Flight Lab: Lift, Drag & Density Altitude · Guide · glidepathscience.org
GuideFlight Lab: Lift, Drag & Density Altitude
Contents
This Guide explains the ideas and equations behind every Walkthrough step, Concept Challenge question and Predict round, in the order a pilot’s handbook would take them. Part 1 prepares the Intro track and Part 2 the Advanced track; each takes about a quarter of an hour to read. Each section ends with a worked example that is not one of the sim’s own problems. Every performance number belongs to this lab’s illustrative trainer: for learning only, never for flight planning.
Part 1 · The essentials
Math you’ll need: algebra, right-triangle trigonometry and reading graphs; no calculus.
Four forces, and a wing that throws air down
An airplane in flight is a body with four forces on it. The Earth pulls it down with its weight. The air pushes on it in two ways that we name separately: lift, perpendicular to the oncoming air, and drag, along it. The propeller, pushing air backward, is pushed forward by that air: the thrust. Everything in this lab is about how those four forces are made and how they balance.
The surprising one is lift. A wing is not a sail held up by impacts, nor a shape whose top air must hurry. It is a device for turning the oncoming air downward. Turning a flow takes a pressure difference across it, and that difference, low above the wing and high below it, is the lift. It grows with the air’s density, the airspeed, the wing’s area and its angle to the air, until the flow can no longer follow the upper surface and the wing stalls. Drag comes with lift, and thin air on a hot, high day asks for more speed and more runway.
Steady flight and the angle of attack
Newton’s first law says that a body moving in a straight line at a steady speed has no net force on it. An airplane cruising straight and level is such a body, so its forces balance: lift equals weight and thrust equals drag,
$L = W$, $T = D$
The tempting idea that steady flight needs a little extra thrust to “keep it moving” is the old belief that motion needs a push. Extra thrust would make the airplane speed up; extra lift would make it climb faster and faster. In steady flight, climbing or not, the forces cancel.
Lift and drag are measured against the relative wind: the air the airplane meets, which blows straight along the flight path from ahead. The angle of attack $α$ is the angle between that relative wind and the wing’s chord line, the straight line from its leading edge to its trailing edge. It is not the pitch, which is the angle of the airplane’s axis above the horizon. A climbing airplane with its nose 10° up, on a path 6° up, meets the air at 4°; a glider sinking along a descending path meets the air from below even with its nose level.
With the angle of attack defined, we can ask how the wing turns it into lift.
How a wing makes lift
Watch smoke flow past a wing at a few degrees of angle of attack. The streamlines ahead of it rise slightly toward the leading edge, curve over the top and leave the trailing edge heading downward: the wing has deflected the stream down. Air follows a curved path only if something pushes it toward the center of the curve, and in a fluid that push is a pressure difference across the streamlines. So above the wing, where the flow curves over the upper surface, the pressure falls below the surrounding air’s; below it, where the flow is slowed and turned, the pressure rises. That difference, integrated over the wing’s area, is the lift, and the low pressure on top supplies most of it.
The same story can be told with momentum. Each second the wing gives a large mass of air a downward velocity, and by Newton’s third law the air pushes the wing up. The pressure picture and the momentum picture are not rival theories: they are the same forces counted at the wing’s surface and far from it. Bernoulli’s equation along each streamline,
$p + ½ρv^2 = \text{constant}$
ties the speeds to the pressures: where the air over the top runs faster, its pressure is lower. Two wrong stories are common. One says that air molecules bounce off the underside like skipping stones; that would give far too little lift, and none from the upper surface. The other, equal transit, says that the air going over the longer top must meet the air going under the bottom at the trailing edge, so it must go faster. Nothing makes the parcels meet. The pressure field round a lifting wing speeds the upper air far more than the geometry would suggest, and smoke released together upstream shows the upper pulses arriving well ahead.
Equal transit also fails by numbers: this section’s upper surface is only 0.81 % longer than its lower one, so the top air would be just 1.0081 times as fast, and the pressure difference only $½ρv^2(1.0081^2 - 1)$.
How much lift a wing makes, at a given angle, is the subject of the lift equation.
The lift equation and the angle it takes
The pressures round a wing all scale with one quantity, the dynamic pressure of the oncoming air,
$q = ½ρv^2$
where $ρ$ is the air’s density and $v$ the true airspeed. It is the pressure the air would gain if brought to rest, and it is what an airspeed indicator measures. The lift is that pressure times the wing’s area $S$ times a number that says how hard the wing turns the flow, the lift coefficient $C_L$:
$L = qSC_L$ $L = ½ρv^2SC_L$
The lift coefficient depends on the angle of attack and on the wing’s shape, not on the speed or the size. Below the stall it grows nearly in a straight line with the angle. For the trainer’s wing it grows by 0.0795 for each degree, starting from zero lift at −2.0°: the wing is cambered, so it lifts a little even at zero angle.
Read the other way, the lift equation tells a pilot what lift coefficient level flight needs. With $L = W$,
$C_L = 2W/(ρv^2S)$
Slower flight means a smaller dynamic pressure, so a larger $C_L$, so a larger angle of attack: to fly slowly the nose comes up, not down. The usual slips are to forget the ½ in $q$, which doubles the answer, or to leave the speed unsquared.
The straight line cannot go on forever. At some angle the flow gives up.
The stall, at any speed
As the angle of attack grows, the pressure over the top falls further near the leading edge and must rise again toward the trailing edge. The thin layer of air next to the surface, slowed by friction, eventually cannot climb that rising pressure. It stops, turns back and lifts the outer flow off the surface: the flow separates, starting at the trailing edge and creeping forward. Past the critical angle of attack the lift falls and the drag rises sharply. That is a stall. It has nothing to do with the engine, and a stalled wing still lifts, just less, with much more drag.
This lab’s section reaches its largest lift coefficient, $C_{L,max}$ = 1.572, at 14.7°. Ideal flow, which draws the sim’s streamlines, has no friction and never separates, so the stall here comes from a model fitted to wind-tunnel behavior, and the picture labels its separation point “modeled”.
In level flight the stall comes at the speed where even $C_{L,max}$ only just carries the weight, the stall speed:
$v_S = \sqrt{2W/(ρSC_{L,max})}$
Because the airspeed indicator reads $½ρv^2$, the stall comes at the same indicated airspeed at any altitude. And because a stall is an angle, not a speed, it can happen fast. In a level turn banked at $φ$, the lift tilts with the wings and its upward part must still hold the weight, so the wing carries the load factor $n$ times the weight,
$n = 1/\cos φ$, $v_S\sqrt n$
and the stall speed rises as $\sqrt n$, the square root because lift grows as $v^2$. A sharp pull-up does the same. The mistake to avoid is thinking that a fast airplane cannot stall.
Lift is never free. The next section counts what it costs.
Drag, the region of reversed command, and gliding
An airplane’s drag has two families. Parasite drag is the cost of pushing a body through air at all: skin friction, the form drag of the landing gear and struts, the interference where parts meet. It grows with the dynamic pressure, as $v^2$. Induced drag is the cost of lift itself. The wing’s downwash tilts the relative wind it meets, and with it the lift, slightly backward; the backward part is a drag. It is largest when the wing works hardest, at a high lift coefficient, so in level flight it falls as $1/v^2$.
Add the two and the total drag in level flight has a minimum, the least-drag speed $v_{md}$, where parasite and induced drag are equal. Faster than that, the total grows with speed, as a driver would expect. Slower, it grows again, because induced drag climbs steeply near the stall: on this back side of the drag curve, holding altitude at a lower speed takes more thrust, not less. The belief that slower always means less drag is true only on the fast side.
Thrust is not power. The power level flight needs is the drag times the speed, $P_r = Dv$, and since the speed falls as the drag rises, its least comes at a slower speed, the least-power speed $v_{mp}$, a propeller airplane’s speed of best endurance. The Pilot’s Handbook draws the region of reversed command on this power curve: below $v_{mp}$, holding altitude at a lower speed takes a higher power setting. Between the two speeds the answers differ. At 1,100 kg in sea-level air the trainer’s $v_{mp}$ is 28.5 m/s and its $v_{md}$ 37.5 m/s; level at 34 m/s it needs 908 N of thrust and 30.9 kW, and at 30 m/s more thrust, 980 N, but less power, 29.4 kW.
Neither curve says by itself what happens when a pilot slows down and leaves the throttle alone: that also depends on how the propeller’s thrust changes with speed. This trainer’s propeller gives more thrust as it slows, so at a set throttle a slowdown from 36 to 34 m/s meets a net force of 7.1 N forward, though 36 m/s is on the drag curve’s back side. Four Forces’ nudge (Advanced) works such cases out; it is a force balance at one instant, not a model of the airplane’s pitch, trim and stability.
The speed of least drag is also the speed of the best lift-to-drag ratio, since the lift there equals the same weight. It matters most when the engine stops. In a steady glide the lift and drag together balance the weight: across the path $L = W\cos γ$, along it $D = W\sin γ$. So the path descends at the angle whose tangent is drag over lift: for every meter lost the airplane covers $L/D$ meters forward,
$d = h\,(L/D)$, $\tan γ = D/L$
A heavier glider has the same best $L/D$, flown faster: weight sets the speed, not the angle, so it glides just as far. A windmilling propeller adds drag that a clean polar leaves out.
All of these numbers assumed sea-level air. On a hot day at a high airport they change.
Density altitude and the hot, high takeoff
A wing’s lift, a body’s drag and an engine’s power all depend on the air’s density, and density depends on both pressure and temperature. For dry air the ideal-gas law gives
$ρ = p/(R_sT)$
with $R_s$ = 287.05 J/(kg·K) and $T$ in kelvins. The pressure altitude is the height in the standard atmosphere with the field’s pressure. The density altitude is the height in the standard atmosphere with the field’s density: the height the airplane “feels”. On a hot day it can be thousands of feet above the field. Heat, height, a low altimeter setting and humidity (water vapor is lighter than dry air) all raise it. Pilots estimate it with a rule of thumb,
DA ≈ PA + 120 ft/°C × $(T - T_{ISA})$
where $T_{ISA}$ is the standard temperature at the pressure altitude, 15 °C less about 2 °C per 1,000 ft. A common slip is to measure from 15 °C at every height; another is to mistake density altitude for the field’s elevation.
An airplane lifts off at the same indicated airspeed at every field, since the indicator reads the dynamic pressure that holds it up. In thin air that is a higher true airspeed,
$v_{TAS} = v_{IAS}/\sqrt σ$
with $σ = ρ/ρ_0$. The airplane must reach it with an engine that, breathing thinner air, makes less power. Both effects lengthen the roll, and drag and friction take a larger share of the smaller thrust. So the same indicated speed does not mean the same runway.
Part 2 looks behind these results: why the circulation round a wing carries its lift, why a finite wing pays induced drag, and how the takeoff roll is computed exactly.
Part 2 · Going further
Math you’ll need: the same algebra and trigonometry, in longer chains of formulas; no calculus.
Circulation, the vortex system and thin-airfoil theory
The flow round a lifting section can be split into a uniform stream and a circulating flow about the section, of strength $Γ$, the circulation: the line integral of the velocity round any loop enclosing it. A sharp trailing edge fixes $Γ$, because real air leaves it smoothly rather than whipping round it (the Kutta condition, which Fluid Flow & Bernoulli develops). The Kutta–Joukowski theorem then gives the lift per meter of span, $L′ = ρvΓ$, and for a whole wing of span $b$ with mean circulation $\bar Γ$,
$L = ρvΓ̄b$
Circulation cannot simply stop at the wingtips (Helmholtz’s theorem): the bound vortex turns at each tip and trails downstream as a tip vortex. And because the air’s total circulation started at zero, the wing left an equal and opposite starting vortex on the runway when it began to lift (Kelvin’s theorem). The tip vortices push the air between them down, which is the downwash, and they are strongest when the circulation is largest: behind a heavy airplane flying slowly with its flaps up.
How much circulation does a section develop at a given angle? Thin-airfoil theory treats a section as its mean line and finds
$c_l = 2π(α - α_{L0})$
with the angles in radians: a slope of 2π per radian, 0.110 per degree, whatever the shape, and a zero-lift angle $α_{L0}$ set by the camber. For the NACA 2412’s mean line, 2 % camber at 40 % of the chord, the theory gives $α_{L0}$ = −2.077°. Real sections measure a little less slope, about 0.105 per degree, since the boundary layer takes a little circulation; the slips to avoid are mixing degrees into the 2π, and dropping or flipping the sign of $α_{L0}$.
The downwash of the tip vortices changes what the wing as a whole does, which is the next section’s subject.
The finite wing: lift slope, induced drag and ground effect
Behind a finite wing the tip vortices make a downwash, so the wing meets a relative wind tilted down by the induced angle. For an elliptic loading, and with a span efficiency $e$ for real wings, lifting-line theory gives $α_i = C_L/(πeAR)$, where $AR = b^2/S$ is the aspect ratio. The wing needs that much more angle for the same lift, so its lift curve is less steep than its section’s:
$a = a_0/(1 + a_0/(πeAR))$
with both slopes per radian. In this model the wing’s largest lift coefficient is still the section’s, reached at a larger angle; real wings reach about 90 % of it.
The same tilt of the lift gives the induced drag, $D_i = Lα_i$, so
$C_{D_i} = C_L^2/(πeAR)$
In level flight this is $D_i = 2W^2/(πeρv^2b^2)$: it falls as the square of the speed and of the span, which is why sailplanes have long wings. The slips are to leave out $e$ or to forget to square $C_L$.
Near the ground the downwash has nowhere to go. The ground acts like a mirror image of the wing’s vortex system, which weakens the downwash and with it the induced drag, by a factor McCormick fits as
$φ = (16h/b)^2/(1 + (16h/b)^2)$
for a wing a height $h$ above the ground. In ground effect an airplane can lift off at a speed too slow to climb once it leaves it: a hazard on a hot day.
With both drags in hand we can build the whole airplane’s polar and its performance.
The polar, power and climbing
An airplane’s drag polar is its drag coefficient against its lift coefficient: a parasite part that hardly changes, $C_{D0}$, plus the induced part,
$C_D = C_{D0} + C_L^2/(πeAR)$, $(L/D)_{max} = ½\sqrt{πeAR/C_{D0}}$
The best lift-to-drag ratio is where the line from the origin just touches the polar, which is where the two parts are equal, at $C_{L,md} = \sqrt{πeAR\,C_{D0}}$. A pilot finds it as a speed, the least-drag speed $v_{md}$, which depends on the weight and the density: its indicated value depends on the weight alone.
Climbing needs power, not lift. In a steady climb at angle $γ$ nothing accelerates, so across the path $L = W\cos γ$, a little less than the weight, and along it $T - D = W\sin γ$: the climb is paid for by the thrust left over after drag. Multiplying by the speed turns forces into powers. The power the engine and propeller deliver is $P_a = Tv$; the power level flight needs is $P_r = Dv$; the rest lifts the weight at the rate of climb,
$\text{ROC} = (P_a - P_r)/W$
The best rate of climb is at the speed of the largest excess power, not at the top speed (where there is none) and not at the least-drag speed. As the airplane climbs into thinner air the power available falls, and where the two curves only just touch, the climb ends: the absolute ceiling.
Everything so far assumed that the flow follows the wing. The next section asks why, past a point, it does not.
Why the flow separates, and how the stall is modeled
Over a lifting wing the pressure on the upper surface falls to a sharp minimum near the leading edge, then rises toward the trailing edge. The outer flow can climb that adverse pressure gradient by trading speed for pressure, as Bernoulli’s equation allows. The boundary layer, the thin layer next to the surface slowed by friction, has less speed to trade. At a large angle the gradient is steep, the slowest air stops and reverses, and the flow lifts off the surface. Separation is not the air running out of speed, nor a vacuum, nor a failure of Bernoulli’s equation, which still holds outside the boundary layer.
Ideal flow has no boundary layer, so it shows the gradient but never separates, and d’Alembert’s paradox says it has no drag at all. A model must supply the stall. This lab uses the one helicopter-blade engineers use (Leishman and Beddoes): Kirchhoff’s solution for a flat plate with the flow attached over the front fraction $f$ of the chord,
$c_l = a_0(α - α_{L0})((1 + \sqrt f)/2)^2$
with $f(α)$ fitted to wind-tunnel behavior: nearly 1 at small angles, falling past about 12°. With $f$ = 1 it is the straight line; the factor, not $f$ itself, multiplies the lift. The same model gives the separated flow’s pressure drag, which is zero while the flow is attached. It is honest about what it is: a fit, labeled “modeled” wherever it shows.
The last section returns to the runway, with the exact chain from the altimeter to the liftoff.
The takeoff, from the altimeter to the runway
A field reports its elevation $H$ and an altimeter setting $A$, the pressure reduced to sea level by the standard atmosphere. The station pressure follows from $A^n = p^n + p_0^nLH/T_0$ with $n$ = 0.190263, and the pressure altitude from the standard atmosphere’s troposphere, $h_p = (T_0/L)(1 - (p/p_0)^n)$, where $T_0$ = 288.15 K, $p_0$ = 101,325 Pa and $L$ = 0.0065 K/m. Exactly, the pressure altitude is the elevation plus the pressure altitude of the setting itself, about 925 ft per inch of mercury near sea level; pilots use 1,000 ft per inch. The density altitude needs only the density ratio:
$σ = (p/p_0)(T_0/T)$, $h_ρ = (T_0/L)(1 - σ^{1/4.25588})$
On the runway every force is quadratic in the airspeed: thrust falling off as the airplane speeds up, drag and lift growing as $v^2$, and rolling friction on the weight the wing has not yet taken. So the net force is $A - Bv^2$, and the acceleration falls as the speed builds. The distance to liftoff integrates in closed form, and the classic estimate takes a constant acceleration, evaluated where the groundspeed is 0.7 of its liftoff value:
$s ≈ v_{LOF}^2/(2a_{0.7})$
It is good to about 5 % while the acceleration at liftoff is at least half the starting one, and errs more on a marginal day. A headwind gives the airplane airspeed before it moves, so it needs less groundspeed; with a nearly constant acceleration the roll shrinks as the square,
$s_w/s_0 ≈ (1 - v_w/v_{LOF})^2$
and a tailwind lengthens it by more than the same headwind shortens it.
Reference
Common mistakes
Part 1
- Equal transit. Nothing makes the air over the top meet the air below at the trailing edge; the top air arrives first.
- Lift from impacts. Most lift comes from the low pressure over the top, not from air bouncing off the underside.
- The angle of attack as the pitch. It is measured from the relative wind along the flight path, not from the horizon.
- A stall as a speed. A stall is an angle: past the critical angle a wing stalls at any speed, in a steep turn or a pull-up.
- Slower always means less drag. Below the least-drag speed, slowing down takes more thrust.
- Thrust and power as one. Power is drag times speed: between the least-power and least-drag speeds, slower flight takes more thrust but less power. The region of reversed command is the power curve’s.
- A heavier glider glides less far. It glides as far, faster: weight sets the speed, not the angle.
- Density altitude as the field’s height. It is the standard height with the air’s density, often far above the field on a hot day.
Part 2
- The same indicated airspeed means the same runway. In thin air it is a higher true airspeed, reached with less power.
- Climbing takes extra lift. In a steady climb lift is a little less than weight; the climb comes from excess thrust.
- Ideal flow shows the stall. Ideal flow never separates; the stall needs a model, or a real boundary layer.
The equation sheet
| Part 1 | |
|---|---|
| steady, straight, level flight | $L = W$, $T = D$ |
| a steady climb at $γ$ | $L = W\cos γ$, $T - D = W\sin γ$ |
| a steady glide at $γ$ (no thrust) | $L = W\cos γ$, $D = W\sin γ$ |
| power level flight needs | $P_r = Dv$ |
| dynamic pressure | $q = ½ρv^2$ |
| lift | $L = qSC_L$ |
| level flight | $C_L = 2W/(ρv^2S)$ |
| stall speed | $v_S = \sqrt{2W/(ρSC_{L,max})}$ |
| a level turn | $n = 1/\cos φ$, $v_S\sqrt n$ |
| glide distance and angle | $d = h\,(L/D)$, $\tan γ = D/L$ |
| density | $ρ = p/(R_sT)$ |
| the rule | DA ≈ PA + 120 ft/°C × $(T - T_{ISA})$ |
| true airspeed | $v_{TAS} = v_{IAS}/\sqrt σ$ |
| Part 2 | |
|---|---|
| circulation | $L = ρvΓ̄b$ |
| thin airfoil | $c_l = 2π(α - α_{L0})$ |
| finite wing | $a = a_0/(1 + a_0/(πeAR))$ |
| induced drag | $C_{D_i} = C_L^2/(πeAR)$ |
| the polar | $C_D = C_{D0} + C_L^2/(πeAR)$, $(L/D)_{max} = ½\sqrt{πeAR/C_{D0}}$ |
| rate of climb | $\text{ROC} = (P_a - P_r)/W$ |
| separation | $c_l = a_0(α - α_{L0})((1 + \sqrt f)/2)^2$ |
| ground effect | $φ = (16h/b)^2/(1 + (16h/b)^2)$ |
| density altitude | $σ = (p/p_0)(T_0/T)$, $h_ρ = (T_0/L)(1 - σ^{1/4.25588})$ |
| the roll | $s ≈ v_{LOF}^2/(2a_{0.7})$ |
| wind | $s_w/s_0 ≈ (1 - v_w/v_{LOF})^2$ |
Symbols: $L$, $W$, $T$, $D$ lift, weight, thrust, drag; $q$ dynamic pressure; $ρ$ density ($ρ_0$ = 1.225 kg/m³ at sea level, standard); $v$ true airspeed; $S$ wing area; $C_L$, $C_D$ lift and drag coefficients ($c_l$, $c_d$ for a section); $α$ angle of attack, $α_{L0}$ the zero-lift angle; $φ$ bank (in a turn) or McCormick’s factor (near the ground); $n$ load factor; $γ$ flight-path angle; $Γ$ circulation; $AR = b^2/S$ aspect ratio, $b$ span; $e$ span efficiency; $C_{D0}$ parasite drag coefficient; $P_a$, $P_r$ power available and required; $f$ the attached fraction of the chord; $h$ height; $σ = ρ/ρ_0$; $R_s$ = 287.05 J/(kg·K); $T_0$ = 288.15 K, $p_0$ = 101,325 Pa, $L$ = 0.0065 K/m in the density-altitude formulas; $g$ = 9.81 m/s²; 1 kt = 0.5144 m/s, 1 ft = 0.3048 m, 1 inHg = 3,386.4 Pa.
Sources and further reading
- Federal Aviation Administration, Pilot’s Handbook of Aeronautical Knowledge (FAA-H-8083-25), a U.S. government work in the public domain: Part 1 follows its topics, their order and its terms, in this lab’s own words; its figures are not reproduced, and no endorsement is implied.
- OpenStax, University Physics Volume 1 (Ling, Sanny and Moebs), chapter 14, “Fluid Mechanics,” §14.1 and §14.6, CC BY-NC-SA 4.0: the notation for pressure, density and Bernoulli’s equation. Fluid Flow & Bernoulli develops the Kutta condition and the pitot tube.
- H. Babinsky, “How do wings work?”, Physics Education 38, 497 (2003); D. F. Anderson and S. Eberhardt, Understanding Flight; D. McLean, Understanding Aerodynamics (Wiley, 2012); J. D. Anderson Jr., Fundamentals of Aerodynamics; J. G. Leishman and T. S. Beddoes (1989) on the separation model; B. W. McCormick on ground effect. The Field Manual (H) describes the model and what it leaves out.
Flight Lab: Lift, Drag & Density Altitude v1.1 · © 2026 Dr. David Meier · CC BY-NC-SA 4.0 · glidepathscience.org