On the 'Wing... #117, published in RCSD July 1998
Steve Morris' Computer Stabilized Flying Wing Project
In our December 1996 column, we included the following quote from
Hans-Jurgen Unverferth of Germany:
"Why do we use radio controls? To build constructions characterized
by very high 'own-stability'? It's a joke! We have to be creative; fantasy
has to rule our thoughts! Think about the F-16, B-2, all the modern fighters.
There is no 'own-stability,' there is a computer! This is the future of
model sailplaning. And there is one geometry waiting for this time - the
In July of 1987, at Dillon Beach California, an actively controlled
unstable flying wing aircraft was successfully flown. This month's column
is devoted to an in-depth description of the aircraft and systems which
made that success possible.
The actively controlled unstable flying wing aircraft project was completed
by Steve Morris (mentioned in a previous "On the 'Wing..." column), Rick
Miley, and Dave Larkin, collectively called The Palo Alto Shipping Co.,
under the direction of Prof. Ilan Kroo of Stanford University and Dr. R.T.
At the time of project inception, Steve had already been involved in
the design, construction, and flying of a number of tailless RC models.
He had written a rather sophisticated computer program to aid in the design
process, and had, in fact, designed, built and flown a preliminary model
of the S.W.I.F.T. (Swept Wing with Inboard Flap Trim) rigid wing hang glider.2
One problematic aspect of tailless flight which intrigued Steve can be
directly related to "planks" and planforms with moderate rear sweep. When
the elevator on a plank is deflected upward, it is to increase wing lift.
Yet the upward deflected elevator creates a severe downforce which limits
overall lift. Additionally, drag increases significantly more rapidly than
lift. See Figure 1.
What Steve was looking for was a way to deflect the elevator downward
to increase lift. One way of achieving this is to use a swept wing planform
in which the wing sweep angle is such that the elevator can be placed inboard
and ahead of the center of gravity (CG). See Figure 2. While a standard
swept wing with elevator outboard can be envisioned to be similar to a
conventional tailed sailplane (Figure 3A), this highly swept configuration
with inboard elevator is similar to a canard configuration (Figure 3B)
in that the control surface for pitch is forward of the CG.
Moving the elevator inboard normally requires a substantial increase in
the sweep angle. In the accompanying diagrams, Figures 2 and 3B, the wing
sweep angle is around 45 degrees. It should be noted wing sweep angles
of over 30 degrees are not usually considered viable for subsonic flight
because of severe cross-span flow and excessive effective dihedral at high
coefficients of lift.
An alternate method of solving "the elevator problem" is shown in Figure
4. In this case the CG is placed behind the aerodynamic center (AC). This
makes for an unstable aircraft which cannot be flown for a sustained period
by a human pilot, but otherwise solves the elevator problem, as well as
making for a more efficient airplane. As Hans-Jurgen stated, an unstable
airplane requires computer control. This is the route The Palo Alto Shipping
Co. chose in order to achieve their goal.3
The basic aircraft was designed using a vortex lattice code.4
This provided stability and control information, and defined an optimal
level of instability. The design was formulated to explore flight characteristics
at 6.5% static instability in pitch. An overview of the aircraft exterior
is shown in Figure 5.
Control of the aircraft was handled by an onboard computer consisting of
a Motorola 68000 CPU with a floating point coprocessor - essentially a
Macintosh motherboard. The computer combined the input from an angle of
attack sensor (a vane mounted near the nose) and a pitch rate sensor (an
RC helicopter gyro) with the pilot commands transmitted from the ground,
and sent appropriate signals to the flap servos in the wings. The aircraft
hardware layout can be seen in Figure 6.
The computer control algorithm was determined by "flying" the wing on a
single axis gimbal. The gimbal was mounted on a Jeep which was then driven
down a quiet road. Because of the relatively short time to double in pitch,
just 0.3 seconds, special high speed servos were needed to keep up with
the feedback cycle and avoid unwanted excursions in pitch.
In addition to controlling the flap surfaces, the onboard computer also
collected data from eight channels at 20 Hz. and stored the information
in RAM for later downloading to a conventional Macintosh computer on the
ground. Two minutes of data could be collected before RAM was filled.
Because of the computer and associated battery supplies, the 12 ft.
span glider weighed 20 lbs., ready to fly.
At Dillon Beach, the aircraft was hand launched from the top of a sand
dune and directed by control inputs from a standard RC transmitter. The
first flight was made with the CG forward of the aerodynamic center; the
CG was moved rearward for subsequent flights.
After launch, the glider was flown through an "S" turn and flared for
landing. Collected flight data indicated that the time to double in pitch
was 0.298 seconds when the aircraft was 6.5% unstable. This closely matched
the data collected during ground testing. Yet the flight characteristics
were so unremarkable the videotape retains the comment, "Boy, if you didn't
know that thing was unstable... you wouldn't know!"
The aircraft was finally flown at 9.0% static instability, well beyond
the design instability point of 6.5%. In this condition, flap deflections
were extreme while turning and during the flare for landing, and flight
data showed a marked decrease in performance.
Given the low cost of small computers and the ease with which various
peripheral data acquisition devices can now be constructed and connected,
we anticipate similar and more advanced experiments involving unstable
tailless sailplanes in the near future.
Since that successful series of flights at Dillon Beach, Steve has been
involved in a number of other tailless projects:
As previously mentioned, Steve is co-designer, along with Ilan Kroo, of
the S.W.I.F.T., a foot-launchable flying wing sailplane now being produced
by BrightStar Gliders of Santa Rosa California. Steve was involved in the
Doctoral program at Stanford University at the time.
In 1990 he created an autorotation system for swept wing tailless aircraft
for use in vehicle recovery.5
Developed an oblique wing demonstrator, powered by a standard model airplane
engine and propeller combination. The wing-fuselage angle was not adjustable.
CNN carried a news story on the model in 1991, complete with commentary
by Dr. R.T. Jones.5
Steve designed, built and flew an oblique wing demonstrator aircraft for
NASA in May 1994. The wing sweep angle varies from 35 degrees at takeoff
to 68 degrees at cruise, necessitating rotating the pylons on which the
two Viojett ducted fan engines are mounted. The model has a span of 20
feet and weighs 80 pounds. It is constructed of a foam and Kevlar sandwich,
and has an aluminum spar and steel landing gear supports. There are ten
trailing edge control surfaces and two moveable fins. Eighteen servos are
required to fly the airplane and steer it on the ground. (Yes, all four
landing gear struts are steerable as well.) An onboard computer system
reads the pilot's radio commands and combines this information with data
collected from six onboard sensors. Eleven data channels are recorded in
RAM for downloading after landing.6
Steve is currently involved in the design and production of several "spy
planes," including the winner of the first University of Florida Micro-Aerial
Vehicle Flyoff. A more recent design, The Bat, an 18 inch span, 16 ounce
swept wing aircraft, is powered by a small off-the-shelf internal combustion
model airplane engine. Able to carry two video cameras, this small aircraft
is currently undergoing study as a viable surveillance vehicle.7
Kuhlman, Bill and Bunny. "Achieving the Potential of Tailless Planforms."
Soaring Digest, Judy Slates Ed., December 1996, pp. 8-9. Also available
in "On the 'Wing... the book, Volume 2." B2Streamlines, Olalla
Washington, 1997, pp. 181-182.
--. "Steve Morris and the 'S.W.I.F.T.'" RC Soaring Digest, Judy
Slates Ed., June 1994, pp. 3-5. Also available in "On the 'Wing... the
book, Volume 2." B2Streamlines, Olalla Washington, 1997, pp.
Morris, Stephen, Rick Miley and Prof. Ilan Kroo. Flight Research Using
Actively-Controlled, Small-Scale Prototypes. Materials provided at presentation
given at NASA Dryden, date unknown.
LinAir Pro. Desktop Aeronautics, Inc., P.O. Box A-L, Stanford CA 94305.
Morris, Stephen. Personal correspondence. September 6, 1993.
--. Personal correspondence. August 7, 1994.
Chandler, Jerome Greer. "Micro Planes." Popular Science, January
1998, pp. 54-59.
Links to GIF images of NASA Dryden presentation viewgraphs related to the actively controlled unstable wing:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Other photos of the oblique wing demonstrator:
take-off run, and in the air
return to the unconventional aircraft links page
B2Streamlines home page