Predict when handling balance becomes a speed boundary

Generated from content/lms/vehicle-dynamics-ii-theory/02-steady-state-handling-balance/03-characteristic-speed-and-critical-speed.md; edit the source file, not this page.

Source path: content/lms/vehicle-dynamics-ii-theory/02-steady-state-handling-balance/03-characteristic-speed-and-critical-speed.md

Course: Read the forces that steer the car

Module: Balance the car with forces and moments

Estimated duration: 60 minutes

Principle

Steady-state handling balance becomes dangerous when you stop treating understeer and oversteer as vague feelings and start treating them as speed-dependent response gains. At low speed, a car can feel manageable even when its basic balance is not friendly. As speed rises, the same understeer gradient changes how much path curvature you get for a given steering input. Positive understeer gradient reduces the response gain compared with neutral steer. Negative understeer gradient increases it. If the negative case is extended far enough in the linear model, the denominator of the response relationship reaches zero and the gain tends toward infinity. That speed is the critical speed.

The clean rule is this: positive K gives you a characteristic speed; negative K gives you a critical speed. Characteristic speed is an understeer measure. It is the speed where the steer angle required for a turn is twice the Ackerman angle. That does not mean the car has reached a loss-of-control boundary. It means the understeer term has grown until the extra steering demand equals the geometric steering demand. Critical speed is different. In an oversteer vehicle, the speed-dependent term subtracts from the denominator. At the critical speed the steady-state model predicts infinite gain. In practical language, a small steering input or external lateral disturbance can produce an unbounded response unless the driver actively catches it.

That distinction is the lesson. Characteristic speed tells you how quickly an understeer car asks for more steering as speed rises. Critical speed tells you where an oversteer car stops being a self-settling steady-state system. The first is a calibration of benign but growing push. The second is a stability boundary.

What you should already have from the sibling lessons

The previous lessons in this module build the bicycle model, calculate understeer gradient from tire and geometry data, and map yaw rate and sideslip to steering and speed. This lesson uses those results. We are not re-deriving axle side-force balance here. We are using the measured or calculated understeer gradient to answer a practical question: at what speed does the sign and magnitude of that gradient become the main fact about the car.

The measurement language is simple. Plot steering angle against lateral acceleration for a known path radius and speed condition. The slope of that steer-angle curve is the understeer gradient. An upward slope means understeer. A flat slope means neutral steer. A downward slope means oversteer. The Ackerman steering requirement is the geometric reference. If the vehicle steer-angle gradient is steeper than the Ackerman gradient, the vehicle is understeer in that region. If the slopes match, it is neutral. If the vehicle gradient is less than the Ackerman gradient, it is oversteer. For an oversteer vehicle, the point where the steer-angle curve slope reaches zero is tied to the stability boundary.

You also need radius discipline. Gillespie gives the practical measurement route: measure speed and steer angle, then determine radius from lateral acceleration or yaw rate. Radius can be derived as R = V^2 / ay or R = V / r when the units are consistent. Without radius, you do not know whether a steering change is a handling-balance change or simply a different path. Without speed, you cannot separate a true speed-dependent response from a different corner. Without lateral acceleration or yaw rate, you are guessing.

The mechanism: why speed turns balance into a boundary

In a steady turn, the steer angle can be understood as a geometric term plus a balance term. The geometric term is the Ackerman steer angle needed for the radius. The balance term grows with lateral acceleration through K. At constant radius, lateral acceleration grows with speed squared. That is why the same K becomes more important at higher speed. A small positive K may be barely noticeable at parking-lot pace but obvious in a long medium-speed sweeper. A small negative K may feel lively at low speed and become a stability problem as speed rises.

For a neutral steer vehicle, the lateral acceleration gain is governed by the numerator and is directly proportional to speed squared. For an understeer vehicle, the denominator includes a positive speed-dependent term, so the gain is always less than neutral. For an oversteer vehicle, the term subtracts from the denominator, so gain increases as speed rises. At the critical speed, that denominator becomes zero.

This is why fixed-radius testing is so revealing. On a fixed radius, an understeer car needs more steering as speed increases. A neutral steer car needs no change in steering input for that fixed path. An oversteer car needs less steering as speed increases. That is not merely a driver feel phrase. It is the operational expression of the sign of K.

The characteristic-speed case

If K is positive, your boundary word is characteristic, not critical. The car asks for more steering as the lateral acceleration rises. At characteristic speed, the total steer angle required for any turn is twice the Ackerman angle. You can think of the steering demand as half geometry and half understeer tax. Below that speed, the geometry term dominates. At that speed, the understeer term has become equally large. Above it, the car may still be stable, but the steering you must add to hold the same curvature becomes a major part of the job.

For a driver, characteristic speed is useful because it turns a vague complaint into a scale. A car with a high characteristic speed has mild understeer across the normal operating range. A car with a low characteristic speed has strong understeer sooner. The speed itself is not the speed where the tires must let go. It is the speed where the understeer contribution has doubled the steering requirement compared with the ideal geometric angle.

For a setup engineer, characteristic speed is useful because it links K, wheelbase, and response. Dixon describes the speed dependence as controlled by kU divided by wheelbase. That matters because two cars with the same understeer gradient but different wheelbases will not feel identical in curvature response. It also matters because K is not only a number from the front axle. Smith frames understeer gradient as front cornering compliance minus rear cornering compliance. If the front axle is more compliant relative to the rear, the car asks for more steering with lateral acceleration. If the rear contribution grows larger, the balance can move toward oversteer.

The critical-speed case

If K is negative, do not try to force the characteristic-speed idea to fit. Dixon states that the characteristic speed would be imaginary in the negative-K case, so the square root is discarded and the result is called critical speed. That is a precise way of saying that an oversteer vehicle has a different kind of speed marker. You are no longer measuring how much extra steering an understeer vehicle needs. You are measuring the speed at which the response relationship becomes singular.

Gillespie ties this directly to straight-ahead disturbance behavior. The same understeer gradient derived from turning behavior determines response to disturbances in straight-ahead driving. Oversteer vehicles have a stability limit at critical speed due to normal disturbances. That means the issue is not only what happens when you deliberately turn in. The same balance can matter when the car is running straight and receives a side force, road input, steering pulse, or wind disturbance.

Above critical speed, Gillespie describes divergent instability in practical terms: any steering-wheel input places the vehicle in a turn of ever-decreasing radius unless the driver makes compensating wheel motions to maintain equilibrium. That is the key driver translation. The car no longer damps the path error in the steady-state sense. The driver becomes the stabilizing element. In a race car with room, grip, skill, and anticipation, a driver may catch a lot. In a driver-development lesson, you do not chase that condition on track.

Critical speed should be predicted on paper, in simulation, or from conservative measured data. It should not be found by accelerating until the car stops behaving.

Trim: the warning label on every answer

A predicted characteristic speed or critical speed is only valid for the trim and environment that produced K. Gillespie is explicit that stability must be examined separately for each environment and trim. Trim includes steer angle, forward velocity, and lateral acceleration. Vehicle behavior also changes with parameters affected by temperature, shock absorber damping, slippery surfaces, and tire cornering properties. In practice, a vehicle can be stable under one set of operating conditions and unstable under another.

This is not a minor caveat. It is the difference between useful prediction and false confidence. A car can be understeer for small inputs and oversteer for large inputs because it is nonlinear. Gillespie and Dixon both point to that reality. Gillespie describes understeer at small inputs and oversteer at large inputs as possible. Dixon says a real nonlinear case may have different characteristics at different lateral accelerations. The result is that one K from one gentle test does not describe the whole track event.

Your answer must therefore include a range. Do not say the car is understeer. Say it is understeer from this lateral-acceleration range to that one, on these tires, at this approximate temperature, on this surface, with this setup. If the steer-angle curve changes slope at higher lateral acceleration, write that down. If the car is understeer at low ay and moves toward oversteer near the limit, the critical-speed question may appear only in the upper trim range.

Aerodynamics can move the boundary

The road-vehicle aerodynamics chunks add another important boundary condition. Steady-state stability can be analyzed as a response ratio whose denominator is a function of speed. The velocity at which that denominator becomes zero is the critical speed. Aerodynamic side-force and yaw-moment derivatives can be included in that denominator. Scibor-Rylski and Sykes argue that, for an average saloon car, those aerodynamic derivative values can be of the same order as tire friction derivatives, so including aerodynamic characteristics is important from the stability point of view.

For the driver and club engineer, the practical lesson is not to treat critical speed as a tire-only number when the car operates in the speed range where aerodynamic forces matter. A car that is benign in a low-speed skidpad test can still have a high-speed stability problem if the aero side-force or yaw-moment behavior changes the denominator. A car that is stable in still air can respond differently with side wind. The aerodynamic discussion also explains why production-road designs are checked so the critical speed is much greater than the maximum speed the car may achieve. The safety margin is the speed spectrum that stays well away from the critical-speed region.

Technique: predict the speed boundary from data

Start with the steer-angle curve, not with a feeling. Use a repeatable radius or a section of track where the path is consistent enough to compare. Record speed, steering input, and either lateral acceleration or yaw rate. Derive radius from the logged channels when you can. The aim is to compare the measured steering requirement with the Ackerman steering requirement for the same path.

Next, identify the slope region you are actually using. If the measured steering-angle curve has a positive slope with lateral acceleration, you are in an understeer region. If it is flat, you are near neutral. If it slopes downward, you are in an oversteer region. Do not average across an obvious slope change. If the car is positive-K at low ay and negative-K at high ay, split the analysis. One value of K across both regions hides the thing you most need to see.

Then decide which speed marker is meaningful. For positive K, calculate or estimate characteristic speed using the same unit system as the K and wheelbase data. If your calculator or worksheet is built from the textbook relationship, the interpretation is still the same: the characteristic speed is where required steer angle is twice Ackerman. For negative K, calculate critical speed. Do not report a characteristic speed for negative K just because your spreadsheet has a cell for it. Dixon gives the correct interpretation: the characteristic-speed expression becomes imaginary and the useful result is critical speed.

After that, compare the predicted speed with the actual speed envelope. If the car is understeer, ask whether the characteristic speed falls below, inside, or above the speed range where you drive constant-radius or long-radius corners. If it is far above your operating range, the understeer may be mild in practice. If it sits inside the range, you should expect steering demand to rise quickly with speed. If the car is oversteer, compare critical speed with maximum vehicle speed and with the speeds of high-speed corners, straights, and side-wind exposure. A critical speed close to the operating range is a stability concern, not a heroic target.

Finally, attach the trim label. A good note reads like this: positive K in the 0.25 to 0.55 g region on current tires; characteristic speed above current medium-corner speed range. Or: K trends negative above the higher lateral-acceleration region; critical-speed calculation should be repeated with high-ay fit only and reviewed before high-speed running. The exact numbers depend on your data, but the structure of the note is what keeps you honest.

Sub-skills that make the prediction trustworthy

The first sub-skill is separating geometry from balance. The steering you hold in a corner is not all understeer or oversteer. Some of it is the Ackerman steer angle required by the radius. The understeer gradient is the part that appears as lateral acceleration rises. If you do not subtract or compare against the geometric term, you will call a tighter line understeer or call a larger radius better balance.

The second sub-skill is preserving a consistent radius. The textbook measurement procedure depends on knowing the radius. On a skidpad, that is direct. On track, you need a corner segment where the path is repeatable enough that the derived radius from speed and lateral acceleration or yaw rate means something. If the driver changes entry, apex, and exit every lap, the steering curve becomes a map of driver inconsistency.

The third sub-skill is sign discipline. Positive K and negative K lead to different speed markers. This is where intermediate drivers often get turned around because both markers have speed in the name. Characteristic speed belongs to understeer. Critical speed belongs to oversteer. If you use the wrong word, you will make the wrong safety decision.

The fourth sub-skill is trim discipline. The car is nonlinear. Tires saturate. Shock damping changes with temperature. Road friction changes. Aero load and side-force behavior can change with speed and yaw. The K that describes one region can be wrong in another. Your prediction is only as honest as the range you attach to it.

The fifth sub-skill is response-gain thinking. Do not reduce the lesson to whether the car pushes or rotates. Ask how much path curvature the car gives you per steering input as speed rises. Understeer lowers that gain relative to neutral. Oversteer raises it. Critical speed is the limiting case where the gain is no longer a usable finite number in the steady-state model.

Calibration cues in the car and in the data

In an understeer region, the cockpit cue is extra steering for the same path as speed rises. The data cue is an upward steer-angle slope with lateral acceleration and a measured gradient above the Ackerman reference. The instructor cue is that your hands keep adding angle while the car does not tighten the path as much as expected. This does not automatically mean the front tires are hopelessly saturated. It may simply mean the car has positive K and you are climbing the speed-squared part of the response.

In a neutral region, the steering needed for a fixed radius does not change much with speed. The data cue is a steer-angle gradient close to the Ackerman reference. This is not a permanent identity for the car. It is a statement about the tested region.

In an oversteer region, the cockpit cue is that less steering holds the same path as speed rises. The car may feel eager or efficient at first. The data cue is a steer-angle slope lower than Ackerman and possibly negative. If the slope heads toward zero in the oversteer case, you are approaching the stability-boundary logic described by Gillespie. At and beyond critical speed, the model says a finite steady-state steering relationship is no longer available. The driver may feel that small corrections produce more yaw response than expected and that the car needs continuous compensating motions rather than settling.

For external disturbances, the cue can appear before a corner. A side force at the vehicle can produce yaw response depending on where it effectively acts relative to the neutral steer line. Gillespie connects static margin and the neutral steer point to whether a lateral disturbance produces steady-state yaw velocity. Understeer behavior tends to be restoring in the straight-ahead disturbance sense. Oversteer behavior has the critical-speed stability limit.

How this lesson should change your driving decisions

Use characteristic speed to size your expectations in fast constant-radius corners. If you know the car has meaningful positive K and the characteristic speed is near your corner speed, you should expect to add steering angle for small speed increases. That affects how you approach a long sweeper. You do not keep adding entry speed and then blame only your hands when the steering demand grows. You recognize that the car's balance is asking for more angle as part of its steady-state character.

Use critical speed as a no-go warning. If an oversteer trim predicts a critical speed anywhere near real operating speed, the right action is to change the car, reduce the speed envelope, or gather safer data at lower trim. It is not a bravery test. Gillespie's description of divergent instability above critical speed is enough: the turn radius keeps decreasing unless the driver makes compensating motions. That is not the condition you deliberately seek at an HPDE or club-racing test.

Use trim labels to decide whether a setup change fixed the problem or just moved it. A car targeted like the C5 example can have no significant change in understeer gradient until very high lateral acceleration, where tire saturation dominates and terminal understeer appears. That is a very different personality from a car that is understeer at low ay and oversteer at high ay. Both may produce the same driver's sentence after one corner, but their speed-boundary predictions are different.

Cross-references

Use the bicycle-model lesson when you need to explain where the axle side-force balance came from. Use the understeer-gradient calculation lesson when you need to calculate K from tire and geometry data. Use the yaw-rate and sideslip lesson when you want to connect the steering input to the vehicle motion channels. This lesson sits after those: it turns K into an operating-speed judgment.

Worked example: the textbook cornering-properties car

Gillespie includes a worked problem built around a car with 1901 lb on the front axle, 1552 lb on the rear axle, and a 100.6 inch wheelbase. The requested outputs include Ackerman steer angles for multiple turn radii, understeer gradient, characteristic speed, and lateral-acceleration gain at 60 mph. That is exactly the workflow this lesson is teaching.

Start by noticing the order. You do not begin with characteristic speed. You first need the geometric steering requirement for the radius. Then you need the understeer gradient from the tire and load information. Only after K is known does characteristic speed have meaning. If K is positive, the characteristic-speed calculation tells you where the understeer term grows until the total steering demand is twice the Ackerman demand. If K were negative instead, you would stop calling it characteristic speed and move to the critical-speed interpretation.

The useful driving lesson from this example is that the speed marker is not a standalone fact about the car. It is downstream of axle loads, wheelbase, tire cornering stiffness, and the measured or calculated steering-gradient behavior. If you change tires, load distribution, or the lateral-acceleration range, you have changed the prediction inputs. A characteristic-speed result printed in a notebook without the tested trim beside it is incomplete evidence.

Worked example: C5-style linearity versus terminal understeer

Smith's C5 discussion gives a different kind of worked example. The target was not merely some understeer. The target was no significant change in understeer gradient until very high lateral acceleration, where tire saturation becomes dominant and terminal understeer results. That is a sophisticated goal because it separates the usable range from the saturation range.

For a driver, the C5 example teaches you not to treat every increase in steering demand as the same problem. In the intended range, a stable positive understeer gradient can make the car readable. The car asks for more steering as speed and lateral acceleration rise, but it does so with consistent balance. Near very high lateral acceleration, tire saturation dominates and the car ends in terminal understeer. That is a limit-state behavior, not the same thing as the characteristic-speed definition.

If you were analyzing this car, you would not fit one careless K across the whole tire-saturation region and then announce one universal characteristic speed. You would identify the main operating range where K is intended to be stable, calculate the characteristic speed for that region, and separately note the terminal-understeer behavior when the tires dominate. The prediction is useful because it is tied to the range where the model applies.

Worked example: the average saloon car and side-wind stability

The road-vehicle aerodynamics material gives a high-speed example that an HPDE driver should respect even if the car is not a prototype. The analysis looks at steady-state response to steering or external force as a ratio of functions, with a denominator that depends on car speed. The speed where the denominator becomes zero is the critical speed. In practical terms, response becomes uncontrollable.

The important addition is that aerodynamic side-force and yaw-moment derivatives can be included. For an average saloon car, the text says those values are of the same order as the tire friction derivatives. That means a high-speed stability prediction that ignores aero can be misleading. Side wind is not just weather trivia. A small side wind can be part of the external-force problem used to examine yaw-angle and turning-radius response with speed.

The practical result is conservative. If a club car has bodywork, ride-height changes, yaw sensitivity, or high-speed straight exposure, do not assume a low-speed skidpad balance completely describes the high-speed boundary. Road-vehicle designs are checked so the critical speed is much greater than maximum speed. Your test plan should keep the same idea: the operating speed spectrum should stay well away from the predicted critical-speed region.

Drill: build a steer-gradient-to-boundary map

Do this as a three-session data drill, not as a speed dare. The goal is to classify the car's tested balance region and write the correct speed-boundary sentence. Success means you can say whether the tested region leads to characteristic speed or critical speed, identify the data that supports the sign of K, and state the trim limits of the answer.

Session one is the baseline. Pick one safe, repeatable, medium-radius corner or a skidpad-style constant-radius exercise if your event provides one. Run six to eight comfortable laps or repetitions without chasing speed. Your job is to make the path repeatable. Log speed and steering angle if available. Log lateral acceleration or yaw rate if available. If you do not have steering-angle data, this becomes an observation drill only and you should not calculate K from it.

Between sessions, mark the segment. For each clean lap, keep the speed, steering, lateral acceleration, and yaw-rate data over the same part of the corner. Derive radius from V^2 / ay or V / r when your channels are reliable. Reject laps where the line changed enough that radius is clearly different. Your first quality check is not speed. It is repeatability.

Session two is the spread. Repeat the same segment at three controlled pace levels: comfortable, moderate, and brisk but well below the limit. Do not add traffic pressure to this drill. The point is to see whether steering demand rises faster than Ackerman, tracks Ackerman, or falls relative to Ackerman as lateral acceleration increases. If the car changes behavior near higher lateral acceleration, split the data instead of averaging it away.

Session three is the prediction. Plot steering angle against lateral acceleration for the accepted laps or have your analysis tool do it. Compare the slope to the Ackerman steering-gradient reference. Positive tested K means report characteristic speed for that region. Negative tested K means report critical speed for that region and treat it as a safety boundary. Add the tire state, surface, temperature impression, setup state, and lateral-acceleration range. The final note should be short and specific: positive K in this range, characteristic-speed interpretation applies; or negative K in this range, critical-speed review required before faster running.

Common mistakes

Mistake one is treating characteristic speed as a crash speed. It is not. It is the understeer speed marker where required steering is twice Ackerman. Good looks like using it to understand how strong the understeer is in your operating range, not using it as a limit-speed warning.

Mistake two is treating critical speed as a lap-time target. It is not. It is the oversteer stability boundary where the steady-state response denominator goes to zero. Good looks like keeping the car's real speed envelope away from it or changing the setup before operating near it.

Mistake three is calculating one K for the whole car forever. Real vehicles are nonlinear. A vehicle can be understeer at low lateral acceleration and oversteer at high lateral acceleration. Good looks like fitting and reporting K by range, then tying each speed prediction to that range.

Mistake four is ignoring radius. Steering angle alone does not tell you balance. If you drove a tighter line, you needed more geometric steering. Good looks like deriving or controlling radius and comparing the measured steering requirement against Ackerman for that radius.

Mistake five is mixing trim states. Temperature, shock damping, road friction, tire cornering properties, speed, and lateral acceleration can all change the answer. Good looks like writing the condition beside the prediction and repeating the analysis after meaningful setup or environment changes.

Mistake six is ignoring aero at high speed. The aerodynamic chunks show that side-force and yaw-moment derivatives can matter for steady-state stability. Good looks like treating high-speed stability, side wind, bodywork, and speed range as part of the boundary question rather than assuming the tire-only low-speed result owns the whole car.

When this principle breaks down

The principle does not break down because characteristic speed and critical speed are useless. It breaks down when you pretend the linear steady-state answer covers conditions it did not measure.

The first breakdown is tire saturation. Smith's C5 discussion separates stable understeer-gradient behavior from the very high lateral-acceleration region where tire saturation dominates and terminal understeer results. Once the tire is saturated, the simple K fit from the main range can stop describing what the car is doing.

The second breakdown is nonlinear balance migration. Gillespie and Dixon both support the idea that the same vehicle can have different characteristics at different lateral accelerations. If the steering curve changes slope, the lesson is not to average harder. The lesson is to split the trims.

The third breakdown is environment drift. Shock damping can change with temperature. A slippery surface can change tire cornering properties. The car can be stable under one operating condition and unstable under another. A prediction made on a cool morning session may not be the same prediction after tires, dampers, and surface change.

The fourth breakdown is unmodeled aero. The road-vehicle aerodynamics material shows that aerodynamic derivatives can alter the steady-state denominator used in critical-speed analysis. If aero is relevant and you omit it, your high-speed stability margin may be fiction.

The fifth breakdown is dynamic instability. This lesson is about steady-state handling balance and speed markers. Gillespie also describes oscillatory instability after a pulse input, where the vehicle turns one way, then the other, and the amplitude grows until it spins out. That is a dynamic response problem. It can coexist with the steady-state lessons here, but it should not be collapsed into one K number.

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.	fc278df5-151f-9ea8-67d4-997fc5a7a788	137	1	uio_books_raw_v1
2	Tires Suspension and Handling Second Edition Dixon John C	84e74ba7-58f1-f1ac-47cf-7f017cbc4fbd	362	1	uio_books_raw_v1
3	Fundamentals of vehicle dynamics Gillespie T. D. Thomas D.	84e79a4d-d418-c429-5be0-3ddc51fb846b	149	1	uio_books_raw_v1
4	Fundamentals of vehicle dynamics Gillespie T. D. Thomas D.	f1b50cf5-cd2b-1a40-fc47-8a12e487b515	257	1	uio_books_raw_v1
5	Racing Chassis and Suspension Design Carroll Smith	00b26d75-535c-d08a-b421-332accf53547	240	1	uio_books_raw_v1
6	Road vehicle aerodynamics Scibor-Rylski A. J Sykes D. M	3c68135b-29a1-7fa8-2e74-b5220c7d24b7	176	1	uio_books_raw_v1
7	Road vehicle aerodynamics Scibor-Rylski A. J Sykes D. M	06a5c540-1df4-4bfd-2d22-8f9975565544	197	1	uio_books_raw_v1
8	Fundamentals of vehicle dynamics Gillespie T. D. Thomas D.	69a0a860-e886-7570-4743-ce77fc1e1157	139	1	uio_books_raw_v1