RADIO CONTROL SOARING - 2003/03

FOR THIS MONTH Mark Drela
provided two interesting topics related to
airfoils. First, I asked him to make a list of
the airfoils he has designed for Radio
Control (RC) aircraft and briefly describe
their intended application. Second, Mark
wrote about tail deadband and how it can
be avoided.
Mark Drela is a professor of
aeronautics and astronautics at MIT
(Massachusetts Institute of Technology).
He is a leading researcher in the field of
airfoil design. In recent years Mark has
designed many airfoils for RC gliders,
especially for smaller or lighter ones
which operate at exceptionally low
Reynolds numbers.
In small and light glider applications,
Xfoil predicts up to 30% less drag for
Mark’s AG-series airfoils than the
previously published RC glider airfoils
which have been aimed at the larger or
heavier end of the RC Soaring spectrum.
Pilots who have flown them concur with
the computational predictions, including
me. There really is a pilot-noticeable
performance increase.
The nomenclature in the airfoil list
requires explanation. Most of Mark’s AG
airfoils have variants for use on different
sections of a particular wing. The root
airfoil is tweaked to perform better on the
smaller chord found on the outer sections
of the wing. They are listed as “root -> tip”
or “root -> mid1 -> mid2 -> tip” to suggest
how to blend the family of airfoils. The
target application and a non-exhaustive list
of US glider designs using the airfoils are
listed after each entry.
The letter “a, b, c, d, or e” near the end
of an airfoil name stands for the flap hinge
line location at 60%, 65%, 70%, 75%, or
80% respectively. A “t” on the end means
the airfoil was thickened near the flaperon
hinge line for structural reasons.
AG03 -> AG11
Built-up and solid-balsa small Hand-
Launched Gliders (HLGs); Wood Apogee
Built-up small Electrics needing large
speed range
AG04 -> AG08
Composite HLGs, strongly favor launch
and run; Apogee, Taboo
Composite S8E rocket gliders
AG12 -> AG13 -> AG14
Composite HLGs, roughly equal
emphasis on float and run; XP-3
(polyhedral version)
Mike Garton, 2733 NE 95th Ave., Ankeny IA 50021; E-mail: [email protected]
RADIO CONTROL SOARING
Figure 1. Xfoil results show that MH32 performs well at RRN of 100,000, typical of 3m
gliders. But at 50,000, which is typical of 1.5m DLGs, MH32 has large drag-producing
separation bubble. Under same conditions, AG16 has 33% less drag on smaller glider.
104 MODEL AVIATION
03sig4.QXD 12.20.02 8:35 am Page 104
AG16 -> AG17 -> AG18
Composite HLGs, emphasis on float;
Photon, Watson-Sidewinder
Composite light 2m poly gliders;
Composite Allegro
AG25 -> AG26 -> AG27
Composite heavier 2m poly gliders
AG24 -> AG25 -> AG26
Composite 3m poly gliders; Hallett
Bubble Dancer
AG31 -> AG32 -> AG33
Built-up small aileron gliders; Wind
Dancer (Pole Cat Aeroplane Works)
(hinge at 75-80% chord)
AG36 -> AG37 -> AG38
Built-up 1.5m poly HLGs
AG35 -> AG36 -> AG37 -> AG38
Built-up light 2m and 3m poly gliders;
Allegro-Lite, Bubble Dancer
AG34 -> AG35 -> AG36
Built-up heavier 3m poly gliders
AG45c -> AG46c -> AG47c
Composite 1.5m aileron HLGs;
SuperGee
AG455ct -> AG46ct -> AG47ct
Composite 1.5m aileron HLGs;
SuperGee II, XP-3, TabooXL
AG44ct -> AG45ct -> AG46ct -> AG47ct
Composite light 2m aileron gliders;
Aegea 2m
AG40d -> AG41d -> AG42d -> AG43d
Composite 3m aileron gliders; Aegea 3m
HT08
All-moving small-glider tails; Allegro-
Lite
(can be thickened to 6-7% for larger
gliders)
HT12
Discus-launch glider and light 2m tails;
Allegro-Lite
(hinge at 35-50% chord)
HT13 -> HT12
Heavy 2m tails
HT14 -> HT12
3m tails; new Mantis
HT21
Built-up tails; Bubble Dancer
HT22
Cambered tails; SuperGee
HT23
Cambered discus-launch glider vertical
tails; SuperGee
Design Rationale from Mark Drela: The
most popular modern RC glider airfoils,
Figure 2. Xfoil polar for SD8020 at RN = 80,000, with overlaid UIUC data. Deadband, or
near-zero lift curve slope, between –1 and +1 angle of attack, has detrimental effects on
control response and ability to hold pitch trim. It is also associated with excess drag.
Figure 3. Three overlaid Xfoil polars for the HT13 at RN = (80,000, 50,000, 30,000). No
deadband is predicted for these tail airfoils, no matter how low the Reynolds number.
Drag is also nearly the lowest possible.
such as the SD7037, RG15, MH32, etc.,
have been designed for reduced Reynolds
numbers of 100,000-200,000, which
corresponds to relatively large and heavy
gliders, where they work quite well. But as
indicated by University of Illinois at
Urbana-Champaign (UIUC) data and
numerical simulations, their performance
rapidly degrades below 100,000, producing
performance penalties on HLGs, 2m gliders,
and wingtips of light 3m gliders.
A common fix has been to thin and
decamber these airfoils to give better
performance at the lower Reynolds numbers.
This approach works to some extent, but it is
haphazard and generally unreliable. The
reason is that thinning has unpredictable
effects on the all-important surface-pressure
distributions, and additional reshaping is
almost always necessary to give the best
possible behavior.
The AGxx airfoils have been designed
from the outset for unit-CL Reynolds
numbers (also called reduced Reynolds
numbers) well below 100,000. The upper
surface pressure distributions are carefully
shaped to promote transition more
aggressively than usual, which shortens the
March 2003 105
03sig4.QXD 12.20.02 8:35 am Page 105
separation bubble and reduces bubble drag.
See the MH32 versus AG16 comparison in
Figure 1.
A side effect of such shaping is that the
maximum thickness point ends up farther
forward than usual. Also, the overall
thickness and camber end up reduced as
expected, which consequently incurs some
loss of maximum lift. However, this can be
compensated with a slightly stretched chord
and/or a lower weight relative to thicker
sections. The penetration is still improved
despite the reduction in aspect ratio or wing
loading. The penetration improvement
resulting from the AG airfoils has perhaps
been most noticeable in discus-launch
gliders, which have the lowest Reynolds
numbers among the competition Soaring
classes.
The few AG airfoils, which have been
designed for the largest Reynolds numbers
(approximately 80,000), not surprisingly
show the most resemblance to other popular
sections. The AG24 is very close to the
MH32, and the AG34 is close to the S3021.
But the other thinner sections in the AG2x
and AG3x series are relatively unique and
specifically well suited to lower Reynolds
numbers. The entire AG1x series is likewise
relatively unique compared to the more
common sections such as the MH32, etc.
A number of special features have been
incorporated into some of the AG series.
The AG03, AG11, and all the AG3x airfoils
have intentionally flat bottoms behind the
30% chord location for ease of construction
with built-up or sanded solid-balsa wings.
Theoretical performance relative to
unconstrained sections is compromised only
slightly. Actual performance may be better
because greater built-up accuracy is often
possible with the flat aft bottoms.
The AG3x airfoils have the additional
feature of exclusively flat facets on the
upper surface behind the 45% chord location
(e.g., behind the D-tube sheeting). This
allows open-bay construction with no airfoil
modification from covering sag. The
Allegro-Lite 2m and Bubble Dancer 3m
poly gliders use these features to advantage.
The AG4x series has been specially
adapted for camber control. In the full reflex
position, the bottom airfoil surface is
smooth. This is a favorable feature at high
speeds, where premature transition of the
lower surface is the greatest concern. The
result is exceptional penetration
performance.
In the moderate camber position the
upper surface is smooth, which delays
separation from the hinge line, then delays
drag rise with the flap set at large camber.
This improves float characteristics. Such
independent top and bottom surface
optimization with camber is commonplace
on modern full-scale sailplane airfoils. On
the AG4x series it provides a very wide
speed range despite their small thickness.
Designing airfoils specifically for low
Reynolds numbers gives other benefits
besides reduced drag. A common problem
with tail airfoils on RC gliders is deadband,
which is the loss of lift response to angle of
attack within a small range. One example is
the SD8020 tail airfoil, which works fine on
very large gliders, but it runs into deadband
difficulties below 100,000 where the
majority of RC glider tails operate. Such
deadband can be readily seen in the UIUC
data and in Xfoil simulations (Figure 2).
The simulations indicate the culprit:
laminar separation at the trailing edge
because of excessive pressure gradient and
no transition/reattachment because of the
low Reynolds number. On the HT tail
airfoils, laminar separation is eliminated
with suitable pressure distribution
shaping much like on the AG airfoils. The
result is 100% attached laminar flow, so
transition is not required for good lift
behavior. This eliminates deadband no
matter how low the Reynolds number (see
Figure 3).
The airfoil shape resulting from the
reshaping has the characteristic forward
maximum thickness location at 18% chord
and a relatively small maximum thickness of
5-8%. The HT airfoils also fortuitously have
a nearly triangular shape with flat sides over
the back 60% of chord, which is attractive
for simple solid-balsa or built-up
construction. The 100% laminar attached
flow has the advantage of giving the lowest
possible tail drag, as can be seen by
comparing the drag values between Figure 2
and Figure 3 at 80,000. MA
Sources of more information:
Apogee, Bubble Dancer, Allegro series,
many airfoils:
www.charlesriverrc.org/
SuperGee, some airfoils:
www.monkeytumble.com/hlg/supergee.htm
Allegro Lite newsgroup (much
information):
http://groups.yahoo.com/group/Allegro-
Lite/
XP-3, Wind Dancer (electric):
www.polecataero.com/
Taboo:
http://olgol.com/taboo.html
Xfoil:
http://raphael.mit.edu/xfoil/
UIUC:
www.aae.uiuc.edu/m-selig/