Tune incidence as a measured tradeoff
Generated from
content/lms/race-aerodynamics/03-make-wings-and-devices-earn-their-drag/01-tune-incidence-without-pretending-it-is-free.md; edit the source file, not this page.
Source path: content/lms/race-aerodynamics/03-make-wings-and-devices-earn-their-drag/01-tune-incidence-without-pretending-it-is-free.md
Course: Engineer downforce you can actually use
Module: Make wings and devices earn their drag
Estimated duration: 55 minutes
Principle: incidence has to earn its keep
Incidence is one of the most tempting aero adjustments because it feels direct. You change the working angle of a wing or device, the car may feel more secure in a fast corner, and it is easy to tell yourself that you have simply added grip. That is the trap this lesson is built around. You are not tuning a free grip knob. You are making an aerodynamic configuration change, and the change is only worth keeping if the measured gain in the parts of the lap where the device helps is larger than the measured cost in the parts of the lap where drag, balance change, sensitivity, or loss of attachment hurts you.
The useful rule is simple: tune incidence by evidence, not by confidence. Evidence means baseline runs, one controlled change at a time, average lap and sector results, speed traces where drag matters, driver feedback on handling balance, and a deliberate return to the baseline when conditions may have drifted. McBeath describes aero configuration testing in exactly that spirit: change the configuration, collect useful information, combine lap and sector data with driver feedback, and keep the test disciplined enough that the result means something. That discipline matters more than the sophistication of the tool.
For an intermediate driver, the most important shift is mental. You are not asking whether more incidence made the car feel stuck to the road. You are asking whether the whole car completed the job faster, repeatably, and with a balance you can drive at speed. A wing can add useful load and still be the wrong setting if it takes too much speed away on the straight. A setting can feel calmer and still be slower if it masks poor minimum speed, slows the car in long sectors, or gives you confidence in one corner while costing you in two others. A setting can also work once in clean air and become a problem in traffic, because aerodynamic interactions are part of racing rather than a rare laboratory exception.
This is also why there is no universal incidence number in this lesson. The bonded sources warn against generalising competition car aerodynamics too casually. What works on one car may not work on another apparently similar car, and trial and error remain part of development at every level. That does not mean you guess. It means you make your trial narrow, reversible, measured, and honest. The skill is not knowing the magic angle. The skill is running the test so the car tells you which setting is better.
What incidence changes in practice
When you alter incidence, you change how hard the wing or device is being asked to work. The supported mechanism in the corpus is attachment. McBeath discusses wing twist as a way to keep flow across the span attached for longer, allowing more downforce before large-scale separation and stall. That tells you the useful boundary for this lesson: an incidence change is only useful while the airflow remains controlled enough for the device to keep producing useful force. Once the flow starts separating at scale, the extra angle is no longer a clean gain. You have crossed from making the device work into making it wasteful or inconsistent.
That attachment boundary is one reason visual evidence matters. The corpus points to seeing what the air is doing around competition cars, especially near wings, spoilers, diffusers, cooling intakes, and outlets. You do not need to become a professional aerodynamicist to use that idea. You need enough visual feedback to avoid lying to yourself. If tufts, flow-visualisation fluid, photos, or video show that a portion of the wing is losing attached flow after an incidence change, then a better lap time on one run may not mean the setting is robust. It may mean you found a narrow condition where the driver made it work once.
The other side of the mechanism is drag. The corpus does not give a table of incidence angles and drag coefficients, so you should not pretend it does. It does, however, make drag one of the two forces you care about in aero testing, discusses downforce and drag values against lap time in simulation, and describes indirect drag measurement through sector or lap times and speeds. For trackside work, that is enough to set the method. If you change incidence and the car gives up speed where it should be accelerating or pulling down a straight, that is part of the bill. You can accept the bill only when the cornering or braking-area benefit pays it back.
The last mechanism is interaction. A wing is not sitting in a private air stream, and a race car is not a collection of isolated aero parts. The corpus is explicit that aerodynamic interactions are a fact of life when cars are racing. That matters even when the lesson is narrowly about incidence. A setting can change how the car behaves with other devices, how it behaves near another car, and how the driver reports balance. The sibling lessons handle ground clearance, attachment before stall, front versus rear aero balance, and moving wings as system components. This lesson stays on the incidence decision, but your test notes should still record when a setting seems to be sensitive to traffic, ride state, or balance change.
The testable question
Before you touch the adjuster, write the question in one sentence. A useful question is not: does more incidence make more grip. A useful question is: does this incidence setting improve the average time through the aero-limited sector without giving back more time on the following straight or upsetting the balance enough to slow corner entry and exit. That wording forces you to measure both the gain and the cost.
Pick the part of the lap where the device should matter most. For a wing or similar device, that will usually be a faster corner, a fast transition, or a braking-to-corner phase where aerodynamic load changes the driver's confidence and balance. Then pick the part of the lap where drag cost is most likely to show up. The corpus supports indirect measurement through sector and lap times and speeds, so use that. If the data logger is simple, terminal speed at the end of a straight, sector time, and lap average are enough to begin. If the logger is better, use speed trace overlays, throttle position, and repeatability, but do not let extra channels replace the test question.
The correct question also includes the driver. McBeath explicitly pairs configuration testing with driver feedback on aerodynamic handling balance. That feedback is not a substitute for time, but it is not noise either. If the data says the setting is slightly faster but the driver reports a balance that is nervous, abrupt, or only manageable with a correction, that belongs in the decision. If the driver says the car feels planted but the sectors show no improvement and the straight is slower, that also belongs in the decision. The test is the combination, not one favourite piece of evidence.
Build the baseline before you build the change
A baseline is not just the old setting. It is the reference run that lets you know whether the change did anything. The corpus is firm on this. The Carroll Smith example described by McBeath compared two wing configurations over five laps each, changed only the wing configuration, averaged lap times, discarded abnormal high or low times, and used the data to understand handling balance and performance in track sectors. McBeath then adds the key track-day reality: if weather or track conditions change, return to the baseline periodically, because tyres and other variables can move the target.
For your next event, that means your first job is not to adjust incidence. Your first job is to make the current setting measurable. Warm the car, get tyre pressures and temperatures into the normal window for the day, and drive a clean run at a pace you can repeat. Do not use a lap where you missed an apex, got held up, had a yellow flag, or experimented with a different braking point. Your baseline is the best honest picture of the car in the condition you are actually testing.
Write down the setup state before the run. Include the current incidence setting in whatever units your hardware actually uses, the hole position or turn count if that is how the adjuster is marked, tyre set and pressure state, fuel load if it is meaningfully different, ambient and track condition if available, and any obvious traffic notes. This is not bureaucracy. It is how you avoid confusing an aero change with a tyre change, a driver improvement, or a track-temperature swing.
Data logging helps, but only when it is installed, calibrated, and used to extract useful information. The corpus includes that practical point directly. A basic logger can answer the most important questions if you use it consistently. An expensive logger can mislead you if channels are miscalibrated, if the driver does not repeat the same procedure, or if you compare one hero lap against one compromised lap. The baseline is where you prove the test system is good enough for the decision.
The incidence test loop
Run the loop in a way you can defend later. First, make a baseline run. Second, change only incidence. Third, run enough laps to separate signal from noise. Fourth, compare the average and the sector shape, not just the fastest lap. Fifth, return to baseline when conditions may have moved. Sixth, decide whether the new setting earned its drag and complexity.
The five-lap structure from the Carroll Smith example is a practical model. You do not need to treat five laps as a law, but it is a good starting count because it is long enough to reveal repeatability and short enough to fit into a club test session. If your event format gives you short sessions, use the clean laps you can get and be honest in the notes about the limitation. Do not turn one clean lap into a conclusion just because the paddock is closing.
Change only one thing. This is the easiest rule to understand and one of the easiest to break. If you change incidence and tyre pressure, or incidence and ride height, or incidence and driver line, you have created a story rather than a test. The sibling lesson on ground clearance exists because ride height can be part of aero setting. That is exactly why you do not mix it into this lesson's incidence test unless the task explicitly says you are testing the combined system.
When you make the incidence change, make it reversible. The corpus does not provide degree increments, so this lesson will not invent them. Use the smallest repeatable hardware step that your wing mount or device allows and record it. If the adjuster has holes, record the hole. If it uses a turnbuckle, record turns or flats. If it uses a shim, record the shim. If the hardware cannot be returned to the same position reliably, the first job is improving the marking or measurement method, not chasing another setting.
Then drive the comparison run with the same purpose as the baseline. Same warm-up logic, same intended pace, same shift points unless safety or traffic requires otherwise, same commitment to clean laps. The driver should not try to prove the new setting by overdriving the section where it should help. Overdriving produces exactly the abnormal high or low lap times that Smith's method discards. Your job is to reveal the car, not win the test sheet.
How to read the result
Start with the part of the lap where the setting was supposed to help. If the wing incidence change was aimed at a fast corner or fast sector, look for repeatable sector improvement there. Repeatable means the sector is better across the usable laps, not merely better once. If the best sector improves but the average does not, that is a warning. You may have found a setting that works only when the driver catches it perfectly.
Next look where drag is likely to charge you. The corpus identifies sector or lap times and speeds as valuable indirect measurements of configuration effects, and it specifically frames drag as the easier aerodynamic force to measure directly with simple tools on a long, straight, flat, smooth road. On track, your practical drag clues are speed at the end of the relevant straight, acceleration shape on the trace, and sector time through straight-heavy parts of the lap. If those numbers get worse, the incidence setting has a cost. That cost may still be acceptable, but it is no longer invisible.
Now bring in driver feedback. Ask for specific balance language, not a mood. Did the car accept throttle earlier, or did it just feel calmer while exiting slower. Did it let you hold the intended line with less steering correction, or did it add a push or a nervous rear that changed your entry speed. Did the car become easier to place in the fast section but worse in traffic. The corpus supports driver feedback on aerodynamic handling balance as part of configuration testing, so use it precisely.
Finally, check whether conditions moved. If tyres degraded, weather changed, traffic changed, or the driver simply got more comfortable, a later run can look better for reasons that are not the incidence setting. That is why the baseline return is so powerful. When you return to the original setting and the old performance comes back, your confidence in the comparison rises. When it does not, the day has shifted and the result needs caution.
A setting earns its keep when the whole evidence stack points the same direction: the target sector improves, the straight or drag-sensitive sector does not give too much away, the average lap or meaningful segment improves, the driver can repeat the balance, and a baseline return does not explain the gain away. A setting fails when it wins only one column while losing the rest.
Calibration cues: what improvement looks like
In the car, a good incidence decision feels quieter rather than merely heavier. You should be able to repeat the same fast-corner placement with fewer corrections and less steering hesitation. The car should not demand a special trick to make the lap work. If the driver has to enter slower, wait longer, or use extra correction to protect the car, the extra aero feeling may not be buying speed.
In the data, a good decision shows up as a repeatable shape. The target sector improves across usable laps. The speed trace through the target corner is cleaner or higher where the driver was previously limited. The loss on the straight, if any, is smaller than the gain in the corners that motivated the change. The result survives averaging. This is the point of discarding abnormal high and low times in the Smith-style method: you are trying to hear the car through the noise.
In lap time, improvement is not automatically the single fastest lap of the day. It is the setting that makes the useful performance easier to reproduce. For an intermediate driver, this matters because your own adaptation is part of the test. You may get faster through the day simply because you are learning the track or trusting the car. The baseline return tells you whether the setting or the driver produced the improvement.
In the paddock conversation, improvement sounds specific. The driver can say which sector improved, which straight paid the bill, and what the balance did. The crew can point to the runs, the setting marks, and the baseline comparison. If the whole explanation is that it felt better, the test is not finished.
Worked example: the five-lap wing comparison
Use the Smith-style comparison as the model for a simple incidence test. You have a current wing setting and one candidate setting. The test goal is to decide whether the candidate setting is faster over the lap and whether the improvement comes from the intended sector rather than from noise.
Start with configuration A, your current baseline. Run five laps if the session allows. Mark traffic, flags, mistakes, and any laps that are clearly abnormal. Record lap times, sector times, end-of-straight speed, and a short driver balance note. Do not change anything else.
Come in, change only incidence to configuration B, and repeat the same run plan. Use the same driver intent. If the car feels better in the fast corner, the driver still has to drive the rest of the lap normally. If the car feels worse, the driver should not rescue the test by changing line, braking point, or throttle strategy unless safety requires it.
After the run, average the usable laps. Remove laps that are obviously abnormal for a real reason, but do not remove laps merely because they make your preferred setting look bad. Compare total lap average, target sector average, straight-heavy sector average, and driver balance notes. If B gains three tenths in the target fast sector but loses four tenths on the straight-heavy sector, B did not earn its drag. If B gains two tenths in the target sector, loses almost nothing on the straight, and the driver reports a repeatable balance, B is a serious candidate.
Now return to A if conditions may have moved. This is the step that separates disciplined testing from paddock storytelling. If A returns to its earlier numbers, the comparison is stronger. If A is now slower too, tyres, weather, traffic, or driver condition may have changed. You may still learn something, but you should label the result as lower confidence.
Worked example: measuring the drag side on a straight
McBeath notes that drag can begin to be measured with surprisingly little instrumentation, and that the basic requirement is a long, straight, flat, smooth piece of road rather than a racetrack. Treat that as a concept, not an invitation to do anything unsafe or illegal. In a track environment, the same principle becomes a controlled straight-line comparison where speed and acceleration reveal the cost side of the incidence change.
Suppose your candidate incidence setting made a high-speed corner feel better, but the lap time barely moved. The next question is whether drag is eating the gain. Pick the straight or straight-heavy sector most likely to show the cost. Compare entry speed onto the straight, throttle application, shift points, and end speed. If entry speed is the same but end speed is lower with the higher-incidence setting, the cost is likely in the aero configuration rather than the corner before it. If entry speed is lower, the corner exit may be the real problem, and blaming drag alone would be too simple.
The important habit is separating where the gain happens from where the bill arrives. Incidence can make the car nicer in the fast corner and still lose the lap if the straight after it is long enough. It can also lose a little speed and still win the lap if the cornering gain is bigger. You do not decide by ideology. You decide by the trace, the sectors, and the repeatability.
Worked example: race interaction check
The MIRA statement in the corpus is short but important: aerodynamic interactions are part of racing. That means a clean-air incidence setting is not automatically a race setting. If you run alone in testing, choose the setting that wins the test. But before you call it final, ask whether the event situation exposes the car to traffic, drafting, passing, or disturbed airflow.
A practical interaction check is simple. During a session where traffic naturally occurs, keep the setting fixed and record whether the car's balance changes noticeably when following another car or being passed. Do not make risky moves to create data. Just listen when the situation happens. If the setting is excellent alone but becomes abrupt, vague, or hard to place near other cars, you have a race-use concern. That does not automatically reject the setting, but it changes the decision. The fastest qualifying setting and the most usable race setting can be different because the air around the car is different.
Drill: the incidence earns-it session
Use this drill at your next test day or open-lapping event when you have permission and time to adjust aero hardware. The count is three runs: baseline A, candidate B, and baseline A again. The target duration is one session for A, one session for B, and a short return-to-baseline check when the schedule allows. The success criterion is not choosing B. The success criterion is leaving with a defensible answer and notes that another driver or engineer could understand.
Before run one, write the test question. Name the target sector, the suspected drag-cost sector, and the exact setting marks. Run A and collect at least three clean laps if five is not possible. Record the same data channels and a driver balance note immediately after the run.
Before run two, change only incidence to B. Record the hardware mark. Run the same plan. If the car feels different, describe the difference in terms of placement, corrections, confidence, entry, mid-corner, exit, and straight speed. Avoid vague words unless you pair them with a place on the track.
Before run three, return to A. The point is to check whether the baseline still behaves like the baseline. If A now matches its earlier performance, the comparison between A and B is stronger. If A has drifted, mark the test as inconclusive or lower confidence. The drill is complete when you can say which setting improved which sector, what it cost elsewhere, whether the driver could repeat it, and whether the baseline return supported the conclusion.
Common mistakes
The first mistake is treating planted feel as proof. A car that feels calmer may be faster, but the feeling alone does not pay for drag. Good looks like a driver comment matched to sector improvement and acceptable speed cost.
The second mistake is changing two things at once. Incidence plus ride height, tyre pressure, or wing position may be a valid system test, but it is not an incidence test. Good looks like one controlled configuration change with the rest of the car held steady.
The third mistake is trusting a hero lap. The corpus-backed method averages laps and discards abnormal results for a reason. Good looks like a repeatable average across usable laps, with outliers explained by real events such as traffic or mistakes.
The fourth mistake is forgetting to return to baseline. Track conditions, weather, and tyre deterioration can move the baseline during the session. Good looks like an A-B-A structure whenever conditions are likely to have changed.
The fifth mistake is ignoring attachment evidence. If visualisation shows the flow is separating badly after the incidence change, the setting may be fragile even if one lap looked good. Good looks like matching performance data with some check on what the air is doing near the device when practical.
The sixth mistake is pretending every car wants the same answer. The corpus is clear that generalisation is difficult and that apparently similar cars can respond differently. Good looks like a local conclusion for this car, this track, this driver, this condition, and this rule set.
The seventh mistake is using tools without common sense. Simulations, data loggers, and visualisation methods can all help, but the corpus's closing advice on analysis tools is to use them carefully and sensibly. Good looks like simple questions answered cleanly, not complex plots used to defend a messy test.
When to stop adjusting
Stop when the evidence stops improving. If more incidence does not improve the target sector, if it creates a larger speed cost than it pays back, if the balance becomes harder to repeat, or if visual evidence suggests separation or stall, the setting has stopped earning its keep. Do not keep adding angle because the previous step helped. Development is trial and error, and one of the hardest skills is accepting that a trend ended.
Also stop when the test quality collapses. If traffic ruins the session, if tyres are falling away too quickly, if weather changes, if the driver is inconsistent, or if you cannot return the hardware accurately, more laps may create false confidence. Write down what you learned, return to a known safe setting, and plan a cleaner test.
Cross-references inside this module
Use the ground-clearance lesson when the incidence change seems tied to ride state or underbody behavior. Use the attachment-before-stall lesson when tufts, flow fluid, or speed traces suggest the device is no longer working cleanly. Use the front-aero-versus-rear lesson when the driver feedback is mostly about balance rather than total grip. Use the system-components lesson when moving the wing, changing endplates, or changing another device becomes part of the question.
This lesson's boundary is the incidence decision itself: what you changed, what it gained, what it cost, and whether the evidence is strong enough to keep it. The best amateur aero testing is not mystical. It is narrow, repeatable, skeptical, and practical. You make the device earn its drag, and you keep only the setting that proves it.
Worked example: the five-lap wing comparison
Use the Smith-style comparison as the model for a simple incidence test. You have a current wing setting and one candidate setting. The test goal is to decide whether the candidate setting is faster over the lap and whether the improvement comes from the intended sector rather than from noise. Start with configuration A, your current baseline. Run five laps if the session allows. Mark traffic, flags, mistakes, and any laps that are clearly abnormal. Record lap times, sector times, end-of-straight speed, and a short driver balance note. Do not change anything else. Come in, change only incidence to configuration B, and repeat the same run plan. After the run, average the usable laps, compare total lap average, target sector average, straight-heavy sector average, and driver balance notes, then return to A if conditions may have moved.
Worked example: measuring the drag side on a straight
Use a controlled straight-line comparison to read the cost side of the incidence change. Pick the straight or straight-heavy sector most likely to show drag cost. Compare entry speed, throttle application, shift points, and end speed. If entry speed is the same but end speed is lower with the candidate setting, the cost is likely in the aero configuration rather than the corner before it. If entry speed is lower, the exit may be the real problem, so do not blame drag alone.
Worked example: race interaction check
A clean-air incidence setting is not automatically a race setting. During normal traffic, record whether the car's balance changes when following another car or being passed. Do not create unsafe situations for data. If the setting is excellent alone but becomes vague, abrupt, or hard to place in disturbed air, label that as a race-use concern rather than pretending the solo test answered every question.
Drill: the incidence earns-it session
Run three controlled comparisons: baseline A, candidate B, and baseline A again. Before run one, write the test question, target sector, suspected drag-cost sector, and exact setting marks. Run A and collect at least three clean laps if five is not possible. Change only incidence to B and repeat the plan. Return to A for the third run. The success criterion is a defensible answer: which setting improved which sector, what it cost elsewhere, whether the driver could repeat it, and whether the baseline return supported the conclusion.
Common mistakes
Common errors are treating planted feel as proof, changing two things at once, trusting a hero lap, forgetting to return to baseline, ignoring attachment evidence, pretending every car wants the same answer, and using tools without common sense. Good work pairs driver feedback with sector data, holds the car steady except for the tested setting, averages usable laps, checks the baseline again, looks for attachment problems when practical, and treats the result as local to the car and conditions tested.
When to stop adjusting
Stop when more incidence no longer improves the target sector, when the straight-speed cost exceeds the cornering gain, when balance becomes harder to repeat, when visual evidence suggests separation or stall, or when the test quality has collapsed because of traffic, tyre deterioration, weather, driver inconsistency, or unreliable hardware marks.
Author Review
No quiz questions are attached to this lesson.
Sources
| # | Document | Chunk | Pages | Score | Collection |
|---|---|---|---|---|---|
| 1 | Competition Car Aerodynamics 3rd Edition McBeath Simon | c0cd0f54-6d9c-7f08-e9af-37c31b3421d3 | 345 | 1 | uio_books_raw_v1 |
| 2 | Competition Car Aerodynamics 3rd Edition McBeath Simon | 2abb3a1a-1abc-3549-8f79-9fce704061d6 | 334 | 1 | uio_books_raw_v1 |
| 3 | Competition Car Aerodynamics 3rd Edition McBeath Simon | 6edca499-2988-7702-ccc8-3d17b516edff | 385 | 1 | uio_books_raw_v1 |
| 4 | Competition Car Aerodynamics 3rd Edition McBeath Simon | 9f0edfc1-9e8c-3a96-a48d-b0d658513db3 | 385 | 1 | uio_books_raw_v1 |
| 5 | Competition Car Aerodynamics 3rd Edition McBeath Simon | c7d0125c-8080-dbcc-df83-3b96d0b84bab | 477 | 1 | uio_books_raw_v1 |
| 6 | Competition Car Aerodynamics 3rd Edition McBeath Simon | d788f877-dfdc-2c41-96e0-e6a0de38e907 | 412 | 1 | uio_books_raw_v1 |
| 7 | Competition Car Aerodynamics 3rd Edition McBeath Simon | 4b5e1aa7-14cf-aacf-908a-c47094ea7ba5 | 504 | 1 | uio_books_raw_v1 |
| 8 | Competition Car Aerodynamics 3rd Edition McBeath Simon | 09774fa8-5f4f-bd8e-8c79-d57ffe9e2cf2 | 9 | 1 | uio_books_raw_v1 |
| 9 | Tune To Win Carroll Smith | 0766022a-ad3f-df5d-e231-0251100491c7 | 5 | 1 | uio_books_raw_v1 |
| 10 | Tune To Win Carroll Smith | 7bf253a7-13f8-d33f-024c-a15c6451b4a7 | 7 | 1 | uio_books_raw_v1 |