Since 2003 I have taken numerous drag and performance measurements on Standard Cirrus glider #60 with various deturbulator configurations.
Also, I have done a great deal of soaring with deturbulated wings. In addition, independent performance measurements were taken by the
legendary Richard H. "Dick" Johnson in December 2006
(see A Fight Test Evaluation Of The Sinha Wing Performance Enhancing Deturbulators).
Altogether, these efforts have produced a large body of evidence
for performance improvements using the deturbulator flow-control method.
Much has been learned about the potential benefits, but much research and engineering remains to be done.
Deturbulator Research is committed to encouraging and assisting the efforts of others to understand and use deturbulation methods.
A full deturbulator configuration consists of (1) a thin (.003") deturbulator tape wrapped around the leading edge of the wing that produces a
very small referse-facing step as seen by the boundary flow soon after it splits at the stagnation point and (2) a deturbulator panel
located at .6 chord on the upper wing surface. More will be said about these below.
Performance with Full Configuration
Because deturbulation on the upper wing surface depends critically of precise control of several variables,
consistency is hard to achieve. Thus, flight-to-flight differences are often extremely large. This means that the traditional
means of reducing scatter in sink rate measurements by averaging data points
obscures real performance changes that reveal the true potential of the deturbulator method. For this reason, it is often necessary
to study individual measurements in order to see the true effects of deturbulation. Examining individual measurements
is not unreasonable, as long as one keeps in mind that the results are approximate.
Confirmation of this may be seen in repeating patterns in Figures 1-3 below.
The Johnson measurements in 2006 assumed consistent performance.
So, Johnson averaged all six of his deturbulated flights and reported a 13% improvement at
50 kts. But, upon examining the individual flights, he found huge deviations (4-5 times normal) and decided to throw out the three
worst offenders. This lead to an 18% improvement at 50 kts. However,
a close look at the six flights revealed non-random patterns.
Furthermore, an onboard flight data recorder made it possible to
corroborate his manually acquired data and to identify a few errors that
made two of his individual polars appear completely unreasonable.
Figures 1-3 below illustrate some selected polars with repeating patterns indicating that the large deviations are indeed real.
Figure 1. Peaked Polars
Figure 1 shows three individual measurements of a "peaked" polar pattern in which performance reaches a high peak at 50 knots indicated air speed (KIAS) and
degrades on each side. Two secondary peaks occur at higher speeds. This is the typical polar when the deturbulator panel
skins are free to move without restriction and properly ventilated.
Figure 2. Notched Polars
Figure 2 shows four measurements of a "notched" pattern in which performance bottoms out to a consistent value
near 50 KIAS and improves considerably on each side. This is a consistent pattern when the deturbulator panel skins are stuck down by condensation,
are too tight or the panels are removed entirely.
This has been confirmed by measuring the performance with the deturbulator panels removed, but leaving the deturbulator tapes in place.
For more on these measurements see
Leading-Edge-Tape-Only Performance Measurements.
These two patterns, peaked and notched, were
first seen in December, 2006
when Richard H. "Dick" Johnson independently took measurements.
(Read his report: A Fight Test Evaluation Of The Sinha Wing Performance Enhancing Deturbulators.)
Figure 3 shows another repeating pattern, the trend over time of performance, with deturbulator tapes and panels installed and the panels working,
while holding the airspeed constnt at on 50 KIAS. This is a 36 second moving average of performance plotted in 4 second intervals.
These plots indicate that, with the onset of extreme performance, the lift coefficient tries to increase but cannot since the pilot is
holding the airspeed constant. This pitches the nose down, to keep the lift equal to the weight of the glider.
With the reduced AOA and ideal flow configuration, induced, skin friction and form drag are reduced for a high level of performance.
However, the reduced AOA changes the top surface flow trajectory so that the reattachment point shifts away from the deturbulator panel.
The lift coefficient then tries to reduce. So, the nose pitches up again to keep the lift equal to the weight.
This causes the performance to subside. It is thought that this cycle will repeat every 4 or 5 minutes if the pilot can hold the airspeed precisely constant.
Figure 3. Moving Average at Fixed Airspeed
Changing Flight Dynamics at Onset of Extreme Performance
Sometimes the onset of extreme performance is dramatic. Following is a video clip of such an event.
It shows a large downward pitch change while the glider is flying in smooth air.
Once the event begins, it can be seen that the airspeed does not change appreciably, yet the nose pitches down greatly as the
vario shows the sink rate dropping. At first, the glider is slowing from 60 kts to the magic 52 kt airspeed
(blue hash mark below 3 o'clock position) where extreme performance occurs.
During this time, the nose slowly pitches up, as it should, and the glider gains 29 feet of altitude.
Then, precisely at 52 kts the nose quickly pitches down sharply and holds the new attitude
while the variometer moves from 4.5 to 1.7 kts. This is dramatic evidence of a sudden increase in the lift
coefficient. With the pilot holding the airspeed steady, the angle of attack must reduce to spill off the additional lift so
that the lift force continues to match the weight of the glider.
The deturbulator device itself is a 2 inch wide strip (panel) consisting of two layers. The bottom layer is a rigid substrate with
shallow spanwise grooves on the surface that are spaced 2 mm apart. The substrate is covered with a thin membrane (skin)
that separates air trapped between the ridges and the surface flow. Panels are usually one to two feet long and
are installed perpendicular to the flow (spanwise) just behind the flow reattachment point where attached turbulent flow
normally begins. The skin must be light and flexible enough to have essentially no influence on the flow-surface
interaction modes. Conceptually, the boundary flow in contact with the skin reacts to compressible rows
of air that are fixed by the substrate ridges and the overlying skin.
Figure 4 is a cross section view of a deturbulator panel.
Figure 4. Cross Section of a Deturbulator Panel
To date, substrates have been made from 2 inch wide aluminum tape. Lengths of this tape are passed through rollers
that impress the grooves into the surface. The latest material that I have been using for skins is a polyester
(polyethylene terephthalate) mesh called PETEX, that is produced by Sefar AG
and can be obtained from
Small Parts, Inc.
This mesh is woven from 34 um threads and has 1% open area. The porosity serves mainly to allow evaporation of condensed humidity
under the skin which would replace air in the ridges and cause the skin to become immobile.
Dr. Sinha will provide substrates to seriously investigators. The only other materials needed
are various tapes that are commonly available.
The jury is still out on whether or not the PETEX mesh actually works. This is because the measurements so far give a notched polar resembling
the performance when removing the deturbulator panels all together and leaving the leading edge tapes in place. This, however, may be due
to the vent design that may suck the skins down too tightly.
Deturbulator panels are attached directly to the wing surface using the adhesive on the back side of the aluminum
substrate. A certain amount of art is required to press out wrinkles that form on the edges because
of stretching from the rolling process.
Ideally, there should be a perfectly smooth transition from the wing surface to the deturbulator skin. However, reality being what it is,
we must shoot for the thinnest and smoothest possible transition. For attaching the skins, I use thin 1/2 inch Scotch Permanent Double Sided Tape.
I place it on the wing surface abutting both edges of the substrate. The skins are cut 2 3/4 inches wide, allowing 3/8 inch overlap onto the double
The skins are placed flat over the substrates without tension.
If the boundary flow is to interact with rows of trapped air beneath the skins, it is critical to have the skins lying in contact with
the ridges but not tensioned. It appears that the skins work better loose rather than tight. Too tight skins appear to reduce the
performance and also produce the notched polar pattern seen in Figure 2.
PETEX mesh was chosen largely because of it's dimensional
stability with changes in temperature and humidity.
It may be best to install the skins a little loose and then rely on the ventilation system to maintain proper contact with
the substrate ridges. In any case, vents are needed to keep the skins on the top surface of the wing from ballooning up and to
deal with static pressure variations from altitude changes.
Discrete vent holes between the panels operate more efficiently exhausting air than admitting it, since with increasing static pressure,
the skins press down between the substrate ridges and restrict flow. This leads to a spanwise variation in skin tightness and, presumably,
a spanwise variation in deturbulator function.
Next, you will need to join the panel skins in a way that provides appropriate ventilation. More will be said about this critical issue
later. I apply a short strip of double sided tape into the substrate joint to hold down the skin at the joint. But, depending
on the vent design, I leave a place near the leading edge (where the pressure in the adverse pressure gradient is lower) open
to allow air movement in and out. Figure 5 shows a rear-facing vent design that I think produces too much suction and
so causes the notch pattern in Figure 2. Also, it introduces a speed dependency that may affect high-speed performance.
Figure 5. Installation Detail of Deturbulator Panels
The last step is applying a thin single sided tape over the leading and trailing edges of the skins. This holds the
top side of the skins, covers the exposed double sided tape and smooths the transition from the wing surface to the skins.
For this, I prefer 3/4 inch Scotch Gift Wrapping Tape. Ordinary transparent tape works too, but Gift Wrapping Tape is easier
to remove. Do not use Scotch Magic Tape (invisible) because it is almost impossible to remove in one piece.
I have taken oil-flow images that show the flow tripping at the leading edge of deturbulators that were too thick and had a
sharp leading edge step. As a precaution you might smooth the leading edge of the covering tape with paste wax or something else
that will not harm the wing surface.
Basic Deturbulator Operation
The basic idea of the deturbulator panel is to intercept the separated flow where it reattaches to the surface,
and begins transitioning to turbulent attached flow, and to modify the transition process in a way that results in mostly small, high-frequency
turbules rather than the full range of sizes. That is, we wish to cascade incipient turbules into
spin off turbules immediately as they leave the surface. Whether this is in fact what happens remains to be seen, but that was the original the goal.
For this to happen, the separated flow must return to the surface at a grazing angle, not steeply on the back side of a circulation bubble
as normqally occurs. Preconditioning the flow this way is the function of the leading edge deturbulator tape that creates a
small rear-facing step that prematurely detaches the laminar flow and steers it on a trajectory that follows the surface closely
until it reaches the reatqachment point. This mechanism alone offers major efficiencies that the deturbulator panel further enhances.
To illustrate a deturbulator panel working, consider Figure 6 below. This oil-flow image was chosen because it offers stark
evidence of modified boundary flow. In this case, the deturbulator panel is located on
the lower wing surface where deturbulator operation is much easier to achieve.
First, notice the normal flow pattern on the left
side of the image. The flow direction is from bottom to top. At the bottom, you can see a thickening of the
oil streams as the flow moves toward the bubble area. This is where the laminar flow separates from the surface, reaching its
highest point over the circulating bubble that traps the line of oil that is clearly evident. On this wing, the bubble on the
lower surface moves fore and aft with small changes in airspeed. This blurs the bubble edges, but the bubble can be seen
anyway. The most important feature to notice on the left side is the light line at the rear of the bubble. This is
where the flow strikes the surface at a fairly steep angle and the violent force wipes the surface almost clean.
Behind that line, the flow becomes turbulent over a short distance, as can be seen by the thickening streams. This is the
place where we wish to modify the transition to turbulent flow.
Figure 6. Panel on the Bottom Side of Wing Behind the Reattachment Point
In the middle of the image you see a silver deturbulator panel. This panel has an aluminized Mylar film membrane.
The striking thing about this image is the way the panel causes the surface ahead and behind to be wiped clean. This is
definitely not the kind of behavior an ordinary piece of tape would produce. The surface looks like the leading
edge of the wing (lower right corner of the image) where laminar flow wipes the oil away.
It is interesting to note that the flow ahead of the panel is modified, even encroaching on the bubble. Another point to note
is that the attached turbulent flow leaving the wing behind the panel leaves oil streams that are noticeably thinner than the unmodified streams
on the left side of the panel. This indicates that the boundary layer is not as thick behind the panel.
Now, what happens when fully turbulent attached flow strikes the panel. The image shows an example of this too. A patch of
thick tape was used to cover a tie-down attachment hole. We can see that this is tripping the flow at its leading edge.
This kills the bubble, since the flow is already turbulent and attached to the surface. The streams
following the patch indicate turbulent flow approaching the deturbulator panel. It is plain that in this region, the panel
is inoperative. In fact, the thick streams leaving the panel indicate a thicker turbulent boundary layer.
The panel is enlarging the turbulent boundary layer.
Now notice that the aileron gap seal behind the deturbulator panel is tripping the flow. This can be seen from the
streams initiating at the leading edge of the seal. I take this as evidence that the regions of cleaned surface surrounding
the deturbulator panel do indeed indicate laminar attached flow.
One final note, this instance is not illustrative of a panel on the top surface operating in a mode that
produces large performance increases. It only shows that deturbulator panels, located in the reattachment
zone, may in fact function dynamically to modify the transition to turbulent flow.
At the beginning of this project,
drag reductions were easily achieved on the lower surface.
I recommend that other researchers begin in this way for early evidence of deturbulator performance.
Top surface results are much more difficult to achieve.
When we began working on the top surface, the drag only increased.
I surmised that on the suction side, the skins were lifting up and flapping in the flow.
Since the pressure gradient is adverse (increasing in the flow direction) where the panels
are installed, ventilation ports between the panels at their leading edges. where the pressure is lowest,
nullified the excess drag. But still there was no drag reduction.
Finally, we placed 2" (5 cm) wide packing tape around the leading edge of the wing.
The idea was to provide a small down-step in a region where the boundary layer is very thin and is moving
under the influence of high pressure and a strong negative pressure gradient in order to initiate premature flow detachment.
The thickness of the tape had to be small enough not to trip the flow. After that, we began measuring drag reductions.
For more leading-edge deturbulator tapes, see Modify Your Glider with Leading Edge Tapes.
Bottom Surface Installation
The first successful panels used .00025" Mylar film for the skin.
They were attached to the bottom of the wing just behind the flow reattachment point as illustrated in Figure 6 above.
When used on the lower surface, the skin was attached with Scotch Tape all the way around with no venting holes.
The skins would sag under the force of gravity, but produced consistently good drag reductions.
Top Surface Installations
The basic top surface installation is pictured in Figure 7. Heavy black lines indicate the locations of the deturbulator tape and the
deturbulator panel. The green boundary layer thickness lines and the upper and lower surface pressure plots represent the
unmodified airfoil flying at 55 kts.
It is noteworthy that the Wortmann airfoil produces a transition bubble on the top surface that is stationary for
normal airspeeds. This is largely responsible for success with the top surface configuration.
Figure 7. Position of Leading Edge Tape and Deturbulator Panel on Top Surface
We used the same configuration when we began deturbulating the upper surface, but quickly found that
venting was needed to pull the skins down into contact with the subrtrate ridges. For this we left an open gap between
the panels near the leading edge where the pressure over the panels is lowest. This worked to pull excessive air out from beneath the skins.
Figure 8 shows a typical vent used for the
Johnson Flight Tests (12/2006).
It was crude, but worked well.
Figure 8. Vent Used for Johnson Flight Tests in 2006
This configuration produced the green curves in Figures 1. However, it had a persistant problem with condensation that
limited flow-surface interaction modes.
After the Johnson tests,
we had to install a new set of deturbulators for further testing. After two failures,
we installed a set using fiber reinforced Mylar, a material made for racing yacht sails. Figure 9 illustrates a
vent hole punched in the aluminum tape joining two of these panels.
Figure 9. Fiber-Reinforced Mylar Skins with Vent Hole
Aluminum tape was used to attach the skins directly to the substrate edges. This made a uniformly flat joint,
but the step height from two layers of aluminum tape was too square and too high so that it tripped the flow
before reaching the skins. The solution was to smooth the sharp
leading edge with thin tape. After that, this configuration produced the blue curve in Figure 1, which replicated
Johnson's third flight, green in Fig. 1. This was the first confirmation that Johnson's measurement was real.
This configuration, however did nothing for the humidity problem, so it often failed because the skins were stuck down with condensation.
To eliminate the condensation problem, I began testing porous membranes. The idea was to allow water vapor to pass directly
through the skins so that the humidity level would quickly equalize between the flow and the air under the membrane.
Condensation would have an extremely short evaporation path to the outside.
Rather than trying to keep moisture out, this method allows it to freely communicate through
the skin. I also thought that the pores would constitute a distributed ventilation system. It now appears
that a non-uniform porosity pattern with greater porosity near the leading edge is needed for this.
For my first tests, I chose Nylon parachute fabric with a weight of 1.2 oz/yd2.
It was installed without vent holes.
Unfortunately, Nylon is not dimensionally stable enough for consistent results.
Figures 11, 12 and 13 show the Nylon before two flights on 5/22/2009 and afterward.
The red curve in Figure 1 above shows the performance measured on the 2nd flight.
It produced the 3rd instance of a peaked pattern at 50 KIAS.
(click image to enlarge)
Figure 11. Before 1st Flight
Figure 12. Before 2nd Flight
Figure 13. After 2nd Flight
On 6/2/2009, a second measurement produced the second notched pattern at 50 KIAS. On that occasion, the Nylon was
tight before launching and very tight upon landing.
The difference in skin tension for these two measurements suggests that little or no tension produces a peaked pattern,
while taut skins produce a notched pattern. This is consistent with
two measurements by Dick Johnson in 2006.
Recent measurements with the deturbulator panels removed and the leading edge tapes in place also produced the notched polar pattern.
This coroberates the idea that notched patterns are produced by inoperative deturbulator panels while peaked patterns are produced by
To achieve better dimensional stability, I installed polyester mesh skins.
For this, I chose a mesh woven from 34 um threads with 1% open area.
This mesh is 80 um thick with a mass density of 1.9 g/m^2. This material is manufactured by
Sefar and goes by the name
As expected, this material exhibited no visible dimensional change with large changes in temperature and humidity.
Also, water trapped beneath the mesh skin evaporated rapidly with air blowing over it.
Figure 14. PETEX Deturbulator Installation (click to enlarge)
On 4/2/2010, I took sink-rate measurements with the PETEX deturbulators. The results are shown in Figure 15 along with prior
measurements also showing a notched at 50 KIAS.
Figure 15. PETEX Sink-Rate Measurements
The red curves show the PETEX data. After initially taking data over a range of speeds, I took the 52 KIAS point a second time,
holding the speed for four minutes.
During that time, the performance started out low, increased, leveled off and fell off after about 1.2 minutes. Figure 3 shows this performance trend
along with two prior measurements that yielded the same performance sequence. Figure 15 shows a point above the Glide Ratio curve representing
the best 37 seconds of the four minute 52 KIAS speed run. In addition, I took data at 49 KIAS.
This point appears below the Glide Ratio curve and indicates the true bottom of the performance notch.
The red curve in Figure 2 replaces the 52 KIAS point with the second (37 second) measurement and adds the 49 KIAS point
to define the bottom of the performance notch. The steepness of the notch is amazing, especially on the low speed side where
a one knot difference takes the performance from 28:1 to over 50:1. On the high speed side, the slope is not as steep.
The 48 and 52 KIAS performance airspeeds are very hard to hold. The glider wants to slide away from them and tends to seek the notch.
In May 2010, I began soaring with the deturbulator panels removed, but with the leading edge tapes in place.
I wanted to get an idea how much of the deturbulator performance changes were due to the leading edge tapes alone.
I had the impression that the 49 kt performance notch was consistently present and that performance was generally improved
between 52 and 65 kts in smooth air. An opportunity to demonstrate this occurred on 5/27/2010 when an out-and-return
cruise of 20 nm (37 km) in smooth air under overcast skies yielded about 25% overall improved performance. I knew that something good
was happening on the return leg. When I turned for the trip home, the glide computer indicated 731 feet
below glide slope to arrive 1000 feet above ground level (agl); i.e., I expected
to land out. However, without interruption, the glider gained steadily on the glide slope so that arrived 917 feet agl.
Figure 16 shows the glide path. Since then, the glider has performed similarly on long glides in smooth air.
Since then, three sets of measurements have been taken without the deturbulator panels installed and only the leading edge tape in place.
These are shown in Figure 2 above. All of these indicate a notched pattern with a severe sharp loss in the center of an otherwise
remarkable performance hump. This appears to be too much of a good thing, an uncontrolled, clean detachment at the rear of the tape
allowing the accelerated flow to shoot upward, heightening the bubble instead of flattening it. It is thought that moving the tape edge
aft slightly will tame this unwanted behavior. We will see.
Figure 16. Flight Path of an Out and Return Glide
To see the effect of the lower surface tape alone, I removed the upper half of the leading edge tape. That left the original tape in place from the
nose of the wing downward. This configuration produced a great surprise. I thought the notch would disappear and the polar would have a
nearly normal shape, but this did not occur. Instead, I got the red curve in Figure 17 below:
Figure 17. Leading-Edge Tape Only from the Nose of the Wing Downward
The great surprise was the large loss at all airspeeds below 60 kts (incicated). I realized immediately that this crossover point corresponded to the
zero AOA point and the for all lower speeds the the flow above the stagnation point was seeing the forward edge of the tape as a step up. The question was
whether this was tripping the flow or detaching it. The second surprise was that the notch was still there. I was not expecting it,
so I did not take points one kt appart near the notch. Also, for some reason, I omitted the 55 kt point that would have helped define the
shape of the curve in that region. Therefore, I simply connected the points in the region with a dotted line. Nevertheless, it is clear that the
notch is still there and occurs precisely at the speed measured on 1/15/2011, shifted slightly to the right because the glider is now flying about 35 pounds
lighter than before.
After some thought, I decided that, as unlikely as it seems, the continued presence of the notch indicated that the flow was detaching
precisely at the nose of the wing for all positive angles of attack. To verify this, I decided to move the forward edge of the tape down
below any point where the relative wind could see it as a step down. I chose 45 degrees below the leading edge and chose to use 12mm wide
Test 4104 tape because it had the right
thickness...about .003" (76um). Figure 17a below shows the results of this change. It performed
exactly as anticipated. The severe loss at positive AOAs was replaced with a nice improvement of approximately 15% and
the notch disappeared. This is the first time something has turned out exactly as I expected, with no surprises.
And it is the first truly practical outcome!
Figure 17a. Leading-Edge Tape Only from 45 Degrees Below Nose of the Wing
One method of visualizing boundary flow patterns is to paint a thin layer of used, high-viscosity engine oil on the wing
surface and then fly the aircraft at a constant airspeed. Afterward, the painted areas can be photographed and studied.
If the oil is viscous enough and the air temperature is not too warm, the oil patterns will remain intact for a long time.
Figures 18 and 19 show normal and deturbulated oil visualizations of top surface boundary flows. Click an image for an expanded view.
Click here for a primer on reading oil images.
Figure 18. Normal Top Surface Oil-flow Pattern
Figure 19. Deturbulated Top Surface Oil-flow Pattern
The difference in these images is immediately obvious. The deturbulator strip in Figure 19 is located behind the normal transition bubble
that is so obvious in Figure 18. Click on Figure 19 to expand it and you will see that in the normal transition bubble region the oil
has been smoothed, indicating gentle reverse flow. Ahead of that region, a stagnant pattern indicates a broad region of detached flow
without the usual streaming from air flow on or very near the surface. The end of thick runs near the leading edge witness significant
early detachment, well beyond normal. Behind the deturbulator panel, again thick oil runs end within two inches of the panel, again
indicating detached flow that can no longer force the oil into streams. Normally, this region has thick oil streams that indicate
attached turbulent flow. These images indicate (1) greater detachment from the surface of the bubble, (2) a flattened bubble with reduced
circulation flow and (3) detached flow behind the deturbulator strip. It appears that the boundary flow approaches the deturbulator
at a grazing angle and then skips off of it to produce detached trailing flow.
Figures 20 and 21 show normal and deturbulated oil visualizations of bottom surface boundary flows.
Since there is no deturbulator panel on the bottom surface, it appears that the back-facing step presented by the
leading edge tape is enough to prematurely detach the flow, flatten the bubble and reattach the flow gently at a grazing angle
so as to produce less turbulence in the trailing edge flow.
Figure 20. Normal Bottom Surface Oil-flow Pattern
Figure 21. Deturbulated Bottom Surface Oil-flow Pattern