Keep the floor honest at running ride height

Generated from content/lms/race-aerodynamics/04-control-the-aero-platform/03-keep-underbody-assumptions-honest.md; edit the source file, not this page.

Source path: content/lms/race-aerodynamics/04-control-the-aero-platform/03-keep-underbody-assumptions-honest.md

Course: Engineer downforce you can actually use

Module: Control the aero platform

Estimated duration: 60 minutes

The skill

Keeping the floor honest means you do not treat the paddock ride height, the setup sheet, or a static aero model as proof of what the underbody is doing at speed. You make the car show you its real running attitude. Then you decide whether the floor, splitter, diffuser, undertray, or rear wing relationship is working in the height and rake window the car actually uses on track.

This is an intermediate aero skill because it sits between two simpler jobs. One sibling lesson teaches you to measure the car's running ride height. Another teaches you to define the window before you chase more downforce. This lesson is the bridge: once you can measure the car, you use those measurements to keep your underbody assumptions honest. The goal is not to become a wind tunnel engineer in the paddock. The goal is to stop making setup decisions from a car that only exists on stands.

The core rule is simple: every underbody claim must be checked at the ride height and attitude the car carries while it is accelerating, braking, cornering, and running fast in clean air. Aerodynamic balance changes with speed. It changes with pitch and rake. It changes with ride height. It can also change in yaw. On track, the car does not hold a fixed attitude long enough for a static assumption to be safe by itself. The car goes from straight running, to braking, to cornering, to accelerating again, and both the mechanical loads and the airflow are changing while that happens.

So when you say the floor works, be precise. Works at what front ride height? At what rear ride height? At what speed? In a straight line or with steering angle? With the rear wing in which position? With tires new or late-session? With the car under braking pitch, cornering roll, or exit squat? If you cannot answer those questions, the floor may still be helping, but you do not yet know where or why.

Why static aero assumptions drift away from the real car

Static testing and simulations are useful, but they freeze a car that is not frozen on track. The bonded aero material is explicit about this limitation: downforce and balance are affected by yaw, pitch, rake change, and ride height, while the real car's attitude is transient. Even if a wind tunnel or CFD run teaches you something true about a given ride height, the on-track car may spend very little time at that exact attitude. The useful question becomes whether the real car repeatedly passes through a workable window, not whether one static height looked good in isolation.

For the underbody, this matters because the floor lives close to the ground. A small front or rear height change can alter local pressure, effective rake, diffuser behavior, and the balance relationship between the floor and the wings. The supplied corpus does not give a specific stall height for a particular floor, and you should not invent one. What it does support is the method: measure the car under speed load, observe the local flow or pressure where you can, and compare performance changes with disciplined testing.

The other reason static assumptions drift is that aero load and mechanical grip coexist. Carroll Smith's suspension material cautions against pretending suspension is unimportant just because aerodynamic grip exists. The tire is still the contact path for acceleration, braking, steering, and driver feel. At the apex of many racing corners, aero download may be secondary to mechanical grip, and aero load is additive to the mechanical platform rather than a replacement for it. That means a floor test must separate low-speed mechanical behavior from higher-speed aero behavior. If a change makes the car worse in a slow corner but better in a fast one, the floor may be helping while the mechanical platform or balance compromise is hurting somewhere else. If you do not separate those cases, you will blame the wrong thing.

The mechanism you are trying to prove

A floor, undertray, splitter, diffuser, or wing-underbody interaction creates useful evidence in several ways. First, it can change suspension deflection at speed. More downforce pushes the car down through the suspension, so front and rear damper travel can become a practical proxy for aerodynamic load when it is logged at constant speed and calibrated back in the garage with known mass on each axle. Track surface noise must be filtered, but a reasonable assessment of download through the suspension can be obtained.

Second, the underbody can show its pressure behavior locally. The corpus notes that mapping static pressure over many locations takes many runs and can be disturbed by environmental fluctuation, so for a club or track-day program the better use is often localized pressure profiles. That may mean pressure taps or surface-mounted pressure transducers on an underbody area you are trying to understand. The key is restraint. You are not trying to map the whole car in one afternoon. You are trying to answer one local question: did this floor or wing-position change alter the pressure in the area where you expected it to?

Third, the air can be made visible. Tuft testing is simple, cheap, and track usable. Small pieces of yarn attached to the surface tell you whether local flow is attached or separated. When attached, tufts align in neat rows with only a little flutter. When separated, they whirl and point randomly. For a floor-honesty lesson, tufts can be useful around the diffuser exit, cooling outlets, wheel openings, spoiler edges, and other visible surfaces that interact with underbody flow. You cannot see through the floor with yarn, but you can often see whether the flow leaving or feeding an aero device behaves consistently.

Fourth, the stopwatch and driver can still help if the test is disciplined. Lap times, sector times, higher-speed corner entry, apex and exit speeds, straight-line speeds, and driver feedback all belong in the evidence set. The danger is not using lap time. The danger is using one lap time from a sloppy test and calling it aero truth. The bonded material points to a practical Carroll Smith method: run one configuration for five laps, change only the wing or aero configuration, run the other configuration for five laps, average the laps, discard abnormal high or low outliers, and return periodically to baseline because weather, track condition, and tire deterioration can move the reference under your feet.

The technique: make the floor prove itself

Start by stating the assumption you are testing in a way that can be wrong. A weak assumption is that lower is better. A useful assumption is that the car gains front and rear downforce when the front ride height is reduced by a small amount, without losing straight-line speed enough to hurt sector time, and without making high-speed balance worse. Another useful assumption is that the rear wing position should improve underbody pressure rather than simply add rear load. A useful assumption has a place, a speed range, an expected direction, and a way to be disproved.

Next, define the baseline. Record static front and rear ride height, wing settings, splitter or undertray configuration, tire pressures, session timing, fuel state if you are tracking it, and the driver feedback from the previous run. If you have damper potentiometers, zero and check them. If you are doing pressure work, check that the ports or transducers are attached and reading. If you are doing tuft work, mount the tufts so they can be seen on video and will not peel off. The lesson is not about expensive instrumentation; it is about disciplined measurement. Low-cost testing can produce useful results, but careless testing is worthless.

Then run the car in conditions that answer the question. If you are measuring downforce from suspension deflection, you need an adequately long, flat, smooth straight where the car can accelerate to and hold a reasonably high constant speed safely. You log front and rear suspension deflection at that speed, then calibrate the deflection back in the garage by loading known mass on each axle through the travel range you are using. If the wheel rate is non-linear, calibration through the range matters because the same millimeter of movement may not represent the same force everywhere in travel.

If you are comparing configurations with lap and sector data, keep the change isolated. Do not change ride height, wing angle, tire pressure target, brake bias, and line all at once. Run the baseline. Run the change. Return to the baseline. Use enough laps that one traffic lap or one overdriven corner does not dominate the result. The Carroll Smith example in the corpus used five-lap configuration blocks and discarded abnormal outliers. That is not a magic statistical law, but it is a practical habit for a club program: make the result survive more than one lap.

If you are using corners as evidence, split the track in your head. Low-speed corners tell you more about the mechanical platform and tire use. Higher-speed corners are where aero effects become easier to see, though the exact threshold depends on downforce level. The McBeath material gives lap time, sector time, straight-line speed, and higher-speed corner speeds above roughly 60 mph or 100 km/h as useful traditional testing parameters. Use that as a practical starting point, not a rigid boundary. The more modest the downforce, the more careful you must be before assigning a change to aero rather than driver, tires, or track condition.

Finally, write the conclusion in conditional language. Do not write that the floor is better. Write that with this ride height, this wing setting, and this session condition, the car showed more front compression at speed, held a repeatable high-speed sector gain, did not lose unacceptable straight speed, and the driver reported a balance change in the same part of the lap. That is an honest conclusion. It tells the next engineer or driver what was actually proven.

Sub-skill 1: translating setup ride height into running ride height

Your first sub-skill is to stop treating static ride height as the aero ride height. The setup pad tells you where the car begins. The track tells you where it runs. Under speed, downforce compresses the suspension. Under braking, pitch changes the front and rear heights differently. Under acceleration, the platform changes again. In cornering, roll and yaw can change the underbody's relationship to the track. The car's floor sees all of that.

The practical move is to log front and rear suspension movement and convert it into ride-height movement with calibration. You are not just watching travel because travel is interesting. You are answering whether the floor still sits in the window you thought you had. If the front is lower than expected on the straight, or the rear is lower than expected through a fast corner, your paddock rake number may be misleading. If the car compresses into a range where your flow visualization looks worse or the pressure response changes in the wrong direction, the floor assumption is not honest yet.

This also protects you from chasing a setup that only exists for a moment. A configuration can look promising in static measurement and still spend the lap outside its useful window. Conversely, a configuration that looks conservative in the paddock may carry the right running attitude once aero load builds. The lesson is not lower or higher. The lesson is measure the running condition before naming the floor good or bad.

Sub-skill 2: separating front and rear aero sensitivity

The underbody rarely changes only total grip. It can move balance. McBeath's material emphasizes that aerodynamic balance can be markedly affected by pitch, ride height, yaw, and speed. For the driver, that shows up as a balance shift: the car may be more secure on entry, more nervous in fast transitions, more planted on exit, or slower on the straight. For the tester, it means front and rear evidence must be kept separate.

Use front and rear damper traces separately when you can. If a change compresses the rear more than the front, the balance story is different from a change that compresses the front more than the rear. Use corner speeds and driver feedback the same way. A higher fast-corner entry speed with a driver report of stable rear support is not the same finding as a higher entry speed with the driver saying the rear was barely hanging on. Both may be fast for one lap. Only one may be a trustworthy floor platform for a race stint.

This is also why low-speed corner comparison matters. If the car loses time only in slow corners after an underbody change, the aero may not be the main problem. The change may have forced a ride-height, spring, or balance compromise that hurt the mechanical car. If the car gains in fast corners and loses on straights, you may have bought downforce with drag. If it gains nowhere except the driver's confidence, that confidence is still useful feedback, but it is not enough to call the floor proven.

Sub-skill 3: using local pressure without pretending you mapped the whole car

Pressure measurement can be powerful, but the corpus warns that measuring static pressure at many locations over a racecar takes many runs and is vulnerable to environmental changes. That is the reason to use pressure locally and with a narrow question. You might compare pressure at a few underbody points before and after a rear wing location change, because the source material notes that rear wing presence or location can affect underbody pressures. You might compare a diffuser-area pressure profile before and after a ride-height change. You might check whether a suspected improvement appears at the surface you expected.

The honest pressure question is small. Did this change move this local pressure profile in the expected direction at the tested speed and attitude? If yes, you have supporting evidence. If no, you have learned that the story is more complicated than the setup sheet. Either result is useful. The error is to take a few pressure points from one run and declare the whole underbody solved.

Sub-skill 4: using airflow visualization as a diagnosis, not decoration

Tuft testing gives you a fast way to see flow behavior, especially around visible surfaces related to the floor: diffuser exits, rear bodywork, spoiler regions, cooling exits, and external devices that may affect underbody flow. The bonded material describes the basic interpretation clearly. Aligned tufts with only a little flutter suggest attached flow. Random swirling tufts suggest separated flow.

For this lesson, use tufts to answer timing and location questions. Does separation appear only at a higher speed? Does it appear after a ride-height change? Does it show up at the diffuser exit when you alter rear wing location? Does it improve when you return to baseline? Video makes the result reviewable, which matters because the driver cannot watch the rear of the car while driving at speed. Keep the test simple enough that the camera view and the tuft field answer one question.

Do not overread tufts. They show local surface flow, not total downforce. A tidy tuft field is not automatically a fast car. A messy tuft field is not automatically a failed floor. Tufts are one witness. Damper movement, pressure, speeds, and driver balance complete the picture.

Sub-skill 5: running an A/B test that survives the afternoon

The most important discipline is baseline control. The corpus is blunt about why. Conditions change. Tire deterioration changes the baseline. Weather and track state move underneath you. If you run baseline at 9 a.m., make three changes, then compare the last run of the day to the first, you may be measuring tires and track more than aero.

A useful A/B pattern is baseline, change, baseline. If time allows, add a second change run. Use the same driver task. Use the same target laps. Use sector data and straight speeds. Discard laps that are obviously abnormal from traffic, mistakes, or other interruptions. Record driver balance feedback immediately. The discipline is not glamorous, but it is what keeps the floor honest.

This is where many club programs lose the thread. They have enough data logger capability to gather useful information, but they do not run the test in a way that lets the information mean anything. The logger is not the method. The method is controlled comparison.

Calibration cues

You are improving when your conclusions become narrower and more repeatable. Early in the process, the note may be vague: the car felt better in fast corners. As the skill improves, the note becomes testable: at the same wing setting and tire state, the lower front ride height produced repeatable extra front compression on the straight, the high-speed sector improved across the average of valid laps, the straight-line speed loss was measurable but acceptable, and returning to baseline returned the balance trend.

A good damper-travel signature is not perfectly smooth, because tracks are not perfectly smooth. But after filtering surface noise, the speed-related compression should make sense and should repeat at comparable speeds. If the front and rear traces jump around in ways that do not correlate with speed, track surface, or driver input, you do not yet have a clean aero measurement.

A good sector signature separates where the change helped and where it hurt. If the fast sector improves but slow corners do not, that may be an aero gain with a mechanical cost. If the straight speed falls and high-speed corner speed rises, you may be seeing a drag versus downforce trade. If lap time improves only because one braking zone was unusually good, you have not proven the floor.

A good driver-feedback signature names a place and a balance. Better is too broad. Better support in the fast right, more entry understeer above 70 mph, or rear secure on the throttle in the high-speed exit is more useful. The driver's body is part of the measurement system because the tires deliver most sensory information through the car. But the driver's feel has to be tied to speed, location, and configuration to become evidence.

A good tuft or pressure signature repeats when the setup repeats. If you return to baseline and the tuft field or pressure profile returns with it, your confidence rises. If it does not, look for wind, temperature, mounting, tire state, or driver-run variation before you build a theory around one image.

Failure modes and recovery

The first failure mode is the paddock-height myth. The car looks right on stands, so you assume the floor is right at speed. Recovery is straightforward: measure suspension deflection at speed and translate it into running attitude. Until you do that, the setup sheet is only the starting condition.

The second failure mode is the one-change-in-name-only test. You call it a floor test, but you also change wing angle, tire pressures, ride height, and driving target. Recovery is to isolate the variable. If you must make a support change for safety or clearance, write it down and treat the result as a package test, not a pure floor result.

The third failure mode is mistaking tire deterioration for aero balance. The corpus specifically warns that tire deterioration can change the baseline during a session. Recovery is to return to baseline periodically and compare valid laps rather than relying on first-run versus last-run impressions.

The fourth failure mode is overvaluing a single fast lap. A single lap can be driver execution, traffic luck, temperature, or a tow. Recovery is to use blocks of laps, average the valid ones, and discard obvious abnormal laps. The goal is a repeatable tendency, not a souvenir lap.

The fifth failure mode is reading low-speed behavior as underbody proof. The car may push in a slow corner because of mechanical balance, not because the floor failed. Recovery is to compare low and higher speed corners separately. If the same change behaves differently with speed, you have learned something useful about aero versus mechanical grip.

The sixth failure mode is pressure-map overreach. A few points do not map the whole car. Recovery is to treat pressure as local evidence and combine it with ride-height, speed, sector, and driver data.

The seventh failure mode is copying another car's floor answer. The McBeath material explicitly values trying ideas on your own car rather than merely copying others. Recovery is not to ignore other cars, but to test whether the idea works on your car, with your ride heights, your wing relationship, and your track-speed profile.

How this connects to the rest of the module

Measure the car's running ride height gives you the raw platform measurement. Map front and rear aero sensitivity at speed teaches you to separate balance changes by axle and speed. Separate the setup sheet from the car at speed is the broader habit this lesson applies to the floor. Define the window before you chase more downforce gives you the guardrails: before adding load, know the ride-height and balance range where the car behaves. This lesson says that the floor earns belief only inside that measured window.

The practical standard is modest but strict. You do not need a wind tunnel. The supplied material emphasizes that useful aero testing can be done on track or on suitable roads with low-cost methods, but only when performed carefully and rigorously. That is the heart of the skill: make the underbody answer one clear question at a time, at real running ride height, with enough repeatability that you would be willing to reverse the change if the evidence tells you to.

Worked example: five-lap A/B floor-height check on a mixed-speed circuit

Start with a car that has a stable mechanical setup and a known baseline aero configuration. Your assumption is that a small front ride-height reduction will help the floor and splitter package in higher-speed corners without costing too much straight-line speed or creating an unacceptable balance shift. You are not trying to prove that lower is always better. You are testing whether this car, at this track, in this session condition, improves in the speed range where aero should matter.

Run five laps at baseline. Record lap time, sector time, straight-line speed, and higher-speed corner entry, apex, and exit speeds. Ask the driver for balance comments tied to corner phase and speed. Do not accept a general note that the car felt good. You want where the car changed: entry support, mid-corner stability, exit traction, straight speed, or nervousness.

Make only the planned ride-height change. If the change requires a related adjustment for clearance or safety, record that and stop pretending the result is pure ride height. Run another five-lap block. Discard laps that are clearly abnormal because of traffic, a mistake, or an interruption. Compare the valid averages, not the single best lap.

Then return to baseline. This is the move that keeps the result honest. If the baseline behavior returns, your confidence grows. If the baseline does not return, conditions have moved or something else changed. Tire deterioration, weather, and track state can all move the reference. In that case, the correct conclusion may be that the test was inconclusive, not that the floor worked.

Interpret the result by speed band. If the higher-speed sector improves and straight speed falls slightly, you may have bought useful downforce with manageable drag. If the slow sector gets worse while fast corners improve, you may have found an aero gain paired with a mechanical compromise. If everything gets worse, revert. If the result is mixed and the driver reports an unstable balance in the fast corners, do not chase the average lap time blindly. A floor setting that is quick only when the driver is holding their breath is not an honest race setup.

Worked example: straight-line downforce run with damper travel and local pressure checks

Use this when your main question is whether the floor or wing-underbody relationship is producing the load you expect at running ride height. You need a long, flat, smooth straight where the car can safely run at a reasonably high constant speed. You also need front and rear suspension deflection measurement, ideally linear potentiometers on the dampers, and a garage calibration method using known mass on each axle.

Before the run, record static ride height and setup. Warm the car consistently, then make constant-speed passes at the target speed. The raw damper traces will include track noise, so you do not read every bump as an aero event. Filter the irregularity and look for repeatable compression at comparable speed. After the run, calibrate the deflection back to load with known axle mass through the relevant range of travel. If the wheel rate is non-linear, do not assume one simple conversion across the whole range.

Add local pressure only if it answers a narrow question. For example, if you changed rear wing location and expect it to alter underbody pressure, place a small number of pressure measurements in the underbody region you are studying. The corpus supports this kind of localized use and specifically notes that rear wing presence or location can affect underbody pressures. Avoid the trap of trying to map the whole car in one test day.

The honest conclusion might be that the revised wing location increased rear suspension compression at speed and changed the local underbody pressure profile, but also reduced straight-line speed. That is useful. It is not a final setup answer by itself. It tells you the change has a real aero effect, and now you decide whether the lap-time, sector, balance, and drag trade are worth it.

Common mistakes: what wrong looks like and what good looks like

Mistake 1: believing the setup pad. Wrong looks like announcing the floor is in its window because the static ride height is correct. Good looks like measuring how much the front and rear move at speed, then judging the floor at the actual running attitude.

Mistake 2: changing too many things. Wrong looks like lowering the car, adding wing, changing pressures, and driving harder, then calling the result an underbody gain. Good looks like one deliberate change, a recorded baseline, a repeatable run plan, and a baseline return.

Mistake 3: using the best lap as proof. Wrong looks like celebrating one quick lap while ignoring four ordinary ones. Good looks like valid-lap averages, abnormal lap removal, and sector analysis.

Mistake 4: forgetting tire deterioration. Wrong looks like comparing the first baseline run of the day to a late-session change and blaming aero for the difference. Good looks like returning to baseline periodically and writing down tire state and session condition.

Mistake 5: diagnosing aero from slow corners only. Wrong looks like rejecting a floor change because the car was worse in a hairpin. Good looks like separating low-speed mechanical behavior from higher-speed aero-sensitive corners, then deciding whether the total package is worth keeping.

Mistake 6: treating tufts as total downforce. Wrong looks like seeing tidy yarn and declaring the floor fast. Good looks like using tuft behavior as local flow evidence, then checking it against damper movement, pressure, speed, sector time, and driver balance.

Mistake 7: overclaiming pressure data. Wrong looks like a few taps becoming a full-car aero map. Good looks like a narrow pressure question, repeated conditions, and a conclusion limited to the measured area.

Mistake 8: copying the paddock winner without testing. Wrong looks like installing a popular floor or wing arrangement and assuming it transfers. Good looks like using other cars as idea sources, then proving the idea on your own car under your own ride-height, tire, and track conditions.

Drill: the baseline-return floor honesty drill

Do this over one event day with three short evidence blocks. The count is three blocks: baseline, change, baseline return. The duration is one session if your run group and paddock time allow it, or two sessions if you need more time to make the change safely. The success criterion is not a faster lap. The success criterion is a written conclusion that names the tested configuration, the running ride-height or damper-travel evidence, the speed range affected, and whether the baseline return confirmed the trend.

Block 1 is baseline. Run five clean laps if traffic allows. Capture lap and sector times, straight speed, and higher-speed corner speeds. If you have damper travel, mark the constant-speed portions. If you have tufts, video the target area. The driver gives balance notes by corner phase.

Block 2 is the single change. Make one underbody-related change: a ride-height adjustment, an undertray change, a wing-location change that you expect to affect underbody pressure, or another specific aero configuration. Run the same five-lap plan. Use the same driver target. Do not chase lap time with a different driving style.

Block 3 is baseline return. Revert the change and repeat the same evidence capture. This is the part you will be tempted to skip because time is short. Do not skip it if the day is meant to teach you something. If the trend disappears when you return to baseline, you have a usable result. If it does not, the test is compromised or the condition changed.

After the drill, write one paragraph. It should say what changed, where the car ran differently, what the measurements showed, and what you will do next. If the paragraph contains only feelings and no condition, speed, or baseline return, repeat the drill with a narrower question.

When this principle breaks down

The principle does not break down because measurement is optional. It breaks down when the test environment cannot answer the question honestly. If the straight is too short, rough, crowned, or unsafe for constant-speed downforce measurement, do not pretend the damper trace is clean aero data. If traffic ruins every comparison lap, do not average noise. If weather, track condition, or tire state changes faster than you can return to baseline, write the test as inconclusive.

It also breaks down when the corpus does not support the specific claim you want to make. The bonded material supports testing method, transient platform caution, suspension-deflection measurement, localized pressure measurement, tuft visualization, and disciplined baseline comparison. It does not provide a universal ride-height target, a universal diffuser stall point, or a named car's floor map. Those must come from your car, your measurements, or a future corpus bond. Refusing to overclaim is part of keeping the floor honest.

Author Review

No quiz questions are attached to this lesson.

Sources

#	Document	Chunk	Pages	Score	Collection
1	Competition Car Aerodynamics 3rd Edition McBeath Simon	da0eb061-4403-af2d-783d-5b7d9ae16326	476	1	uio_books_raw_v1
2	Competition Car Aerodynamics 3rd Edition McBeath Simon	4adf8cb4-89c7-1b45-bd4d-9bb03634ecf3	345	1	uio_books_raw_v1
3	Competition Car Aerodynamics 3rd Edition McBeath Simon	0278f848-5839-0a5b-9776-f3dabc163310	350	1	uio_books_raw_v1
4	uio julian edgar car aero testing	06a78fd3-5941-96b8-badf-a5ccb37cf731	6	1	uio_books_raw_v1
5	Racing Chassis and Suspension Design Carroll Smith	148524fa-62af-201e-6dff-3b729c84477a	8	1	uio_books_raw_v1
6	Competition Car Aerodynamics 3rd Edition McBeath Simon	2abb3a1a-1abc-3549-8f79-9fce704061d6	334	1	uio_books_raw_v1
7	Competition Car Aerodynamics 3rd Edition McBeath Simon	c0cd0f54-6d9c-7f08-e9af-37c31b3421d3	345	1	uio_books_raw_v1
8	uio julian edgar car aero testing	ca286b28-e538-f94e-ae4f-aabf087533dd	93	1	uio_books_raw_v1