What’s Washout? Watch out!
By Martin Simons
There has been some discussion in recent issues of the VGC Newsletter, about the relationships between wing planform, aerofoil section change across the span, and washout, all in relation to stalling and spinning. It seems that some clarification is required.
Most sailplanes, if handled clumsily, will ‘drop a wing’ in a turn and some will at once begin to spin. This is a common cause of accidents. Although the phenomenon is commonly called ‘tip stalling’, much more than the wing tip stalls. The stall begins perhaps two thirds of the way out along the semi-span, but spreads rapidly across the whole. Stalling causes the usual nose down pitch, and the aircraft simultaneously rolls sharply because the airflow breaks away on one side first. The drag on that side rises greatly and pulls the aircraft round. The effect is quite sudden and frightening, especially for less experienced pilots. In a severe case, the glider will roll over into a semi-inverted position and if this happens during an approach to landing, it is disastrous. A sailplane which does this sort of thing, even when competently flown, is very dangerous.
To reduce the likelihood of this unpleasantness, the idea of building some twist into the wing across the outer half or third of the span, has been used since the earliest days of gliding. This twist, negative in the sense that the tip is brought to a smaller angle of incidence than the root of the wing, is called washout. If done correctly, it is effective. The outer part of the wing meets the air at a lower angle of attack than the centre section, so the stall begins near the root and a wing drop is much less likely. (It is still usually possible to spin a sailplane with washout off a badly co-ordinated turn.) The gradual change of the angle of incidence can often be seen by looking along a wing from the tip end along the line of the main spar. There is a good deal more to washout however, than meets the eye. In vintage sailplanes the twist is nearly always accompanied by a progressive change of the aerofoil or wing section. This introduces complications.
A test wing in a wind tunnel will stall at some angle of attack to the airflow. The stalling angle varies a good deal from one wing section to another and the character of the stall also is variable, depending on the details. Things like the thickness and camber, the radius of the extreme leading edge and wing surface quality, have an influence. Some profiles exhibit a very sharp and sudden breakdown of the airflow, stalling abruptly with little warning. Others are more placid. Wind tunnel results give a fairly reliable indication of what may be expected.
In the days when most old sailplanes were designed, the angle of attack in the wind tunnel was measured in a variety of ways. One needs to know the reference line from which this angle was measured. Figure 1 should help to make this clear. It is also important to distinguish between the geometric and aerodynamic angles of attack. They are the same only if the wing profile gives zero lift at its geometric zero angle. This applies to symmetrical and to a few special ‘reflex’ profiles, but not to any normally cambered aerofoil.
A great many of the older sailplanes have aerofoils developed at Göettingen University during the 1920’s. Most were undercambered sections and the convention in wind tunnel work at that time, was to measure the angles of attack from a line tangential to the under surfaces of the wing, as shown in Figure 1a, or sometimes from another, more or less arbitrary datum, depending mainly on the convenience of the draughtsman and the local custom. It is easy to see that when the angle is taken from such a line, a geometric zero angle of attack on the test chart corresponds to a positive angle measured from the true chord line, which joins the extreme nose of the profile, to the extreme trailing edge.
Many sailplanes used symmetrical profiles at the wing tips and, the wind tunnel laboratories not being entirely consistent, the reference line used for these was usually the true chord line. When the data for a particular vintage sailplane gives an angle of wheel twist of so many degrees, it nearly always refers to the difference in angle between the root section’s undersurface tangential line, and the tip section’s chord line. What this amounts to is a concealed geometric ‘washout’ in most of the published figures. Even when the nominal figure is zero, a modern engineer, using the true chord line as his reference, would find several degrees of difference in wing incidence across the semi-span from root to tip. An example is the Rhöenbussard, which nominally has no washout, but if the true chord lines are used as a reference, the washout is 2.38 degrees. This remains a geometrical matter, to do with lines on the drawing board and conventional methods of measurement. The aerodynamic behaviour of the wing is another story.
Compare two wind tunnel test charts such as would have been available to a sailplane designer about 1934 (Figure 2). The Göettingen 535 wing section was used at the root on many famous sailplanes. The chart shows it stalling at 16 degrees, measured from its undersurface reference line. Alongside this is another chart for a symmetrical profile, the NACA M3, very similar to the wing tip profiles on most older German gliders. The geometrical stalling angle of this profile is shown also as 16 degrees but measured in this case from the chord line. The M3 does not reach such a high lift coefficient as the 535, but its nominal stalling angle is the same
When the 535 profile is at a nominal angle of zero, it is still developing lift. It does not reach its angle of zero lift until it is at minus 8.7 degrees. The angle of zero lift is a fundamental physical characteristic independent of engineering conventions or drawing board lines. It is a better reference point for the comparison of aerofoils than the nominal angle of attack, and is called the aerodynamic zero. Since the symmetrical section is at zero lift when it is geometrically at zero angle, its aerodynamic zero is also its geometric zero. Since the aerodynamic zero of these two profiles is 8.7 degrees different, a wing using the 535 at the root and the symmetrical section at the tip, with no geometric twist, would be described as having 8.7 degrees of aerodynamic washout. This figure would apply to the Rhöenbussard.
All is not yet said. When looking at the elderly test charts, an allowance has to be made for the aspect ratio of the model wings used in the old fashioned tunnels. The two charts came from different laboratories and as the “Size of model” figures given on each show, the test pieces had different aspect ratios, one six (NACA M3) and the other five (Göettingen 535). It is not satisfactory to make a direct comparison of these two profiles without making allowance for this difference. There were, doubtless, other variations between the tunnels used, such as the methods of measuring forces and fine grained turbulence in the airflow. In addition, such details as the shape of the wing tips on the models had some effect, hard to quantify. Fortunately these technical problems are not of vital concern to the present discussion, but the aspect ratio has to be taken into account.
There is a difference in air pressure above and below any lifting surface such as a wing. This difference is what generates the lift. Behind the wing tips a strong vortex develops which not only increases the measured drag but also generates downwash. Air is a fluid medium and providing the velocities involved are well below the speed of sound, the flow begins to feel the effects of the vortices even before the wing arrives. The trailing vortices of any real wing distort the airflow not merely behind the lifting surface, but all round and ahead of it. This is illustrated in Figure 3. There is a ‘vortex induced’ reduction in the true, aerodynamic angle of attack, compared with the undisturbed flow. Since the vortices are produced by the lift, it is easy to see that the strength of the downwash will be greater if the lift coefficient on the chart is greater, because the strength of the vortices increases. The lift coefficient, other things being equal, depends on the angle of attack, so the induced downwash is large if the angle of attack is large and small if it is small. At the aerodynamic zero, when the wing produces no lift there is no average difference in pressure between upper and lower surfaces, so virtually no tip vortices and no induced downwash. At the stall or near it, downwash is large because the lift coefficient is large and the vortices are strong. This is of considerable importance because at a high lift coefficient, such as the 1.56 maximum reached by the 535 aerofoil, the change of downwash if the aspect ratio is 6 instead of 5 is just under one degree. If the Göettingen 535 were tested in the same wind tunnel and with the same type of model as that used for the M3, it would stall at about 23.75, not 24.7 degrees aerodynamically from the zero lift angle. Geometrically, the 535 stalling angle is one degree less than the M3 at aspect ratio six, measuring these angles in the old, inconsistent manner. The aerodynamic zero, not affected by aspect ratio, remains the same.
(In modern practice, aspect ratio effects are removed from wind tunnel results altogether, the charts being published in standard form with the aspect ratio theoretically infinite. It is much easier to compare wind tunnel results directly in this way and then, during the design process, make corrections for aspect ratio and other planform factors. This is a routine matter nowadays, but it was not standardised until the late 301s.)
This brings us to real wings in flight, rather than wind tunnel tests. As mentioned above, the strength of the downwash caused by the trailing vortices depends on the angle of attack and the value of the lift coefficient. As a rule, the lift coefficient is not constant across the whole span. Indeed, the idea of washout is to keep the outer wing always further away from the stall than the inner panels, so this alone should lead us to expect the local lift coefficient to vary from place to place. If the section changes to a symmetrical or nearly symmetrical form, as it usually does at the wing tips of vintage sailplanes, the maximum possible lift coefficient near the tips will be less because, as the charts show, a section like the M3 reaches a maximum lift coefficient of only 1.08 compared with 1.56 for the 535. If the washout is correctly done with this sort of combination of profiles, when the wing root is trembling on the verge of stalling at lift coefficient close to 1.56, the tip profile will not have reached 1.08.
Since the lift coefficient varies in the spanwise direction, the strength of the downwash should be expected to vary accordingly. The usual explanation of this is that the main vortex behind a wing tip is actually a combination of a very large number of small vortices which leave the trailing edge of the wing all the way along, in a sheet. The wing may be thought of as producing a large number of individual but invisible twisted strands, somewhat like spun threads, all of which stream off behind. Since each small segment of the wing has its own small vortex, each point along the span adds a little to the total of the downwash. This effect is additive. The numerous tiny vortices react with one another, the outer ones wind the inner ones into themselves rather like a spinning machine drawing in numerous fine threads to make a thicker one. All the vortiqes are drawn together to make two large, strongly rotating ‘ropes’ which form some little distance behind, and somewhat inboard of the tips (Figure 4).
Since the lift coefficient varies spanwise and the vortices follow suit, the induced downwash at each point along the span reflects this. Hence, the local angle of attack will be modified to some extent depending on the strength of the vortices summed at each point. The wing at a particular spanwise location will stall if its aerodynamic angle of attack is sufficient, but if the vortex strength at any place is large, induced downwash will reduce the local angle of attack and delay the stall over this segment of the wing. At other places along the span, if the local vortex strength is small, the downwash effect will be weak and the angle of attack will be less modified. Stalling is more likely.
The idea that the vortex strength, and hence downwash, at some spanwise location, is found by adding up all the vortices that have been ‘wound in’ from the inner parts of the wing, as well as the one produced at this place, is very important. If the wing is rectangular in plan, and without any twist or change of aerofoil the downwash near the centreline of the aircraft will be small, because only the innermost vortices have been added. Further out along the wing, the downwash increases because here the inner vortex strength is added to the outer ones. The true angle of attack will thus be less than at the centre. Further out still, as the sum of all the vortices increases still more, the angle of attack locally will be less again and near the tips where all the vortices have made their full contribution, downwash will be strongest and the angle of attack least. It is evident and well borne out in practice, that a rectangular wing is unlikely to suffer from ‘tip stalling’ even if there is no washout. This, apart from ease of construction and repair, is why vintage training gliders like the Zoegling and Cadet, had rectangular wings. To make assurance doubly sure, they sometimes had washout too, but pupils still managed to spin them into the ground sometimes.
Conversely, a tapered wing will spread the vortex strength evenly, tending to keep the downwash more constant across the span. The ideal wing, from the point of view of least vortex drag, is one which has a chord distribution of elliptical kind. With such an outline, again without twist or section change, the lift co- efficient at each station across the span is the same and the vortex strength at each position is graded so that no part of the wing gets more downwash than any other. The sailplane designer has always tried to approach this ideal. At low speeds, as when circling in thermals or weak hill lift, vortex drag accounts for more than half the total drag of a sailplane, so it has always been recognised as desirable to reduce it by trying to achieve the perfect elliptical form and constant spanwise downwash. A close approximation is obtained with a constant chord over the inner wing panels, with moderate taper outboard. The Rhoenbussard, Gull 1 and 3, and many other old sailplanes used this shape. Simply tapered wings like those of the Olympia and Weihe lose a little more, compared with the ellipse. The double or even triple-tapered wing, very common now, loses hardly anything. Each small part of the wing carries a lifting load proportional to its area. This ‘work sharing’ allows the wing to produce its lift with the least wastage of energy. Unfortunately, for safe handling, a constant lift coefficient across the span can lead to dangerous wing dropping tendencies, because even a small amount of yaw or sideslip in a turn can precipitate an asymmetric stall. The modern sailplane designer solves this problem by using wing tip aerofoils which develop high lift coefficients and stall late, exactly the opposite of the old time solution, which was to use symmetrical profiles and large amount of aerodynamic washout.
It will probably already have occurred to the reader to ask why the old-time designers used aerofoils at the wing tips which not only stall at low aerodynamic angles of attack relative to their aerodynamic zero but also could not reach high lift coefficients. This is a good question. There is no doubt that a lot of washout was necessary when the wing was tapered and the tip profile was symmetrical or nearly so. Too little, and the sailplane became dangerous to fly.
Consider the aerodynamic washout associated with advertised geometric figures for such old sailplanes as the Kranich (8.8 degrees) which used similar aerofoils to those of the Bussard. Aerodynamically the Kranich has about 17.5 degrees of twist. This did produce safe handling at low speeds. But at high speeds the wing tips bend downwards because they are working at negative angles of attack. This creates much drag, and throws additional load onto the inner parts of the wing. The price of washout of these magnitudes is paid at the high speed end of the performance curve. By modern standards, practically all the old timers had far too much wing twist, aerodynamicaly, but this was mainly a result of the choice of wing tip profiles. Towards the end of the 1930’s, the better designers were already beginning to realise this, and the symmetrical wing tip gradually began to disappear.
(a) Aerofoils designed and tested in most early research were measured from their conventional reference lines, usually a line tangential to the under side. The more fundamental reference datum is the aerodynamic zero, which is the angle at which the aerofoil gives no lift.
(b) The line used in most modern work is the true chord line, which joins the extreme leading edge to the trailing edge. There is a difference of several degrees between this and the old reference line. The aerodynamic zero is unaffected.
(c) A symmetrical aerofoil such as those used on many vintage sailplanes at the wing tip has chord line and conventional reference line coincident with the aerodynamic zero.