SoaringDigest

Radi C  ntr  lled

February 2008 Vol. 25, No. 2

 February 2008
CONTENTS Vol. 25, No. 2

Front cover: A Pike Perfect comes in with flaps down

during the 2007 Tangerine soaring contest sponsored by the
Orlando Buzzards. Photo by Peter Jensen.
Canon EOS 20D, ISO 200, 1/2000 sec., f4.5, 170 mm.

In a future issue...
and a contest announcement 26
Look forward to a Mark Nankivil walkaround of the
3 RC Soaring Digest Editorial Morelli M-200, and an announcement concerning the
Australia F3B Open International, 7-8 March 2008.
4 A Mid-summer trip to Volksrust
Model Analysis with XFLR5 27
A four-day excursion to Tamatie-Berg for some slope
flying, related by Piet Rheeders. XFLR5, a “virtual wind tunnel” allows a detailed
investigation of the aerodynamics of a specific
20 A photo by David Copple airframe. Francesco Meschia demonstrates all
Two Tangent ASH-26 models play in the sun. of the procedures necessary to evaluate the
X-Models 1.9 m Blade V-tail sloper.
21 Have Sailplane - Will Travel
Sloping at Armadillo World Headquarters
Tom Nagel (without a sailplane this time) finds out that
there actually is a slope site in Austin Texas - Mt. Bonnell.
24 At Last... The Troodons are finished! Back cover: Piet Rheeders' photo of the incoming
Not a small bird-like dinosaur, but a 3.65 m span F3J/F5J thunderstorm which brought an end to their flying at
sailplane. Simon Nelson gives a quick overview of his “Tamatie-Berg” Volksrust, South Africa, back in December.
most recent creation. (See Piet's article on this adventure starting on page 4.)
Panasonic DMC-LZ5, ISO 100, 1/250 ,sec, f`5.6, 37mm

2 R/C Soaring Digest

 R/C Soaring Digest In the Air
Managing Editors, Publishers B2 Kuhlman

Contributors Don Berry

Francesco Meschia
Tom Nagel
T his issue of RC Soaring Digest has some truly
spectacular photography. Although photos make the
resulting PDF substantially larger, reader feedback is very
Simon Nelson positive. We plan to continue incorporating a large number
Piet Rheeders
  Jerry Slates of photos in each issue, and presenting them at a resolution
which displays them to best advantage.
Photographers Dave Garwood
Dave Beardsley For those desiring articles of a more technical nature, we're
Dave Copple sure Francesco Meschia's detailed description of his use of
Mark Nankivil
XFLR5 <http://sourceforge.net/projects/xflr5/> to evaluate
Contact [email protected] the X-Model 1.9 m Blade will be a most welcome addition to
Web: http://www.rcsoaringdigest.com your library.
Yahoo! group: RCSoaringDigest
AIM screen name: RCSDigest During our recent windy and rainy weather here in the
Microsoft Messenger: rcsdigest
Northwest, we've been able to devote some time to our
————————————————————
Redwing XC project. All that's left to do is get some lead in
the nose to bring the CG forward to the appropriate location.
R/C Soaring Digest (RCSD) is a reader-written monthly publication Unfortunately, it looks like this may take the overall weight to
for the R/C sailplane enthusiast and has been published since January more than 11 pounds and over the FAI weight limit.
1984. It is dedicated to sharing technical and educational information.
All material contributed must be exclusive and original and not infringe On a similar front, our granddaughter Alyssa is building
upon the copyrights of others. It is the policy of RCSD to provide accurate
information. Please let us know of any error that significantly affects the
a composite sailplane for summer flying. This is to be an
meaning of a story. Because we encourage new ideas, the content of each entirely scratch-built enterprise, so there's been a lot of
article is the opinion of the author and may not necessarily reflect those of construction hardware to set up. She started by laminating
RCSD. We encourage anyone who wishes to obtain additional information
to contact the author.
tows of carbon fiber for the outer wing panel spar caps,
and is excited to move on to building the fuselage and tail
Copyright © 2007 R/C Soaring Digest structures. Alyssa's planning a multi-part article which will
Published by B2Streamlines <http://www.b2streamlines.com> document the entire construction process. Watch for that in
P.O. Box 975, Olalla WA 98359
All rights reserved
future issues of RCSD.
Time to build another sailplane!

February 2008 3
A Mid-summer trip to
Volksrust
Piet Rheeders, [email protected]
BERG club, South Africa

Mike May lets loose his Jart with a mighty heave. Note the whiplash on TX aerial.

4 R/C Soaring Digest

4 R/C Soaring Digest
T he long awaited summer holidays
normally starts in the beginning
of December in South Africa. Most
everything is green, and sitting on top
of the mountain breathing in some fresh
mountain air, admiring the stunning views
hoping that we would somehow get our
fair share of good winds.
This turned out to be exactly the case,
companies and businesses here come to that surround you, and in company of but most of us took a selection of models
girding halt for three to four weeks over some fellow modelers, is alone worthy of that could cope with just about any wind
Christmas, and only come to life again in the effort and expense to go there. condition. This included electric park
the first or second week in January. The BERG club had two groups visiting flyers, HLG gliders, Gentle Ladys, F3B
Although this might not be the best “Tamatie-Berg” recently, the first group glass slippers, Zagis, Weasels, Hill Billys,
time of the year to go slope soaring at drawing more members than the second, Jarts, scale gliders, IC-power gliders,
“Tamatie-Berg” Volksrust, the temptation but opting to stay over for only two days. and last but not least, some serious “Go
to just go there and fly, even if the wind Myself and six other fellow R/C modelers Big Or Go Home” stand-off scale giant
is going to be variable and light, is stayed for four days, knowing that the gliders.
far too strong. At this time of the year wind condition would not be ideal, but

February 2008 5
February 2008 5
6 R/C Soaring Digest
 A selection of models that we took along.

Opposite: The stunning views as seen from the top of

“Tamatie-Berg” Volksrust, South Africa.

February 2008 7
8 R/C Soaring Digest
Mike, Glen and myself
(from Durban) set out
in the early morning on
Wednesday the 2nd of
January 2008, and arrived
at Volksrust at around
10 AM. We had to wait,
however until 1 PM before
we could get to the top
of “Tamatie-Berg” due
to a locked farm gate.
This was no problem as
there was no wind from
any direction. This gave
us the time to go to our
overnight guest farm,
situated directly under the
northwesterly slope, to
unpack our baggage and
personal belongings.

Our overnight accommodation situated

directly under the northwest slope.

The 7.5 meter span, 15 Kg, stand-off scale DG500 Elan of Mike May.

February 2008 9
 Glen about to launch his weasel as high as possible in the week lift we had on Day 1.

10 R/C Soaring Digest

Once we got to the top,
a quick glance of the
northwest slope showed no
promise of any flying there,
and we proceed to the SE
slope and waited for the light
wind to pick up somewhat.
Glen with his light weight
Weasel was the first to
test the air and sometimes
managed to maintain height
for a reasonable time before
he was forced to turn and
land. Other than that there
was not anything else that
could stay up in the light to
no wind conditions.
By the end of the day we had
very little flying time, and Izak
Theron from the ETB club
also now joined us.
At round about 4PM we
packed up as the little
wind we had died away
completely. As the sun set
we left the mountain with a
hope that Thursday (Day 2)
would yield better conditions, No wind on Wednesday, Day 1.
and we returned to our
overnight quarters.
As per normal, and after we refresh ourselves, the braai
(barbeque) fire was started, followed by the by many R/C
flying stories until we could not keep our eyes open any more.
Once you hit the sack you drift of into dreamland just about
instantaneously.

February 2008 11
On Day 2 the morning broke with a partly This time round we had no problem As we arrived at the SW slope we found
cloudy sky and the SMS I received from with the locked farm gates. Once on top some more glider pilots joining us —
my friend Evan Shaw predicted SE wind we quickly assembled our gliders and Herman and Izak from the ETB club, and
later on in the day. then waited as the early morning mist Charl and Peter from the club BERG.
After breakfast, and just in case the wind disappeared and as the sun got higher in
did not work, I had two flights with my the sky.
IC-power glider before we departed for
the south slope.

Piet launching his

IC-power glider at our
home base before
departing for the day’s
flying on the slope.

12 R/C Soaring Digest

Most of us, when we come to
“Tamatie-Berg,” have at least one
glider that needs to be maidened, as
we only come here three to four times
a year.
Mike May had a new Jart and also his
big 7.5 meter span DG500 to fly their
first flights, and likewise my Hill Billy.
Mike had his DG500 inspected by
some other big scale R/C glider pilots
to make sure that it would fly first
time.
At around 2 PM the wind pickup
nicely and I did one short maiden
flight with my Hill Billy. Then as the
wind go stronger Mike let lose his
Jart on its maiden flight. The sky was
now getting busier as the conditions
improved as Zagi’s and other smaller
slope ships took to the sky. At last
this was the wind we waited for and
now were enjoying every moment of
it. I got my second flight in on my Hill
Billy and this time around got it on the
step as she cut through the sky.
From left to right: Herman, Izak and his son Shane trimming his HLG. An HGL is ideal for
Although the air was now good to
no wind and low visibility/misty conditions in the early morning on top of the mountain.
launch Mike’s DG500 but he opted
not to fly it because of the limited
landing space on the south slope.
Needless to say we returned to our
accommodation at the end of the day, well
satisfied and happy as can be. For us a half
a day of good slope flying is worth many a
day’s waiting for the right conditions.

February 2008 13
 Glen’s pink Jart in a blue sky makes for a pretty sight.

14 R/C Soaring Digest

 Izak waiting for the breeze to pick up before he launches his F3B ship on NW slope.

February 2008 15
Pre-flight check on Mike May’s new monster 7.5 meter DG500 Elan.

Day 3 (Friday) we had similar conditions ships and managing some reasonable This happened to be so on Day 4
as Day 1 (Wednesday) and we ended flights. The indications, however, were (Saturday), with the wind direction from
up with light winds turning the full 360 that the next day (Saturday) was going to the northwest and picking up ever so
degrees before ending the day on the be the best day of our visit. gradually until it got strong enough for
northwest slope flying our light thermal Mike to maiden his big DG500.

16 R/C Soaring Digest

 Charl launch Mikes’ 3M Swift.
After this flight, Mike was confidant
that the air was now good enough to
fly his 15 Kg monster DG 500 Elan.

February 2008 17
Mike had one flight with his 3M Swift
and was confidant that the air was now
good enough to fly the 15 Kg monster
DG 500.
The launch went well with Glen on the
left tip, Charl in the middle and Peter on
the right tip. Mike was quick to pick up
the left wing that dropped a bit because
of a late release by Glen.
The DG500 then climbed high above
the slope and soared for eight minutes
before Mike became aware of a
threatening thunderstorm on the lee-
side of the hill and decided to land
before the wind changed direction.
The landing was also good as the
powerful flaps slowed the DG500 down
and Mike had to tuck them away for a
moment to retain some speed before
touchdown.
The thunderstorm came closer and
we had to dismantle and pack up
our models in a hurry. [A photo of the
approaching thunderstorm serves as
the back cover for this issue.]
I don’t think that any one of us would
regret this outing, and as for myself I
have flown six of seven models that I
took with me, and so as we drove down
this magic mountain we are already The Berg team just before we departed for home on Sunday morning. From left to
planning the next trip to Volksrust and right: Piet, Glen, Jenny, Charl, Peter, Mike. Middle center: Blake
“Tamatie-Berg.” n

Launching Mike’s 7.5 meter DG500 Elan.

18 R/C Soaring Digest

February 2008 19
 David Copple captured this photo of two Tangent
ASH-26 models against a sun ring at Cuesta Ridge
in San Luis Obispo, California.
FujiFilm FinePix S5000, ISO 200, 1/2000 sec., f8.0

20 R/C Soaring Digest

 Have Sailplane
-
Will Travel
Tom Nagel, [email protected]

Sloping at Armadillo World Headquarters

HSWT in Texas

M y brother Terry has a daughter

working on her graduate degree
at the University of Texas in Austin. He
County looked pretty darn flat.
Then, just before we left town, Terry
took us up to Mt. Bonnell Park, on the just to the north (to your right facing
talked me and the wife into joining with northwest side of town. Mt. Bonnell the slope) of the actual “peak” of Mt.
him and his wife for a trip down to Austin turned out to be a limestone cliff rising Bonnell. “Peak” seems a little pretentious
for Christmas. We’d get a chance to almost vertically some 350 feet above the in this setting. What I really mean is, try
play a little music together, sample the body of water variously known as Lake launching 50 yards or so to the right of
local margaritas and meet up with his Austin, or Town Lake, or the Colorado the highest point of the escarpment.
daughter’s fiancé and family. River.
Mt. Bonnell Park is a relatively small
I didn’t count on doing any sloping, so Mt. Bonnell isn’t much of a mountain, but Austin City Park. There are no facilities,
I didn’t take a plane. Accordingly, you it looks like a pretty nice place to slope. and no rangers. The barriers at the edges
might notice a total absence of anything The escarpment faces almost due west, of the cliffs are sketchy and trails (and
resembling an RC slope plane in the with maybe a little West-South-West a little trash here and there) indicate the
accompanying photos. curve to it at the northern end. You can barriers are not frequently observed.
At first, Austin did not look too slope access the cliff top by walking up a fairly
There is no direct route down to the
friendly. The locals kept telling me that gentle slope from the north end of the
base of the cliff, and if you ootch
Austin was at the south end of Texas park, or up 150 or so stone steps from
out to the edge and look down, you
Hill Country. I thought they meant Hank the south end of the park.
will see upscale, you might even
Hill, because as far as I could see Travis The best launch sites appear to be say Austintaceous, houses along

February 2008 21
 the waterfront, built on a series of man-made lake front
parapets. You do not want to send a plane down there.
This would be a good place to slope something with an
electric motor.
The landing zone comes in two varieties: full of tourists
and full of trees. Luckily, the trees are of the Texas scrub
vegetation variety, and not particularly thorny. Landing will
require either dumping it in the trees on the downwind side
of the access trail, or enlisting a spotter to keep the tourists
at bay while you slide in on the gravel and stone access
path.
This would be a good place to slope something foamiferous
with an electric motor.
The very south end of the park affords a great view of
downtown Austin and the Gozer Building. Austin has only
one really tall building downtown, the proper name of which

22 R/C Soaring Digest

is the Frost Bank Tower, but it looks a lot like the TopoZone - Mount Bonnell, USGS Austin West (TX) Topo Map http://www.topozone.com/print.asp?lat=30.32076&lon=-97.77334...

Gozer Building from Ghost Busters. You can see the

Gozer Building from almost anywhere in Austin, so
it is easy to get oriented again if you get lost driving
around town.
Here are some other things to do and see in the
Austin Area:
Armadillo World Headquarters, a restaurant and bar
now known as Threadgill’s, which in the years 1970
to 1980 was home to more bands and more live
music than can be easily contemplated. Everybody
from AC/DC to Frank Zappa played there, with
groups like Asleep at the Wheel, Jimmy Buffet, Dire
Straits, Genesis, Charlie Mingus, Willie Nelson, Linda
Ronstadt, Earl Scruggs and Ravi Shankar filling in the
middle. It is still a great live music venue.
The Broken Spoke: the same, for country music.
The Congress Street Bridge Mexican Free-tailed Bat
flock: bug eating ornithopters by the millions.
The Texas State Capitol Building, which, because
it is in Texas, is actually bigger than the US Capitol
Building.
The Barton Springs swimming hole, a series of warm
springs re-engineered into a year-round naturalistic
two hundred yard long outdoor pool.
South Congress Street: the hip part of town to eat,
drink, shop or be seen. No RC stores, though.
Austin Segway Tours, a gyro-stabilized livery that for
$59 will liberate your inner geek and give you a two-
wheeled tour of Austin. UTM 14 617931E 3354966N (NAD83/WGS84)
Mount Bonnell, USGS Austin West (TX) Quadrangle
A QuickTime VR 360 degree view from the top Projection is UTM Zone 14 NAD83 Datum
of Mt. Bonnell is available at <http://www.edb.utexas.
edu/teachnet/QTVR/MtBonnell.htm>. n
1 of 1 1/17/08 11:52 PM

February 2008 23
 At last...
the Troodons are finished!
Simon Nelson, [email protected]

T he Troodon (TROE-odon) is a 3.65 meter

span F3J sailplane which can also be
configured as an F5J machine.
The sailplane version weighs 2.2 Kg (77.6
ounces) ready to fly.
The wing chord is 300 mm, with the area
about 35 % more than the Inkwazi, similar to
the Eish, the local F3J plane, described in a
previous issue of RC Soaring Digest.
The laser-cut ribs were done by Paul at
Lasercore in Pinetown. The wing structure
is composed of carbon tube spars with I
beams. The fuselage is of balsa and full
carbon.
The stabilizer is from AMT.
The wing center section includes a D-tube
leading edge, with the whole wing having a
geodesic structure.
Transparent covering allows that interior
structure to be featured. Sorry, will wear a t-shirt next time.

24 R/C Soaring Digest

The V-tail model is the electric
F5J version. It has a 480-33
motor, a 15 x10 prop, and a 3-cell
double stack 2200 LiPoly battery.
The name comes from a genus of
relatively small, bird-like dinosaur
from the Late Cretaceous Period.
It’s from a long line of similar
constructed planes. I will be
doing plans, as with all the
others. n

February 2008 25
In a future issue...

Australian F3B Open International

Milang, South Australia
7th 8th 9th March 2008
Three days of F3B action with pilots from
Germany, Japan, New Zealand and Australia

Enjoy a summer holiday

Morelli M-200 in South Australia

Great conditions with great blokes
Serious early season F3B flying
walk-around Further information and registration of interest
Manufactured by CVT in Italy, this is an 18.15 meter span please contact Mike O’Reilly at:
two-place staggered side-by-side sailplane which is [email protected]
stressed for aerobatics and has a 32:1 glide ratio.
Wood construction and a large cockpit make it a good
candidate for RC aerotowing with vario or GPS installed.

26 R/C Soaring Digest

Model analysis with XFLR5
Francesco Meschia, [email protected]

Virtual wind tunnels and simulators long time, and a numerical method for the aeromodeling realm, thanks to the
aircraft performance analysis, known increase in processing power of modern
Many aeromodelers are already familiar
as vortex lattice method, or VLM, was PCs and to a few suitable software
with XFoil, the famous virtual wind tunnel
developed since the 1930’s. I am not packages which are now available. Again
developed by Mark Drela and Harold
an aerodynamicist and so I cannot Mark Drela and its research group have
Youngren at MIT. XFoil, like many of its
describe the VLM in great detail, but released AVL, a powerful and complete
siblings, simulates with good accuracy
from the Internet I learned that VLM program with a “family feeling” with XFoil
the airflow around a “two dimensional”
involves dividing the lifting surfaces into in its command-line user interface. Much
airfoil, i.e. the purely ideal case, not likely
a fine mesh of panels. Each panel is like what happened to XFoil with Stefano
to be attained in practice, of an airfoil
surrounded by a horseshoe vortex, that Duranti’s “Profili 2,” VLM has become
section traveling through the air by itself,
extends chord wise to infinity; with a few more accessible to the non-specialists
without an associated wing, therefore
boundary conditions one can calculate thanks to XFLR5, a program developed
neglecting induced drag and any other
the lift and drag contribution of each by André Deperrois. Mr. Deperrois’s
planform induced effect.
vortex and so, by summing the individual work includes an interesting contribution:
This kind of analysis lends itself well to contributions, one eventually evaluates whereas the “classical” VLM analysis
the comparison of different airfoils but, the performance of the whole surface. assumes a purely inviscid flow around
on the other hand, requires additional the lifting bodies and is therefore a bit
Because of its numerical nature, the
work if one wishes to analyze the unrealistic for the Reynolds numbers
VLM didn’t really “take off” until enough
performance of a wing or a complete used by model aircraft, XFLR5
number crunching power became
aircraft, full-size or model, a task that postulates that the viscous and inviscid
available, with the advent of computers
involves many other variables and contributions to aerodynamic forces are
in the 1960’s; it proved since then to be
requires careful modeling of many other linearly independent, so that an inviscid
a very powerful tool and it was used to
factors. VLM output may be complemented
study and develop a large number of
Because of its extreme importance different aerodynamic configurations. by a viscous XFoil analysis to get a
for the whole aviation industry, the more realistic mathematical model.
In the last few years VLM has landed in Deperrois warns the users, though, that
latter problem was investigated for a

February 2008 27
 the “independence hypothesis” is not
supported by a theoretical model, and
so XFLR5 results need to be considered
preliminary and experimental work and
model validation still needs to be done.
XFLR5 users need to keep this
philosophy in mind when approaching
the program. XFLR5 must be “fed”
with both a geometrical model of the
lifting surfaces to be analyzed and
a set of polars derived from viscous
analysis of the adopted airfoils, for
a range of Reynolds number and lift
coefficient broad enough to cover all
flying conditions. In other words, before
even starting XFLR5 the user must find
the coordinate files for all the airfoils
which will be used, and a 3-view of the
aircraft. Also we must remember that
XFLR5 is merely a simulator for a set
of physical behaviors, and can’t tell us
anything about an aircraft but what we
are prepared to ask; it’s a handy tool, but
teaching us how to design a successful
model is entirely out of its scope.
Much better, in my opinion, is to
approach XFLR5 with a practical touch,
to show what the program is capable of
and also what to ask the program for.
I have therefore chosen a well known
model, the 1.9m Blade from X-Models,
and I will use it as a test case to look for
X-Models 1.9m Blade answers to some common questions
such as lift distribution, stability analysis,
Photo by Stefano Bisio
and speed polar determination.

28 R/C Soaring Digest

Modeling the Blade
The airfoil used in the Blade wing
is a modified RG-15, which was
thinned from the original 8.9%
thickness to a sleek 7.8%. According
to many modelers, it is likely that
the airfoil evolves into some other
variation near the wing tips, but I
could get no final word about this
and so I will neglect it in my analysis.
With some juggling with cardboard
templates I have determined that
the V-tail relies upon a 7% thick
symmetrical airfoil, a NACA 0007.
The coordinates for the 4-digit
NACA foil may be calculated with
a well-known algorithm, but the
RG-15 coordinates must be found
somewhere in literature. I would
suggest using the airfoil database by
M.Selig and the University of Illinois
at Urbana-Champaign <http://www.
ae.uiuc.edu/m-selig/ads/coord_
database.html#R> for that.
Once the coordinate file has been
downloaded, we can start XFLR5
and begin our journey by importing Figure 1
the foil data. From the “File” menu
we choose “Load File” and we
command in the “Design” menu, as
select the DAT file; XFLR5 will show an
shown in Figure 2.
outline of the airfoil in the lower pane of
the window (Figure 1). As we can see The airfoil for the tail is calculated by
the thickness is 8.9% by the book, so it a routine that may be invoked with
must be thinned to 7.8% if we want to the “Naca Foils” command under the
model the actual Blade foil. This is done “Design” menu. We need to tell XFLR5
via the “Scale Camber and Thickness” the NACA number of our choice (0007) Figure 2

February 2008 29
 Figure 4

different Reynolds if needed. Care must

be taken so that the Reynolds mesh
coverage is tight enough and XFLR5
can use it. (We’ll see later how to tell
if this is correct and how to extend
coverage if necessary.)
Polar computation might be carried
out manually, Reynolds number by
Reynolds number, but there’s a handy
Figure 3 utility that lets us process a whole
polar family at once. We will begin by
and the number of points (100 points For each airfoil we must compute a selecting from the airfoil drop-down list
will do). The program will calculate the set of polars that covers a range of one of the foils (e.g. RG-15 7.8%) and
coordinates and will draw the airfoil Reynolds number representative of then choose “Run Batch Analysis” from
profile (Figure 3). the “flying” conditions. XFLR5 will take the “Polars” menu. A new window will
those polars and will add the viscous pop up (Figure 4) where we shall enter
Now that we have the wing and tail
contribution to the VLM analysis by the parameters for the batch run. First
airfoils it’s time to use the embedded
interpolating between computations at we select the desired analysis type (e.g.
XFoil module and compute their polars. Type 1, that stands for fixed airspeed

30 R/C Soaring Digest

Figure 5

Figure 6

and chord, and variable lift coefficient); routine will start working and in a little any plot we’re interested in (Cl vs. Cd, Cl
then we enter the desired Reynolds while will crank out the polars. When vs. alpha, Cm vs. alpha, and so on) by
range (I’d suggest to use a list of Re’s, the computation is over we may close choosing the appropriate item in “Polars
like the one shown in Figure 5) and the the window, then we open the “View” > View.”
angle of attack range (e.g. AoA ranging menu and choose “Polars.” The newly- The same procedure must be re-iterated
from -3° to +9°, in 0.25° steps). When computed polar family will be shown for the NACA 0007 tail section (Figure 7),
we click the “Analyze” button the XFoil (Figure 6), and we’ll be able to choose at first for the same range of Reynolds

February 2008 31
Figure 7

numbers used for the main wing section Now we have the 2D polar families for its planform and decalage. For this I will
and for AoA between -4 and +4 degrees; both wing and tail sections, it’s time to use actual geometric data I measured on
theory suggests that we should shift enter the realm of 3D simulation with the my own Blade.
the Reynolds range towards lower Re’s, vortex lattice method. First we must set We can enter the XFLR5 plane geometry
because the tail chord is shorter than the up in XFLR5 a simple three-dimensional routine first by selecting the “Wing
wing chord, but we’ll see later if this will model of the Blade, re-creating the wing Design” option from the “Application”
actually be necessary. planform, the wing dihedral, the tail with menu, then by choosing “Define a Plane”

32 R/C Soaring Digest

 Figure 8
from the “Wing/Plane” menu which will geometry definition sub-windows. In purposes. A finer mesh means more
appear. A window similar to the one these sub-windows (see Figure 9 for accuracy, but we don’t want to take
shown in Figure 8 will pop up. In this wing geometry and 10 for the tail) we this too far because we could run into
window we can assign a name to the will model the lifting surface planform by XFLR5 internal memory limitations. I
model, we can set the wing and tail dividing it up in a series of trapezoidal suggest using the “Reset VLM Mesh”
rigging angles (via the “tilt” parameter) panel, and we’ll also need to define a button after defining all the panels so
and, most important, we can enter the suitable mesh for each panel for VLM that the program defines a mesh it sees

February 2008 33
 fit. Each panel is defined
by its airfoil, its root and
tip chords, its span,
its offset, its dihedral
relative to the horizontal
plane, and its twist;
XFLR5 uses metric units
by default, but it may
be set to use the U.S.
system via the “Units”
option in the “View”
menu. (I am from Europe
so I will use the metric
international system in
data and calculations.)
As the data are entered,
the program immediately
calculates the wing
surface, the mean
aerodynamic chord, and
the wing aspect ratio.
The wing of the Blade
has three degrees
central dihedral, that is
1.5 degrees per each
half-wing relative to the
horizontal plane. The
V-tail is modeled in
XFLR5 as an elevator
without fin: each half-
elevator has 35 degrees
dihedral to the horizontal
plane. I measured a
longitudinal dihedral
Figure 9 (decalage) of one

34 R/C Soaring Digest

degree, which I modeled
by setting the tail at zero
degrees and rigging
the wing at 1 degree,
and the wing has a
negative 0.5 degrees
twist at the outer aileron
edge relative to the root
section.
When we have defined
wing and tail geometry,
I suggest we make sure
that the “Check Panels
on Exit” option is NOT
checked (this option
may sometimes cause
a program crash in
XFLR5 v3.21) and we
can dismiss the window
with the OK button. If we
select the “3D” option
in the “View” menu, a
three-dimensional view
of the model will be
shown, like in Figure 11.
We can also request any
of the three orthogonal
views with the X, Y and
Z buttons in the window
pane at the right.
The 3D polar of the
model
Figure 10
We now are just one
step from obtaining the
3D polar, that is the polar

February 2008 35
 of the whole model. The polar computation requires a set
of boundary conditions, and so we must decide whether
we want to either have constant airspeed, or constant lift,
or constant angle of attack. In the gliding task the wing
must generate, at any given moment, just enough lift to
balance the weight of the model, therefore we are mostly
interested in the second type of boundary condition. We
can tell XFLR5 about our choice by choosing “Define a
Polar Analysis” from the “Polars” menu, and by choosing
a “Type 2” analysis from the window that will pop up
(Figure 11). We must also tell XFLR5 the weight of our
model (my Blade weighs 1450 grams) and the position
of the center of gravity (“moment reference location” in
XFLR5 jargon) relative to the leading edge of the wing
root that is taken as the coordinates’ origin (79 mm for
my Blade). From the same window we notice that VLM is
the only applicable analysis method (the other one, the
lifting line theory method or LLT, can only be used for
an isolated wing, i.e. without tail) and we need to make
sure that the “Viscous” option is checked. That said,
we can dismiss the window with the OK button and we
are ready to run the computation task. In the rightmost
window pane we set up a sequence of angles of attack,
for instance from -2°to 7° in steps of 0.25°, we check the
“Store OpPoints” and “Store points outside the polar
mesh” boxes (see Figure 12) and we can start the job
by clicking the “Analyze” button. A window will pop up
with a processing log and maybe some final messages
warning us that some operating point fell outside the
Figure 11 polar mesh. We’ll see later what those log messages
mean, for now we want to jump directly to the results.
We choose (if not already done) the “3D” option in the
“View” menu and we are presented with a beautiful color
rendering of the distribution of the pressure coefficient
over the lifting surfaces (Figure 13). By altering the angle
of attack via the drop-down menu in the upper right part
of the window we can see how the pressure distribution

36 R/C Soaring Digest

Figure 12

Figure 13

changes its shape accordingly. and maybe easier to understand if a list of variables we can assign to X and
Undoubtedly impressive, but somewhat little less appealing. For instance, we Y axes. If we assign vertical velocity (Vz)
hard to decipher because on the same may choose “Polars” from the “View” to Y axis and airspeed (Vinf) to X axis
plot we can find lift vectors, drag vectors menu, then we right-click on the graph we’ll get the speed polar for our 1.9m
and also downwash vectors. A number that will appear and last we choose Blade (Figure 14). This graph pictures
of different output modes for the same “Graph” and then “Variables” from the the relationship between airspeed and
data are available, equally interesting pop-up menu. We’ll be presented with a sinking speed, which is a true snapshot

February 2008 37
Figure 14

of the performance of a glider, be it airspeed of about 10 m/s), and another the pilot who wants to stay aloft in dead
full-size or a model. It is worth noticing point, different from the former, for which calm air for as long as possible must trim
that the plot has a minimum sink point the horizontal-speed-to-vertical-speed the sailplane to keep the minimum sink
which corresponds to an airspeed value ratio (also known as efficiency and glide airspeed, whereas the pilot who wants
slightly above stall speed (theoretically ratio) has a maximum (at 11 m/s airspeed to travel as far as possible must trim for
speaking, the 1.9m Blade could attain a for our Blade). These two points picture maximum glide ratio.
sinking speed as low as 40 cm/s at an two very important attitudes of a glider: Another, clearer, way to study the glide

38 R/C Soaring Digest

 Figure 15

ratio versus airspeed relationship is to which is theoretically above 25 at 11 trim speed is in normal conditions. The
ask XFLR5 to assign the “Glide Ratio Cl/ m/s airspeed. The peak is well defined, answer is readily found by asking XFLR5
Cd” to the Y axis while keeping airspeed that means that any drift from optimum to plot a pitching moment coefficient
(Vinf) on the X axis. The resulting chart airspeed translates into a marked loss of versus airspeed chart (Figure 16). We
(Figure 15) immediately displays not efficiency. see that the curve crosses the horizontal
only the maximum glide ration point, We should now ask ourselves how fast axis at an airspeed of about 18 m/s. This
but also the numerical value of the ratio, the Blade wants to fly, that is what its point is the only balance point for the

February 2008 39
 stability later).
It is rather conspicuous that 18 m/s is
significantly faster than both 10 m/s
(minimum sink speed) and 11 m/s
(maximum glide ratio speed). This is
consistent with the kind of flight the
Blade is made for. It’s a fast glider
and not a thermal duration model, a
runner and not a floater. It wants to
“sting like a bee” rather than “float
like a butterfly.” It does not make
much sense to force it to fly at the
minimum sink airspeed, and its pilot
is probably willing to trade some
efficiency for more thrill. At 18 m/s
(40 mph) the Blade sinks in the air
at a rate of a 1.1 meter per second.
That means that a slope updraft with
a vertical velocity component of just
above 1 m/s (fairly common for most
slopes) will be enough to keep it
flying.
The lift distribution
We will now talk about another
way to look at the results of XFLR5
computations. So far we have
Figure 16 looked at charts that are a picture of
the behavior of the whole model as a
function of one of the flight variables.
whole flight envelope of our glider, since also suggests that any higher speed Each point in the curve is an “operating
here the model does not experience would cause the Blade to pitch up and point” of the model, that is a collection of
any pitching moment, neither upward decelerate, and any lower speed would data for lift, speed, efficiency, moment,
nor downward. The model simply make it pitch down and accelerate. That etc., computed for one particular value
flies undisturbed at 18 m/s: this is its means that we have chosen a center of of the independent variable we have
trim speed with neutral elevator trim gravity location that allows for a stable chosen for our analysis (i.e., in our
and 1 degree of decalage. The curve equilibrium (but we’ll see more about examples, the angle of attack). But

40 R/C Soaring Digest

the vortex lattice analysis does
not merely yield a “point,” a scalar
value representing the entire model,
but rather it produces a far richer
collection of data, spread across the
many panels that make up the lifting
surfaces. XFLR5 lets us “drill down”
in every operating point by looking at
the aerodynamic parameters in their
distribution over the lifting surfaces.
This kind of study is very interesting,
a true fundamental exercise in
computational aerodynamics. If we
look at the lift coefficient at which
every wing “station” is working,
from the root all the way to the tip,
we can easily see what parts of the
wing are working “harder” (higher
lift coefficient) to produce lift and,
therefore, will be the first to come to a
stall according to the viscous analysis
we performed with the XFoil module
at the very beginning of the journey.
Let’s pull down the “View” menu and
choose “Operating Point.” We should
get a plot like the one shown in Figure
17, that is a set of curves that picture
local lift coefficient versus position Figure 17
along the half-wing span, for a range
of angles of attack. If we prefer, we can When we inspect the family of lift higher and higher lift coefficient, so that
also examine just one curve, at a defined distribution curves we see that for small eventually we shall have the dreaded tip
AoA, by right-clicking the chart area and AoA’s (up to 2 to 3 degrees) the portion of stall.
choosing “Show Only Current Opp” and the wing that generates the highest lift is To see this in greater detail, we need to
then choosing the desired AoA from the located at about one foot (300 mm) from take one step back and take a look at the
drop-down list in the far left end of the the wing root; but if we further increase error messages in the polar computation
toolbar. the AoA the tips are forced to work at a log file (“Operating Point” menu, then

February 2008 41
 the wing root) ought to generate a
lifting coefficient of 0.86 at Reynolds
23308. According to the XFoil polars
in its memory, XFLR5 doesn’t think
that our modified RG 15 airfoil can
accommodate such requirements.
In other words, XFLR5 foresees the
beginning of a tip stall at this angle
of attack. XFLR5 may not always
be right (in this particular case,
if we examine the family of airfoil
polars we can see there’s such a
wide a gap between the Re=20000
and Re=30000 polars that XFLR5
interpolation might not be able to
bridge reliably), but it is an interesting
clue nevertheless. It is also worth
mentioning that, despite XFLR5’s
opinion, the “real” Blade is not
particularly prone to tip stall problems
(or, at least, it hasn’t yet played
such bad tricks to me). This might
support our initial suggestion that, at
the wing tips, the RG 15 foil evolves
into a different section, one that can
generate more lift at low Reynolds
numbers.
Figure 18 Another variable we might want to
examine the distribution of is local
“View Log File”); at the end of the file we Cl = 0.86 could not be interpolated lift (not lift coefficient). This is the force
should see a similar message: Span pos = 937.49 mm, Re = 23 308, that is generated by every wing portion;
Cl = 0.86 could not be interpolated it is important because, according
...Alpha=5.75
Calculating elevator... to Prandtl’s theory, the induced drag
Calculating induced angles... contribution for a given wing area is
Calculating aerodynamic coefficients... What this message is telling us is that, minimized whenever the lift distribution
Calculating wing... at an AoA of 5.75°, the wing tip station shape is elliptical. We can ask XFLR5 for
Span pos = -937.49 mm, Re = 23 308, (937 mm, or about three feet, away from this distribution by right-clicking on the

42 R/C Soaring Digest

graph area, then choosing “Variables”
and finally “Local Lift C. Cl/M.A.C.”
The resulting curves (Figure 18)
closely follow an elliptical arc (a true
elliptical arc may be superimposed to
the plot via a specific option from the
right-button menu), and prove that
the Blade wing planform is backed by
a good amount of careful design and
study.
Analysis of pitch stability
We now want to use XFLR5 to
approach another classical problem
in aircraft performance, that is the
study of pitch behavior and static
stability conditions. Stability, in this
context, is the tendency of the aircraft
to keep its pitch attitude against any
disturbances that may arise (wind
gusts, for instance).
We start this chapter by returning
to the polar plotting mode (View >
Polars) and by instructing XFLR5
(right-click > Graph > Variables) to
plot the pitching moment coefficient
versus AoA. We will get a chart like
the one shown in Figure 19 - a simple Figure 19
monotonically decreasing curve.
The slope of the curve is “negative” (i.e.
the curve is decreasing) for every point the chart. First of all, the chart is telling degrees. The chart shows that at zero
we have considered. This is of extreme us that the model glides in balance angle of attack the model develops a
importance because it is means that (i.e. CM is null) at an angle of attack of negative pitching moment coefficient,
the configuration we have chosen for approximately -0.5 degrees. Let’s now i.e. it “feels” a nose-down moment that
the model is stable for all the AoA’s of suppose that the flight is disturbed by tries to push the wing to its original AoA,
our interest. This is easily understood a gust that forces the model to take a and therefore the model is longitudinally
if we review the physical meaning of greater angle of attack, for instance zero stable. Every configuration for which the

February 2008 43
 the CG at 100 mm from the leading
edge. The resulting CM vs. AoA curve
will be like the green curve in Figure
20: it is a monotonically increasing
curve, and therefore the model will
be unstable for any AoA in the flight
envelope.
We can take this even further by
tracing other Cm vs. AoA curves for
other CG locations. For instance,
in Figure 20 I have also plotted (in
blue) the curve that corresponds to
a Blade balanced at 95 mm from the
leading edge. This one is no longer
monotonic but has a maximum at
about two degrees. Balanced in this
way, the model is neutrally stable.
Strictly speaking, it will be neutral
only at the angle of attack for which
the dCM/dAoA derivative is null, but
in practice the CM shows such a
slight variation, and for such a broad
range of AoA’s, that the stabilizing
or de-stabilizing effect may be
negligible. The model is not willing to
react to any disturbance, but will fly
at and maintain any angle of attack
Figure 20 we can trim it to. As a consequence, it
will require a pilot’s constant attention
moment coefficient versus AoA curve The steeper the slope of the CM vs. AoA and input, and personally I would not
is decreasing from left to right is stable; curve, the greater the stabilizing or de- enjoy flying it at a slope with a brisk
on the other hand all configurations with stabilizing action will be. breeze.
an increasing CM vs. AoA curve are If we want to decrease a model stability, The use of camber flaps and
unstable, because every disturbance or even make it unstable, we just need variable camber
results in a pitch unbalance that tries to to move the CG aft; we can test that The Blade, like many other gliders, has
further increase the disturbance effect. by computing a new polar and placing control surfaces (inboard flaps and

44 R/C Soaring Digest

Figure 21

outboard ailerons) along the whole

wingspan. This configuration allows
for changing the wing camber in flight
by taking advantage of the transmitter
mixing capabilities.
What’s the use of this feature?
Generally speaking, any variation in
wing camber results in a modification
in the polar curve, and in particular
results in an upward or downward
shift of the Cl vs. AoA curve while
keeping the difference in drag at a
minimum. Some airfoils are more
suitable for this task than others,
but the idea is that variable camber
lets the pilot optimize the model
Figure 22
for different flight conditions, thus
extending its envelope. the best setup for different conditions. that will pop up (Figure 21) we check the
It goes without saying that we are Our first step will be to go back to the “T.E. Flap” box, then we input the desired
extremely interested in investigating this XFoil routine (Application > XFoil Direct deflection angle (we can start with +2
kind of optimization. We will therefore Analysis), then choose the RG-15 7.8% degrees, i.e. 2 degrees of downward
use XFLR5 to analyze different camber airfoil from the airfoil drop-down list and deflection of the trailing edge flap) and
configurations, and then we will compare invoke the “Set Flap” command from the hinge position as a function of the
them against each other to find what is the “Design” menu. From the window chord (approximately 77.5% for the Blade

February 2008 45
 versus AoA plots (View > Polars, then
right-click > Graph > Variables) we
will probably see a tangled mess of
curves, but if we clean up things a bit
(Polars > Hide all polars) and then we
plot only a single Reynolds number
for each airfoil modification (by
choosing the airfoil and the desired
Re from the drop-down lists in the
toolbar, then checking the “Show
Curve” box) we should see a graph
like the one shown in Figure 22. This
graph shows that, in a limited AoA
range, different flap settings result in
a conspicuous variation in Cl, with a
limited influence on Cd.
We might now feel the temptation
to see what flap setting allows for
maximum Cl/Cd ratio, but we must
always keep in mind that these
are the 2D polars, and the results
may not be immediately translated
into performance prediction for the
complete model. My suggestion
is to spend some more time and
“persuade” XFLR5 to simulate the
complete model with differently
Figure 23 cambered airfoil sections, so that our
investigations will be based upon a
wing flaps) and of the airfoil thickness the polar family exactly like we did for 3D model of the plane behavior.
(100% in our case, because all wing the unmodified airfoil, and we might To set up a 3D model of the Blade
control surfaces in the Blade are top also want to repeat the whole process with flapped airfoil, first we shall enter
hinged); then we click the “OK” button for a broader range of flap deflection the 3D module of the program (via the
and we store the modified profile under angles (in addition to 0 and +2 degrees, “Application > Wing Design” menu).
a suitable name (for instance, “RG-15 I suggest +4, -2 and -4 degrees). Now, Then we recall the already-processed
7.8% +2”). Now we must pre-process if we look at the Cl versus AoA and Cd Blade model via the drop-down list

46 R/C Soaring Digest

in the toolbar, and we duplicate it
(“Wing/Plane > Current Wing/Plane >
Duplicate”), we assign a new name to
the new copy and, via the geometry
definition routine, we modify the
wing by choosing one of the flapped
airfoils (for instance, “RG-15 7.8%
+2”) for all the wing stations from
the root through the 850 mm station
(which corresponds to the aileron
outer tip). After we’ve finished, we
dismiss the window, we process the
3D polar with the same procedure
we’ve already been through, and we
jump to the polar graphs. The speed
polar (Figure 23) is now telling us that
two degrees of flap deflection result
in a small improvement in sink rate for
airspeeds below 10 m/s, but things
get rapidly worse for higher airspeed
values. Glide ratio is also worse in all
the speed envelope but in a small arc
between stall speed and10 m/s. This
means that dialing in two degrees of
flap deflection to increase the wing
camber of the Blade is a good idea
only if we plan to slow the model
down to about 10 m/s. Should we Figure 24
want to do so? The answer is not so
easy... it is common knowledge that the
sticking the model up in some tree top; the model will stall; if we fly faster than
Blade doesn’t like to be slowed down,
in these cases we may want to slow the 10 m/s we will not be attaining minimum
and undoubtedly it is no F3J model,
model down and dial some flaps, even sink. Increasing camber makes this more
and flying at the minimum sink rate
if flying at the minimum sink airspeed and more critical, because it makes the
is not its intended best. But in some
is not a funny exercise. Why? Just look polar curve “narrower” and “sharper.”
cases, sinking at 50 cm/s instead of 70
at how close to stalling is the minimum There is also another effect to be
cm/s may make the difference between
sink airspeed. If we slow below 9 m/s considered. If we look at the CM vs. Vinf
returning safely to the landing zone and

February 2008 47
 tendency to oscillate around the
desired airspeed rather than keep it
throughout the flight.
This behavior can be investigated via
XFLR5 by creating a modified NACA
0007 airfoil with a leading edge flap at
75% chord, deflected by -1.5 degrees
(upward deflection, that is), and
instructing the program to simulate
a Blade endorsing this airfoil section
in its V-tail. We will get a CM vs. Vinf
graph like the one shown in Figure 25.
The model with “up trim” (magenta
curve) is now balanced at an airspeed
of less than 10 m/s, but the slope
of the curve near the balance point
is much steeper than the slope of
the curve that describes the “zero
trim” Blade (blue curve) at the 15
m/s balance point. This means that,
for instance, any drift from the trim
airspeed for the “trimmed up” Blade
will result in a pitch response three
times stronger than we would have
with the neutrally trimmed model;
therefore, we may expect a stronger
tendency to pitch oscillations.
Figure 25
Obviously there is a way to get the
model to balance at 10 m/s without
graph, Figure 24, the curve shows that as slow as required. But there’s a catch: requiring so much up trim; we just need
the zero trim airspeed is about 15 m/s, as we trim the model in this way, the to move the CG aft. If we re-compute the
slower than the trim speed for when pitching moment to airspeed relation model polar by setting the CG at 88 mm
no flaps were used, but faster than our will change, and in particular when from the leading edge, we see that CM
target speed for which more camber is we trim for “nose up” the model will goes to zero at the desired airspeed of
useful. This means that we should also become more pitch-sensitive to airspeed 10 m/s (Figure 26), and that the slope of
dial some up trim to make the Blade fly changes. This will result in a stronger the curve is not too step, as predicted.

48 R/C Soaring Digest

Flying with a such a rearward CG,
though, may be challenging and
may even hide any improvement (the
model can really fly “by the polar”
only if it’s not disturbed by pilot input,
and this is not compatible with too
aft a CG). This means that the “best”
combination of CG position and
pitch trim (or decalage) results from
a trade-off between aerodynamic
behavior and pilot skills. A model
set up with a rearward CG is more
pitch-sensitive and may be trimmed
for a broad range of airspeeds, but it
demands the hands of a keen pilot;
a more forward CG likes to fly at an
airspeed only, is less versatile but is
easier to fly. When one is involved in
competition flying, the temptation of
moving the CG more and more aft is
always strong. At first the model will
become more versatile and suitable
for different tasks, but at the same
time it will be less easy to fly, and
sooner or later the pilot will that he
is losing performance instead of
improving.
The very same method we have just Figure 26
applied may be reiterated for more
flap settings, yielding a broad family these graphs are very interesting. For at the minimum sink rate (or, in other
of model polars. If we complement the instance, we can see that there is no words, flying as long as possible) is not
two polars we already have with others, single flap setting which is consistently the same as flying at the best glide ratio
calculated for flap deflections of +6°, +2°, better than others in the entire speed (or flying as far as possible), even in still
-2°, -4°, and then we ask XFLR5 to plot envelope. And, perhaps less obvious, air. It is part of a pilot’s skills to tailor his
all the speed polar and glide ratio curves we also see how difficult it is to design flying style to the particular task and to
in a single chart, we will get something and fly a true “multi-task” model. Flying the particular weather conditions.
similar to Figure 27. If we look closely,

February 2008 49
 airspeed, especially at a turbulent
slope where it would be virtually
impossible to tell the former from
the latter. But, just for fiction, let’s
suppose we want to bring our Blade
to a thermal duration contest on a
flat field. In this case it’s easier to
distinguish the two flight conditions,
each in its own particular “task.”
We will seek the best glide ratio
when we want to explore as much
air mass as possible while hunting
for a thermal, and we will struggle to
fly at the minimum sink rate in dead
air, where there are no thermals to
be found, or while we are trying to
gain as much altitude as possible
in a weak thermal. In addition, if we
need to penetrate headwind or to
escape from a sinking air zone, it is
also important to fly faster than the
best glide ratio airspeed. In other
words, it becomes important for a
pilot to exploit the “extended polar”
of its model, that is the curve arising
from the envelope of the many
different polar curves associated with
Figure 27 variable camber. To be successful
in this, we need to study the curves
and match the theoretical capabilities
Trade-offs and conclusions need to deflect the leading edge flaps
of the model with our piloting skills. For
by 4 degrees downwards and keep an
By looking at Figure 26 we see that example, from the charts we see there
airspeed of 9 m/s (20 mph). We already
the 1.9-m Blade has its peak efficiency is hardly any gain, either in sink rate
mentioned that the Blade is a slope
(best glide ratio) when flying at about or in glide ratio, when we deflect the
model, designed to fly fast and that
11 m/s (25 mph) with “flat” wing (no leading edge upwards by more than two
probably nobody would want to fly it at
flaps deflected), and that when we want degrees; this may be translated in a radio
the minimum sink or at the best glide
to descend as slowly as possible we setup that will allow for a “speed” flight

50 R/C Soaring Digest

phase with no more than two degrees of pull on the yoke of a glider, in fact, we elevator deflections required for minimum
up deflection for both flaps and ailerons. are shifting the operating point towards sink and best glide airspeeds, but I later
Conversely, the charts suggest that we lower airspeeds, and we have seen that found that these are consistent with the
may improve the sink rate by deflecting increasing wing camber at the same time not-so-good accuracy (half degree) of
the leading edge downwards by 2, 4 may be a good idea. The snap-flap mixer the homebuilt incidence meter I used to
or even more than 6 degrees, but... the may be seen as a semi-automatic way measure the rigging angles of the main
more deflection we use, the narrower to choose the most suitable polar for wing and the V-tail group.
the polar curve will become, and the the conditions we are “requesting” with My conclusion is that a simulator such
closer to stall the operating point will the pitch control. Setting up the snap- as XFLR5 may be a valuable and useful
be. The pilot must find out how close flap mixer is not trivial, but again a good tool, provided that the user knows what
to the limit he is still comfortable, and simulation of the plane behavior, paired to look and ask for, and that he always
able, to fly; in any case we have seen with a good understanding of what we keeps in mind the limitations of the
that slowing down the model either are asking the model for in terms of underlying mathematical model and the
requires a good amount of up trim (or, airspeed and load factor, may be helpful. finite precision both of our tools and of
equivalently, a rather large decalage), or The last thing I’d like to discuss is how our beloved planes. n
requires moving the CG aft. In the former XFLR5 simulation compares with reality.
case we’ll have a more marked tendency I’ve had the opportunity to test a flight
to oscillation, in the latter we will have data recording system by Eagle Tree
a more pitch-sensitive model. Only Systems. Unfortunately I could not install
the pilot, with his own judgment, can XFLR5
it in my Blade (too narrow and crowded
tell the difference between “sensitive,” fuselage) but I used it in my F3J Stork 2 Author: André Deperrois
“responsive,” “critical” and “dangerous.” Pro model. With its static and dynamic Link: http://xflr5.sourceforge.net
In my Blade I have decided I see very pressure ports (the latter connected to License : open source GNU General
little improvement in more than two a dynamic probe outside the boundary Public License
degrees of downward flap deflection, and layer), the data recorder samples Current version as of print date: 3.21e
so I have setup a “thermal” flight phase airspeed and barometric altitude ten
accordingly. times a second, so that the collected Blade 1.9
There is an additional possibility which information may be analyzed offline on Maker: X-Models
is worth mentioning, that is using the the PC. I’ve been glad to discover that Wingspan: 190 cm (75 in)
so-called “snap-flap” mix. This mix, the minimum sink setup (very important Length: 107 cm (42 in)
available in most modern radios, causes in an F3J model), that I had honed in Airfoil: modified RG-15 (thickness
a variation in the “camber” signal (which several test flights, is consistent with the 7.8%)
is then mixed and sent to flap and aileron Stork behavior simulated with XFLR5. Wing surface: 32 dm² (3.44 sq ft)
servos) whenever the pilot acts on the Camber settings for thermal flight, fast Typical ready-to-fly mass: 1400-1500 g
pitch stick. In my opinion it is a good flight and towing were also comparable (49-53 oz)
idea, and not only for fast F3F models to those suggested by XFLR5. I was
in which it is typically used. When we less satisfied with the predictions of the

February 2008 51