had a lower maximum lift
coefficient (which would lead to a
higher stall or minimum speed), an Figure 4. VariEze calculated wing lift coefficient distribution during climb.
abrupt stall of the canard, and an
undesirable pitch-up tendency in
the post-stall range.
One more consideration in select-
ing a canard airfoil is to under-
stand what the effects of rain or
other surface contamination may
have on the aircraft’s flying quali-
ties. Rain, bugs, or even paint stripes
near the leading edge can prema-
turely trip the laminar flow on the
canard surface and cause an increase
in minimum flying speed and a
change in the trim requirements to
maintain the desired speed. NASA
investigated the effect of water and
a fixed transition on the canard as
part of its VariEze wind tunnel test.
Burt Rutan wrote “Effects of Rain
or Surface Contamination on Pitch
Stability and Control” in EAA Sport
Aviation, March 1983. The article
Figure 5. Drag area comparison for several canard aircraft.
provides some additional detailed
flight test results for having an early,
fixed transition on several Rutan air- with the R1145MS airfoil designed size to the wing. Though the larg-
craft. Sometimes rain effects require by John Roncz. According to RAF’s er canard results in the canard air-
a modification to the aircraft, as quarterly newsletter, the Canard foil operating at a higher Reynolds
was the case when vortex genera- Pusher, this optional canard offered number, care still needs to be exer-
tors were added to the canard of the a negligible trim change in rain and cised in selecting an airfoil. In his
around-the-world Voyager. RAF also a 2-knot increase in stall speed under article “Quickie-Type Aircraft Design
tested several canards with differ- the same conditions. Origins” (EAA Sport Aviation, Octo-
ent airfoils for the Long-EZ to better Some canard designs like the ber 1981), Burt Rutan discussed the
understand this phenomenon. The Quickie, Q2, Dragonfly, and PAT-1 background of the development of
end result was a canard equipped Pugmobile have a canard similar in this tandem-wing arrangement. The

EAA Sport Aviation 57

article “Wind-tunnel investigation of
an advanced general aviation canard
configuration,” NASA TM-85760,
written in 1984 by Joseph Chambers,
Lon Yip, and Thomas Moul, presents
the wind tunnel results for a powered
one-third scale model of the PAT-1.
The wing design for a canard-
equipped aircraft also requires careful
attention. The canard downwash has
a noticeable effect on the wing’s angle
of attack along its span. Figure 4
illustrates this, which is the calculated
span-wise lift coefficient (CL) distri-
bution for the VariEze wing during
58 NOVEMBER 2009
climb. The green curve shows the CL
distribution for the wing without the
presence of the canard, with the dip
in the curve in the first 30 percent of
the semi-span due to the wing strake.
Outboard of the strake the CL distri-
bution drops off toward the wingtip,
but does not go to zero at the tip like
it would on a typical wing. The reason
is that the VariEze is equipped with
winglets that are the aircraft’s vertical
tail and rudder. These winglets also
act like a wingtip extension and make
the wing behave as though it has a
longer span than it actually does.
The blue curve shows the wing’s
calculated CL distribution as affect-
ed by the canard’s downwash. The
downward shift of the distribution
for the inboard 40 percent of the
semi-span is caused by this por-
tion of the wing operating at a
lower angle of attack due to the
canard downwash. This downwash
reduces near the canard’s tip, caus-
ing the local CL to increase until the
two curves cross near the location
of the canard’s tip (at 53 percent
of the wing semi-span). You can
also see that the canard’s influence
also continues beyond its tip, as
the wing’s CL distribution is now
above that for the wing alone.
This is caused by a vortex occur-
ring at the canard’s tip created
by the equalizing of the pressure
differences above and below the
canard. The tip vortex creates an
upwash that increases the angle of
attack on the outboard portion of
the wing, resulting in the raised
CL distribution. This increase can
These winglets also act like
a wingtip extension and
make the wing behave as
though it has a longer span
than it
actually
does.
be problematic if it causes prema-
ture stalling of the outboard wing,
and the VariEze had a segmented,
outboard-wing leading-edge droop
added to prevent this from occur-
ring. Later, it was found that several
small fences located below the lead-
ing edge (called vortilons) provided
the same benefit and had less drag,
so they were incorporated into the
VariEze and Long-EZ designs in lieu
of the outboard droop.
While the canard’s influence on
the CL distribution can be adjusted by
varying the wing’s airfoil incidence
along the wingspan, the canard air-
craft designer also needs to take into
account the wing-lift distribution as
affected by the canard when sizing
the wing spar. The altered lift distri-
bution could impact the sizing of the
spar web and caps.
 Speed has always been one of the References:
big selling points for aircraft. A fair “A look at handling qualities of canard aircraft,” NASA TM-88354, Seth Anderson, 1986.
question would then be, “How does
Canard: A Revolution in Flight, Andy Lennon, Aviation Publishers, 1984.
the performance of a canard air-
“Tale of Three EZs,” Burt Rutan, EAA Sport Aviation, February 1980.
craft compare to a similar conven-
tional aircraft?” There have been “Wind-tunnel investigation of a full scale canard configured general aviation airplane,”
quite a few records set by canard NASA TP-2382, Lon Yip, 1985.
aircraft, which is a good indication “Effects of Rain or Surface Contamination on Pitch Stability and Control,” Burt Rutan,
that they can do very well. AIAA EAA Sport Aviation, March 1983.
Paper 84-2507, “Design and Anal- “Quickie-Type Aircraft Design Origins,” Burt Rutan, EAA Sport Aviation, October 1981.
ysis of Optimally-Loaded Lifting “Wind-tunnel investigation of an advanced general aviation canard configuration,” NASA TM-85760,
Systems,” by Ilan Kroo is a theoreti- Joseph Chambers, Lon Yip, and Thomas Moul, 1984.
cal look at the big debate, and its “Design and Analysis of Optimally-Loaded Lifting Systems,” AIAA Paper 84-2507, Ilan Kroo.
conclusion gives the performance Available online at http://Aero.Stanford.edu/Reports/MultOp/multop.html.
edge to a conventionally config- NASA reports are available online at http://NTRS.NASA.gov/search.jsp.
ured aircraft. Figure 5 provides a EAA Sport Aviation articles are available online in the members-only section at www.Oshkosh365.org.
parasite-drag-area comparison of
several high-performance canard
and conventional homebuilt air-
craft designs. The data came from AIRCRAFT TOOL
David Lednicer and various CAFE
Foundation flight tests. It has been SUPPLY COMPANY
adjusted to remove the estimated Toll Free: 1-800-248-0638
landing-gear drag area in order to
provide a fair comparison. If we
take the drag area for a particular
design and divide the value by its
exposed surface (wetted) area, we
get its wetted drag coefficient. This ATS DELUXE HAND/BENCH RIVET
coefficient is an overall indication
of how clean a design is. Looking
SQUEEZER KIT
at Figure 5, we can see that the
canard aircraft have a wetted drag
coefficient around 0.0050 (50 “drag
counts” in aerodynamic speak).
This drag area is comparable to that
of the T-18 and Glasair, but higher
than a few of the other high-per-
formance aircraft. The higher value
is likely due to the higher drag of
the canard airfoils used and the
relatively blunt after-body on the
VariEze and Long-EZ. p/n 5022-DX
Depending on the designer’s goals,
it is likely that the canard configura-
tion will continue to be used on some
future designs. As Rutan stated, “The
designers’ database for these types of
designs is extremely limited, and the Our newest Rivet Squeezer Kit is designed to take
importance of understanding their on any riveting task. The kit contains the new ATS
aerodynamics is great.”
Heavy Duty Rivet Squeezer, a Bench Mount Kit, plus
An EAA member since 1981, Neal an extra 1-1/2” yoke head to reach into tight ar-
Willford learned to fly in an ultra- eas. You’ll also receive a complete set of squeez-
light in 1982 and received his pilot
certificate in 1987. He has done
ers and dimple dies housed in a convenient hold-
design work on a variety of aircraft at er. Our Deluxe Rivet Squeezer Kit also carries the
Cessna, from the 172 to the Citation Ats Pro Lifetime warranty.
X. In recent years he has been heav-
ily involved in the development of the
Cessna NGP and 162 SkyCatcher. In
his spare time he is finishing a Thorp
T-211 Sky Scooter.
www.aircraft-tool.com
EAA Sport Aviation 59