The balance that sets cornering speed
Travelling a curve requires a force pointing at the centre of the curve, of magnitude mv²/R. On flat pavement the only thing that can supply it is friction at the tire contact patches, and the most friction available is μmg. Set the two equal and the mass cancels from both sides, leaving v² = μgR.
That cancellation is the most useful fact in the equation. A heavy car and a light car with the same tires and the same grip coefficient corner at the same speed. Mass demands more centripetal force and supplies proportionally more friction, and the two effects exactly offset. Weight matters enormously for acceleration, braking distance, tire wear and heat, and for how grip degrades under load transfer — but not for the point-mass limit speed.
Speed grows with the square root of radius. Double the radius and the limit rises by only √2 = 41.4%. That is why a driver gains far more from straightening a corner's radius by taking a wide entry than the geometry suggests at first glance, and why long sweeping corners are where aerodynamic grip pays off — the speeds are already high enough for downforce to matter.
Lateral acceleration and grip coefficient are the same number for a point mass at the limit on flat ground. When a data logger reads 1.05 g in a steady corner, the tires are operating at an effective friction coefficient of 1.05. That equivalence is what makes the skidpad the standard grip measurement: drive a circle of known radius at a steady maximum, time a lap, and the arithmetic gives you μ directly.
Each form of the relationship, and where banking enters
Solving for speed gives vmax = √(μgR) with g = 32.174 ft/s². Work in feet and feet per second, then divide by 1.466667 to reach mph. In SI, use 9.80665 m/s², metres, and multiply m/s by 3.6 for km/h.
Solving for acceleration gives a = v²/(gR) in g. Highway engineers write the same relation as e + f = V²/(15R), with V in mph and R in feet, where e is the superelevation rate and f the side friction factor. The 15 is 32.174 divided by 1.466667², which is 14.96 — the AASHTO Green Book rounds it.
Solving for radius gives Rmin = v²/(μg), the tightest curve a given speed and grip permit. That is the form a course designer or an accident reconstructionist uses.
Banking tilts the road so that part of the normal force points toward the centre of the corner. Resolving forces along and perpendicular to the banked surface, and taking friction at its limit, gives v = √(gR(μ + tanβ)/(1 − μ tanβ)). Two things are worth reading out of it. The numerator adds tanβ to μ, so the bank contributes directly. The denominator subtracts μ tanβ, because the friction force on a banked surface also has a component pointing inward — a second-order gain that becomes large as the product grows.
The denominator can reach zero. When μ·tanβ ≥ 1 the expression has no finite solution: friction on the banking is sufficient to hold the vehicle at any speed within the model. At that point the point-mass model has stopped describing reality, and rollover, suspension travel or aerodynamic limits govern instead. Steep oval banking with modern tires sits close to this regime, which is why real limits there are set by other things.
The skidpad is a special case. On a circle of radius R lapped in time t, speed is 2πR/t, so acceleration is 4π²R/(gt²). Note that it depends on the radius the vehicle's centre of gravity actually follows, not the radius of the cones, which is why the measurement convention matters when comparing published figures.
Worked example: a 300 ft corner with 5° of banking
A track corner has a 300 ft radius on the racing line and 5° of banking. The car has been measured at 1.00 g on a skidpad. You want the limit speed, and you want to know what 80 mph would demand.
- Flat-pavement limit. v = √(1.00 × 32.174 × 300) = √9,652.2 = 98.25 ft/s, which is 98.25 ÷ 1.466667 = 66.99 mph.
- Banking term. tan 5° = 0.087489.
- Banked limit. (1.00 + 0.087489) ÷ (1 − 1.00 × 0.087489) = 1.087489 ÷ 0.912511 = 1.191754. Then v = √(9,652.2 × 1.191754) = √11,502.7 = 107.25 ft/s = 73.13 mph.
- What the banking bought. 73.13 − 66.99 = 6.14 mph, a 9.2% gain from 5° of tilt.
- Demand at 80 mph. 80 mph is 117.33 ft/s, so a = 117.33² ÷ (32.174 × 300) = 13,767 ÷ 9,652 = 1.426 g. That exceeds the 1.00 g available, so 80 mph is not achievable on this corner without more grip.
- Minimum radius at 80 mph. 13,767 ÷ (32.174 × 1.00) = 427.9 ft. To hold 80 mph at 1.00 g you would need a line with at least that radius.
- Skidpad cross-check. On a 200 ft circle (100 ft radius) in 11.50 s: v = 2π × 100 ÷ 11.50 = 54.64 ft/s = 37.25 mph, and a = 4π² × 100 ÷ (32.174 × 132.25) = 3,947.84 ÷ 4,255.01 = 0.928 g.
Notice the skidpad result of 0.928 g does not match the 1.00 g assumed for the corner. That is exactly the kind of disagreement worth chasing: either the skidpad time was not on the limit, the skidpad radius was measured to the cones rather than to the car's path, or the corner assumption was optimistic. The measurement wins.
What the result is telling you, and what it hides
Compare the demanded g against the available g, not against a feeling. The single most useful output here is the lateral acceleration your chosen speed requires. If it exceeds the grip coefficient you have measured, that speed is not available — no amount of technique changes it, and the vehicle runs wide.
Grip coefficients are not tidy. A worn all-season tire on cold pavement may produce well under 0.8 g; a good summer performance tire on warm dry asphalt commonly exceeds 0.9; racing slicks exceed 1.4 and can pass 1.6 in ideal conditions. Wet pavement roughly halves it, and standing water can take it far lower still. These are orientation figures, not specifications: measure your own on a skidpad, because the number depends on the tire, its temperature, its pressure, the surface texture and the load.
Downforce breaks the mass cancellation, which is the point. Aerodynamic load adds vertical force without adding mass, so the friction available rises while the centripetal demand does not. The effective coefficient becomes μ(1 + Fdown/W), which is why an aero car's cornering g climbs with speed. Compute the download at speed with the downforce and drag calculator and enter the boosted coefficient here.
The point-mass model ignores load transfer. Real tires lose grip coefficient as vertical load rises, so shifting load from the inside tires to the outside ones reduces the total force an axle can make. A real vehicle therefore corners slightly below the point-mass prediction, and the shortfall grows with CG height and shrinks with track width — see the weight transfer calculator for how much load actually moves.
Steady state only. Everything here describes a constant-radius corner at constant speed. Turn-in, trail braking and corner exit all combine lateral and longitudinal demand, and the tire's total capability is shared between them along a friction circle: using 0.7 g of braking from a 1.0 g tire leaves √(1 − 0.49) = 0.71 g laterally. The brake bias calculator and stopping distance calculator cover the longitudinal side of that budget.
Maximum cornering speed on flat pavement
| Radius (ft) | μ = 0.70 | μ = 0.85 | μ = 1.00 | μ = 1.30 | μ = 1.60 |
|---|---|---|---|---|---|
| 50 | 22.9 | 25.2 | 27.3 | 31.2 | 34.6 |
| 100 | 32.4 | 35.7 | 38.7 | 44.1 | 48.9 |
| 200 | 45.8 | 50.4 | 54.7 | 62.4 | 69.2 |
| 400 | 64.7 | 71.3 | 77.3 | 88.2 | 97.8 |
| 800 | 91.5 | 100.8 | 109.4 | 124.7 | 138.4 |
| 1,500 | 125.3 | 138.1 | 149.8 | 170.8 | 189.5 |
The coefficients span roughly a worn tire on cool pavement through a performance road tire to a racing slick. Treat them as orientation for reading the table, and measure your own vehicle's figure on a skidpad.
Measuring grip on a skidpad without getting it wrong
Lay out a circle and drive it at a steady speed on the limit, timing complete laps in both directions and averaging — most vehicles differ slightly left to right because of driver weight, fuel position and camber. Three details decide whether the number is comparable to anyone else's. First, the radius must be the one the vehicle follows, usually taken at the centre of gravity, not the radius of the cone circle; a 200 ft cone circle driven with the left tires at the cones puts the car's path several feet larger. Second, the lap must be steady state: any acceleration or braking is using part of the friction budget and understates the lateral figure. Third, tire temperature and pressure must be at their working values, since a cold or over-inflated tire will read low.
Limits and assumptions of this model
- It is a point mass. All four tires are treated as one contact patch with one coefficient. Real vehicles lose capability to load transfer and to front-rear balance mismatch.
- Grip coefficient is treated as constant. Tire friction falls as vertical load rises, varies with temperature, pressure, slip angle and surface, and is not the same in both directions of a corner.
- Steady state, constant radius. No braking, no acceleration, no radius change. Any longitudinal demand comes out of the same friction budget through the friction circle.
- Rollover is not checked. A tall vehicle can reach its static stability limit before its friction limit; compare the demanded g against track ÷ (2 × CG height) as well.
- Banking is assumed constant through the corner. Real transitions into and out of banking are themselves a source of grip loss.
- Aerodynamic load is not included unless you fold it into the grip coefficient yourself, which requires knowing the download at the speed in question — a circular problem best solved by iterating once or twice.
- The racing line radius is not the road radius. A driver using the full width of a corner travels a noticeably larger radius than the centreline, and it is the driven radius that belongs in the formula.
Where these numbers are used outside racing
Highway design uses the identical relation in the form e + f = V²/(15R). A designer picks a design speed, chooses a superelevation rate within the maximum permitted for the climate — snow and ice cap it, because a stopped vehicle must not slide down the banking — and then checks that the side friction factor required is below the comfort-based limits published in the AASHTO Green Book. Those limits are far below tire capability, because the design criterion is passenger comfort and driver behaviour, not the friction limit.
Accident reconstruction runs the relation backwards. A vehicle that departed a curve of measured radius must have exceeded √(μgR), and a drag sled or a test vehicle establishes μ on the actual surface. The result is a minimum speed, not the speed, because the vehicle may have been travelling faster and lost control for another reason.
Autocross and track driving use it to decide where speed is available. Compute the limit speed for each corner radius on the map, and the corners where your car is slowest relative to that limit are where technique or setup is costing time — not the ones that simply feel slow.
Vehicle testing publishes skidpad g as the standard summary of lateral grip, which is why the skidpad section of this calculator matters: it lets you compare your own measurement against published figures, provided you use the same radius convention.
To turn any of this into lap time you also need what the car can do in a straight line, which is where the power-to-weight ratio and quarter mile calculations come in.
Key terms
- Lateral acceleration
- Acceleration directed toward the centre of the corner, conventionally expressed in multiples of g. Numerically equal to the effective friction coefficient at the limit on flat ground.
- Grip coefficient (μ)
- Peak lateral tire force divided by the vertical load producing it. Not a constant: it falls as load rises and varies with temperature, pressure and surface.
- Superelevation
- The banking of a roadway curve, quoted by highway engineers as a rate e in feet of rise per foot of width — numerically the tangent of the bank angle.
- Skidpad
- A marked circle of known radius used to measure steady-state lateral grip by timing laps at the limit.
- Friction circle
- The representation of a tire's total force capability as a circle of radius μ, within which longitudinal and lateral demands must jointly fit.