           "A" & "D" series airfoils by Don Ayers

I have been designing slope glider airfoil sections for a number
of years.  My first series was published in 1995 in Sailplane
Modeler, and has been made available on this forum, on commercial
plotting program library diskettes, and more recently, on two web
pages.

The first series is simply identified by the letter "A" followed
by four digits.  The first two digits denote the thickness
percentage in tenths of a percent, while the last two digits
denote the camber percentage in tenths as well.
Example: the A7018 is 7.0% thick with 1.8% camber.

The new "D" series is designated in the same manner, in order
that flyers may easily relate to the thickness & camber used,
without using a decoder ring.  However, it has a new thickness
profile & camber line, hence the "D".

My design goals have focused specifically upon thin airfoils
which can be constructed accurately using foam cores sheeted
with 1/32" plywood, hand-shaped hardwood leading & sub-trailing
edges, and triangular balsa ailerons.  This "old-fashioned"
method of composite construction can be employed by anyone
using rudimentary shop tools and facilities, and results in
extremely strong thin wings, particularly if the finished wood
is over-laminated with glass.

This wing design will withstand the severe slope abuse under high
wing loadings (+50 oz/sq/ft) which exceeds the structural limits
and landing impact resistance of typical carbon-fiber
vaccuum-bagged wings.

However, this construction technique is not conducive to the use
of most modern computer-designed airfoils, which generally
feature a precisely & delicately undercambered lower trailing
edge, and/or a slightly-reflexed upper trailing edge contour.
These concave & sharpened trailing edge profiles are impractical
if not impossible to create accurately in solid balsa ailerons,
flaps and fixed trailing edges.

In fact, the accurate replication of these trailing edges is
challenging even when using vaccuum-bagging techniques!  In many
cases, the true contour of the trailing edge is inaccurately
rendered, despite the best efforts of the builder.

There are so many variables inherent in the composite process:
printing the templates on the laser printer; cutting the upper
& lower templates, and indexing them to each other; hot-wire
cutting the foam cores; then, the pressure, temperature and
material interactions induced during the lamination process.

It is therefore not surprising when a planned 8" root chord
wing comes out to be 7.9" or 8.1" after final finishing!
An unintended vertical thickness variation of only one-thousandth
of an inch will greatly change the trailing edge contour and the
airfoil's performance, although this is erroneously viewed simply
as a "chord length mistake" to be disregarded.

I feel that the inherent difficulties in precisely replicating a
modern computer-designed thin airfoil are significant.
Wind-tunnel testing of presumably accurate wing panels may yield
performance results which vary greatly between samples of the
same section, and more importantly, may diverge significantly
from the performance parameters predicted by computer analysis.

The divergence beween real-world wing performance and
computer-modeled wing performance is greatest when ultra-low
drag thin sections are involved.  Case in point:  the S6063 shows
remarkable performance parameters in computer modeling, but the
sample(s) tested by U.I.U.C. in the wind tunnel did not meet the
rosy predictions of the computer.

Now, this one example could mean that the test panel was
shoddily-built, but I feel that is unlikely.  Even if built
"slightly off", that variance range was likely typical of that of
any experienced builder.  But, since the section is supposed to
be built and used, and not just admired on paper, achieving
sub-nominal results from a meticulous building effort is
particularly unsatisfactory!

I suspect that the S6063 section requires extreme precision in
fabrication to fully realize the expected results.  The problem
is, the precision required may be beyond the reasonable
expectations of the typical hobbyist or small business
manufacturer.  As such, less-than-stellar performance may be
obtained in a significant percentage of builds.

My theory is supported to some extent by experiments I've
performed on Calcfoil; first using a "stock" S6063 section, and
then comparing it to the same section when modified "on paper"
to approximate typical builder's shortcuts and errors.  For
example, if the tiny aft undercamber is removed from the computer
profile, the polars are dramatically disturbed for the worse.

This type of sensitivity to slight building distortions and
errors leads me to label this particular section, and others like
it, as "hypercritical", for lack of a better term.  I feel that
95% of builders should reap the maximum performance from a given
airfoil on every build, not just those high-budget perfectionists
who cut a solid aluminum wing plug on a CNC mill, and use it to
make a hollow wing mold and perfect wings!

Hopefully, you are now understanding my interest in the
traditional "old-fashioned" type of ply-sheeted wings described
earlier!  When wings feature a triangular trailing edge, and lack
any concave surface features, a uniformity of build quality &
performance is more readily anticipated and obtainable.  The
aforementioned 1/1000" thickness variation error in overall
thickness profile has little or no practical impact on the
performance of these "simpler" geometry airfoils.  As a result,
the average builder can expect maximum performance for the many
hours and dollars expended on his thin wing construction, and not
just hope that he will be "the lucky one" who ends up with a
perfect S6063 wing, without a vicious tip stall or a
narrower-than-expected max Cl.

In response to the need for a wing that not only fits the
"ply/balsa" building parameters, but also offers a wider
performance envelope than my "A" series, I have recently
completed my "D" series of thin airfoils.

The "A" series was designed for an extremely narrow drag bucket
and very low drag at a zero angle of attack, while maintaining
the high cambers found necessary for heavy-wingloaded PSS type
models.  This results in a high speed but "difficult to turn
sharply" glider, requiring a large slope and broad sweeping
manuevers to maintain momentum.

The "D" series, in contrast, was designed with a much wider drag
bucket now extending well into negative lift, at the expense of
slightly increased drag.  The camber has been reduced in order to
achieve lower drag at zero lift than that of the "A" series; this
aids in terminal dive speed and in efficiency during negative G
maneuvers.

I expect that the "D" series will not handle massive wingloadings
as well as the "A" series, but that is something that most
slopers do not have the luxury to play with anyway!

[On a related note, I am interested in determining the record for
the highest wingloading ever flown in a glider model less than
56" in span.  By "flown", I don't mean "tossed off into the
brush"; I mean sloped in great conditions, beating all lighter
planes for speed & style points for at least an hour, and then
landed without damage to roaring applause!  We have gone as high
as 55 oz/sq/ft, which is over double the FAI racing limit.
Let me know what you've done!]

The "D" series should be used for the same range of high-speed
projects usually handled by the S6063 or thinned versions of the
RG-14, MH22, MH30, etc.  The difference is that the computer
projected performance should be easily attainable, unlike a
typical attempt at a S6063, and the performance should be far
better than that of the typical "thinned" airfoils mentioned
above; again, as indicated by Calcfoil.

This time, I have departed from my usual schedule of
"design/build/test/crash/re-design/rebuild/retest/etc", and have
gone out on a limb....  I have decided that I would release this
series of eight airfoils now, before I have tested even one of
them on a real glider, thus allowing any other brave and
experimental souls to try one during this winter's building
season.

"WHY??" you ask, "would we risk a wing build on a section that
no one has flight tested yet???  Are you nuts?"

Perhaps, but I have two good reasons.

First, someone's gotta do it.  That's why they call it "testing."
No guts, no glory!

Second, this series produces great performance numbers on
Hepperle's "Calcfoil" computer program, and so it looks like a
low-risk experiment.  Unlike the "A" series, the "D" series is
based on a very conservative design philosophy; the thickness
high point is moved further forward, the nose contour is widened,
and the camber is reduced.  While retaining my trademark
"triangular trailing edge" contour which promotes the simple
build techniques, I have at last obtained performance "on paper"
which rivals or exceed that of more complex "undercambered"
sections, and definitely exceeds the typical "thinned" versions
of the RG-14, etc.  See for yourself!

In summary, I do not dispute that the S6063 has been optimally
designed for the smoothest Cl curve and the lowest drag bucket
possible in a thin airfoil; I believe that "on paper" it is the
finest high-speed airfoil extant, and Dr. Selig should be
commended for his superior work, as usual.  However, I suspect
that the average builder will not experience those superlative
results on a real plane unless he is either very lucky or is a
brilliant craftsman.

For example, the S6063 has been employed on and off since its
inception on a large variety of PSS models built by an assortment
of local flyers who frequent Point Fermin, Bluff Cove and Parker
Mountain in Southern California.  These guys live for high
performance. If their S6063 planes gave the very best results,
everyone would be using the S6063 today.  They aren't.

Based on my experience, I feel that my new "D" series will be
easy to build effectively due to the "more forgiving" accuracy
tolerances inherent in the design.  The "D" should actually
deliver the same great performance that is predicted on paper,
by avoiding "rolling the dice" on a delicate and precisely
concaved trailing-edge section.

My main motivation for releasing the airfoils prior to flight
testing is that I'm moving out of Southern California in six
months, and I will not be able to adequately shake down the
series before my departure. To make matters worse, I'm relocating
to a dismal slope area of the country, and will never be able to
enjoy the fruits of my labors until I take a trip back to the
Coast!  The worst of it is...I may have to learn to... (gasp)...
THERMAL... at my new home!  The spectre of future events has made
me consider that I might lose momentum on this project, and
resutling delays in my personal flight testing might then push
the release of "field-tested sections" into the next century!

So, I figure that now's the best time to release the foils to all
of my sailplane-flying buddies out there, both known and unknown,
with the hope that they will prove worthy of both my design
efforts and YOUR efforts in building & testing!

                               TEST DATA
I have prepared the following charts using the test output of
"Calcfoil", in order to pique the interest of the technically
astute readership of RCSE:

==========================================
AA airfoil series / D. Ayers / copyright 1998.
S6063 shown for comparison.
Data generated by Calcfoil using 2nd order panel method.
==========================================
Re = 200000
Data at MAXIMUM LIFT COEFFICIENT:
------------------------------------------
       Alpha  MAX Cl    Cd      L/D 
------------------------------------------
D6213:  4.8   0.65   0.0141   46.42  
D6214:  5.0   0.68   0.0144   47.36 
 
D6515:  5.5   0.74   0.0154   48.42  
D6516:  5.7   0.77   0.0158   49.02
 
D7016:  6.1   0.81   0.0166   48.77  
D7017:  6.4   0.85   0.0172   49.22
 
D7517:  6.8   0.89   0.0183   48.58
D7518:  7.0   0.92   0.0188   48.62

S6063:   5.2   0.72   0.0143   50.64
==========================================
==========================================
Re = 200000
Data at MINIMUM DRAG:
------------------------------------------
       Alpha    Cl    MIN Cd    L/D 
------------------------------------------
D6213:  0.0   0.13   0.0076   17.11
D6214:  0.0   0.14   0.0076   18.31
 
D6515:  0.0   0.15   0.0077   19.42
D6516:  0.0   0.16   0.0078   20.60
 
D7016:  0.0   0.16   0.0079   20.43
D7017:  0.0   0.17   0.0079   21.58
 
D7517:  0.0   0.17   0.0083   20.64
D7518:  0.0   0.18   0.0084   21.73

S6063:   0.0   0.15   0.0078   18.90

(Notes: The D6515 surpasses the S6063 in both minimum drag AND
L/D at zero A.O.A.  The D6516 & S6063 show equal miminum drag,
but the D6516 shows better Cl & L/D.)
==========================================
==========================================
Re = 200000
Data at ZERO LIFT:
------------------------------------------
       Alpha  ZERO Cl   Cd     L/D 
------------------------------------------
D6213: -1.1   0.00   0.0085    0
D6214: -1.2   0.00   0.0085    0
 
D6515: -1.3   0.00   0.0086    0
D6516: -1.4   0.00   0.0087    0
 
D7016: -1.4   0.00   0.0088    0
D7017: -1.5   0.00   0.0088    0
 
D7517: -1.5   0.00   0.0089    0
D7518: -1.6   0.00   0.0089    0

S6063:  -1.3   0.00   0.0079    0

(Notes:  There is no free lunch!  The S6063 has the lowest drag
of all at zero lift. This means that from identical altitudes,
and with all else being equal, the S6063 would dive fastest of
all.
The question is: using a S6063, can you GET a heavily-ballasted
speed plane to the same altitude as a plane using a "D" section
with a higher max Cl, and if so, will you pull out of that dive
with more or less retained energy on the vertical climb back up?
==========================================

I hope this introductory article and data table has been of some
interest to the readership.  Thank you for your consideration of
this latest series, and please let me know of any projects
undertaken and results obtained with any of my sections!

Best regards,

Don Ayers
c/o chetpado@myowndomain.com
