Map transient handling with steering frequency response
Generated from
content/lms/vehicle-dynamics-ii-theory/04-transient-vehicle-response/04-steering-pulse-and-frequency-response.md; edit the source file, not this page.
Source path: content/lms/vehicle-dynamics-ii-theory/04-transient-vehicle-response/04-steering-pulse-and-frequency-response.md
Course: Read the forces that steer the car
Module: Model the car's reaction to sudden inputs
Estimated duration: 45 minutes
A steering frequency response map answers one practical question: when you ask the car to change direction at different rates, does the car answer in a controlled, predictable way, or does the answer change character as the steering gets quicker?
That is different from asking whether the car has steady-state understeer or oversteer. Steady-state balance matters, but it is not the whole handling story. Handling is the quality that lets you control the vehicle safely and predictably, hold the desired course, and still control it when longitudinal and lateral accelerations are high. The word predictably is doing a lot of work here. A car can feel balanced in a long, constant-radius corner and still be hard to place during a fast lane change, an abrupt correction, or a quick turn-in after braking. The difference is transient response.
In this module you have already treated the car as a planar model, written body-axis equations, linearized the car, and studied the first answer to a step steer. This lesson uses that foundation differently. A step steer asks what the car does after one sudden change in steering. A frequency response test asks how the car answers a repeating steering input as the input rate changes. You are not trying to memorize a new equation. You are learning how to read a response map as a driver and data user: gain, delay, overshoot, damping, and the point where the car stops feeling like one clean system and starts feeling like a stack of separate lags.
The key principle is simple. Steering response depends not only on how much steering you add, but on how fast you add it. Dixon makes the same point in the damper context by noting that handling frequency depends on the rate of control application. A slow steering input and a rapid steering input can involve very different suspension stroke, damper velocity, and damper force ranges. A lane-change event has its own stroke and velocity signature. A single bump or rough road can produce a different one again. So if you judge the car only from a slow corner, you may be judging the wrong frequency band for a fast correction. If you judge it only from a quick transient, you may be judging the car in a band where dampers, compliance, and driver timing dominate.
Think of the steering wheel as the input signal. The car answers with yaw rate, sideslip behavior, lateral acceleration, roll motion, pitch motion if you are also changing brake or throttle, and tire load variation. In a simple linear exercise, you can imagine applying a small sinusoidal steer input and watching those outputs repeat. At low input frequency, the car has time to build lateral acceleration and body motion in a relatively settled way. At higher input frequency, the steering is changing before the whole vehicle has finished answering the previous part of the cycle. The output can shrink in gain, lag in phase, overshoot, ring, or become messy because road, tire, suspension, and driver inputs are no longer separated.
For an intermediate driver, the useful skill is not to produce a formal Bode plot from memory. The useful skill is to map which band of steering input gives a clean car, which band gives a delayed car, and which band gives an unsettled car. You should be able to say: at slow hands the car takes a set and tracks; at medium hands the yaw rate builds cleanly but the body arrives a little late; at fast hands the car needs a pause before the second input; over bumps the same steering rate is not comparable because the road is now an input too. That is frequency response thinking.
Section 1 - What a steering frequency response map contains
A steering frequency response map has one input and several possible outputs. The input is steering angle over time. The outputs can include yaw rate, sideslip behavior, lateral acceleration, roll angle, roll rate, pitch response, suspension stroke, damper velocity, and tire load variation. The exact channels depend on the instrumented car, but the logic is the same: compare how much output you get, how late it arrives, and how cleanly it settles.
The cleanest textbook input is a small-amplitude sinusoidal steer. Small amplitude matters because the lesson is about response to input rate, not about driving the tires deep into nonlinear saturation. Dixon frames unsteady-state handling around models with yaw and sideslip behavior, steer response, power steering, and disturbance response. That is the family of problems you are entering. You are asking how a steering disturbance travels through the vehicle.
There are three basic readings.
Gain is how much output you get for a given steering input. If the same small steering angle produces less yaw rate as frequency rises, the car is filtering the input. That can be good or bad depending on context. A race car that answers every tiny hand tremor at high speed can be tiring or nervous. A car that filters too much can make quick placement corrections feel late.
Phase is how late the output arrives compared with the steering input. A little lag is normal because tire force, yaw acceleration, roll motion, and suspension load transfer do not all appear instantly. Too much lag is the source of the familiar complaint that you turned, waited, added more steering, and then the car finally arrived all at once. That is not merely a driver confidence issue. It is a timing problem in the driver-car loop.
Damping and settling describe what happens after the main answer. If the car gives one clean yaw response and then quiets down, the transient is well controlled. If it oscillates, needs a second correction, or keeps moving after your hands have already unwound, the response is undercontrolled for that input band. Dixon emphasizes that acceleration variations cause pitch and roll angles to develop, and that this development must occur in a controlled way. That is exactly what you are watching in a frequency response map: not only whether the car turns, but whether the body and tire system arrive under control.
Section 2 - Why slow steering and rapid steering are not the same test
The damper table in Dixon is useful because it refuses to let you pretend that all steering is one event. It separates slow steering, rapid steering, lane change, single bumps, brake or accelerate, smooth road, and rough road by suspension stroke, velocity, damper force, effective force per velocity, and effective force per stroke. The exact numbers are less important than the pattern. The car lives in different motion regimes depending on the task.
A lane change in the table is associated with a large suspension stroke compared with small smooth-road motion, but the damper velocity is not the same as a single bump or rough road. A single bump can have much smaller stroke than a lane change but much higher velocity. Rough road can combine meaningful stroke with high damper speed. Brake and acceleration events bring pitch into the picture. This is why the same damper, spring, tire, and chassis package can feel composed in one exercise and busy in another.
For steering frequency response, this means you must name the input band before judging the car. Slow steering is a low-frequency control application. You are giving the chassis time to build roll and lateral acceleration. Rapid steering is a higher-frequency control application. You are asking the car to create yaw rate before all the slower body motions have fully settled. A lane change is not just two corners next to each other; it is an alternating steering task with timing. A bump is not steering, but it can inject suspension velocity into the same time window as your steering input and corrupt the result.
On track, this is why a car can feel fine in a long sweeper and vague in a quick esses section. It is also why a driver can misdiagnose a car after a rough corner entry. If the road surface is exciting the suspension at the same moment you are testing steering response, you are not mapping pure steering frequency response. You are mapping steering plus disturbance response.
Section 3 - The driver problem: open-loop input versus driver variability
Dixon includes a discussion prompt asking whether definitive objective transient handling tests are possible given driver variability. That question matters because the driver is usually the least repeatable part of an HPDE frequency response test. If you change steering amplitude, rate, entry speed, brake release, throttle pickup, and line all at once, you have not learned the car. You have changed the experiment.
A useful steering response exercise is as close to open-loop as practical. Open-loop does not mean you stop being responsible for the car. It means the input is planned before the car responds. You decide the steering waveform first: small amplitude, smooth sine-like motion, fixed speed range, fixed location, fixed number of cycles. Then you observe the car. If you instead steer until the car feels right, you are running a closed-loop correction exercise. That is useful for driving, but it hides the response map because your hands are correcting the very behavior you are trying to observe.
The discipline is to separate three phases. First, make the planned input. Second, let the car answer. Third, evaluate whether the answer was clean. When you are still learning, you will want to fix the car halfway through the input. Resist that during the drill, within safety limits. The goal is not to be passive in real driving. The goal is to collect one honest sample.
This is also why small amplitude is important. A small steer input lets you stay inside a range where the linearized model is still meaningful. Carroll Smith notes that steady-state equations can ignore tire and bushing compliance and linearize roll resistance in ways that are reasonable for road racing vehicles, but that these simplifications can be removed for vehicles with larger roll angles. The same warning applies here in driver language. If you make the input big enough to produce large roll, large tire saturation, or large compliance steer, the clean small-signal map is gone. You may still learn something, but it is a different lesson.
Section 4 - How to read the response in the seat
Start with the steering wheel. In a clean low-frequency response, you turn in with a measured input, the car builds lateral acceleration progressively, and the body takes a set without a second heave or delayed roll. You do not need to add a surprise second steering input to make the car arrive. You also do not need to unwind early because the car suddenly catches up.
At a medium frequency, the useful cue is the gap between hand motion and yaw response. You add steering and feel whether the nose starts rotating while your hands are still moving, just after your hands stop, or only after you have already decided the car missed the first part of the request. A small delay is normal. A delay that makes you add more steering before the car answers is the danger sign. That is how drivers create too much input: the car is late, the driver stacks on more steering, then the stored response arrives and the car rotates more than intended.
At a higher frequency, watch for reversal quality. In an alternating input, the first half of the motion is only half the test. The second half asks whether the car can stop building yaw one way and begin building yaw the other way without a sloppy pause. If the first response is acceptable but the reversal feels heavy, delayed, or oscillatory, the car may be fine in single turn-in events and poor in linked transitions.
Now add the body. Dixon points out that handling is affected not only by cornering and braking directly, but by ride motions that cause problems during cornering or braking. Pitch and roll angles develop as acceleration varies, and they need to develop in a controlled way. In the seat, uncontrolled development feels like the platform is still moving after the steering request has changed. You turn left, the car begins to roll, you ask it back right, and the body is still finishing the left answer. The steering wheel may be precise, but the whole vehicle is late.
Finally, add the tire contact feeling. A frequency response map is not only a yaw map. Dixon's ride discussion says that a basic ride simulation showing tire discomfort is not a complete representation of suspension effect on handling, but it is a useful guide to trends. The tire still matters. If high-frequency steering creates a skipping, pattering, or inconsistent contact feeling, you may be seeing the interaction between damper speed, road input, and tire deflection rather than pure steering geometry. The response map should tell you when the car is answering through tire force and when it is being disturbed by vertical motion.
Section 5 - How to read the response in data
If you have data, use it to discipline your memory. The simplest useful channels are steering angle, lateral acceleration, and yaw rate. Add speed so you are not comparing different tests. Add brake and throttle so you can reject samples where longitudinal input changed the pitch state. If you have suspension potentiometers or damper velocity, they can help you understand why a clean steering trace still produced a messy car.
For each steering frequency band, ask five questions.
First, was the input repeatable? The steering trace should show the intended amplitude and rate. If the input changed every cycle, driver variability is the dominant signal.
Second, did yaw rate gain change? If the yaw response gets weaker as you steer faster, the car is filtering the input. That might be acceptable at very high frequency, but it matters if the weak band overlaps actual track tasks such as a lane change or quick transition.
Third, did lateral acceleration lag yaw rate or arrive with a delayed body set? Yaw can begin before the tire load and roll state are fully settled. The driver feels that as a nose that points before the car is fully loaded.
Fourth, did the response overshoot or ring after your hands stopped? A clean steering input followed by a second yaw movement is evidence that the vehicle still had stored motion to release. In driver terms, the car did not finish when you finished.
Fifth, did suspension velocity spike because of road or brake input? Dixon's table separates smooth road, rough road, bumps, braking, and steering because they are different motion sources. Data lets you avoid blaming steering response for a vertical disturbance.
This is where time-domain thinking remains useful. Dixon describes time-domain analysis as considering vehicle behavior as time passes, with velocities following from position changes over a time step or acceleration over a time step, and with forces deduced from positions and velocities. You do not have to run a full simulation to use the idea. When you look at a data trace, you are reading positions, velocities, accelerations, and forces in time order. Frequency response is not an escape from time. It is a way of organizing repeated time histories by input rate.
Section 6 - Worked example: lane change as a mid-frequency handling task
Use the lane change as the first worked example because the corpus names it directly and because it is closer to a real driver task than a pure laboratory sine wave. A lane change is not a single turn. It is an input, a response, a reversal, and a second response. The car must accept the first steering demand, build yaw and lateral acceleration, then stop that build and answer the opposite direction.
Dixon's damper table lists lane change separately from slow steering, rapid steering, single bumps, brake or accelerate, smooth road, and rough road. The lane-change line shows meaningful suspension stroke and a defined damper speed and force range. That tells you the exercise is not just about the steering rack. The suspension is moving enough that body control and tire load control become part of the steering answer.
Suppose you run a small, approved lane-change style exercise on a skidpad or controlled test area. The first pass is slow. You move the wheel left and right with a long, smooth count. The car follows with little drama. Gain is high enough, lag is easy to manage, and the body has time to take a set. If you judged only this pass, you might call the car predictable.
The second pass is the same amplitude but quicker. Now the timing matters. If the car begins yawing left just as you are already preparing to steer right, the phase lag has become large enough to affect the task. If you add more left steering because the car feels late, and then the car finally rotates left as you begin the right-hand input, you have created a driver-induced stack of inputs. The car may feel like it suddenly oversteered, but the map tells a more precise story: your input frequency entered a band where the response was late, so you added amplitude to compensate for phase.
The third pass is not faster for the sake of being faster. It is a confirmation pass. You keep the same speed and amplitude, repeat the quicker input, and see whether the response repeats. If it does, you have learned something about the car. If it does not, driver variability or surface variation is still too large.
A good lane-change response has three signatures. The yaw response starts promptly enough that you do not chase it. The reversal does not require an extra pause after the first input. The body stops moving in the old direction before it compromises the new direction. A poor response has the opposite signatures: late yaw, stacked steering, and body motion that arrives after the steering command has changed.
Section 7 - Worked example: rapid steering on smooth road versus rough road
The second example is the trap that catches many drivers. You test quick steering on a smooth section and then assume the same answer will appear on a rough section. Dixon's table separates smooth road and rough road from steering events, and it shows why the assumption is dangerous. Road input can produce suspension motion and damper velocity that occupy the same time window as your steering input.
On smooth road, a small rapid steering input is mostly a steering response test. You can watch the chain from hand motion to yaw rate to lateral acceleration to roll control. If the car is late, you can reasonably investigate steering response, tire response, damping, roll control, or compliance.
On rough road, the same hand motion is no longer the same test. The suspension may already be moving because the road is driving it. Damper force may be coming from vertical velocity rather than from steering-induced roll alone. Tire load can be changing under you before the steering input has finished. If the car feels nervous or delayed, that may be a combined disturbance response.
The practical lesson is to tag your samples. If you are mapping frequency response, do your cleanest samples on the smoothest available controlled surface. Then run a separate disturbance comparison on rougher pavement. Do not blend the two and announce a single handling conclusion. A car that is clean on smooth road and poor on rough road may need a different diagnosis than a car that is poor in both places. The first may be dominated by ride and tire contact control. The second may be dominated by steering, tire, yaw, or roll response even without road disturbance.
This also explains why damper tuning language can be misleading when used loosely. A change that improves the rough-road feel may not improve the pure steering response. A change that sharpens the smooth-road steering response may make the tire less settled over a bump. Dixon's ride discussion warns that a basic ride simulation is not a complete representation of handling because it does not include the effect of stiffer springs and suspension in reducing body pitch and roll. The reverse is also true in driver practice: a pure handling impression is incomplete if you ignore how the tire is being disturbed vertically.
Section 8 - The calibration cues that tell you the map is improving
You are improving when your descriptions become frequency-specific. Early drivers say the car understeers, the car is nervous, or the car is lazy. Better drivers say the car is clean with slow hands, late in the first half of a medium-speed transition, and unsettled only when the road adds vertical input. That is a better map.
In the seat, the first cue is reduced surprise. The same input rate produces the same answer. The car may still have delay, but the delay is consistent enough that you can place the car without guessing.
The second cue is less stacked steering. If the car is late, you learn to wait for the response rather than doubling the steering input. Better still, you shape the original input so the car receives the rate it can use. That does not mean slow hands everywhere. It means matched hands: quick enough for the corner task, not so quick that you excite a response the car cannot finish cleanly.
The third cue is cleaner reversals. In linked transitions, you can feel the old yaw and roll state finish before the new one dominates. You do not need a rescue correction after the second input. If you have data, the yaw trace looks less like a delayed wobble and more like one controlled answer per steering half-cycle.
The fourth cue is better separation of causes. You stop blaming every transient on balance. You can distinguish slow-speed understeer from high-frequency lag, and you can distinguish steering lag from rough-road disturbance. That distinction matters because the fixes are different. A line or hand-timing fix may solve one. A setup or damper direction may be needed for another. A surface-specific compromise may be the honest answer.
Section 9 - Common mistakes
Mistake 1: treating steady-state balance as the whole handling map. A car can have acceptable steady-state balance and still answer poorly to a quick input. Good looks like naming the input rate before naming the balance. Say the car is balanced in a settled corner but late in a medium-frequency reversal, not simply good or bad.
Mistake 2: using too much steering amplitude. If you make the input large enough to create large roll angles, tire saturation, or big compliance effects, you have left the small-signal exercise. Good looks like a small, repeatable input that lets you compare rate without changing the whole operating condition.
Mistake 3: changing speed, brake, throttle, and steering together. That turns one response test into four overlapping tests. Good looks like holding speed and longitudinal inputs as constant as the venue safely allows, then rejecting samples where brake or throttle changed the pitch state.
Mistake 4: correcting during the sample. If your hands fix every delay as it happens, the trace shows your correction strategy more than the car. Good looks like a planned input, a short observation window, and then a correction only if safety requires it.
Mistake 5: mapping rough-road disturbance and calling it steering response. Rough road can drive suspension velocity and tire load variation independently of steering. Good looks like labeling surface condition and separating smooth-road steering samples from rough-road disturbance samples.
Mistake 6: ignoring the body because the steering wheel feels sharp. A sharp initial nose response is not enough if roll or pitch arrives late and compromises the next input. Good looks like evaluating the full chain: hands, yaw, lateral acceleration, roll or pitch control, and tire contact.
Mistake 7: asking the data to be more objective than the driver input. Dixon's driver variability question is real. If your steering trace is inconsistent, the output cannot be treated as a clean vehicle signature. Good looks like repeatable input first and interpretation second.
Section 10 - Drill: three-band steering response map
Do this only in an approved controlled environment such as a skidpad, autocross-style test area, or instructor-approved exercise space. Testing vehicle performance can be dangerous, and this drill is not for open public roads. If your event does not permit this type of exercise, use existing data from normal laps and label it as observational rather than controlled.
The drill takes three runs after warm-up. Use a constant moderate speed that leaves a large safety margin. Use small steering amplitude. The goal is not maximum lateral acceleration. The goal is a clean response map.
Run 1 is the slow band. Make six smooth left-right steering cycles with a long count, roughly one full left-right cycle every four seconds. Keep amplitude small and consistent. Success means each cycle feels the same, the car builds lateral acceleration progressively, and you do not need a correction after the final unwind.
Run 2 is the medium band. Make six cycles at about twice the rate, keeping the same amplitude. Do not add steering just because the car feels late. Success means you can identify whether yaw starts during the steering motion or after it, and whether the reversal is clean.
Run 3 is the high band. Make four smaller cycles at a quicker rate, still below any speed or amplitude that would make the car unstable. Success is not that the car feels great. Success is that you can tell whether the output gain has dropped, the lag has increased, or the body motion has become the limiting factor.
After the runs, write three sentences. Slow band: what did the car do? Medium band: what changed? High band: what became the limit? If you have data, mark steering angle, yaw rate, lateral acceleration, speed, brake, throttle, and any suspension channels. Reject any sample with a speed jump, brake overlap, throttle change, or visible surface disturbance that changes the meaning of the test.
Repeat the drill once more later in the day only if conditions are similar. The success criterion is repeatability, not bravery. You have a usable map when the same band produces the same description twice.
Section 11 - When this principle breaks down
The steering frequency response idea is powerful, but it has boundaries. The first boundary is amplitude. The bonded material points toward small-amplitude sinusoidal steering for this kind of response discussion. Once you ask for large tire forces, large roll angles, or abrupt combined inputs, the small-signal response map no longer explains everything.
The second boundary is model simplicity. Steady-state equations can ignore tire and bushing compliance and linearize roll resistance. That can be reasonable in some road-racing modeling contexts, but the omitted effects become important when the vehicle has larger roll, flexible bushings, or strong compliance steer. If the real car behaves differently from the simple model, do not force the model to win. Use the model to ask better questions.
The third boundary is ride and disturbance. A ride simulation that helps show tire discomfort trends is not a complete handling representation unless it includes the pitch and roll effects that matter in handling. Likewise, a steering response impression is not complete if it ignores bumps, rough road, braking pitch, and acceleration pitch. The car does not know which textbook chapter your input came from. It only sees forces, velocities, positions, and time.
The fourth boundary is the driver. A formal map assumes repeatable input. A real driver brings variability, anticipation, fear, correction, and adaptation. That does not make the exercise useless. It tells you to be humble. Your first job is to make the input repeatable enough that the car's answer can be seen.
The final boundary is purpose. You are not trying to turn every HPDE session into a laboratory. You are learning a language for transients. When the car feels late, ask at what frequency. When it feels nervous, ask whether the road is part of the input. When it feels balanced, ask whether that balance survives a reversal. That is the practical value of steering frequency response. It turns vague handling adjectives into a map you can drive, test, and improve.
Worked example: lane change as a mid-frequency handling task
Use the lane change as the first worked example because the bonded corpus names it directly and separates it from slow steering, rapid steering, bumps, braking or accelerating, smooth road, and rough road. A lane change is an alternating steering task: input, response, reversal, and second response. The driver should watch whether yaw starts promptly, whether the reversal is clean, and whether the body finishes the old motion before the new steering demand arrives. A good result is not maximum aggression. It is repeatable timing, consistent gain, and no extra rescue correction after the second half of the input.
Worked example: rapid steering on smooth road versus rough road
A quick steering input on smooth pavement is close to a pure steering response sample. The same hand motion on rough pavement is a combined steering and disturbance sample because the road can drive suspension stroke, damper velocity, and tire load variation inside the same time window. Treat these as different tests. If the car is clean on smooth road but unsettled on rough road, the diagnosis points toward ride, damping, and tire contact as much as steering response. If it is poor in both places, the delay or undercontrol is more likely part of the steering-yaw-roll response itself.
Common mistakes
The most common mistakes are treating steady-state balance as the whole handling map, using steering amplitude large enough to leave the small-signal range, changing brake or throttle during the sample, correcting so much that the input is no longer open-loop, blaming rough-road disturbance on steering response, ignoring late body motion because the steering wheel feels sharp, and trusting data from inconsistent driver inputs. Good work looks like small repeatable steering, labeled surface condition, constant speed and longitudinal input, and descriptions that name the frequency band before judging the car.
Drill: three-band steering response map
In an approved controlled environment, make three small-amplitude steering response runs at one moderate speed. Run six slow left-right cycles, then six medium cycles at about twice the rate, then four smaller quick cycles. The success criterion is not speed. It is repeatability and diagnosis. After each run, write one sentence describing gain, one sentence describing lag, and one sentence describing body control or disturbance. Reject samples with speed changes, brake overlap, throttle changes, or visible rough-road disturbance.
When this principle breaks down
The steering frequency response map is strongest for small, repeatable inputs. It becomes less definitive when tire saturation, large roll angles, bushing compliance, rough-road disturbance, braking pitch, acceleration pitch, or driver variability dominate the sample. In those cases, keep the frequency-response language but stop pretending it is the only explanation. Use it to ask better questions about which motion source is controlling the car.
Author Review
No quiz questions are attached to this lesson.
Sources
| # | Document | Chunk | Pages | Score | Collection |
|---|---|---|---|---|---|
| 1 | The Shock Absorber Handbook Wiley-Professional Engineering Publishing Series - 2nd edition John Dixon | 64ad1bc5-5b1d-6427-e5e1-4b5925322867 | 71 | 1 | uio_books_raw_v1 |
| 2 | Tires Suspension and Handling Second Edition Dixon John C | 273f4a0a-0338-757b-3176-7f9f84b4b071 | 501 | 1 | uio_books_raw_v1 |
| 3 | The Shock Absorber Handbook Wiley-Professional Engineering Publishing Series - 2nd edition John Dixon | 55b05470-d10c-7a99-4f4b-67b8abbe4ed3 | 139 | 1 | uio_books_raw_v1 |
| 4 | Tires Suspension and Handling Second Edition Dixon John C | 1c80aab5-0b3d-be93-e678-120497031a2c | 642 | 1 | uio_books_raw_v1 |
| 5 | The Shock Absorber Handbook Wiley-Professional Engineering Publishing Series - 2nd edition John Dixon | 04552de3-aaa9-77ee-5c50-4ff44e8a9978 | 132 | 1 | uio_books_raw_v1 |
| 6 | The Shock Absorber Handbook Wiley-Professional Engineering Publishing Series - 2nd edition John Dixon | 5fe51cac-899e-690e-e2f2-1650d42d298b | 147 | 1 | uio_books_raw_v1 |
| 7 | Racing Chassis and Suspension Design Carroll Smith | 1ac1a126-b9d2-24ff-6133-1843c3554108 | 213 | 1 | uio_books_raw_v1 |
| 8 | Tires Suspension and Handling Second Edition Dixon John C | 562d076c-1a7b-c26a-a846-bb72d168644b | 574 | 1 | uio_books_raw_v1 |
| 9 | Tires Suspension and Handling Second Edition Dixon John C | 0db56ae7-2c85-d4cd-3084-2ce0416f663a | 576 | 1 | uio_books_raw_v1 |
| 10 | Tires Suspension and Handling Second Edition Dixon John C | 6f9ed2d1-018f-8140-8455-54c40d69b43a | 14 | 1 | uio_books_raw_v1 |
| 11 | The Shock Absorber Handbook Wiley-Professional Engineering Publishing Series - 2nd edition John Dixon | e524037f-d8f0-77aa-6658-5b831877c564 | 7 | 1 | uio_books_raw_v1 |