Calculate understeer gradient from tire and geometry data
Generated from
content/lms/vehicle-dynamics-ii-theory/02-steady-state-handling-balance/02-understeer-gradient-and-neutral-steer.md; edit the source file, not this page.
Source path: content/lms/vehicle-dynamics-ii-theory/02-steady-state-handling-balance/02-understeer-gradient-and-neutral-steer.md
Course: Read the forces that steer the car
Module: Balance the car with forces and moments
Estimated duration: 50 minutes
The skill in this lesson is not naming a car understeer or oversteer after you have already felt it. The skill is calculating the balance number before you drive the car, then checking whether the number makes sense against measured steering data. You are turning tire cornering stiffness, axle loads, wheelbase, and simple geometry into an understeer gradient.
Understeer gradient is the rate at which the steering needed for a steady corner changes as lateral acceleration rises beyond the pure geometric steering requirement. In the low-speed bicycle model, the car is simplified to zero track width, small steering angle, and no front or rear slip angle. At that condition, the steer angle required for a constant-radius path comes from wheelbase and radius. In degree form, the geometric steer angle is 57.3 times wheelbase divided by turn radius. That is the clean baseline. It is what the car would need if tire slip did not have to be part of the job.
At real cornering speed, the tires do develop slip angle. The front and rear axle slip angles are not just side effects; their difference is what changes the steer angle required to hold the same path. If the front axle needs more slip angle contribution than the rear axle for each g of lateral acceleration, the car needs extra steer as speed rises on a fixed radius. That is positive understeer gradient. If the front and rear contributions are equal, the extra steer term is zero and the car is neutral by this steady-state definition. If the rear axle contribution is larger, the required steer angle decreases as speed rises and the gradient is negative.
The calculation has two halves. The first half is geometry: wheelbase and radius tell you the Ackermann or kinematic steer angle. The second half is tire and load data: axle load divided by axle cornering stiffness tells you how much slip angle that axle needs per unit of lateral acceleration. Carroll Smith describes the synthesis version as front cornering compliance minus rear cornering compliance. That is the shortest useful sentence in the lesson. Calculate the front axle compliance, calculate the rear axle compliance, subtract rear from front, and the result is the understeer gradient.
Use this sign convention throughout the lesson. K greater than zero means understeer. K equal to zero means neutral steer. K less than zero means oversteer. Do not reverse the sign because the car feels tight or loose in your favorite corner. The sign is tied to the slope of required steer against lateral acceleration, after the geometric steering requirement has been accounted for.
The governing steady-state relationship is steer angle equals geometric steer angle plus K times lateral acceleration. With radius written directly, that is steer angle equals 57.3 times wheelbase divided by radius, plus K times lateral acceleration. If you measure speed and lateral acceleration instead of radius, the radius can be derived as speed squared divided by lateral acceleration. If you measure speed and yaw rate, radius can be derived as speed divided by yaw rate. Those substitutions matter because track data rarely hands you perfect radius directly.
For the tire-and-geometry calculation, the inputs are simple but must be used carefully. You need wheelbase for the geometric side. You need front and rear axle load, or the mass and CG position on the wheelbase that imply those loads. You need tire cornering stiffness at the prevailing vertical load on each tire. The phrase prevailing load is important. Gillespie does not calculate understeer gradient from a generic tire number; the example first interpolates cornering stiffness at 950 lb per front tire and 776 lb per rear tire. The tire is a different lateral-force device at different vertical loads, so the stiffness row you use is part of the answer, not clerical setup.
Once you have the tire stiffness values, be clear about whether the data is per tire or per axle. Many tire tables give cornering stiffness per tire. The axle sees two tires. If the front tire stiffness is 232 lb per deg per tire, the front axle stiffness is 464 lb per deg when both front tires are represented by one bicycle-model front axle. If the rear tire stiffness is 195 lb per deg per tire, the rear axle stiffness is 390 lb per deg. Using the per-tire values directly with axle loads doubles the compliance and gives the wrong magnitude. It may leave the sign unchanged in a simple example, but it teaches you the wrong scale.
Now calculate axle compliance. In the linear range, one g of lateral acceleration requires lateral force equal to the axle weight at that axle. If the front axle weight is 1901 lb and the front axle cornering stiffness is 464 lb per deg, the front axle needs about 4.10 deg of slip angle contribution per g. If the rear axle weight is 1552 lb and the rear axle cornering stiffness is 390 lb per deg, the rear axle needs about 3.98 deg per g. Understeer gradient is front compliance minus rear compliance, so the result is about 0.12 deg per g. That is a small positive understeer gradient.
That worked number is useful because it shows what the calculation is actually comparing. You are not asking which axle has the higher stiffness by itself. In the example, the front axle has higher total cornering stiffness than the rear axle. But the front axle also carries more weight and therefore must generate more lateral force per g. The balance number comes from load divided by stiffness at both ends. A heavily loaded but stiff axle can still have more compliance than a lighter axle if the load-to-stiffness ratio is larger.
This is also why CG position belongs in the lesson. Dixon states that, for the linear suspensionless vehicle, understeer gradient can be related to tire characteristics, mass, and the position of the center of gravity on the wheelbase. If the CG moves forward, more of the lateral-force demand belongs to the front axle. If the front tire cornering stiffness does not rise enough to match that added demand, the front compliance grows relative to the rear and K moves positive. If the rear carries more of the demand without matching stiffness, K moves negative. The calculation is a load-to-stiffness balance, not a front-tire-versus-rear-tire beauty contest.
Keep the scope linear unless you have data that supports a map. The single-number calculation assumes tire cornering stiffness is usable as a local slope at the operating load. Real vehicles do not keep one perfect K everywhere. Smith notes that the magnitude varies as lateral acceleration increases because of chassis tuning and nonlinear tire characteristics, and Dixon makes the same point for real vehicles beyond the linear handling regime. That means a calculated K from one operating point is a local estimate. It is the right number for early setup reasoning and for low-to-moderate steady-state comparison, not a promise that the car will keep the same balance all the way to saturation.
The practical workflow is five steps. First, define the trim condition. Decide whether you are calculating the car as static, at a known steady lateral acceleration, or at a test point with measured loads. Second, determine front and rear axle loads at that trim. Third, obtain the tire cornering stiffness at the tire loads you will use. Interpolate if the table gives stiffness at nearby loads. Fourth, convert tire stiffness into axle stiffness if the table is per tire. Fifth, compute front compliance, rear compliance, and K as front minus rear.
The trim condition is not a formality. If you calculate from static axle weights but then compare the answer to data taken near the limit, the mismatch may come from nonlinear tire stiffness, load transfer, roll stiffness distribution, or saturation rather than an arithmetic error. Sergers notes that suspension tuning changes cornering balance through load distribution between the front and rear axles, with spring rates, rollbar rates, damping, tire pressures, and similar setup changes influencing that distribution. Gillespie also explains that suspension and steering factors influence the cornering forces developed at each wheel. So when you compare calculation to track data, ask whether you calculated the same car state you measured.
The radius side is the sanity check. At very low speed, where acceleration effects are negligible, the steer angle should be explained by wheelbase and radius. As speed rises on the same radius, the departure from that geometric angle is the balance term. SAE J670 definitions, as presented by Sergers, frame neutral steer, understeer, and oversteer by comparing the steering wheel angle gradient divided by steering ratio against the Ackermann steering gradient. In plain workshop language, get from steering wheel angle to road-wheel steer angle, remove the geometric steering slope, and read the remaining slope against lateral acceleration.
If the measured steering gradient is greater than the Ackermann gradient, the car is understeer at that trim. If it is equal, the car is neutral. If it is smaller, the car is oversteer. Gillespie states the same plot-reading rule: the slope upward to the right is understeer, zero slope is neutral steer, and a negative slope is oversteer. That is why the calculated K should eventually be compared to a plot, not just kept in a spreadsheet. The calculation tells you what the tire and geometry data predicts. The steering plot tells you what the car did.
Do not confuse Ackermann steer angle with mechanical Ackermann steering. The chunks use Ackermann steer angle as the kinematic angle required by wheelbase and turn radius in the bicycle model. It is a geometry baseline for the whole vehicle. It is not the same thing as the steering linkage layout that gives different inner and outer wheel angles. The calculation here works with the bicycle-model steer angle, steering ratio, radius, lateral acceleration, and axle slip-angle requirements.
The axle-compliance view is especially powerful for setup thinking because it tells you which side of the car is consuming the extra slip angle. Smith describes front and rear cornering compliances as axle contributions to understeer gradient. If front compliance is high, the front axle needs more slip angle per g. If rear compliance is high, the rear axle needs more slip angle per g. Understeer gradient is the difference, but the two terms are the diagnosis. Two cars can have the same K while having very different front and rear compliance levels, and those cars can feel different in response time and damping even if their steady-state balance number matches.
For this lesson, keep that diagnosis bounded. The sibling lessons cover building the bicycle model, speed boundaries, and yaw-rate or sideslip mapping. Your job here is to calculate the steady-state balance term from tire and geometry data. You should know that lateral acceleration response time and yaw velocity damping can be synthesized with understeer gradient and cornering compliance on a handling map, as Smith describes, but you do not need to turn this lesson into a transient-response lesson. K is the steady-state balance number. It is not the whole handling character of the car.
A good result looks boring in the best way. The units are consistent. The tire stiffness values match the loads. Per-tire stiffness has been converted to axle stiffness. The sign convention is clear. The geometric steering term is separated from the balance term. The final number is reported in deg per g. If you also have measured data, the road-wheel steering plot versus lateral acceleration should show a slope that agrees in sign with the calculated K in the same operating region.
When the number and the driver impression disagree, do not immediately throw away the math. Check whether the driver is describing entry, middle, or exit. This calculation is for steady-state cornering balance. Entry feel can be dominated by brake release, yaw response, and transient load movement. Exit feel can be dominated by power application. The chunks here support steady-state cornering, constant radius, steering gradient, and tire cornering stiffness. Use the number in that lane. If the car is tight only on throttle exit, this lesson is not enough evidence to call the steady-state K wrong.
The mechanism behind the sign is tire slip angle demand. Gillespie describes the simple physical result: if the front tires must assume a greater slip angle to maintain the needed lateral force, the front ploughs out and the vehicle understeers. If the rear tires are the ones forced into the larger slip, the rear slips out and the vehicle oversteers. The calculation is a clean way to predict which axle is closer to that extra slip-angle burden in the linear range.
A final caution: do not worship the single number past its evidence. Smith describes a Corvette C5 target where understeer gradient would not change significantly until very high lateral acceleration, where tire saturation dominates and terminal understeer results. That is a target for linearity across a range, not a claim that every car has one constant K. If you have tire data at multiple loads or measured steering data at multiple lateral accelerations, calculate or fit K locally across the range. The shape of K versus lateral acceleration is part of the handling story. The single K number is the first page, not the full book.
Worked example: Gillespie axle-load and tire-stiffness calculation
Start with the sample vehicle data in the bonded chunks. The car has 1901 lb on the front axle, 1552 lb on the rear axle, and a 100.6 inch wheelbase. The tire table is used at the prevailing tire loads, not at an arbitrary catalog value. The front tire load is 950 lb per tire, and interpolating the tire data gives 232 lb per deg per front tire. The rear tire load is 776 lb per tire, and interpolation gives 195 lb per deg per rear tire.
Because those stiffness values are per tire, convert them to bicycle-model axle stiffness. Front axle stiffness is 2 times 232, or 464 lb per deg. Rear axle stiffness is 2 times 195, or 390 lb per deg.
Now calculate the front axle cornering compliance. At one g, the front axle must generate lateral force equal to its axle load, so the front slip-angle contribution is 1901 divided by 464, which is about 4.10 deg per g. Rear compliance is 1552 divided by 390, which is about 3.98 deg per g. Understeer gradient is front minus rear, so K is about 0.12 deg per g.
The result is positive, so the car is understeer by this calculation. It is not a large number in this example, but the sign is not ambiguous. On a fixed radius in the same linear region, this car should require slightly more road-wheel steer as lateral acceleration rises than the geometric steer angle alone would require.
The important lesson is not the arithmetic; it is the comparison. The front axle is stiffer in absolute terms, but it also carries more load. The rear axle is softer in absolute terms, but it carries less load. The balance comes from each axle load divided by each axle stiffness. You should be able to look at the two compliance terms and say which axle is asking for more slip angle per g before you ever interpret the final K.
Worked example: C5-style handling-map thinking
Smith describes using understeer gradient and front and rear cornering compliances as synthesis targets for a production performance car, with the C5 intended to avoid significant change in understeer gradient until very high lateral acceleration, where tire saturation dominates and terminal understeer appears. Treat that as a lesson in how to use the calculation across a range.
Suppose you have tire data at several vertical loads and a suspension model or measured data that lets you estimate front and rear tire loads at several lateral accelerations. You do not calculate one K and declare the whole car solved. You calculate front compliance and rear compliance at each trim point. Then you subtract rear from front at each point and look at the curve.
A C5-style target would not be judged only by whether K is positive at one point. It would also be judged by whether K changes smoothly and modestly through the intended range. If K rises early, the driver may feel the car asking for more and more steering in a nonlinear way before the limit. If K crosses from positive to negative at higher lateral acceleration, the car may move from understeer to limit oversteer. Gillespie notes that some vehicles are understeer at low lateral acceleration and change to oversteer at high lateral acceleration. That is exactly why the plotted gradient matters.
This worked example is less about a named corner and more about a named development method. A handling map is a way to keep the calculation honest. Front compliance, rear compliance, and K together tell you whether the balance number is stable, whether one axle is driving the change, and whether the driver will feel the car as linear or as changing personality with lateral acceleration.
Worked example: Hockenheim data as a measurement cross-check
Sergers discusses Formula One data at Hockenheim in the context of roll angle and suspension channels, and the same data-analysis discipline applies when you check understeer gradient from logged steering data. The calculation from tire and geometry data predicts a balance term. The logged channels let you see whether the steering gradient agrees in the steady-state portions of the lap.
For each steady cornering segment, you need speed, lateral acceleration or yaw rate, and steering angle. If you have lateral acceleration, derive radius from speed squared divided by lateral acceleration. If you have yaw rate, derive radius from speed divided by yaw rate. Then calculate the geometric steer angle from wheelbase and radius. Convert steering wheel angle to road-wheel steer angle using the overall steering ratio. The difference between road-wheel steer angle and the geometric angle is the understeer angle for that point.
Plot that difference against lateral acceleration for the steady portions. A positive upward slope means the car is asking for more steer as lateral acceleration rises. A flat line near zero means the car is close to neutral by this definition. A downward slope means the road-wheel steer demand is falling relative to the geometric requirement, which is oversteer by the steady-state definition.
This Hockenheim-style cross-check also protects you from mixing different mechanisms. If roll stiffness distribution, tire pressure, damping, or other setup changes alter load distribution, the measured balance can move away from the original tire-table calculation. That does not make the calculation useless. It tells you the original inputs no longer describe the car state you are measuring.
Common mistakes
The first common mistake is using steering wheel angle as if it were road-wheel steer angle. SAE J670 definitions compare steering wheel angle gradient divided by overall steering ratio against the Ackermann steering gradient. If you skip the steering ratio, your measured gradient is in the wrong scale. Good work converts the steering signal before comparing it to the geometric term.
The second mistake is confusing Ackermann steer angle with Ackermann steering geometry. The Ackermann steer angle in this lesson is the kinematic bicycle-model angle set by wheelbase and turn radius. It is a baseline for steady-state path following. Good work treats it as geometry, not as a claim about the mechanical steering linkage.
The third mistake is using tire cornering stiffness at the wrong load. Gillespie explicitly finds the stiffness at 950 lb per front tire and 776 lb per rear tire before calculating K. Good work interpolates or otherwise obtains stiffness at the prevailing loads, then states what load each value represents.
The fourth mistake is mixing per-tire and per-axle stiffness. If the tire table gives 232 lb per deg per tire, the bicycle-model axle stiffness is two tires together. Good work converts per-tire values to axle values before dividing axle load by stiffness.
The fifth mistake is reading the sign backward. Positive K is understeer because the required steer angle rises faster than the geometric Ackermann requirement as lateral acceleration rises. Negative K is oversteer. Good work states the sign convention next to the result.
The sixth mistake is treating one linear K as a whole-limit truth. Smith and Dixon both describe the gradient changing with lateral acceleration in real vehicles because tires and chassis behavior are nonlinear beyond the simple linear regime. Good work reports the operating region of the calculation and uses multiple points or a map when the data supports it.
The seventh mistake is using the calculation to explain transient complaints. This lesson is about steady-state balance. Good work compares it to constant-radius or steady-corner data, not to a driver comment about brake-release entry rotation or power-on exit behavior unless those phases have been separated from the data.
Drill: one-session understeer-gradient notebook
Do this drill at your next test day only where the event format and safety rules allow steady-state data collection. The goal is not to set a lap time. The goal is to make one tire-and-geometry K estimate and one measured steering-gradient check for the same car state.
Before the event, prepare the calculation sheet. Enter wheelbase, front axle load, rear axle load, tire cornering stiffness at the intended tire loads, and steering ratio. If your tire data is per tire, create a separate line that doubles it into axle stiffness. Calculate front compliance, rear compliance, and K. Write the expected sign in plain language: positive means more steer with rising lateral acceleration, zero means no extra steer beyond geometry, negative means less steer.
At the event, collect six to eight steady-state samples from an approved constant-radius area or from steady portions of similar-radius corners. For each sample, log or record speed, steering wheel angle, and either lateral acceleration or yaw rate. Keep the car in a stable trim. Do not mix trail braking, throttle application, curb strikes, or obvious transients into the samples.
After the session, convert steering wheel angle to road-wheel steer angle. Derive radius from speed squared divided by lateral acceleration, or from speed divided by yaw rate. Calculate the geometric steer angle from 57.3 times wheelbase divided by radius. Subtract the geometric angle from the road-wheel steer angle. Plot that remainder against lateral acceleration.
The success criterion is a plot that lets you identify the sign of the slope and compare it to your calculated K. You succeed if you can explain whether the measured trend agrees with the calculated sign, and if you can name at least one input that would need refinement if it does not. You do not need a perfect match on the first attempt. You need a clean separation between geometry, tire compliance, and measurement.
When this principle breaks down
The principle breaks down when the inputs stop representing the car state. Tire cornering stiffness is a local slope, and the chunks make clear that real vehicles change behavior as lateral acceleration rises. At high lateral acceleration, nonlinear tire behavior and saturation can dominate. A car can be understeer at low lateral acceleration and move toward oversteer at high lateral acceleration, or it can build terminal understeer near saturation.
It also breaks down when suspension and steering effects dominate the simple tire-and-load picture. Gillespie states that many design factors influence the cornering force developed at a wheel, with suspension and steering systems as primary sources. Sergers points to roll stiffness distribution and setup changes such as springs, rollbars, damping, and tire pressures as ways the load distribution between axles is changed. If those effects are changing the tire loads or effective cornering stiffness during the test, a static tire-table calculation is only the starting estimate.
The recovery is to turn the single calculation into a local map. Recalculate at the loads and lateral accelerations you actually care about. Compare front and rear compliance separately, not only K. Then validate against measured steering gradient in steady-state data. That keeps the lesson honest: calculate first, measure second, and update the model only when the evidence says the operating point has changed.
Author Review
No quiz questions are attached to this lesson.
Sources
| # | Document | Chunk | Pages | Score | Collection |
|---|---|---|---|---|---|
| 1 | Fundamentals of vehicle dynamics Gillespie T. D. Thomas D. | 84e79a4d-d418-c429-5be0-3ddc51fb846b | 149 | 1 | uio_books_raw_v1 |
| 2 | Fundamentals of vehicle dynamics Gillespie T. D. Thomas D. | 6a17ba6e-e469-096e-5da7-bc3b382deedb | 150 | 1 | uio_books_raw_v1 |
| 3 | Analysis Techniques for Racecar Data Acquisition (Jorge Sergers) | 982a8f4b102b95cc277b15bba765c60e | 10 | 1 | uio_books_raw_v1 |
| 4 | Racing Chassis and Suspension Design Carroll Smith | 00b26d75-535c-d08a-b421-332accf53547 | 240 | 1 | uio_books_raw_v1 |
| 5 | Tires Suspension and Handling Second Edition Dixon John C | 22d6eebc-229f-5f45-1d5b-ce2e9209d742 | 366 | 1 | uio_books_raw_v1 |
| 6 | Fundamentals of vehicle dynamics Gillespie T. D. Thomas D. | 88cbdbfe-237b-b2ed-0f8b-09605b9839e3 | 139 | 1 | uio_books_raw_v1 |
| 7 | Analysis Techniques for Racecar Data Acquisition | 8992f0b7-7dd2-a28d-21ec-332277b778fa | 11 | 1 | uio_books_raw_v1 |
| 8 | Analysis Techniques for Racecar Data Acquisition (Jorge Sergers) | 59fe115fc09f34b0eca5bfdc4d4b4f1a | 12 | 1 | uio_books_raw_v1 |