Category AIRFOILS AT LOW SPEEDS

• NACA 6409-PT (Fig. 12.66)

The NACA 6409 is considered more of a free-flight airfoil than one for RC soaring. The actual wind tunnel section was the only model that had open bay construction from leading to trailing edge. The ribs were | in (l% chord) thick and had a spacing of 3 in (25%), giving an open-bay cell aspect ratio of 1:4. The sagging of the covering was about 0.025 in (0.2%) worst case, and generally much less—0.005 to 0.015 in. Due to the lack of torsional rigidity, data was taken only up to 200k.

As Fig. 12.66 shows, the airfoil stalls at 8° at 60k; however, in Fig. 12.67, which shows lift data only, the premature stall is not found. The lack of re­peatability may have been caused by the gusty weather at the time of the lift run.

As for the performance, the NACA 6409 is an excellent low speed, floater airfoil, but, much like the AQUILA airfoil, the large camber severely limits the high speed performance.

Also see: AQUILA, S2091, SD7043 Digitizer plot: Fig. 10.25 Polar plot: Fig. 12.66 Lift plot: Fig. 12.67

Thickness: 9.00% Camber: 6.00%

• NACA 64A010-PT (Fig. 12.64)

Compared to the two symmetric airfoils discussed so far (J5012, NACA 0009), the 10% thick NACA 64A010 is the worst for low-Rn applications. The shape of the polar, with the large decrease in drag just before the stall, is caused by a long run of a favorable pressure gradient followed by a steep recovery region. In the design of this airfoil, the intention was to achieve long runs of laminar flow for low drag. While this design approach works for high Rn’s, the same line of reasoning cannot be applied at low Rn’s.

Figure 12.65 shows the lift characteristics for J£n’s of 100k, 80k and 60k. At 60k the presence of a long bubble can be inferred from the nonlinearities around zero angle of attack. It is interesting that for i2n’s of 100k and 80k the nonlinearities almost vanish. This does not mean that the separation has vanished; instead, the separation on the upper and lower surfaces apparently have the combined effect of cancelling each other out.

Also see: NACA 0009, SD8020, J5012 Digitizer plot: Fig. 10.24 Polar plot: Fig. 12.64 Lift plot: Fig. 12.65

Thickness: 10.00% Camber: 0.00%

• NACA 2.5411-PT (Fig. 12.62)

The NACA 2.5411 is a four-digit NACA section with 2.5% maximum camber at 40% of chord, and 11% thickness.

The drag polar shows that the overall drag is low, and this gives a good first impression. However as shown in Fig. 12.63, the stall characteristics are undesirable. There is a smooth, continuous increase in lift up to a sharp and abrupt stall. Flight near maximum lift with this section would be exceedingly difficult because there is virtually no angle of attack margin between maximum lift and stall.

As mentioned in the discussion of the MB253515, a lift plateau with the asso­ciated high drag is desirable, as it provides warning that stall is imminent. Unless measures such as stall strips on the leading edge are used to alleviate this prob­lem, the NACA 2.5411 and closely related sections (NACA 2412, NACA 24159) are not recommended for slow, thermal duration flying. However, its perfor­mance at high speed, such as is needed in windy conditions, is very good, and stall strips make the thermalling performance at least acceptable.

Also see: E374, CLARK-Y, DF101, SD5060 Digitizer plot: Fig. 10.23 Polar plot: Fig. 12.62 Lift plot: Fig. 12.63

Thickness: 11.00% Camber: 2.50%

• NACA 0009-PT (Fig. 12.60)

The NACA four-digit symmetric airfoils are often used on tail surfaces, most commonly the stabilizer. As expected, the drag of the 9% thick NACA 0009 is lower than the 12% symmetric J5012. It is interesting to note, however, that the lift range of the thinner NACA 0009 is on a par with the J5012.

Again, as with the J5012, there is a deadband in lift about 0°, but as shown in Figs. 12.61 the deadband for the NACA 0009 is worse than the J5012. As mentioned before, any substantial nonlinearity in the lift-curve slope is caused by some type of laminar or turbulent separation. Also, between approximately — 2° and 2°, the lift characteristics are slightly asymmetric. For the NACA 0009 at low Rn’s, a long, thin bubble forms on both the upper and lower surfaces. Any slight asymmetries present (either in contour or surface finish) will alter the boundary layer behavior, thereby affecting the symmetry of lift. In fact, as Fig. 10.22 reveals, the NACA 0009-PT is not perfectly symmetric which leads to the asymmetric lift and drag characteristics about 0°.

Also see: SD8020, NACA 64A010, J5012 Digitizer plot: Fig. 10.22 Polar plot: Fig. 12.60 –

Lift plot: Fig. 12.61

Thickness: 9.00% Camber: 0.00%

In 1972, S. J. Miley30 received his Ph. D. from Mississippi State University with a thesis titled “Analysis of the Design of Airfoil Sections for Low Reynolds Numbers.” His thesis culminated with an example airfoil, the M06-13-128, de­signed for Rn’s greater than 600k. As the title may suggest, there has been no consensus as to how the term low Rn’s should be defined. A simple way would be to state a cut-off Reynolds number number below which everything is considered low Rn’s. But low Rn’s usually implies laminar separation and high drag. Thus another definition is Rn’s where laminar separation is significant. However, with this definition the Rn would be different for each airfoil. As with any definition, there is some degree of ambiguity, and it may simply be best to state first what Rn’s are being considered, and second, whether or not laminar separation is included in the particular problem under study. In Miley’s case, laminar and turbulent separation were assumed negligible and were ignored in the anaylsis.

• M06-13-128-PT (Fig. 12.56)

As shown in Fig. 12.56, the MILEY airfoil was tested below its design Rn of 600k. At all Rn’s the MILEY airfoil has a high-drag laminar separation bubble as indicated by the large bulge in the polar. In fact for a Rn of 200k, the drag between the stall limits was the highest of any airfoil tested. Consequently we do not recommend it for model sailplanes.

• M06-13-128-PT u. s. bumps xjc = 31%, type A (Fig. 12.57)

• M06-13-128-PT u. s. trips, xjc = 31%, Rn = 200,000 (Fig. 12.58)

(А-trip) h/c = .ll%,wjc = .52%

(В-trip) hjc = .08%,wjc = .52%

(C-trip) zig-zag tape, type В (D-trip) bumps

An upper-surface trip on the MILEY airfoil at 31% produces a dramatic drag reduction—up to 73% at Rn = 200k. For all the cases shown, the А-trip works the best, with the zig-zag tape (C-trip) being next. The В-trip is apparently not high enough (half that of А-trip), as the drag is slightly greater. As for the bumps (D-trip), they do not work as well as the others at high lift, although elsewhere they are as effective as the A – and В-trips. The exact mechanism of these performance differences and how it relates to the shape of the trip is still not well understood.

Also see: SD7003, FX63-137, SD7062 Digitizer plot: Fig. 10.21 Polar plot: Figs. 12.56-12.58 Lift plot: Fig. 12.59

Thickness: 12.81% Camber: 5.16%

• MB253515-PT (Fig. 12.53)

The MB253515 (designed by Michael Bame) was one of the most intriguing of all the airfoils tested. The relatively high drag of this 15% section leaves much to be desired; nevertheless, it is favored by some.

It may be that the attraction has more to do with the lift characteristics than the drag. Under most types of flying conditions the RC sailplane spends considerable time climbing in thermals, with the wing operating very near the maximum lift coefficient. Unfortunately for most airfoils, just beyond C(max the airfoil stalls, posing handling problems. In a thermal the turbulence is quite high and, for a sailplane operating close to its stall angle of attack, the turbulent conditions can cause portions of the wing to stall intermittently. The problem is further aggravated by the lower tip chord Rn’s because with most airfoils the stall angle of attack decreases with Rn. The net effect of the local stalling and tendency to tip stall makes efficient thermalling difficult.

Although the drag characteristics of the MB253515 are hardly dazzling, the airfoil may make up for this deficiency with good handling. (See also Section 5.2.) In Fig. 12.55, the lift characteristics of the MB253515 are shown for Rn from 30k to 100k. Note that the stall angle of attack is at least 18°—very far from the thermal operating point. This large angle of attack margin gives the MB253515 section a docile feel in thermals and ultimately helps the thermaling efficiency, not through low drag but through handling—by decreasing the work load of the pilot.

This characteristic of the MB253515 separates it from most low-Rn airfoils. Usually the lift increases smoothly with angle of attack and finally breaks away, with a stall following shortly thereafter. The highly desirable stall characteristics of the MB253515 may explain why it is favored by some flyers.

Referring to Fig. 12.55 for Rn = 100k, the lift increases rapidly between —2° and 0°, flattens between 0° and 3°, then becomes more typical above 3°. When separation begins to take place, maximum lift (Cimai = 1.0) is reached at 10°, followed by a dip and a long plateau which is maintained down to as low as 30k.

Note also that as the Rn is decreased the lift characteristics change markedly. Comparing the 100k, 90k and 40k cases, two important observations can be made. First, below the angle of attack at Cimax (a « 10°), the lift decreases with Rn, which indicates the presence of a large bubble or large trailing edge separation or both. Some type of separation may also be deduced from the high drag below Cimax shown in Fig. 12.53. The second observation is the well-defined hysteresis loop in lift which is clear indication of laminar separation. Hence the relatively high drag can be attributed to a laminar separation bubble, suggesting that the airfoil could be improved with a trip.

• MB253515-PT u. s.t. xjc – 20%,hjc = .17%,w/c = 1.0% (Fig. 12.54)

Figure 12.54 shows that a trip at 20% chord improves the performance for Rn less than 200k. For airplanes using this airfoil, therefore, a trip should be used at least on the wing tips, and possibly on the entire wing, depending on the expected speed range.

Also see: S4233, SD7062, E193MOD, WB135/35, WB140/35/FB

Digitizer plot: Fig. 10.20

Airfoil comparision plot: Fig. 11.11

Polar plot: Figs. 12.53, 12.54 Lift plot: Fig. 12.55

Thickness: 14.96% Camber: 2.43%

• J5012-PT (Fig. 12.52)

The J5012 is a 12% thick, symmetric section designed by Jef Raskin. The airfoil is intended for aerobatic slope soaring where inverted flying is as important as normal flight.

Beside this application, the airfoil is a candidate for use on tail surfaces. In this regard, however, the J5012, as well as some other symmetric airfoils (NACA.0009, NACA 64A010), has an undesirable characteristic if it is to be used as a full-flying surface, e. g. a stabilator. At Rn of 60k near zero angle of attack, the lift curve is nearly flat. Consequently a stabilator deflection around zero lift produces little response. This “deadband” is a common characteristic of symmetric airfoils, although it is not always present (compare the SD8020). As an aside, the aerodynamic characteristics are not quite symmetric about zero angle of attack because, as shown in Fig. 10.19, the airfoil profile itself is not symmetric.

Also see: SD8020, NACA 0009, NACA 64A010 Digitizer plot: Fig. 10.19 Polar plot: Fig. 12.52

Thickness: 12.00% Camber: 0.00%

• HQ2/9A-PT (Fig. 12.42)

• HQ2/9A-PT u. s.t. xjc – 20%,h/c = .17%, w/c = 1.0% (Fig. 12.43)

• HQ2/9A-PT u. s.t. xjc — 40%, h/c = .17%,wjc = 1.0% (Fig. 12.44)

• HQ2/9A-PT l. s.t. xjc = 50%, h/c = .17%, tv/c = 1.0% (Fig. 12.45)

Before discussing the lift-drag characteristics of the HQ2/9, one very impor­tant point needs to be made. The differences between the nominal HQ2/9, RG15, and S2048 are small. Moreover, these differences are of the same order as the differences between the nominal airfoils and the models of the HQ2/9, RG15, and S2048 actually tested.

To illustrate, Figs. 11.3 and 11.4 compare the nominal HQ2/9 with the nom­inal RG15 and S2048. As for the actual airfoils tested, Figs. 11.5, 11.6, and 11.7 compare the HQ2/9B-PT with the HQ2/9A-PT, RG15-PT, and S2048-PT. Looking at the figures, one arrives at the conclusion that the nominal airfoils were not really tested; rather a group of very similar airfoils was tested instead.

Another important point, which immediately follows, is that the performance differences between these models are probably not meaningful (at least as they apply to RC sailplane performance) despite the fact that some differences in the overall shape and appearance of the polars are apparent. When all the other components of drag are factored into the performance of the sailplane—induced drag, wing/body interference drag, fuselage drag, etc.—the subtle differences between similar polars tend to be lost. And even though the airfoils were not tested with flaps (though flaps are usually used) it is doubtful that any one of them would have a decided advantage, even with a flap.

Two versions of the HQ2/9 were tested. The first model, version A, was found to be inaccurate and was later modified to become version B. The polar taken on the plain A model is shown in Fig. 12.42, while Figs. 12.43-12.45 show data taken with trips.

What stands out is that the drag is quite low, especially when compared with the Eppler airfoils. There are two reasons for this:

1. First, this type of airfoil, which is popular in F3B type flying, is intended for use with flaps in order to achieve a wide lift range, while the Eppler airfoils (E205, E193 type) are designed for a wide speed range without flaps. If an airfoil is intended to be used with flaps then a certain amount of lift range (at any given position of the flap) can be traded for lower drag, since the flaps will be used to recover the range. This is the approach taken by Quabeck.

2. The second reason was described in the overview of the HQ-series airfoils; that is, the upper surface pressure gradient is rather gradual, which improves the management of the laminar separation bubble.

Trips were placed on the upper surface at 20% and 40% chord. For the 20% case shown in Fig. 12.43, there is an improvement only for Jin’s of 150k or less. For the 40% case shown in Fig. 12.44 the break-even point seems to be between 150k and 200k. For the HQ2/9, therefore, it is generally advisable to use trips on any portion of the wing that usually operates at Rn less than 150k-200k. With a tapered wing, the trip location near the root should be further aft (in percent of chord) than it is at the tip, since the tip operates at a lower Rn.

Figure 12,45 shows the effect of a trip at 50% chord on the lower surface. If anything, the drag has increased around C of 0.5.

• HQ2/9B-PT (Fig. 12.47)

• HQ2/9B-PT u. s.t. x/c = 50%, h/c = .17%,w/c = 1.0% (Fig. 12.48)

The untripped HQ2/9B-PT has performance much like the (less accurate) HQ2/9A-PT. An upper surface trip at 50% improves performance below Rn of 200k, and only slightly hurts the performance at 300k.

• HQ2/9B-PT u. s. blowing xfc — 50%, type В (Fig. 12.49)

• HQ2/9B-PT trips, Rn = 200,000 (Fig. 12.50)

The A/В model was hollow so that tests involving boundary layer blowing through a row of spanwise holes could be conducted. The blowing configuration is shown in Fig. 5.7. As depicted in Fig. 5.8 ram air from a single “total head tube” fixed to the lower surface of the model was used for the air supply.

A complete polar for the type В blowing is shown in Fig. 12.49. Even though tripping the boundary layer with a trip strip improved the performance of the airfoil (Fig. 12.47) at low Rn, tripping by blowing degraded it. Perhaps the amount of blowing was too large.

Figure 12.50, which compares the blowing data with the cases with and with­out trips, shows that all of the trips are worse than the untripped case at 200k. Since this model arrived late in the experiments, extensive testing was not pos­sible.

Also see: RG15, S2048, SD8000, S2030, S2055, SD2083, S3021

Digitizer plot: Figs. 10.17, 10.18

Airfoil comparision plot: Figs. 11.3-11.7

Polar plot: Figs. 12.42-12.45, 12.47-12.50

Lift plot: Figs. 12.46, 12.51

Thickness: 8.97% Camber: 1.99%

The HQ-series of airfoils are apparently generated in a manner similar to the NACA four – and five-digit airfoils. In this method, the thickness distribution is wrapped around a camber line. Although no details of Quabeck’s method have been given, based on the lack of smoothness of the coordinates it appears that neither the thickness distribution nor the camber are analytical functions, as they are with the NACA sections.

Quabeck’s design philosophy, which has been widely published in the Euro­pean model press, concentrates largely on the effects of the camber distribution on the airfoil pitching moment29. Although it is not often mentioned, Quabeck has been guided by his prolific building and expert flying skills.

The airfoils themselves are characterized by a fair amount of aft loading. For some of them, the upper-surface velocity distribution has a gradual pressure recovery region which lowers drag. In our opinion this feature is the principal key to the success of Quabeck’s more popular sections.

Finally, it should be noted that the airfoil designation “HQ” is also used by Horstmann and Quast, who collaborate in the design of airfoils for full-scale competition sailplanes (for example, the Ventus, ASW-22, and ASH-25).

• FX63-137A-PT (Fig. 12.40)

• FX63-137B-PT (Fig. 12.41)

The high-lift, 13.6% thick Wortmann FX63-137 is often selected in feasibility studies of high altitude, long endurance, unmanned aircraft24. Such aircraft are envisioned for missions where manned platforms are impractical, for example: long-duration reconnaissance, surveillance, search and rescue, meteorology and mapping, relay of radio and television, etc. Because of the strong interest in the FX63-137, it has been widely tested in other facilities. At Princeton the FX63-137 was used to make comparisons with the other tunnels (see Section 2.4).

The FX63-137 was first tested without Monokote covering (version A) and then with covering (B). As it came from the builder, the model was sheeted with heavy grain obeche which was varnished smooth; however, the grain was not completely filled by the varnish. With the Monokote covering the wood grain was effectively filled, but this had little effect on the airfoil performance. It will later be shown (see SD7032) that the differences between smooth and rough surfaces can have measurable effects on performance by influencing transition.

Also see: SD7062, S4233, MB253515, SD7043, SPICA

Digitizer plot: Fig. 10.16

Polar plot: Figs. 12.40, 12.41

Thickness: 13.59% Camber: 5.94%