Build a tire and brake test matrix
Generated from
content/lms/tires-and-brakes-engineering/07-validate-and-handoff-a-working-spec/01-build-a-tire-and-brake-test-matrix.md; edit the source file, not this page.
Source path: content/lms/tires-and-brakes-engineering/07-validate-and-handoff-a-working-spec/01-build-a-tire-and-brake-test-matrix.md
Course: Engineer tire and brake grip that lasts
Module: Validate and hand off a working spec
Estimated duration: 55 minutes
A tire and brake test matrix is not a spreadsheet you fill out after the interesting work is over. It is the plan that keeps the interesting work from lying to you. In tire and brake testing, almost every number you care about is easy to contaminate. Lap time moves with the driver, ambient conditions, track state, tire temperature, brake temperature, fuel load, traffic, and confidence. Brake distance and deceleration move with the tire as much as with the brake system. Pad wear changes with the number of hard stops. A tire that looks healthy around most of its circumference can still be ruined by one locked-wheel flat spot. A test matrix exists to control those moving parts tightly enough that a change in result can be traced back to a change in the car, not to a change in the day.
The rule for this lesson is simple: write the matrix around one question at a time, keep a known control in the loop, and record enough setup, condition, measurement, and driver-comment data that you can explain why each run was better, worse, or invalid. If you cannot say what changed, what stayed fixed, what you measured, and what result would make the change worth keeping, you do not yet have a test matrix. You have a paddock note with extra columns.
This lesson sits after paper testing and before the final handoff spec. The paper-test lesson asks whether the idea is worth taking to the track. The handoff-spec lesson turns the final answer into instructions someone else can use. Here, your job is the middle discipline: turn a tire or brake question into a track test that produces a defensible answer.
Start with the test question, not the part. A weak matrix begins with a part name: front pads, rear pads, rain tires, brake ducts, balance bar, fluid, tire pressure. A strong matrix begins with the behavior you are trying to validate. Do you need more maximum deceleration before lockup? Do you need the car to stop straighter? Do you need less fade after repeated stops? Do you need lower rotor or caliper temperature? Do you need a tire that gives the same pace with better feedback? Do you need to know whether a lining will survive a race distance? The bonded brake material names the track-test families clearly: brake-balance adjustment, deceleration, effectiveness, temperature measurements, and brake-lining comparisons. The matrix should not mash those into one vague test day. It should decide which question is primary, then sequence supporting runs around that question.
For an intermediate driver, the biggest jump is learning that tires and brakes cannot be separated during testing. Brake testing is always partly tire testing because tire grip affects braking. A pad change can look worse because the tire is cold. A balance change can look better because a different tire compound tolerated the load. A deceleration number can improve because the driver attacked the stop harder, not because the system made more usable torque. Puhn's brake-testing guidance explicitly warns that tire grip affects braking and that brake testing may require tire instruments and even different tires. Haney's tire-testing guidance makes the same point from the other side: lap time alone is not an absolute value because ambient changes and driver variation get mixed into it. That is why tire tests use control tires of known specification as the benchmark. Your matrix must therefore state which tires are installed, what pressures and temperatures are being tracked, and when the control configuration returns to the car.
The control is the anchor. If you are testing tires, the control is a known tire specification. If you are testing brake balance, the control is the previous balance-bar or proportioning-valve setting. If you are testing cooling ducts, the control is the current duct layout. If you are testing linings, the control is the known lining material bedded and measured before the comparison. The control run is not ceremonial. It tells you whether the day has moved under your feet. If the same driver produces different times, different comments, and different deceleration behavior on the control each time it comes back, the test is not ready for a conclusion. The answer might be that the track changed, the tire changed, the brake temperature window changed, or the driver is not repeating the input well enough. The matrix should expose that before you spend money on the wrong conclusion.
Build the matrix in five blocks. The first block is the fixed setup. This records the brake-system specification and the tire specification: balance-bar setting, proportioning-valve setting if present, lining material front and rear, brake fluid, rotor diameter and thickness, caliper type, tire size and brand, and road or weather condition. The second block is the run design: run number, control or candidate, laps, miles, number of hard stops, entry speed for the stop if you are doing stop tests, and the exact change made before the run. The third block is the measurement plan: lap time, segment time, deceleration reading, tire pressure, tire temperature, brake temperature at caliper, rotor edge, hub, and lining when those are part of the question, plus lining thickness before and after for wear questions. The fourth block is driver and observer comments. Brake performance versus temperature comments are not optional because a number without a feel report can hide the failure mode. The fifth block is the decision cell: keep, reject, rerun, or isolate with another test.
The fixed setup block matters because brake and tire results are history-dependent. If the linings are not bedded, bed them before testing. That warms the engine and chassis as well, but the main reason is to avoid comparing an unready friction pair against a ready one. Before a lining or brake-effectiveness test, record starting lining thickness. For disc-brake pads, measure in the middle and at each corner because pads can wear on a taper. For drum linings, measure corners and inside and outside edges of the center, and account for the shoe plate. The point is not to turn you into a machinist during a track day. The point is that wear per stop is a useful result only if you know beginning thickness, ending thickness, measurement location, and number of hard stops. Without that, the phrase pad life is just a guess.
The run-design block is where many amateur tests fail. Do not change brake balance, tire pressure, and pad compound in the same row unless the test question is explicitly about the combined race setup and you already know the pieces. For validation work, one row means one intentional change. If you are moving balance, keep the tire, pressure target, lining, fuel state as consistent as practical, and run condition stable. If you are testing tire A against tire B, keep the brake setting stable unless the matrix has a separate tire-plus-balance phase after the tire comparison. If you are testing cooling ducts, keep lining and balance stable, then measure temperature after the same type of stop and after the same measured cooling interval. The matrix protects you from a result that sounds fast but cannot be explained.
The measurement block must be small enough to actually use. Modern data systems can gather almost anything: accelerometers, ride-height sensors, slip-angle sensors, strain gauges, infrared tire-temperature readings, and component sensors across the car. Haney notes the irony that many teams collect more data than they have manpower to analyze. That is a warning for your matrix. More channels do not automatically make the test better. If the question is brake-balance adjustment, your core outputs are deceleration, lockup order or lockup tendency, driver confidence, and whether the car remains controllable. If the question is cooling, the core outputs are brake temperature at the defined locations and how those temperatures change after the same stop and the same cooling time. If the question is tire comparison, stopwatch, segment times, pressure, temperature, and driver comments may be enough unless you already have reliable acceleration or slip-angle data you can analyze. Choose measurements because they answer the question, not because the sensor exists.
The comment block is more technical than it sounds. Driver comments are not soft data when they are collected consistently. They tell you whether the brake pedal changed with temperature, whether the tire recovered after a slide, whether the car rotated predictably under trail braking, whether a front lockup warning arrived early enough to release, and whether the same control setup felt the same on repeated runs. Puhn's brake data sheet explicitly includes comments on each test and comments against temperature. Haney's tire-test description keeps driver feedback beside lap-time and tire data. Your matrix should force a short comment immediately after each run while the feel is fresh. A later memory from the paddock is weaker than a disciplined note taken at the stop.
The decision block prevents endless testing. Decide in advance what will count as a pass. A brake-balance change might pass only if it improves repeatable deceleration without inducing earlier flat-spot risk or driver uncertainty. A cooling change might pass only if it reduces the relevant temperature after the same stop and measured cool-down interval without causing another problem. A lining might pass only if effectiveness, temperature behavior, and wear per stop all stay inside the needed window. A tire might pass only if it beats or matches the control in the relevant segment or lap pattern while giving repeatable comments and pressures. If the matrix does not define pass, reject, and rerun before you start, the loudest anecdote usually wins.
Now build the sequence. A disciplined brake-and-tire validation day usually starts with inspection and starting data, but do not let this lesson sprawl into the full inspection lesson. For the matrix, you need enough starting state to make the runs interpretable: brake setup, tire setup, lining thickness, pressure targets, known tire set, weather and road condition, and whether the brakes have been bedded. After that, run an overall baseline. Record performance comments, deceleration if that is part of the test, tire data if it affects the question, and brake temperatures if heat is part of the question. This baseline is row zero in your matrix. If row zero is messy, the rest of the matrix is already suspect.
After baseline, run the primary question. For brake balance, begin from the known setting, record deceleration before changing balance, then adjust in controlled steps and continue until the matrix shows the best usable deceleration. Usable matters. Excessive wheel locking can flat-spot tires, and the driver may have difficulty identifying which end locked without damaging the tire. The protocol is to ease off the pedal the moment a wheel starts to lock, not to hold the pedal down until the smoke tells the story. That requires a sensitive driver and an observer who is watching for lockup. The matrix should include a lockup note and a tire-damage note because a setting that wins one stop and ruins the tires is not a winning setting.
For deceleration and effectiveness testing, repeatability is the test. Puhn advises at least three runs to verify results and warns that inconsistent results can come from brake-temperature changes or other changes. In your matrix, that means one good stop is not an answer. Use at least three comparable runs when the system is supposed to be behaving the same way. If the numbers scatter and the car truly feels the same, change the test method. If the numbers scatter and temperature changed, the scatter is a result: you are learning the system is temperature-sensitive. The matrix should show which case you are in.
For temperature testing and cooling-duct work, the matrix needs time discipline. Puhn's sample plan includes tuft tests to confirm airflow in a duct, measuring brake temperature after one stop, running the car for a measured time, measuring the cooling effect, modifying ducts, and testing cooling again. Translate that into rows. Row A is current duct after one defined stop. Row B is current duct after the defined cooling interval. Row C is modified duct after the same stop. Row D is modified duct after the same cooling interval. If the stop, cooldown, measurement location, or weather note changes without being recorded, you will not know whether the duct improved or the test moved.
For tire comparison, copy the control-tire logic from proper tire testing. A driver runs a few laps at a time, tires are changed between runs, and control tires return occasionally to validate the day. You may not have a tire-company engineer in your paddock, but you can still use the method. Do not make tire B fight a warmed-up, confident driver after tire A did the first nervous laps. Do not treat the fastest lap on a candidate as proof if the control would also have improved at that point in the day. Insert control rows. Use segment times when full-lap traffic or cool-down laps make the lap time dirty. Record pressure, temperature, and driver comments. If the control rows drift, interpret the candidate rows through that drift.
For lining comparisons, start with thickness and bedding. A lining test that does not record beginning thickness, ending thickness, number of hard stops, and where thickness was measured cannot produce wear per stop. It may still tell you whether the pedal felt better, but it cannot honestly tell you life. Also separate lining material from fluid unless the test question is a combined package. Puhn's sample plan lists alternate lining tests and alternate fluid tests as separate items. Follow that separation unless you are already in final package validation.
For long-distance validation, the matrix shifts from best single result to stability. Puhn's test plan calls for checking temperature, checking wear rate after the test, and recording driver comments and brake inspection results. That is the right lens. A setup that wins a single stop but overheats, tapers the pad, changes pedal feel, or punishes the tire over distance has not passed the race or advanced-HPDE use case. Your long-distance rows should include total laps or miles, number of hard stops, temperature checks, wear delta, driver comments, and post-run inspection notes.
For wet validation, only do it in an appropriate controlled environment. The brake-testing plan includes wetting the track, installing rain tires, and balancing brakes for wet. The point for your matrix is that wet is not a footnote to dry. It is a separate condition with different tires and different grip. If you need a wet answer, write wet rows with wet tires, wet condition notes, and their own balance target. Do not bury wet conclusions inside a dry matrix.
The driver protocol is part of the matrix because the driver is a measurement device. Warm-up matters. Tires do not have peak traction when cold, and brakes and driveline often need temperature as well. Accelerating and braking sharply where safe, riding the brakes with moderate throttle in a lower gear, or swerving when appropriate can all be used to bring components toward operating temperature, but the key matrix lesson is consistency. Record what warm-up method was used, then repeat it. Also remember that a tire can cool quickly at high speed with no load or while sitting in the pits. A run after a long stop is not the same as a run after a short stop unless the matrix marks it.
The driver must also protect the test articles. Flat spots are not just an annoyance. Tires can be lost because a wheel locks under braking or the car slides sideways, wearing one small area to the cord even though the rest of the tire appears to have tread. The person most likely to feel a flat spot during the session is often the driver, especially through the steering wheel with front bias. If the matrix includes aggressive brake-balance or deceleration work, it also needs a stop condition: release when lockup begins, abort the run if vibration suggests a flat spot, and inspect before continuing. A destroyed tire set is not better data.
Pressure, slip, load, and heat are the tire side of the same matrix. Smith's tire-performance material frames tire optimization around heat energy and momentum-loss trends with pressure, slip, load, and related variables. You do not need to reduce the whole tire model at the track, but you should respect the mechanism. If a brake test changes how much slip the tire experiences, the tire temperature and wear story changes. If a tire test changes grip, the brake-balance story changes. If pressure changes between rows, the contact patch and heat behavior may change. Record pressure and temperature because they are not decorations; they help explain why the same brake or tire did not act the same way twice.
Organization is not glamour, but it is part of testing. Puhn's brake-test chapter is blunt about bringing tools, supplies, instruments, spare parts, data sheets, and a notebook. Put the matrix, blank sheets, parts lists, and setup notes in one binder or equivalent record system. Label boxes. Assign one person to keep track of tools if you have help. The engineering reason is simple: when a test day loses a pyrometer, pressure gauge, pad thickness tool, or spare part, the matrix starts losing rows or inventing substitute methods. Bad logistics become bad data.
After the test, analyze by comparison, not by hope. First compare every candidate against the nearest valid control, not against the best memory of the morning. Second, separate clean runs from invalid runs. Traffic, abnormal driver input, unexpected temperature change, pressure drift, lockup, flat-spot vibration, or a missed measurement can all make a row unusable. Third, look for agreement between channels. A brake setup that improves deceleration, keeps comments positive, and stays inside temperature limits is stronger than a setup with one pretty number and nervous driver notes. A tire that is faster but gives inconsistent control comparisons may need a rerun. A cooling change that lowers rotor-edge temperature but raises another concern needs another question, not a victory lap.
Finally, reduce the matrix to a handoff-ready result. This lesson does not write the final handoff spec, but it must produce material the next lesson can use. Your completed matrix should be able to answer these questions cleanly: what was tested, what stayed fixed, what changed, what control was used, how many runs verified the result, what measurements supported it, what comments matched or contradicted the numbers, what failure modes appeared, and what is the recommended setting or part for the operating condition. If those answers are not visible in the matrix, the handoff will inherit uncertainty.
Worked example: adjustable balance-bar race car. Suppose the question is whether a race car with an adjustable balance bar can stop harder and straighter on the current tire. The weak plan is to send the driver out, twist the bar after each complaint, and keep the setting that felt brave. The matrix plan starts with fixed setup: current tires, pressures, lining material, brake fluid, rotor and caliper specification, and current balance setting. The brakes are bedded, starting pad thickness is recorded, and the warm-up routine is defined. Row 1 is the control: three comparable stops from the defined speed or track marker, with deceleration readings, comments, and lockup notes. Rows 2 through 4 move the balance in controlled steps. Each row records the setting, the same stop protocol, deceleration, comments, and any lockup. The driver is instructed to release as soon as a wheel begins to lock. If one setting gives the best deceleration but the driver reports sudden front lockup and the observer sees tire smoke, that row is not automatically the winner. It may be the edge that ruins tires. The keep setting is the one that gives repeatable deceleration with controllable lockup margin and consistent comments. If the best two rows are close, return to the control or repeat the top candidate rather than declaring a winner from one pass.
Worked example: control-tire benchmark comparison. Suppose the question is whether a new tire specification is worth using for an HPDE development weekend. The weak plan is to run tire A in the morning, tire B after lunch, then compare best laps. The matrix plan treats the control tire as a recurring benchmark. Run the known tire for a short stint and record lap time, segment time if available, pressures, temperatures, and comments. Change to the candidate for a similar stint. Then return to the control. If the control is now faster by the same amount as the candidate seemed faster, the day or the driver improved and the candidate did not prove much. If the control is stable and the candidate is faster in the relevant segments with repeatable pressure and temperature behavior, the candidate has evidence. If the candidate is faster once but produces different comments each run, the answer is rerun or reject depending on the risk and cost. This is the core lesson from proper tire testing: a stopwatch matters, but without a control it can flatter the wrong variable.
Common mistakes. The first mistake is changing two variables at once. What it feels like is progress because something changed. What it costs is attribution. Good looks like one intentional change per row unless the matrix is explicitly in final package validation. The second mistake is treating lap time as absolute. What it feels like is certainty because numbers look objective. What it costs is false confidence when ambient condition, driver variation, or track state moved. Good looks like control runs and segment review. The third mistake is skipping starting measurements. What it feels like is saving time. What it costs is losing wear-rate and condition evidence. Good looks like beginning and ending thickness, tire pressure and temperature, and setup state recorded before conclusions. The fourth mistake is chasing lockup until the tire tells you the answer with smoke. What it feels like is commitment. What it costs is flat spots, ruined tires, and unsafe data. Good looks like release at first lockup, observer notes, and inspection if vibration appears. The fifth mistake is collecting data nobody will analyze. What it feels like is professionalism. What it costs is distraction. Good looks like a measurement set matched to the question. The sixth mistake is leaving comments blank. What it feels like is keeping the sheet clean. What it costs is losing the mechanism behind the number. Good looks like short, immediate notes on pedal, lockup, stability, tire feel, fade, recovery, and confidence.
Drill: the three-control matrix. At your next event, do not start with a giant test. Build a nine-row matrix around one low-risk question, such as tire-pressure effect on repeated braking feel or brake-balance feel within a conservative range approved for the car. Use three control rows and six candidate rows. Run control, candidate A, control, candidate B, control, then repeat the best candidate and one rejected candidate if time allows. Each run gets the same warm-up plan, same measurement fields, and same comment prompts. The drill takes one session to plan, one to run, and one paddock break to analyze. Success is not finding speed. Success is being able to explain which rows were valid, whether the controls stayed stable, and what you would test next. If you cannot explain those three things, shrink the matrix and repeat the drill.
Cross-references. Use the paper-test lesson before this one to reject ideas that are not worth track time. Use the tire-inspection lesson before and during this one so a damaged tire does not masquerade as a setup result. Use the handoff-spec lesson after this one to turn the winning row, the control evidence, the measurement caveats, and the operating conditions into a specification another driver or crew member can follow.
Worked example: adjustable balance-bar race car
Suppose the question is whether a race car with an adjustable balance bar can stop harder and straighter on the current tire. The weak plan is to send the driver out, twist the bar after each complaint, and keep the setting that felt brave. The matrix plan starts with fixed setup: current tires, pressures, lining material, brake fluid, rotor and caliper specification, and current balance setting. The brakes are bedded, starting pad thickness is recorded, and the warm-up routine is defined. Row 1 is the control: three comparable stops from the defined speed or track marker, with deceleration readings, comments, and lockup notes. Rows 2 through 4 move the balance in controlled steps. Each row records the setting, the same stop protocol, deceleration, comments, and any lockup. The driver is instructed to release as soon as a wheel begins to lock. If one setting gives the best deceleration but the driver reports sudden front lockup and the observer sees tire smoke, that row is not automatically the winner. It may be the edge that ruins tires. The keep setting is the one that gives repeatable deceleration with controllable lockup margin and consistent comments. If the best two rows are close, return to the control or repeat the top candidate rather than declaring a winner from one pass.
Worked example: control-tire benchmark comparison
Suppose the question is whether a new tire specification is worth using for an HPDE development weekend. The weak plan is to run tire A in the morning, tire B after lunch, then compare best laps. The matrix plan treats the control tire as a recurring benchmark. Run the known tire for a short stint and record lap time, segment time if available, pressures, temperatures, and comments. Change to the candidate for a similar stint. Then return to the control. If the control is now faster by the same amount as the candidate seemed faster, the day or the driver improved and the candidate did not prove much. If the control is stable and the candidate is faster in the relevant segments with repeatable pressure and temperature behavior, the candidate has evidence. If the candidate is faster once but produces different comments each run, the answer is rerun or reject depending on the risk and cost. This is the core lesson from proper tire testing: a stopwatch matters, but without a control it can flatter the wrong variable.
Common mistakes
The first mistake is changing two variables at once. What it feels like is progress because something changed. What it costs is attribution. Good looks like one intentional change per row unless the matrix is explicitly in final package validation. The second mistake is treating lap time as absolute. What it feels like is certainty because numbers look objective. What it costs is false confidence when ambient condition, driver variation, or track state moved. Good looks like control runs and segment review. The third mistake is skipping starting measurements. What it feels like is saving time. What it costs is losing wear-rate and condition evidence. Good looks like beginning and ending thickness, tire pressure and temperature, and setup state recorded before conclusions. The fourth mistake is chasing lockup until the tire tells you the answer with smoke. What it feels like is commitment. What it costs is flat spots, ruined tires, and unsafe data. Good looks like release at first lockup, observer notes, and inspection if vibration appears. The fifth mistake is collecting data nobody will analyze. What it feels like is professionalism. What it costs is distraction. Good looks like a measurement set matched to the question. The sixth mistake is leaving comments blank. What it feels like is keeping the sheet clean. What it costs is losing the mechanism behind the number. Good looks like short, immediate notes on pedal, lockup, stability, tire feel, fade, recovery, and confidence.
Drill: the three-control matrix
At your next event, do not start with a giant test. Build a nine-row matrix around one low-risk question, such as tire-pressure effect on repeated braking feel or brake-balance feel within a conservative range approved for the car. Use three control rows and six candidate rows. Run control, candidate A, control, candidate B, control, then repeat the best candidate and one rejected candidate if time allows. Each run gets the same warm-up plan, same measurement fields, and same comment prompts. The drill takes one session to plan, one to run, and one paddock break to analyze. Success is not finding speed. Success is being able to explain which rows were valid, whether the controls stayed stable, and what you would test next. If you cannot explain those three things, shrink the matrix and repeat the drill.
When to rerun instead of deciding
Rerun when the control moved, the driver comments changed more than the setup did, brake temperature changed in a way the matrix did not plan for, tire pressure or temperature drifted outside the row definition, a lockup or flat spot compromised the tire, a measurement was missed, or the result depends on one unusually good run. A rerun is not failure. In tire and brake validation, it is often the first sign that the matrix is doing its job: protecting the handoff spec from a conclusion the evidence cannot support.
Author Review
No quiz questions are attached to this lesson.
Sources
| # | Document | Chunk | Pages | Score | Collection |
|---|---|---|---|---|---|
| 1 | Brake Handbook Fred Puhn | bae04eac-8408-07b2-538d-33051e7db6c4 | 110 | 1 | uio_books_raw_v1 |
| 2 | Brake Handbook Fred Puhn | 07dade4d-8bb3-cc02-322d-cca272a63945 | 110 | 1 | uio_books_raw_v1 |
| 3 | Brake Handbook Fred Puhn | b428d525-10b5-f995-e2e9-8a064043d69a | 115 | 1 | uio_books_raw_v1 |
| 4 | Brake Handbook Fred Puhn | d4006b5c-f33d-de59-c11f-e8b7df0a22c5 | 114 | 1 | uio_books_raw_v1 |
| 5 | The Racing and High-Performance Tire Paul Haney | 11880aec-933e-aa8f-4b04-34e8fbf40f0e | 168 | 1 | uio_books_raw_v1 |
| 6 | Race Car Engineering Mechanics Paul Van Valkenburgh | 497023f2-2fc5-86df-1857-e91fbf31f847 | 19 | 1 | uio_books_raw_v1 |
| 7 | Brake Handbook Fred Puhn | eec70339-5799-3a15-184a-c384934cec4d | 115 | 1 | uio_books_raw_v1 |
| 8 | Racing Chassis and Suspension Design Carroll Smith | eae9f9ce-0394-b6ca-6779-c954881967cc | 28 | 1 | uio_books_raw_v1 |