At the end of Part 1, I had managed to turn what started as an argument about braided brake hose into a much larger question:

Where does the brake pedal actually go?

That sounds simple until you try to answer it. A hydraulic brake system contains fluid that compresses, tubing that expands, hoses that expand a little more, pads that compress, calipers that open, backing plates that bend, master cylinders that deform, and pedal structures that move.

Then, just to make things interesting, the car starts moving. Now we can add rotor runout, wheel-bearing deflection, hub deflection, seal rollback and pad knock-back. Those consume pedal travel too, although, as we established in Part 1, they aren’t really compliance. They’re clearance recovery. The brake pedal remains blissfully unaware of this distinction. It just moves.

So this time we’re going to stop talking about all of this qualitatively and actually put some numbers on the worksheet. And before anyone gets too emotionally attached to any particular number, an important warning:

This is not an attempt to define the universal brake system. Because there isn’t one. The objective is to establish orders of magnitude. If one mechanism costs us 0.1 mm of pedal travel and another costs 5 mm, it isn’t terribly important whether the first number is actually 0.08 or 0.13 mm. We already know which one deserves our attention. That’s the entire point of a compliance budget.

First, We Need a Brake System

Let’s invent a reasonably plausible race-car brake circuit. We’ll use:

• - 0.750-inch master-cylinder bore

• - 5:1 pedal ratio

• - 1,500 psi hydraulic pressure during a - - hard brake application

• - roughly 12 feet of 3/16-inch steel tubing in the complete system

• - short flexible connections at the wheels

• - conventional four-wheel disc brakes

I’m intentionally not modeling a particular car. We’re building a yardstick.

A 0.750-inch master cylinder has a piston area of about 285 mm². That gives us a wonderfully useful conversion. Move about 0.285 cc of additional brake fluid into the system and the master-cylinder piston has to move approximately 1 mm.

With our 5:1 pedal ratio, that becomes approximately: 5 mm of pedal travel for every 0.285 cc of hydraulic volume consumed.

Or, turned around: 0.057 cc of volume consumption costs about 1 mm at the pedal. Suddenly very small volume changes don’t look quite so small. That’s our conversion factor for almost everything that follows.

Steel Brake Line

We might as well begin with the argument that started this entire mess. Steel tubing does expand under pressure. It has to. There is no such thing as an infinitely rigid pressure vessel.

For thin-wall tubing, hoop stress is approximately:

σₕ = Pr/t

where:

• P = internal pressure

• r = internal radius

• t = wall thickness

From that stress we can estimate circumferential strain using the elastic modulus of steel and account for Poisson effects. From the resulting dimensional change we can calculate the increase in tube volume. Run the numbers for roughly twelve feet of ordinary 3/16-inch steel brake tubing at 1,500 psi and the resulting hydraulic volume increase is tiny. Depending on wall thickness and exact assumptions, we land in the neighborhood of only several thousandths of a cubic centimeter.

Put that back through our 0.750-inch master cylinder and 5:1 pedal ratio and we get roughly:

Steel hard-line expansion: ~0.1 mm of pedal travel.

That number is remarkably close to what I arrived at when I first started this exercise. And it remains one of my favorite numbers in the entire project because it establishes scale.

Steel tubing absolutely expands. The pedal absolutely moves because of it. But we’re talking about something on the order of a tenth of a millimeter.

Keep that number in your pocket. We’re going to need it.

Brake Fluid Isn’t Incompressible

Here’s where things start getting more interesting. We usually treat brake fluid as incompressible because, for most hydraulic calculations, that’s an excellent approximation. But we’re no longer doing most hydraulic calculations. We’re hunting fractions of a cubic centimeter. The relevant relationship is:

ΔV = V ΔP / K

where:

• V = hydraulic fluid volume

• ΔP = pressure rise

• K = bulk modulus of the fluid

Published brake-system models use effective brake-fluid bulk moduli in roughly the 1.7–2.0 GPa neighborhood; one experimental SAE study reported about 1.95 GPa in its fluid model.

Let’s use 1.9 GPa as a reasonable illustrative value. Now we need to know how much fluid we’re compressing. This is where brake-system architecture matters enormously. A compact formula car may contain surprisingly little hydraulic volume. A production-based race car with long plumbing runs, ABS hardware, large calipers and generous internal passages may contain considerably more.

Let’s use 40 cc of pressurized fluid volume for our example circuit.At 1,500 psi, or about 10.3 MPa:

ΔV ≈ 40 cc × 10.3 MPa / 1900 MPa

That works out to approximately:

0.217 cc

Now send that through our master cylinder.

0.217 / 0.285 ≈ 0.76 mm master-cylinder movement.

Multiply by our 5:1 pedal ratio:

Brake-fluid compression: ~3.8 mm of pedal travel

Wel…that escalated quickly. Remember our steel tubing? About 0.1 mm. Fluid compression in this illustrative system. Nearly 4 mm. And this leads to an interesting observation from Part 1.

Changing from 3/16-inch hard line to smaller-bore -3 PTFE hose can actually reduce the amount of brake fluid contained in that portion of the system. Goodridge, for example, lists its -3 600-series brake hose with a 3.5 mm bore. So replacing hard line with PTFE hose produces two competing effects: The hose wall becomes more compliant. But the fluid volume may decrease. One moves the pedal in the wrong direction. The other moves it in the right direction. And as always, engineering remains deeply inconsiderate of simple arguments.

Now Let’s Talk About Hose (Not hoes)

Flexible brake hose unquestionably expands more than steel tubing. This is not controversial. The interesting question is how much pedal travel does that expansion buy us? SAE J1401 actually specifies a volumetric-expansion test for hydraulic brake hose, measuring the increase in hose volume under pressure.

Older published limits are useful simply for establishing scale. For small brake hose at 1,500 psi, allowable expansion values can be measured in tenths of a cubic centimeter per foot, with low-expansion hose permitted substantially less than ordinary hose. That sounds enormous compared with steel tubing. And it is. But we have to be careful. A race car generally doesn’t contain twelve feet of flexible hose at each wheel. It may contain only a few feet total, particularly if flexible hose is used primarily to accommodate suspension and steering movement.

Modern PTFE/stainless hose is specifically intended to limit this expansion; Goodridge describes its 600-series construction as PTFE lined with stainless braid and intended to provide consistent pedal feel under pressure.

Unfortunately, detailed pressure-versus-volume curves for racing PTFE hose aren’t commonly published.So rather than pretend we know the exact answer, let’s bound it. Suppose our race car has four 18-inch flexible hoses. That’s six feet total. If a high-quality PTFE hose consumed, for the sake of illustration, 0.01 cc per foot at our operating pressure, we’d add:

0.06 cc total.

That translates to:

0.06 / 0.285 = 0.21 mm master travel

or roughly:

~1.05 mm at the brake pedal

If the hose were twice as compliant:

~2.1 mm

If it were half as compliant:

~0.5 mm

This is the first place where I want an actual hose sample on a test fixture rather than another spreadsheet. And that is probably where this project is ultimately headed.

But even with our intentionally broad estimate, something useful is emerging. Braided hose isn’t imaginary. Its contribution can absolutely be perceptible. But it also doesn’t automatically dominate the system. This may seem counterintuitive to some readers given the raised hackles I often observe when discussing flexible brake hose in the paddock. but in our example, the brake fluid itself may already consume several times as much pedal travel.

Then We Arrive at the Pads

This is where the spreadsheet starts fighting back. Brake-pad compressibility is sufficiently important that SAE has an entire test procedure devoted to measuring friction-material and pad-assembly deflection. SAE J2468 specifically identifies compressive response as relevant to brake-fluid volume consumption and pedal travel. And the phrase pad assembly matters. We aren’t merely squashing friction material. We’re loading:

• - friction material

• - backing plate

• - shims

• - piston contact areas

• - pad geometry

Research into brake-pad compressibility specifically notes that measured deflection depends on backing-plate thickness, pad shape, piston position and loading geometry in addition to the friction material itself. This makes assigning a universal number somewhere between difficult and dishonest. But let’s explore the sensitivity.

Suppose each opposed pair of pads shortens by an effective combined 0.05 mm under hard braking. That doesn’t sound like much. But the caliper pistons must move far enough to accommodate it. Now imagine a four-piston caliper with a total hydraulic piston area on one side of roughly 2,500 mm².

A 0.05 mm displacement consumes:

2,500 × 0.05 = 125 mm³

or:

0.125 cc

per caliper.

Two front calipers:

0.250 cc

And suddenly our master cylinder must move:

0.250 / 0.285 = 0.88 mm.

And at the pedal:

4.4 mm

And recall, that’s from only fifty microns of effective pad-stack compression. If the real displacement is 0.10 mm? Double it. Now we’re approaching:

9 mm of pedal travel

That’s why pad compressibility deserves considerably more respect in this conversation than it usually receives. The displacement is tiny. The loaded area is enormous. Volume is area times distance. The brake pedal does the necessary accounting.

Caliper Stiffness

Now take that same logic and apply it to the caliper. When hydraulic pressure generates clamp force, the caliper wants to open. A stiff monoblock caliper opens very little. A less rigid design opens more. But once again, the relevant quantity isn’t simply caliper deflection. It’s the hydraulic volume required to follow that deflection.

Suppose the loaded caliper structure effectively opens by just 0.04 mm. Using that same approximate 2,500 mm² piston area gives:

2,500 × 0.04 = 100 mm³

or:

0.100 cc per caliper

Two front calipers:

0.200 cc

Translated through our master cylinder and pedal ratio:

0.200 / 0.285 × 5 ≈

3.5 mm of pedal travel

Again, forty microns…🧐

The thickness of a human hair is roughly in that general dimensional neighborhood. And we’re already several millimeters into the pedal. This is why high-end brake manufacturers spend so much engineering effort on caliper stiffness. The absolute deflections look laughably small when measured with a dial indicator. Through hydraulic area and pedal leverage, they aren’t small at all.

The Master Cylinder and Pedal Box

At this point we need to remember something embarrassing. We’ve spent all this time analyzing the components downstream of the master cylinder while quietly assuming the part we’re pushing on doesn’t move. Of course it does. The master-cylinder body expands. The piston and pushrod deflect. The balance bar and bearings move. The pedal bends. The pedal box bends. The bulkhead or chassis structure holding the pedal box bends (sometimes a LOT). And unlike the hydraulic components, some of these movements can appear almost directly at the pedal.

Imagine the master-cylinder mounting structure deflects just 0.25 mm under load. Depending on geometry, a substantial portion of that displacement gets multiplied by pedal ratio. A quarter millimeter at the master-cylinder reaction point can therefore become something around:

1–1.5 mm at the pedal

And we’ve all seen production-based race cars where the firewall or pedal structure moves considerably more than a quarter millimeter. This is one of those areas where a GoPro, a dial indicator and twenty minutes in the shop may teach us more than another three pages of equations.

Let’s Add It Up

Now we have enough rough numbers to do the thing we came here to do. Let’s create an illustrative race-car compliance budget at 1,500 psi.

ILLUSTRATIVE RACE-CAR COMPLIANCE BUDGET — 1,500 PSI

Steel hard-line expansion

Approx. pedal travel: 0.1 mm

Brake-fluid compression

Approx. pedal travel: 3.8 mm

PTFE hose expansion

Approx. pedal travel: 1.0 mm

Pad-stack compression

Approx. pedal travel: 4.4 mm

Caliper opening

Approx. pedal travel: 3.5 mm

Master/pedal structure

Approx. pedal travel: 1.2 mm

TOTAL ELASTIC TRAVEL: ~14.0 mm

Don’t fall in love with the individual values. Several of them are deliberately estimates. Fall in love with their relative magnitude. Because that is the interesting part. In this illustrative system, hard-line expansion accounts for less than one percent of the elastic pedal travel. PTFE hose expansion matters more, but it still represents only around seven percent. Fluid, pads and calipers dominate.

Now imagine we replace all of the PTFE hose with perfectly rigid tubing. We don’t actually get perfectly rigid tubing, of course, but let’s grant ourselves magic. Pedal travel falls from roughly:

14.0 mm to 13.0 mm.

That’s an improvement. A sensitive driver may even notice it. But the transformation from a 14 mm pedal to a 13 mm pedal is very different from the folklore version of the story in which braided hose is responsible for the entire soft brake pedal.

And now consider another modification. Suppose we replace the caliper with a design twice as stiff. Instead of saving one millimeter, we might save something approaching two. Improve the pad-stack stiffness at the same time and suddenly we’re making a much larger dent in the budget. That is exactly why I wanted the budget in the first place. Not to prove that brake hose doesn’t matter. To figure out how much it matters relative to everything else.

And We Haven’t Even Hit the Rotor Yet

Everything above describes elastic compliance. The brake pedal starts at zero pressure and each component deforms progressively as pressure rises.

Now let’s introduce 0.10 mm of pad knock-back. This is fundamentally different. The first 0.10 mm of piston movement creates essentially no brake pressure at all. We’re simply moving pads back toward the rotor. Using our same approximate front piston area:

2,500 mm² × 0.10 mm × two calipers =

0.500 cc

Our master cylinder now moves:

0.500 / 0.285 = 1.75 mm.

At a 5:1 pedal ratio:

8.8 mm of pedal travel

Eight.point.eight millimeters. Before the compliance budget even starts.

Now our driver experiences roughly 9 mm of clearance recovery + 14 mm of elastic travel or something approaching:

23 mm total pedal movement

And here’s the truly evil part. That first 9 mm may not be there during the next brake application. Or it may be 4 mm. Or 12 mm. It depends on what the car just did. That is why pad knock-back feels so much worse than ordinary compliance. Compliance can produce a relatively long pedal that is nevertheless completely predictable. Knock-back produces a pedal whose starting point moves around. Human beings are remarkably good at adapting to a consistent mechanical relationship. We’re much worse at adapting to one that changes every corner.

The Brake Pedal Still Doesn’t Know

This brings us back to the sentence that started the first article. The pedal doesn’t know whether its movement came from:

• - compressing fluid,

• - expanding hose,

• - opening a caliper,

• - compressing a pad,

• - flexing a pedal box,

• - or moving a piston across an air gap created by knock-back.

It only knows that your foot moved. But now we know something the pedal doesn’t. Those mechanisms do not deserve equal attention. If we’re chasing tenths of a millimeter in hard-line expansion while several millimeters are disappearing into pad compression, we’ve optimized the wrong thing. If we’re replacing perfectly good braided hose because the pedal gets long after high-g corners, we may be solving a compliance problem that doesn’t exist. If the pedal is consistent but slightly more progressive than we’d like, then hose, fluid volume, pad stiffness, caliper stiffness and structural stiffness all belong in the discussion. And if the pedal is rock hard in the paddock but occasionally heads for the floor on track, stop arguing about hose construction and start looking for clearance recovery.

That’s what the compliance budget gives us. Not an answer. A hierarchy.

Which Leaves One Unfinished Piece

There are two columns in my spreadsheet now. One contains the things we can calculate reasonably well. The other contains the things I really want to measure. And unfortunately for my free time, the second column is becoming much more interesting. I want actual pressure-versus-volume curves for:

• - 3/16-inch steel tubing

• - -3 PTFE/stainless brake hose

• - quality rubber brake hose

• - several race brake-pad compounds

• - several different caliper designs

I’d also like to put a dial indicator on a few pedal boxes and see how much supposedly rigid race-car structure actually moves at realistic pedal loads. Because the calculations have already answered the original question well enough to make the next question unavoidable. Braided brake hose expands more than steel tubing. Fine. How much more? And, more importantly:

Does anyone sitting in the driver’s seat actually care once the rest of the compliance budget gets a vote?

If you really want to know, it’s time to build a test rig. 🤓

I’ve heard some version of that argument for years, and at first glance it seems perfectly reasonable. Steel tubing is unquestionably stiffer than a flexible hose. If you pressurize both to the same pressure, the hose will deform more than the steel tube. Nobody is really arguing that point.

What I’ve always been less certain about is the conclusion people draw from it. Somewhere along the way, “hose expands more than steel” often becomes “hose expansion is one of the primary determinants of brake pedal feel.” Those aren’t the same statement, and I don’t think we should accept the second one simply because the first one is true.

Part of my skepticism comes from experience. Over the years I’ve built several race cars using almost entirely braided PTFE brake hose, with hard line only where it made sense for mounting or routing. None of those cars ever left me chasing a soft pedal. That’s obviously anecdotal, and anecdotal evidence has to be treated carefully, but it also shouldn’t be dismissed out of hand. If hose expansion were consuming a large fraction of the available pedal travel, I’d expect that to have shown up repeatedly over the years. It never really did.

That doesn’t prove braided hose is “better,” just as it doesn’t prove conventional wisdom is wrong. It simply suggests that there may be other contributors to brake pedal travel that are as large - or perhaps much larger - than hose expansion itself.

As I thought about the discussion over the following few days, I realized that I’d never actually tried to quantify that intuition. Like most engineers, I’d spent plenty of time thinking about individual pieces of brake systems: caliper stiffness, master cylinder sizing, pedal ratios, pad compounds, rotor temperatures, hydraulic leverage. But I’d never sat down and asked a deceptively simple question:

Where does the brake pedal actually go?

Not philosophically. Literally.

Imagine a brake system that requires roughly half an inch of pedal movement before full braking force is achieved. We usually describe that pedal as firm, soft, progressive, or long, but those are subjective descriptions of a very objective mechanical process. Somewhere inside the system, that half inch is being spent. Some of it is compressing brake fluid. Some of it is expanding brake lines. Some of it is bending calipers, compressing pads, flexing pedal structures, or deflecting backing plates. If the wheel bearings have a little more compliance than we’d like, perhaps some of it is recovering pad knockback before meaningful braking even begins.

The brake pedal doesn’t know where that movement came from.

It only knows the total.

Once I started looking at the problem that way, the original discussion about braided hose suddenly felt incomplete - not because it was wrong, but because it had isolated one contributor from an entire chain of them. That’s a trap engineers fall into more often than we’d like to admit. We become so focused on the component we’re discussing that we gradually lose sight of the system it’s part of.

Engineering has a way of correcting that habit. Whenever a design becomes complicated enough, we stop talking about individual parts and start talking about budgets. A race car has a weight budget. An electrical system has a power budget. Structures have stress margins and fatigue margins. Mechanical assemblies have tolerance stacks. Every one of those tools exists for the same reason: they force us to account for every meaningful contributor instead of arguing about one in isolation.

It occurred to me that brake pedal travel deserves the same treatment. Not because I think we’ll ever know every number with laboratory precision. In fact, I suspect some of the numbers will remain frustratingly fuzzy. Pad compressibility varies by compound. Caliper stiffness depends on geometry. Chassis flex depends on installation. Rotor knockback depends on bearings, suspension loads, cornering forces, curb strikes, and half a dozen other variables that don’t lend themselves to a single answer.

But uncertainty isn’t a reason to abandon the exercise. It’s a reason to organize it. So I opened a blank worksheet and started making a list. Every component between my foot and the tire that could plausibly consume pedal travel went onto the page. Some of those entries, I suspected, would turn out to be almost irrelevant. Others, I thought, might surprise me. The goal wasn’t to prove that braided hose was good or bad. It was to build a framework that would let me compare all of the contributors using the same units and the same questions.

By the time I finished that first page of notes, I had stopped thinking about brake lines altogether. I had started thinking about a compliance budget.

Building the Budget

Before I could assign numbers to anything, I had to decide what actually belonged on the worksheet. That turned out to be a little more complicated than I expected, because not everything that consumes brake pedal travel is doing the same thing.

Imagine applying the brakes while the car is sitting on jack stands. The rotors are perfectly centered, the bearings are unloaded, nothing has been bumped or flexed, and the pads are already sitting exactly where they should be. As hydraulic pressure builds, every component in the system behaves a little bit like a spring. The brake fluid compresses ever so slightly. The caliper bridge opens by a few thousandths. The pads compress. The backing plates bend. The brake lines expand. Even the pedal box and firewall give up tiny amounts of motion if they’re not infinitely rigid.

Every one of those effects is elastic. If you could somehow freeze the pedal in place and release the hydraulic pressure, every component would spring back toward its original shape. Energy is stored and then returned. That’s classic compliance.

Race cars, however, rarely spend much time sitting on jack stands, and they never do when it counts. Out on the track the brakes are constantly being disturbed by things that have very little to do with elasticity. The rotor might deflect slightly in a high-speed corner because of wheel bearing compliance. A curb strike can push the pads away from the disc. Rotor runout can gently nudge the pistons back into the caliper as the wheel rotates. Seal rollback retracts the pistons by design after every brake application. None of those mechanisms are really behaving like springs in the same sense as a flexing caliper or an expanding hose. They’re simply creating a gap that has to be closed before meaningful braking begins.

From the driver’s seat, though, both phenomena feel almost identical. The first few millimeters of pedal travel disappear, regardless of whether they were spent compressing material or simply moving the pads back into contact with the rotor. The pedal doesn’t know the difference.

The engineer should.

That realization caused me to split my worksheet into two completely different sections. The first became what I’d traditionally think of as compliance: elastic deformation throughout the hydraulic and structural system. The second became clearance recovery: all of the things that force the pedal to spend travel simply re-establishing contact before pressure can even begin to build.

It seems like a subtle distinction, but I think it’s an important one because the engineering solutions are completely different.

Suppose you calculate that braided hose contributes an additional tenth of a millimeter of pedal travel compared with steel hard line. If your complaint is excessive elastic compliance, then replacing that hose may genuinely improve the pedal.

But suppose your pedal is long because the front wheel bearings are allowing enough deflection to generate significant pad knockback in fast corners. Replacing every inch of braided hose with steel won’t change that at all. You’ve optimized the wrong part of the system.

Likewise, imagine replacing a caliper with one that’s twice as stiff. That reduces elastic deformation, but it won’t recover a single thousandth of pad knockback. Conversely, improving wheel bearing stiffness or reducing rotor runout may dramatically shorten the first part of the pedal while leaving the compliance of the hydraulic system completely unchanged.

Those are entirely different problems hiding behind the same symptom. That’s one of the reasons brake feel can be such a frustrating thing to diagnose. Two drivers can both describe a “long pedal,” yet be experiencing completely different physical phenomena. One may have a system that’s elastically compliant but consistent lap after lap. The other may have an exceptionally stiff hydraulic system that simply spends the first half inch of pedal travel chasing pads back to the rotor after every corner. The sensation sounds similar. The solution couldn’t be more different.

Once I started separating those two categories, something else became obvious. The brake line debate had been asking an incredibly narrow question. Not, “Where does the pedal travel go?” Not, “What are the dominant contributors?” Not even, “How much does hose expansion matter relative to everything else?”

Instead, we’d skipped straight to debating one entry on what was turning into a fairly long worksheet.

That’s not unusual. Engineers do it all the time. We naturally gravitate toward the component that’s easiest to see or easiest to change. But good system design usually comes from asking a different question first: where is the largest opportunity?

Only after we have that answer does it make sense to start optimizing individual parts. And with the worksheet finally organized, we can begin doing exactly that. Rather than arguing about braided hose in the abstract, we can start putting numbers next to each contributor and ask a much more interesting question:

How much of the compliance budget does each one actually own?

Putting Numbers on the Worksheet

At this point I had what looked like a reasonable framework, but a framework without numbers is really just a list. If I wanted to understand whether brake line expansion was important, I first needed to know what “important” actually meant.That immediately raised another question. Where do you start?

Some contributors are notoriously difficult to quantify. Brake pad compressibility depends on the friction material, operating temperature, backing plate geometry, and even wear. Caliper stiffness varies dramatically with design, piston spacing, bridge geometry, and mounting. Pedal box stiffness depends as much on the structure it’s bolted to as it does on the pedal assembly itself.

Those are all fascinating problems, but they’re terrible places to begin.

Instead, I wanted to start with the parts of the system that are reasonably well understood. Not because they’re necessarily the largest contributors, but because they give us a stable foundation before we venture into the fuzzier parts of the problem.

The first obvious candidate was the brake tubing itself. Unlike many of the other components in the system, steel tubing is relatively easy to model. We know its dimensions. We know its material properties. We know approximately how much tubing exists in a typical race car. If we assume a representative hydraulic pressure of around 1,500 psi during a hard stop, we can make a surprisingly good estimate of how much the tubing expands.

When I ran those numbers for a fairly typical installation - roughly twelve feet of 3/16-inch steel brake line - the result was almost anticlimactic. After translating the increased hydraulic volume back through the master cylinder and pedal ratio, the tubing accounted for roughly a tenth of a millimeter of pedal travel. Not a centimeter.Not a millimeter.About a tenth of one.

Now, before anyone starts sharpening their keyboards, that number obviously changes with assumptions. A different master cylinder bore changes it. A different pedal ratio changes it. Longer tubing changes it. Higher braking pressures change it. Like almost every engineering calculation, the exact answer depends on the system. But it doesn’t change by orders of magnitude.

Even if my assumptions are off by fifty percent - and I don’t think they are - we’re still talking about a contributor measured in tenths of a millimeter.That was my first surprise. Not because it proved steel tubing was “perfect,” but because it established a useful scale for the rest of the worksheet. Whatever else we discover, a tenth of a millimeter becomes the yardstick against which other contributors can be compared.

The next obvious entry was brake fluid. Unlike steel tubing, brake fluid often gets treated as though it’s incompressible. For almost every hydraulic calculation we perform, that’s a perfectly reasonable assumption. Pascal’s Law works beautifully because brake fluid is nearly incompressible. Nearly, however, isn’t the same as completely.

At typical braking pressures, brake fluid does compress, and unlike the tubing, every cubic centimeter of hydraulic volume in the system participates. Every caliper passage, every hose, every hard line, every master cylinder cavity contributes to the total volume being compressed. That means the amount of pedal travel isn’t determined solely by the properties of the fluid. It’s also determined by how much of it you’re carrying around.

That turned out to be another interesting realization. When we replace a length of 3/16-inch steel tubing with a -3 PTFE hose, two things happen simultaneously. The hose itself generally becomes more compliant than the steel tube it replaced. If that were the only thing changing, the conclusion would be straightforward. But it isn’t.

A typical -3 PTFE hose also has a slightly smaller internal diameter than standard 3/16-inch hard line. That means the hydraulic system now contains a little less brake fluid. Less fluid means less fluid compression. One effect increases pedal travel while the other reduces it. Neither effect is especially large, but it’s a good reminder that engineering problems rarely move in only one direction. Changing one component often changes several variables at the same time.

I hadn’t expected to spend much time thinking about hydraulic volume when I started looking into brake lines.Now it seemed impossible to ignore.By the time I finished the calculations for steel tubing and brake fluid, something interesting had happened. The original brake line discussion had almost disappeared into the background. It wasn’t that brake lines no longer mattered— - they absolutely belong on the worksheet - but they had become just one contributor among many.

And that’s when I realized the next entries on the worksheet were likely to be much more influential. Brake pads. Caliper stiffness. Backing plate deflection. Those weren’t going to be quite as easy to calculate.

The Parts That Refuse to Sit Still

At this point, I had enough of the worksheet filled in to start developing some intuition. A few contributors appeared to be surprisingly small. Others were beginning to look much larger than I would have guessed before doing the math. Unfortunately, that’s also where the easy part ended.

Steel tubing is relatively cooperative. Brake fluid is reasonably well characterized. Hydraulic volume can be measured with a ruler and a calculator. Once you move beyond those items, though, the problem becomes considerably more interesting. Take brake pads, for example.

Everyone knows they’re compressible. In fact, if they weren’t, brake systems would be incredibly harsh and difficult to modulate. That tiny amount of compliance between the rotor and the caliper is part of what gives a brake pedal its progressive feel. Yet when I started looking for published numbers, I discovered that meaningful data is surprisingly scarce.

Manufacturers spend enormous effort developing friction materials, but they don’t generally publish detailed compressibility curves over pressure and temperature. Even if they did, they would only describe one piece of the assembly. The backing plate itself also bends, particularly under the enormous clamping loads generated by modern multi-piston calipers. The piston faces aren’t infinitely rigid. Even the friction material doesn’t behave quite the same way when it’s cold, fully bedded, or glowing after twenty minutes of hard track use. Those effects are real. They’re measurable. They’re just difficult to isolate.

The same thing happens when you look at calipers. It’s easy to say that one caliper is “stiffer” than another, but what does that actually mean? Does the bridge deflect? Do the mounting ears move? Do the piston bores spread slightly? Does the caliper twist because the rotor isn’t perfectly centered between the pistons?The answer, of course, is yes. To varying degrees, all of those things happen simultaneously.

That’s one of the reasons brake development is still an active engineering discipline even at the highest levels of motorsport. The calculations will get you surprisingly far, but eventually you run into behaviors that are difficult to model accurately without finite element analysis, laboratory testing, or careful instrumented experiments.

I don’t see that as a weakness in the compliance budget. Quite the opposite. One of the things I’ve always appreciated about engineering is that it doesn’t require perfect information before you can make good decisions. We rarely know every variable with complete certainty. We estimate. We bound the problem. We identify which assumptions matter and which ones don’t. Then we decide whether the remaining uncertainty is actually important.

That’s a very different mindset from internet debates, where uncertainty is often treated as a reason to stop thinking altogether. In this case, uncertainty actually tells us where to focus. If hard-line expansion contributes roughly a tenth of a millimeter of pedal travel, it almost doesn’t matter whether the exact answer is 0.08 millimeters or 0.14 millimeters. That uncertainty isn’t likely to change any engineering decisions. If pad compression turns out to consume two or three millimeters, however, suddenly we’re talking about one of the dominant entries in the entire budget. That’s worth understanding better.

The same logic applies to caliper stiffness, backing plate deflection, pedal box rigidity, and even the mounting structure that ties everything together. I don’t need to know every value to three decimal places before I can recognize where the largest opportunities probably exist. That realization also changed how I thought about the purpose of the worksheet itself. Originally, I imagined it would produce an answer. Instead, it started producing priorities.

Some entries were clearly worth spending hours refining because they had the potential to move the overall result in meaningful ways. Others were already “good enough” after a first-order calculation because, even if I doubled or halved their estimated contribution, they remained relatively insignificant in the grand scheme of the system. That may be the most valuable lesson the worksheet has offered so far.

Engineers sometimes become obsessed with making tiny numbers more accurate while leaving the biggest unknowns largely unexplored. It’s an understandable trap because small numbers are often easier to calculate than large, messy ones. But precision and importance aren’t the same thing. Knowing a tenth-of-a-millimeter contributor to four significant figures isn’t nearly as useful as roughly understanding a contributor that’s twenty times larger. The worksheet doesn’t eliminate uncertainty. It helps organize it. And once I stopped expecting every number to be perfect, I could finally move on to the question I’d been trying to answer all along. If all of these contributors are added together, what does a realistic compliance budget actually look like?

Three Brake Systems

By this point I had accepted something that probably should have been obvious from the beginning. There isn’t a single compliance budget. There are as many compliance budgets as there are brake systems.

A Formula SAE car, a GT3 car, a Spec Miata, and a forty-year-old production car all consume brake pedal travel in different ways. Some spend almost none of it flexing structures, while others lose a surprising amount before hydraulic pressure has even fully developed. Rather than chasing a single “correct” answer, I found it more useful to think about three representative systems. Not because every race car fits neatly into one of these categories, but because almost every brake system resembles one of them more than the others.

The Idealized System

This is the car every engineer imagines while designing. The wheel bearings are new and properly preloaded. The rotors have almost no runout. The pedal box is mounted to a rigid structure. The firewall contributes essentially nothing. The calipers are stiff monoblock units. The brake pads have substantial backing plates and well-developed friction material. The plumbing is short, thoughtfully routed, and uses quality components throughout. Even the wheel hubs and uprights were designed with brake feel in mindNothing about this system is magic. Every individual improvement is relatively small. But none of the small penalties are allowed to accumulate.

When the driver steps on the brake pedal, almost every millimeter of movement is doing useful work. There is very little wasted motion recovering clearances or flexing structures that shouldn’t have moved in the first place. This is the benchmark.Not because it’s common, but because it shows us what the lower limit probably looks like.

The Typical Club Racer

This is probably where most of us live. The car has good components, but they weren’t all designed together. Maybe it started life as a production car before gradually evolving into something much faster. The calipers are good aftermarket pieces. The pedal box is solid but mounted to a chassis that was never intended for racing slicks. The bearings have a season or two on them. The rotors are straight enough, but not perfect. The brake lines are a practical mix of hard line and braided hose because that’s what fit the packaging. Nothing here is fundamentally wrong. In fact, this car probably has excellent brakes by almost any reasonable standard.

The interesting part is that the compliance budget has become much more democratic. Instead of one dominant contributor, several moderate contributors begin adding together. A little pad compression here. A little caliper deflection there. A touch of fluid compression. Perhaps a small amount of knockback after the fastest corners. No individual entry is alarming. Collectively, though, they define the pedal the driver experiences.

The Problem Child

Every racer eventually owns one. The brake pedal never quite feels the same twice. You bleed the system, and it feels fantastic in the paddock. Three laps later, the pedal is long again. You replace hoses. Then the master cylinder. Then the pads. You start wondering whether there’s still air trapped somewhere because the symptom refuses to stay fixed. Sometimes there is. Often there isn’t. The problem is that this car is spending pedal travel everywhere.

The wheel bearings have enough compliance to generate pad knockback. The rotors have measurable runout. The calipers are older designs with modest bridge stiffness. The pedal box moves more than anyone realized. The firewall contributes a little. The brake fluid is doing exactly what brake fluid always does. The hoses are expanding slightly. The pads are compressing. The backing plates are bending. None of those effects are catastrophic individually, but together they consume enough pedal travel that the driver begins chasing symptoms instead of causes. That’s what makes these cars so frustrating. There usually isn’t a single smoking gun. There are half a dozen small fires.

One of the things I like about these three examples is that they shift the conversation away from individual parts and back toward complete systems. If someone asks whether braided PTFE hose is appropriate for a race car, the honest answer is no longer a simple yes or no. It becomes another question. Which car are we talking about?

On the idealized system, changing brake line construction might represent one of the last remaining opportunities because nearly everything else has already been optimized.

On the problem child, however, replacing perfectly good brake lines while ignoring wheel bearing compliance, rotor runout, and pedal box stiffness is a bit like replacing a titanium bolt on an overweight race car. The change may be real, but it’s probably not where the performance is being lost.

That, to me, is where the compliance budget becomes genuinely useful. It doesn’t tell us what to optimize. It tells us where to look first.

stay tuned for part two where we get extra nerdy with numbers.

The setup is usually recognizable. Experienced builder. Serious car. Proper radiator, correctly sized for the application. Good ducting. Distilled water. Water Wetter. Everything the conventional wisdom checklist calls for. And the car still runs hot or shows signs of overheating.

The responses follow a familiar pattern too. Bigger radiator. More flow. Check for air pockets. Flush the system. Upgrade the water pump. Then occasionally someone suggests a different coolant and gets shouted down by someone citing the heat transfer coefficient of water versus glycol, and the thread eventually dies without resolution.

Here is what almost nobody in that thread is asking: what is the system actually doing under hard use? Not what the temperature gauge reads. Not what the coolant specification calls for. Not whether the radiator is the right size on paper. What is physically happening at the metal surface inside the engine when heat flux is at its peak? And is the air that radiator needs actually making it through the bodywork in sufficient quantity to do the job? Those questions lead somewhere more useful than a coolant recommendation.

The Scale of the Problem

Before getting into coolant chemistry, it is worth understanding how much heat an engine actually produces. Consider a 500-horsepower engine operating at roughly 33% thermal efficiency. Five hundred horsepower is approximately 373 kW of useful mechanical output. At 33% efficiency, the engine must consume around 1,130 kW of fuel energy to produce that. The remaining 757 kW becomes heat.

Not all of that heat enters the coolant. Between one third and one half leaves through the exhaust. Some goes through the oil. Some radiates directly from the engine surfaces. Even so, the cooling system may be responsible for rejecting more than 300 kW continuously at that output. For perspective, that is equivalent to running several hundred household space heaters simultaneously.

Viewed through that lens, the cooling system in a race car that demands that output from the engine consistently stops feeling like a supporting component at that point. It becomes one of the primary engineering problems of the vehicle. And that reframing matters, because it changes what you are trying to optimize. The coolant is one input. The radiator is one component. The ducting, the pump, the system pressure, the water jacket geometry - all of it is the system. And the system is what either solves the problem or doesn't.

The graphic above shows why this matters structurally. A coolant choice at step one doesn't produce a vehicle performance outcome directly. It produces a fluid property change, which changes heat pickup behavior, which changes operating temperature, which changes radiator requirements, which changes system weight and drag, which finally reaches lap time, which is several steps removed from where the argument usually starts. And changes anywhere in that chain feed back upstream. This is not a linear problem. It is a system problem.

The Equation Everyone Quotes

Most cooling discussions eventually arrive at:

Q = hAΔT

Heat transfer equals the heat-transfer coefficient multiplied by area multiplied by temperature difference. The discussion then immediately focuses on h. Water has a higher heat-transfer coefficient than glycol. Case closed. Except the equation contains three variables. Not one. And it is not even the only equation that matters. The coolant transports energy according to:

Q = ṁCpΔT

Where ṁ is mass flow rate, Cp is specific heat capacity, and ΔT is the temperature rise across the system.

Now the problem looks a bit different. Heat-transfer coefficient matters. Specific heat matters. Flow rate matters. Radiator area matters. Temperature difference matters. And underpinning all of it - the radiator only rejects heat in proportion to the mass of air passing through it and the temperature difference between that air and the coolant. Which means the ducting that delivers air to the core, and the exit path that lets it escape, are not incidental details. They are variables in the same equation. This is where the single-metric argument about water starts to break down. Every choice is connected to every other choice.

The Trade Space

As glycol concentration increases, specific heat decreases. Thermal conductivity decreases. Viscosity increases. Boiling point increases.

The curves above make the tradeoff concrete. The properties that make water an excellent coolant - high specific heat, high thermal conductivity, low viscosity - all degrade as glycol fraction increases. The one property that improves is boiling point. Which means the entire case for glycol rests on what that boiling point margin is actually worth in a given system. Sometimes it is worth a great deal. Sometimes it is worth almost nothing. The answer depends entirely on where the system is operating thermally. Some properties improve. Others get worse. There is no free lunch.

The same thing happens with pressure. Increasing system pressure raises boiling margin. But it also requires stronger hoses, stronger tanks, stronger radiator construction, and greater system complexity. And the same thing happens with airflow. Better ducting improves heat rejection. But it also consumes volume, affects aerodynamics, and adds weight and fabrication cost. Every lever has a cost. And none of them are free.

The graphic above shows the two primary levers available for controlling boiling margin - chemistry and pressure - and what each one costs. Chemistry raises boiling point but degrades thermal properties and increases viscosity. Pressure raises boiling point without touching thermal properties, but requires heavier, more complex hardware. The most capable systems use both simultaneously, which is exactly what WWII aircraft engineers and modern F1 teams converged on independently. Neither lever alone was sufficient. The question was always how much of each cost was worth paying for the margin gained.

One useful way to think about this is through a conceptual penalty model:

Ptotal = Pthermal + Pboiling + Psystem

Pthermal represents penalties from poorer heat-transfer properties. Pboiling represents penalties from local boiling, vapor formation, and insufficient thermal margin. Psystem represents penalties from pressure, weight, complexity, packaging, ducting, and reliability.

These penalties move in different directions. The graphic above shows what happens when you sum them. Adding glycol reduces boiling penalties while increasing thermal penalties. Adding pressure reduces boiling penalties while increasing system penalties. The total penalty curve has a minimum - and that minimum is not at either extreme. It is somewhere in the middle, and where exactly depends on the vehicle, the operating conditions, and the rules the system has to operate within.

It is also worth noting that these penalties are not simply additive, they interact. Raising system pressure changes local flow dynamics and heat-transfer behavior. Changing glycol concentration changes viscosity, which changes flow rate, which changes the ΔT term in the transport equation. Improving ducting changes the bulk coolant temperature, which changes the local ΔT at every hot spot in the jacket. The penalties are coupled, which is part of why the optimum is so system-specific and why people running similar engines can arrive at genuinely different answers.

The Part About Boiling That May come as a Surprise

Most people think of boiling as a failure mode. Heat-transfer engineers tend to view it differently.

The graphic above shows what actually happens as surface heat flux rises. In single-phase liquid cooling, h is moderate and stable. As heat flux increases, nucleate boiling begins - tiny bubbles form and collapse at the surface, and h actually rises dramatically. This is the regime where water earns its reputation. Controlled nucleate boiling is an outstanding heat-transfer mechanism. The problem is that controlled nucleate boiling is very difficult to achieve in an internal combustion engine that is asked to operate at widely varying RPMs and what invariably comes next.

Past the critical heat flux point, nucleate boiling transitions into film boiling. A continuous vapor layer forms between the metal and the coolant. Vapor is a terrible conductor. The local heat-transfer coefficient collapses - not gradually, but sharply. Metal temperature rises rapidly even though bulk coolant temperature may still look reasonable on a gauge. Everyone has seen the trick where you can pass your hand quickly through something like liquid nitrogen or a hot stream of extremely hot oil without injury - these are examples of film boiling. The thin layer of air present almost completely isolates you from the huge temperature differential.

This is the part that matters for the frustrated forum poster. Their temperature gauge reads bulk coolant temperature at the sensor location. It does not read metal surface temperature at the exhaust valve bridge or the spark plug boss or the turbocharger water jacket. Those are the locations where local heat flux is highest. Those are the locations where film boiling begins first. And if the bulk coolant temperature is already elevated because of inadequate airflow through a poorly ducted radiator, the margin before local boiling begins at those hot spots is even smaller than it would otherwise be. A system can be rejecting heat adequately at the bulk level while simultaneously cooking metal at the local level. The gauge will not tell you this is happening until it is too late.

One of the more interesting realizations from this topic is that water's greatest strength and its greatest weakness may be the same phenomenon. Controlled nucleate boiling is an outstanding heat-transfer mechanism. Uncontrolled boiling produces catastrophic reductions in heat transfer. Water only gets credit for its heat-transfer coefficient while it is actually touching the metal.

A fluid with a lower coefficient that remains stably liquid throughout a difficult thermal event may outperform a fluid with a superior coefficient that intermittently generates an insulating vapor layer. The coefficient is not the answer. The stability of the coefficient is the answer.

What Is Actually Being Optimized?

Imagine two otherwise identical race cars.

Car A runs a large radiator with properly sealed ducting and maintains coolant temperature at 180°F all day. Car B runs a nominally identical radiator with poor inlet sealing, recirculates hot air back through the core, and operates at 240°F despite having the same coolant specification. Many people looking at those two cars would start asking about coolant. “Are you running water?” The race engineer would start asking about bodywork, or even step back further to take a broader system-level look. Because the race engineer understands something important. Cooling is not the objective. Cooling is overhead. The cooling system consumes resources on behalf of the engine - weight, drag, packaging volume, complexity, reliability margin. Every unit of cooling capacity comes with a price attached, and that price is paid whether the capacity is being used effectively or not.

The objective is not maximum cooling. The objective is sufficient cooling at minimum cost to vehicle performance. And sufficient cooling means the entire chain - air delivery through properly designed ducting, heat rejection at the core, coolant transport at adequate flow and pressure, and stable local heat transfer at the metal surface - is working together. A weak link anywhere in that chain limits the whole system, regardless of what fluid is in it.

Three Engineers, Three “Correct” Answers

What makes these sorts of discussions genuinely interesting is that people often believe they are having the same argument when they are not. One engineer is optimizing coolant properties. Another is optimizing cooling-system architecture - radiator sizing, ducting, pressure, flow distribution. The race engineer is optimizing lap time. All three can be technically correct. They are simply working at different levels of the system hierarchy. The disagreement is rarely about physics. It is about objective functions.

The person advocating water may be minimizing thermal penalties. The person advocating glycol may be minimizing boiling penalties. The person advocating better ducting and a higher pressure cap may be minimizing system penalties without touching the fluid at all. Same physics. Different objective functions. Different answers. This is why the argument about water being the best coolant is simultaneously correct and incomplete. Water probably wins most property comparisons. But property comparisons are not what gets measured at the end of a race.

Back to the Argument

But we digress…years ago I got into a disagreement with someone who knew enough thermodynamics to be dangerous. His position was that water's heat-transfer coefficient is dramatically higher than any glycol-based fluid, and that settled the matter. He was quite confident about that. He wasn't wrong about the coefficient. He was wrong about what the coefficient means.

My counter at the time was intuitive rather than rigorous. I had run Evans waterless coolant on a couple of cars at that point as an experiment and had a sense that something about the local boiling behavior mattered in ways the property table wasn't capturing, because in one of those vehicles, despite the fact that it ran in steady state at a higher temperature, signs of cooling issues like early onset pinging, were not present. At the time I didn't have the framework to say it cleanly. The argument I should have made is the one this article has been building toward.

Waterless coolant does not win because it has better thermal properties. It doesn't. It potentially wins only in engines where the water-based system is spending meaningful time in the unstable part of the boiling curve - at hot spots around exhaust valve bridges, spark plug bosses, turbocharger water jackets, or anywhere local heat flux is high enough to push past controlled nucleate boiling into film boiling. And it only wins there if the rest of the system, the ducting, the pressure, the flow distribution, has already been addressed. Switching to a waterless fluid in a system with poor ducting and a stock pressure cap is optimizing the wrong variable in the wrong order.

In a well-designed, properly pressurized, adequately ducted system with no local boiling problem, waterless coolant normslly provides no meaningful advantage and carries real disadvantages if you’re not operating in conditions that warrant the freezing margin of glycol. Glycol has lower specific heat, higher viscosity, which will reduce pump flow, meaningfully higher cost, and is problematic at racetracks if fluid spills onto the racing surface.

So What Should the Forum Poster Actually Do?

If you are the person with the correctly sized radiator, the distilled water, the Water Wetter, the good ducting, and the car that still runs hot, here is a more useful diagnostic sequence than just changing your coolant.

First, verify you are actually measuring what you think you are measuring. A coolant temperature gauge reads bulk fluid temperature at the sensor location. It does not read local metal temperature at hot spots. If the engine is experiencing local film boiling, the gauge may read normally right up until something fails.

Second, audit your airflow path before touching the coolant. A radiator rejects heat in proportion to the temperature difference between coolant and air and the mass flow of air passing through it. Both of those are heavily dependent on ducting. A radiator with no inlet duct, poor sealing between the core and the surrounding bodywork, or a hot air recirculation path back to the inlet is not performing anywhere near its rated capacity , regardless of what fluid is inside it. Sealed inlet ducting that forces all available air through the core, and an exit path that allows that air to escape without recirculating, is a prerequisite for the rest of this conversation. Many cooling problems that get blamed on coolant selection are actually ducting problems in disguise.

Third, check your system pressure. This is the most underused lever in the trade space and it costs almost nothing to adjust. A higher pressure cap (or even better, a pressure relief valve, as caps can and do degrade over time) raises the boiling point of your coolant without changing its thermal properties at all. You keep water's specific heat, water's thermal conductivity, water's viscosity, and you buy margin before local boiling begins. If you are running a stock pressure cap on a high-output engine, this is where to start.

Fourth, look at your coolant flow rate and distribution. A high-flow water pump is not always better cavitation at high rpm can actually reduce effective flow. And “high flow” pumps have very poor performance with complex pumping system or systems with unavoidable restrictions that add back pressure. More important is whether coolant is reaching the hot spots at adequate velocity. Local overheating often results from flow dead zones in the water jacket, not from inadequate bulk flow.

Fifth, consider your coolant composition. Moving from pure water toward a modest glycol fraction - somewhere in the 20 to 30 percent range - raises boiling point further while accepting a moderate thermal penalty. This is where most well-designed street and competition cooling systems actually live, and for good reason. The penalty curve sum is lower there than at either extreme. You are not optimizing a single property. You are minimizing total system cost.

note that following this approach, only after working through those levers does coolant composition become the primary variable - and even then the question is not which coolant has the best property table. The question is whether your system is spending meaningful time in the unstable part of the boiling curve at conditions where a higher-boiling fluid would stay liquid. A turbocharged engine with repeated hot shutdown cycles, an iron block with aluminum heads running marginal system pressure, a race car with sustained low-speed running and high heat soak - these are the scenarios where moving further along the composition axis, or to a waterless fluid entirely, is worth the thermal penalty it carries. The distilled water and Water Wetter are not wrong. They are optimizing Pthermal while ignoring Pboiling. But the system may be asking for a different balance. Keep your brain nimble.

The Real Question

In the end race engineers do not race heat-transfer coefficients. They race lap times. The best solution rarely emerges from maximizing a single property. It emerges when all the penalties have been accounted for, the objective has been clearly defined, and the question being asked matches the problem actually being solved.

The coolant is one variable in a system that includes pressure, flow distribution, jacket geometry, airflow delivery through properly designed ducting, and the thermal regime the metal surface is actually operating in. Change any one of those things and the optimum moves. Which means the right answer for your engine, in your car, on your track, may be genuinely different from the right answer for someone else's — even if the engines look identical on paper. Water wins the property table. But cars don’t race on a property table.

At the end of the last article, I said I had cheated a little. That cheat was useful, but it was still a cheat. I described the inerter as if it were simply another version of F = ma. That is the right intuition, but it is not quite the right equation. A normal mass produces force according to its acceleration relative to an inertial reference frame and F = ma applies directly. There are linear and rotating versions of the same concept, but both and simple and intuitive.

An inerter produces force according to the relative acceleration between its two terminals:

F = b(a₂ - a₁)

or, written in terms of displacement:

F = b(d²x₂/dt² - d²x₁/dt²).

That little change is everything. The constant b in the above is called inertance. It has the same units as mass - kilograms - but it is not simply the physical weight of the device. It is the effective inertial relationship created between the two attachment points. A mass says: “Accelerate me relative to the world, and I will push back.” An inerter says “Accelerate one end of me relative to the other, and I will push back.” That is the essence of the trick.

Completing the Family

One of the reasons the inerter is such a satisfying (yeah I’m a nerd) device is that it completes a pattern. As we’ve reviewed before A spring creates force from relative displacement: F = k(x₂ - x₁). A damper creates force from relative velocity: F = c(v₂ - v₁). An inerter creates force from relative acceleration: F = b(a₂ - a₁)

The spring is the displacement element.

The damper is the velocity element.

The inerter is the acceleration element.

Viewed this way, the inerter stops looking like a strange Formula One gadget and starts looking like something that should have existed all along. In fact, once you see the pattern, it becomes difficult to unsee it. Springs convert position into force. Dampers convert velocity into force. Inerters convert acceleration into force. The entire device is hiding in that observation.

Where Inertance Comes From

The easiest way to visualize a practical inerter is a ball screw connected to a flywheel. As the suspension moves, the two ends of the inerter move relative to one another. That linear motion turns the ball screw, and the ball screw spins the flywheel. The flywheel resists changes in rotational speed because it has rotational inertia. Relative motion between the suspension components is therefore forced to accelerate a rotating mass. The harder you try to accelerate it, the harder it pushes back. That should sound familiar because it is exactly what ordinary mass does. The cleverness lies in the mechanical advantage of the inerter implementation .

The flywheel can be made to rotate dramatically faster than the suspension itself is moving. Through the screw pitch, gearing ratios, and geometry of the mechanism, a relatively small rotating mass can behave as though it were a much larger linear mass.

The derivation is surprisingly straightforward. If the ball screw has lead p, then: θ = (2π/p)x

Differentiating once: ω = (2π/p)v

Differentiating again: α = (2π/p)a

The torque required to accelerate a flywheel is: T = Jα, where J is the rotational inertia of the flywheel.

The ball screw converts that torque back into linear force. Ignoring losses: F = T(2π/p)

Substituting: F = Jα(2π/p)

and then substituting angular acceleration: F = J(2π/p)²a

Comparing this with: F = ba

gives: b = J(2π/p)².

This is the equation that makes the entire device work. Notice what isn’t in it. Vehicle weight.Spring rate. Tire stiffness. Damper force. None of those things appear. The effective inertance is determined entirely by the inerter flywheel inertia and the geometry used to connect it to the suspension. That is an extraordinarily powerful idea because it means a relatively small piece of hardware can create a very large inertial effect simply by changing the mechanical advantage between the suspension and the flywheel.

A Real Example

It’s always good to put math and physics concepts into terms that we can relate to, so let’s work up a practical example. Suppose the flywheel has: J = 0.001 kg·m²

and the ball screw lead (or pitch) is: p = 0.010 m/rev, or 10 mm per revolution.

Using the governing equation we derived above: b = J(2π/p)²

gives: b ≈ 395 kg

The obvious question then is whether 395 kg is a lot. But unfortunately, that’s a little like asking whether a 600 lb/in spring is stiff without knowing things like motion ratio, and just like that example, the answer depends entirely on what you’re attaching it to. Before we answer that question, it is worth understanding what the flywheel itself might look like.

For a simple solid disk: J = ½mr²

Suppose the flywheel radius is 50 mm.

Solving for mass: 0.001 = ½m(0.05²)

gives: m ≈ 0.8 kg

In steel, that would be a flywheel roughly 100 mm in diameter and approximately 13 mm thick. In other words, a completely ordinary-looking chunk of metal. And yet that flywheel can create nearly 400 kg of effective inertance between the two attachment points.

At this point it is tempting to conclude that we’ve somehow created 400 kilograms out of thin air. We haven’t. The car still weighs exactly what it weighed before. The inerter has simply created a force path that behaves as though hundreds of kilograms of inertial resistance exist between its terminals. That distinction is subtle. It is also the entire reason the device is useful. The squared term in the equation is doing most of the heavy lifting.

Using the same flywheel:

20 mm lead → 99 kg inertance

10 mm lead → 395 kg inertance

5 mm lead → 1,580 kg inertance

This is where the device starts to feel less like a suspension component and more like a mechanical lever operating on inertia itself.

So…Is 395 Kilograms a Lot?

The more useful question is not whether 395 kg is large. The more useful question is what force does it create? An inerter produces force according to F = ba as we’ve determined. At 1 g of relative acceleration:

100 kg inertance ≈ 981 N (or 220lbf)

400 kg inertance ≈ 3,924 N (or 880lbf)

1,000 kg inertance ≈ 9,810 N (or 2200lbf)

These are substantial forces. A few hundred kilograms of inertance can easily generate force levels comparable to spring or damper loads during aggressive suspension events. That immediately explains why engineers became interested. The forces are large enough to matter. Very large.

The Frequency Trick

The really interesting part appears when we stop looking at acceleration directly and start looking at frequency. Most suspension motion is oscillatory. For sinusoidal motion a = ω²x, which means F = bω²x

This can be rearranged as: k_effective = bω²

The inerter begins to look like a stiffness that grows with frequency. This is why inerters are so powerful. They do not influence all motions equally. Their influence grows with the square of frequency. At 3 Hz (roughly in the range of body and platform motions or “suspension frequency” for a stiff race car):

100 kg inertance → 35,500 N/m (203lbf/in)

400 kg inertance → 142,000 N/m (812lbf/in)

At 10 Hz, where tire and wheel-hop dynamics begin to become important:

100 kg inertance → 395,000 N/m (2256lbf/in)

400 kg inertance → 1,580,000 N/m (9022lbf/in)

That is a remarkable change. The same device becomes dramatically more influential as motion frequency increases. This is one of the reasons inerters are so interesting to race engineers. A heave mode at 2–3 Hz, a curb strike at 10–15 Hz, and a tire vibration mode at 20 Hz are all asking different questions of the suspension. The inerter’s “answer” to those different inputs changes dramatically depending on which question is being asked.

A spring responds essentially the same way regardless of how quickly the motion occurs. An inerter does not. As frequency rises, the force contribution from the inerter rises rapidly. That gives engineers another way to influence high-frequency events without necessarily making the entire car stiffer everywhere. Viewed that way, the inerter starts to look less like another suspension component and more like a frequency-selective tool. Suddenly the question is not

“How much inertance does the device have?” The question becomes “Which frequency range am I trying to influence? That is a much more interesting engineering problem.

What Is a Useful Range?

For race-car applications, I would think about inertance roughly in these ranges.

25–100 kg

Relatively subtle. Useful when the goal is to influence higher-frequency wheel and tire behavior without making the inerter a dominant element in the system.

100–500 kg

This is where things become genuinely interesting. Forces are large enough to influence tire-load variation, platform control, and suspension dynamics without completely dominating the response.

500–2,000+ kg

Now the inerter becomes a major player. It can be extraordinarily effective when targeted correctly, but it can also create unintended resonances, fight other suspension elements, or reduce tire consistency if used carelessly.

Like spring rate and damping, there is no universally correct value. Everything depends on the mode being targeted. The practical tuning question is not “How much inertance can we create?” It is “Which motion are we trying to change, at what frequency, and what force contribution do we want the inerter to make?” That is the real tuning problem.

The Mathematical Problem That Became a Racing Solution

What I find most fascinating is that the inerter wasn’t originally invented to solve a racing problem at all! It was invented to solve a mathematical problem.

Malcolm Smith was studying mechanical network theory and noticed something peculiar. Electrical engineers had a complete family of ideal network elements. Mechanical engineers did not. Springs had electrical analogues. Dampers had electrical analogues. Mass sort of did. But not really. The problem was that mass is not a true two-terminal device in the same sense as a spring or damper. A spring connects two points. A damper connects two points. A conventional mass ultimately references the inertial frame of the universe itself. The missing element was a true two-terminal acceleration device. That device became the inerter.

Once it existed, mechanical systems could be synthesized much more like electrical networks. That sounds abstract, but it has profound practical consequences. Instead of simply selecting springs and dampers and accepting whatever compromises resulted, engineers gained another fundamental building block from which to construct dynamic systems. The fact that Formula One eventually found it useful is almost secondary. The original breakthrough was conceptual.

Why Formula One Cared

Formula One cared because tires, suspensions, chassis structures, and aerodynamic platforms are all dynamic systems interacting with one another. The tire itself is a spring. The chassis is a spring. The suspension is a spring-damper network. The aerodynamic platform has its own dynamic behavior. Many of these modes are not ideally controlled by spring rate alone or damping alone. Viewed another way, a traditional suspension engineer only has two direct knobs available. They can decide how much force is generated by displacement, and they can decide how much force is generated by velocity.The inerter introduces a third knob:

Force generated by acceleration.

That is a very different tool to hand a race engineer trying to maintain tire load over a curb, control a heave mode at high speed, or stabilize an aerodynamic platform that is highly sensitive to ride height. An inerter gives the engineer another lever to pull. It allows acceleration-dependent forces to be introduced into specific load paths and specific modes of motion. That is why the famous “J-damper” was simultaneously revolutionary and slightly misleading. It wasn’t really a damper. A damper dissipates energy. An ideal inerter stores kinetic energy in a rotating mass and returns it later. The name helped describe where it was installed. It did not accurately describe what it was, as is often the case with new ideas in F1 - make a name that sounds different than what it actually does, partly in order to confuse your competitors.

The Missing Piece

Looking back, it’s almost surprising that the inerter arrived so late. Once you accept that springs convert displacement into force and dampers convert velocity into force, the next step feels almost inevitable. Of course someone would eventually build a device that converts acceleration into force.

The surprising part isn’t the physics. The surprising part is that nobody packaged it into a practical suspension component until the twenty-first century. For most of automotive history, acceleration was largely an outcome. Engineers adjusted springs. Engineers adjusted dampers. Acceleration changed as a consequence. The inerter changes that relationship. Acceleration itself becomes part of the design space.

And that brings us back to the statement that ended the previous article.

A spring turns position into force.

A damper turns velocity into force.

An inerter turns acceleration into force.

The important shift is that acceleration is no longer merely the result of the suspension setup. Acceleration becomes something the engineer can tune directly. For nearly a century, suspension engineers tuned where things were and how fast they were moving. The inerter gave them a way to tune how quickly that speed could change.

In the last suspension article, we talked about springs and dampers as if they were fundamentally different devices. In reality, they’re both responding to different derivatives of the same motion.

A spring produces force proportional to displacement:

F = kx

A damper produces force proportional to velocity:

F = cv

And that got me thinking about another equation that every engineering student learns early on, probably on the furstcday of a 100 level dynamics class, or even in Calc 1:

F = ma

Force equals mass times acceleration. Viewed that way, springs, dampers, and masses aren’t really three separate concepts at all. They’re three different ways of producing force from three different descriptions of motion.

A spring looks at position and asks, “How far have we moved?”

A damper looks at velocity and asks, “How fast are we moving?”

A mass looks at acceleration and asks, “How quickly is that velocity changing?”

Position, velocity, and acceleration are all connected. Velocity is simply the rate of change of position, and acceleration is the rate of change of velocity. The math is straightforward enough that most of us stop thinking about how profound it really is. Much of engineering consists of repeatedly asking the same question:

How fast is the thing changing?

The first answer is velocity.

Ask again and the answer is acceleration.

Ask again and you arrive at jerk - the rate of change of acceleration. If you’ve ever ridden in a car that felt abrupt or harsh despite relatively modest acceleration levels, or in a really laggy turbo car that suddenly “hits hard” you’ve experienced jerk whether you knew the term or not.

The deeper you go, the more you realize that many systems are simply responding to different derivatives of the same underlying motion. Once you start looking for it, you see the pattern everywhere. Control systems, aerodynamics, electrical circuits, suspension systems, even the way humans perceive motion all end up caring about different derivatives of the same signal.

That perspective becomes particularly useful when thinking about suspension systems. Springs care about where things are. Dampers care about how fast they’re moving. Mass cares about acceleration. None of those statements are especially controversial, but together they reveal something interesting.

Humans don’t really experience suspension position directly. We don’t feel that the wheel moved three inches upward relative to the chassis. We don’t directly perceive suspension velocity either.

What we feel is acceleration. We feel the shove into the seat under throttle. The weight transfer under braking. The impact from a pothole. The sharp kick when a tire strikes a curb. Acceleration is where the suspension system finally becomes something our bodies can perceive. Which raises an interesting question:

If a spring gives us force proportional to displacement, a damper gives us force proportional to velocity, and mass already gives us force proportional to acceleration, why did Formula One engineers spend years developing a completely new mechanical device called an inerter?

After all, Newton solved the acceleration problem centuries ago. We already have mass. Or at least it seems that way.

The Problem With Mass

The problem is that mass is stubborn about where it lives. Suppose I tell you that a race car would perform better if I added fifty pounds to the chassis. Most racers immediately recoil from the idea. We’ve spent decades trying to remove mass from race cars. We buy lighter wheels, lighter brakes, lighter seats, lighter batteries, lighter body panels, and occasionally lighter wallets in the process. Adding mass sounds like engineering malpractice. But that’s because we’re thinking about total mass.

Suspension engineers are usually thinking about where mass is located, how it is connected to other masses, and how it participates in motion. A wheel assembly is a mass. A tire is a mass. A control arm is a mass. A chassis is a mass. The driver’s body is a mass. The driver’s head is a mass connected to another mass through a spring and damper system we call a neck.

Everything in the car is trying to obey F = ma whether we want it to or not. The challenge is that real mass always arrives with side effects. If I physically add fifty pounds to the chassis, I don’t just get the acceleration-related behavior I wanted. I also get fifty extra pounds to accelerate, fifty extra pounds to brake, fifty extra pounds to corner, and fifty extra pounds to drag up every hill. The mass participates in every maneuver whether it is helping or hurting. That’s unavoidable in our universe.

What suspension engineers really wanted was something more selective. They wanted some of the dynamic behavior associated with mass without necessarily carrying around the actual weight penalty. At first glance, that sounds impossible. Mass is mass, right? Right? Newton seems fairly unambiguous on the subject. Yet if you cut open a typical inerter, it almost looks like someone found a loophole.

A Mechanical Loophole

Externally, an inerter often resembles a shock absorber. It has two attachment points and is installed between suspension components much like a damper. The magic happens inside. One common implementation uses a ball screw connected to a flywheel. As the two ends of the inerter move relative to one another, the ball screw converts linear motion into rotational motion. That rotational motion accelerates a flywheel whose rotational inertia resists changes in speed. Relative motion between the suspension components is therefore forced to accelerate a rotating mass. The harder you try to accelerate it, the harder it pushes back. That should sound familiar because it is exactly what ordinary mass does.

The cleverness lies in the mechanical advantage. The flywheel can be made to rotate dramatically faster than the suspension itself is moving. Through the screw pitch, gearing ratios, and geometry of the mechanism, a relatively small rotating mass can behave as though it were a much larger linear mass.

The result is a device capable of producing surprisingly large acceleration-dependent forces while containing only a modest amount of actual physical mass. Unlike a spring, it does not primarily store energy. Unlike a damper, it does not primarily dissipate energy. Instead, it produces force proportional to the relative acceleration between its attachment points. That distinction turns out to be incredibly important.

Virtual Mass

A conventional mass exists relative to an inertial reference frame. It has a position, velocity, and acceleration with respect to the rest of the universe. An inerter doesn’t care about the universe. It only cares about the relative acceleration between the two points attached to it. That may sound like a subtle distinction, but it fundamentally changes what is possible.

An inerter allows acceleration-dependent forces to be introduced between arbitrary points within a suspension system. It behaves somewhat like a mass that exists between those points rather than a mass that exists somewhere in physical space.

Engineers sometimes describe this as creating virtual mass, but I think it’s more useful to think of it as relocating where inertial effects occur.

Instead of adding weight to the car and accepting all the consequences that come with it, an inerter allows some of the beneficial dynamic characteristics of mass to be placed exactly where they are useful.

Viewed through the same lens as springs and dampers, the elegance becomes obvious. A spring responds to displacement. A damper responds to velocity. A mass responds to acceleration.

Put another way, suspension engineers spent nearly a century tuning the zeroth derivative and the first derivative of motion while treating the second derivative largely as a consequence. The inerter was the first device that allowed acceleration itself to become a tuning target. For most of automotive history, suspension engineers only had practical ways to directly tune the first two. The inerter changed that. Suddenly suspension design gained another degree of freedom. Not another spring. Not another damper. A fundamentally different element reacting to a fundamentally different derivative of motion.

The Third Element (and no, it’s not Lee Loo Dallas)

Viewed through that lens, the excitement surrounding Formula One’s famous “J-damper” starts to make more sense. The teams weren’t merely optimizing an existing suspension component. They had effectively discovered a third fundamental reactive element that could be combined with springs and dampers to shape how the car responded to disturbances.

For nearly a century, suspension engineers had been composing variations on the same theme. Different spring rates, different damping curves, different geometries, but always using the same two fundamental tools. Then a third tool appeared. The remarkable thing isn’t that Formula One became obsessed with inerters. The remarkable thing is that it took us so long to invent one as the basic physics and the need for it was there all along. Sometimes things seem obvious in retrospect though…

Of course, understanding what an inerter is and understanding how to use one are two very different things. A spring can improve a suspension or completely ruin it depending on where it is placed and how stiff it is. Dampers are no different. Inerters introduce an entirely new set of tradeoffs, resonances, and tuning opportunities that are not always intuitive. The moment an inerter is added to a system, questions begin to emerge.

How much effective inertance should it provide? Where should it be connected? What modes of motion should it influence? How does it interact with spring rates, damping coefficients, tire stiffness, and unsprung mass? And perhaps most importantly, how can a relatively small flywheel create the dynamic effects of a mass many times larger than itself?

Those questions lead directly into the mathematics of inerters, and the mathematics turns out to be every bit as interesting as the device itself.

And if you’re an engineer, I should probably confess something. I’ve cheated a little in this article. The equation for an inerter isn’t actually F = ma. That’s a useful mental model, but not the whole story.

The real equation introduces a quantity called inertance, and understanding what that means opens the door to how inerters are designed, sized, tuned, and integrated into real suspension systems. We’ll tackle that next time.

A damper doesn’t care where the suspension is. It doesn’t care whether the car is rolled over in a corner, squatting under acceleration, or diving under braking. It only cares about one thing: How fast is the suspension moving? That distinction sounds small, but it fundamentally changes how a race car behaves. The basic physics are surprisingly simple. A spring generates force proportional to displacement:

F = kx

where k is the spring rate and x is displacement.

A damper generates force proportional to velocity:

F = cv

where c is the damping coefficient and v is velocity.

Mass resists acceleration:

F = ma

where m is mass, and a is acceleration.

That’s literally the whole game. Springs react to position. Dampers react to velocity. Mass reacts to acceleration. Everything else is details.

If you remember one piece of math from this article, make it this:

v = dx/dt

Velocity is simply displacement viewed through time, or stated in slightly more mathematical terms - velocity is the time DEI active of displacement. That’s why a damper can see two identical suspension movements as completely different events - the damper only sees the velocity, not what created it. Or mathematically, the damper only sees the slope of the curve, not where on the curve it sits.

Time is Everything

Imagine the wheel moves one inch. If it takes two seconds, that’s one event. If it takes twenty milliseconds, that’s a completely different event. The spring sees one inch. The damper sees radically different velocities. The same displacement can happen slowly or quickly. It can happen in a controlled manner or an uncontrolled manner. The spring doesn’t particularly care how quickly things happen. The damper cares enormously.

That raises an interesting question. Why does the damper exist at all? Imagine a car with springs but no dampers. Hit a bump. The spring stores energy. Then it releases that energy. Then stores it again. Then releases it again. Left alone, a spring-mass system naturally wants to oscillate. The damper’s job is to remove energy from that cycle. Not necessarily to stop motion. But to control it.

And that leads to one of the most misunderstood ideas in suspension tuning. Most people think dampers reduce motion. That isn’t quite right. They REDISTRIBUTE motion. To understand why, it helps to think about the order in which events actually occur.

Forces create motion.

Motion contains frequencies.

Frequencies imply velocities.

Dampers react to velocities.

Tires experience load.

Notice that the damper never directly experiences a curb, a pothole, a braking zone, or a steering input.

The damper only experiences the shaft velocities those events create.

Drivers experience frequencies. Engineers often analyze frequencies. Damper pistons experience velocity. The driver says:

“The car feels nervous over curbs.”

The engineer starts looking for the source of the motion.

The damper only knows it saw eighteen inches per second of shaft velocity. This distinction becomes even more important when you realize the suspension is not the first suspension in the system. The tire is.

Before the wheel ever moves relative to the chassis, the tire has already deformed, stored energy, and dissipated energy. The tire is a spring. The tire is a damper. And it’s the only suspension component that actually touches the road. A race car therefore contains at least two major masses.

The sprung mass is the chassis, engine, driver, and body. The unsprung mass is the wheel, tire, brake assembly, upright, and portions of the suspension. The actual sequence looks more like:

Road → Tire → Unsprung Mass → Suspension → Sprung Mass

Every bump, curb strike, steering input, braking event, and aerodynamic disturbance injects energy into that system. And that energy cannot simply disappear. The wheel can move. The tire can deform. The chassis can move. The damper can convert some of that energy into heat. But the energy has to go somewhere. This is why more damping does not necessarily mean less motion. If the wheel moves less, the tire may deform more. If the tire cannot absorb the disturbance, the chassis may move more. If neither can absorb it effectively, grip may suffer. The motion wasn’t eliminated. It was reassigned.

That’s why I think it is more useful to think of a damper as a motion allocator than a motion suppressor. Its job is not to decide whether motion exists. The road has already made that decision. Its job is to influence where the resulting motion occurs and how quickly it develops.

Let’s Run Through an Example!

A practical example makes this easier to visualize. Imagine a driver coming in after a session and saying:

“The car feels nervous entering fast corners. It takes a set, but then it feels like it keeps moving around underneath me.”

An inexperienced engineer might immediately start thinking about springs or anti-roll bars. A more experienced engineer starts asking different questions. What motion is occurring? Where is that motion occurring? How quickly is it developing? How quickly is it decaying? Notice that none of those questions are asking how much the car rolled. They’re asking how the car behaved through time.

Suppose the data shows the car reaches its target roll angle quickly enough, but then oscillates slightly around that attitude before finally settling. The problem may not be the amount of roll. The problem may be the rate at which energy is leaving the system. The spring is doing exactly what it was designed to do. The spring stores energy. The question is whether the damper is removing that energy in a way that produces the desired behavior. The engineer begins investigating low-speed rebound damping. Not because rebound damping is inherently better. But because the chassis appears unwilling to settle after the initial transient.

This illustrates a useful distinction: Springs solve position problems. Dampers solve timing problems. Suppose the engineer increases rebound damping and the platform becomes noticeably more settled. The driver reports that the car feels calmer and more confidence inspiring in the fast corner. Success? Maybe.

A few sessions later the driver returns with a new complaint.

“The car doesn’t want to rotate into slower corners anymore.”

The original problem was reduced. A new compromise appeared. Why? Because the damping adjustment didn’t eliminate motion. It redistributed motion. The chassis now holds attitude longer. That improved stability in the fast corner, but it also reduced some of the transient weight transfer and rotation that the driver had previously been using to help point the car into slower corners. The engineer hasn’t found the perfect answer. The engineer has simply moved the compromise. Most suspension tuning works exactly this way.

Which brings us back to the knobs.

Low-speed compression damping primarily influences large-scale chassis motions such as roll, pitch, dive, squat, and steering response. Increasing it generally makes the suspension less willing to enter those attitudes. The car takes a set more slowly and load transfer develops more gradually.

Low-speed rebound damping influences how willing the suspension is to leave those attitudes once it gets there. Increasing rebound often causes the chassis to hold attitude longer. Reducing rebound allows it to recover more quickly.

High-speed compression damping largely determines how willing the wheel is to absorb sharp disturbances such as curbs, pavement joints, and bumps. Increasing it generally shifts more of the disturbance into the tire and chassis. Reducing it generally allows more wheel movement.

High-speed rebound damping influences how quickly the wheel recovers after those events. Too little and the wheel may oscillate. Too much and it may struggle to return before the next disturbance arrives.

Notice that none of these adjustments directly change the amount of energy entering the system. Again, the road has already decided that. What they change is where the resulting motion occurs and how quickly that motion develops. The driver knows he’s entering Turn One. The engineer knows he’s entering Turn One. The damper doesn’t know it’s in a corner - it only knows its shaft velocity. And perhaps that’s the most important thing to understand about damper tuning. The spring determines the positions available to the chassis. The damper determines how the chassis moves between those positions. Everything else is negotiation.

Mention golf ball dimples around racers, cyclists, engine builders, or aero nerds and within minutes you’ll hear:

“those definitely work”

“those definitely don’t”

“F1 would use them if they mattered”

“MythBusters proved it”

“textured ports make power”

“textured ports kill flow”

Those comments usually come from people who are all at least partially right and partially full of it at the same time. That’s what makes this topic dangerous to write about. Because the second you discuss dimples, roughness, or surface textures, people immediately want a binary answer: do they work or not? But fluid dynamics absolutely refuses to behave in that kind of emotionally satisfying way. Facts > feelings, and all that…

A golf ball absolutely works because of its dimples. That part is real. The problem is that people then copy that geometry without understanding the scaling laws and instability mechanisms that made the original geometry work in the first place. And that’s how you end up with things like dimpled cycling water bottles with “laminar” molded into them, dimpled intake ports, dimpled diffusers, textured cycling helmets, clay covered Ford Tauruses on TV shows, and racers quietly trying things they usually won’t fully admit publicly until after a couple beers, because deep down, everybody suspects there’s something real hiding in there. And there is.

The problem is that the geometry itself is usually not the important part. The governing flow regime is. Which means this article is either going to clarify things slightly…or irritate people from multiple disciplines simultaneously. Which probably means we’re on the right track.

Reynolds Number and the Golf Ball Cheat Code

One of the most persistent myths in aerodynamics is the idea that some forms of surface roughness itself reducedrag. People see golf ball dimples work, then immediately try to dimple other things. But the dimples themselves are not the magic. The scaling is. That’s the part people almost always skip.

A golf ball works because it happens to live inside a very weird and very narrow aerodynamic window where a rougher surface can actually produce lower total drag than a smooth one. And that only happens because the golf ball sits near a transition threshold governed largely by Reynolds number:

Re = ρVL / μ

Where:

ρ = fluid density

V = velocity

L = characteristic length (object size)

μ = dynamic viscosity

Or more intuitively: Reynolds number describes how dominant inertia is relative to viscosity. High Reynolds number flow tends to keep moving, separating, shedding vortices, and becoming turbulent. Low Reynolds number flow tends to stay smooth, orderly, and heavily controlled by viscosity. And importantly, Reynolds number is not just speed. Scale matters equally.

A tiny object moving very fast can have the same Reynolds number as a huge object moving slowly, which is why aerodynamic behaviors sometimes scale in very unintuitive ways. A golf ball is small, roughly 43mm in diameter, but moving very fast, often 100–180 mph. That places it roughly around:

Re ≈ 100,000–300,000

Which is extremely important because that is right in the neighborhood where a smooth sphere naturally undergoes what’s called the drag crisis. And the drag crisis is one of the strangest and most counterintuitive phenomena in fluid dynamics.

Normally, increasing Reynolds number increases turbulence and drag complexity. But spheres do something weird. At lower Reynolds numbers, the boundary layer around a smooth sphere stays laminar too long. Laminar flow sounds good, but laminar boundary layers contain relatively little momentum near the wall where the fluid meets the object. As the flow moves around the rear half of the sphere and encounters an adverse pressure gradient (decreasing pressure), the low-energy boundary layer simply gives up and separates early. And once that happens, the wake behind the sphere becomes enormous.

And for “bluff bodies” (think visually non-streamlined shapes) like spheres, the dominant drag mechanism is not skin friction. It’s pressure drag caused by that giant low-pressure wake.

But then something fascinating happens. As Reynolds number climbs high enough, the boundary layer naturally transitions turbulent. Turbulent boundary layers contain more momentum near the wall because the turbulence continually mixes higher-energy outer flow downward toward the surface. That extra energy allows the flow to remain attached much farther around the rear of the sphere before separation finally occurs.

The wake suddenly collapses dramatically in size. And when the wake shrinks, pressure drag plummets. That abrupt reduction in drag coefficient is called the “drag crisis”.

The drag coefficient of a smooth sphere may sit around:

Cd ≈ 0.47

before transition, then suddenly drop toward:

Cd ≈ 0.2–0.3

once the drag crisis occurs. Cutting drag basically in half! Golf ball dimples essentially cheat this transition. The dimples intentionally destabilize the boundary layer early, forcing the sphere into post-drag-crisis behavior sooner and more consistently than a smooth sphere naturally would. Or phrased another way: the dimples effectively shift the drag crisis leftward in Reynolds-number space, so depending on how big a ball is and how fast it’s flowing in a specific fluid, dimples can decrease drag. That’s a much more accurate statement than simply saying “dimples reduce drag.” They reduce drag because they manipulate instability and separation timing.

And importantly, even golf balls themselves are sensitive to geometry scaling. Dimple depth, diameter, density, and spacing all influence where flow transition occurs and how stable the wake becomes. There is no single universal “golf ball dimple.” The geometry is tuned around a very specific Reynolds-number regime and instability behavior.That distinction matters enormously because once people see the geometry work in one case, they start copying the geometry instead of asking whether the governing instability is even the same. That’s where things go sideways.

The Dimpled Water Bottle?

Take bicycle water bottles. I own a few of these curiosities - water bottles with golf-ball style dimples molded into them. A few ironically actually have “laminar” stamped on the side, which is pretty funny considering the dimples exist specifically to disturb the flow into a turbulent condition.

Now to be fair in this case, the Reynolds numbers are not completely absurd for them to be doing something to reduce drag. For a normal ~70mm cycling bottle:

15 mph: Re ≈ 31,000

20 mph: Re ≈ 42,000

30 mph: Re ≈ 63,000

40 mph: Re ≈ 83,000

That at least places it in the broader neighborhood where bluff-body surface texture can potentially matter, especially toward the higher-speed end. But importantly, this is already approaching the lower edge of where classic golf-ball-type drag-crisis behavior even begins becoming strongly relevant.

But unlike a golf ball, the bottle is not operating in clean freestream air. It sits behind the front wheel wake, next to the down tube, near the rider’s legs, in steering yaw, and inside already disturbed incoming air. That changes the dominant problem substantially.

For the bottle, the dominant drag mechanisms are probably bluff-body pressure drag, wake interaction, vortex shedding, and rider/frame interference effects, not clean isolated transitional sphere-type behavior. So the dimples are probably not acting as classic golf-ball transition triggers anymore. The flow may already be partially or fully turbulent before it even reaches the bottle.

However, the dimples might be altering vortex shedding, slightly stabilizing separation, disrupting coherent wake structures, or honestly just functioning as aero-themed marketing texture - all hypotheses that would have to be checked in a wind tunnel. The geometry looks similar. The governing physics probably are not.

The Troxel Helmet Era

The same thing happened with a lot of 1980s cycling aero experimentation too. I own a couple mid-1980s Troxel aero/racing helmets - I love my aero oddities - and it’s basically a physical artifact from the era where designers clearly understood airflow behavior mattered, but before modern CFD and tunnel correlation fully matured. Those helmets often had ridges, channels, abrupt transitions, strange bluff-body shaping, and partial texture experimentation. Looking back now, they almost feel like manifestations of the “golf ball logic” era. Not necessarily wrong. But exploratory.

And interestingly, the Reynolds numbers were at least plausible for transitional behavior to matter. Take a helmet characteristic length around 0.25m.

At 30 mph: Re ≈ 270,000

At 50 mph: Re ≈ 450,000

That’s actually extremely interesting territory because now you are entering and crossing Reynolds-number regions where transitional behavior, separation timing, and wake structure can change dramatically depending on surface shape and perturbation strategy. But again, a helmet is not a golf ball. A helmet operates in yaw, attached to a rider, near shoulders and arms, near already-forming body wakes, with constantly changing head angle.

For the helmet, the dominant drag mechanisms become bluff-body pressure drag, rider wake interaction, separation behavior, downstream wake organization, and yaw stability, not simply isolated sphere drag. And for this specific helmet, the dimples on its surface are beyond the point where the flow will have become unstable or separated in the first place. Upstream discontinuities that “trip” flow may work with the dimples, but who knows?

Modern aero helmets reflect that evolution in understanding. The industry largely moved away from obvious texture experimentation and toward controlled trailing edges, truncated tails, shoulder integration, yaw stability, and direct separation management. In other words, the field matured from “roughness might help” toward “control the instability directly.”

Intake Ports and the Boundary Layer Trap

That same mistake of copying geometry instead of understanding mechanism shows up regularly enough in engine porting too. People discovered years ago that slightly textured intake ports often outperform mirror-polished ports, especially on wet-flow engines. Then the internet collectively concluded: “roughness helps airflow.” Not exactly. Again - wrong conclusion from the right observation.

An intake port is not a golf ball either. It’s internal flow, accelerating flow, pulsatile flow, pressure-wave-driven flow, often carrying suspended fuel droplets. Typical intake-port Reynolds numbers can easily reach Re ≈ 100,000–300,000+ depending on RPM, valve lift, port size, and local velocity. But unlike the golf ball, the flow is already highly turbulent. The dominant mechanisms now become boundary-layer attachment, short-side (curved) separation control, fuel suspension, wet-flow behavior (for port injection), mixture homogenization, and maintaining effective port area.

And now the scaling problem becomes completely different because the relevant dimension is no longer the port diameter itself. It’s the boundary layer thickness inside the port. That’s the important scale in this case.

People love copying literal golf-ball dimples into ports, but most of those dimples are gigantic relative to the active near-wall flow structures actually governing the port behavior. A boundary layer inside a port occupies real area. If you aggressively texture the wall, you can absolutely increase displacement thickness enough that the port effectively becomes smaller.

The irony is that a port can gain energy near the wall while simultaneously losing effective area. Sometimes that trade helps. Sometimes it kills flow. That depends on local velocity, fuel behavior, short-side attachment, valve lift, port geometry, Reynolds number, and where separation is occurring. Again: context. Not geometry alone.

Why Dimpled Diffusers Usually Miss the Point

Diffusers expose the exact same trap. People occasionally try dimpled diffusers because the conceptual logic sounds familiar - turbulent boundary layers resist separation better. And that part is true.

But rear diffusers are already operating in absurdly turbulent flow fields full of wheel wakes, floor vortices, suspension turbulence, pitch changes, ride-height variation, tire squirt, and upstream boundary-layer growth.

Typical diffuser Reynolds numbers are enormous: Re ≈ 1,000,000–5,000,000+ depending on speed and characteristic length. At those Reynolds numbers, the flow transitioned into turbulence a very long time ago. So now the dimples are no longer functioning as transition-control devices. They can’t.

For diffusers, the dominant aerodynamic mechanisms become pressure recovery efficiency, separation resistance, wake management, viscous dissipation, vortex interaction, and stall behavior under ride-height changes. That’s a completely different problem.

The Styrofoam Boat Lesson

Ironically, I think I accidentally learned some version of this as a kid building little pool boats from styrofoam. The weird thing was raw styrofoam worked perfectly fine for model airplanes. Bare foam wings flew great. But for boats, especially once they exceeded some modest speed threshold, the same material became awful - hugely draggy.

One catamaran hull I built was especially strange. The piercing hulls were sanded smooth and felt smooth to the touch, yet once the boat got moving it became unbelievably draggy. It almost felt like the hull wanted to suck itself downward into the water. Then I filled and painted the hull. It was a completely different boat afterward - you could give it a gentle shove and it would glide all the way across a typical indoor school swimming pool if the water was smooth enough. At the time it made no sense to me because the foam already felt smooth. It stuck with me. What I didn’t realize at the time was hydrodynamically, “feels smooth” is basically meaningless. Water is merciless compared to air -driven by a density that is 800X that of air.

Even modest-speed small boats can easily operate around Re ≈ 100,000–1,000,000+ depending on hull length and speed. And unlike aircraft, the dominant drag mechanisms for small boats become wetted skin friction, boundary-layer shear, wave-making drag, pressure recovery, and free-surface interaction - things that are minor, or don’t even figure into objects traveling through air.

The viscous sublayer in water can become incredibly thin, which means microscopic bead structure and pores in the foam can suddenly protrude through the active near-wall region once speed increases enough. And once that happens, turbulence production spikes, wetted drag spikes, pressure distribution changes, wake behavior changes, and the governing flow regime itself changes.

The Actual Lesson

That’s the deeper lesson here. Fluid systems are often nonlinear. A surface is not simply smooth or rough. It is smooth or rough relative to Reynolds number, boundary-layer thickness, viscosity, dominant instability scale, and the actual mechanism controlling the flow. That’s why golf balls work. And why copying golf balls usually doesn’t.

I’ve been thinking a lot lately about internalized confidence while driving, and how strange human learning actually is. Not confidence in the motivational poster sense. More the kind where actions stop requiring constant conscious supervision.

Most very skilled drivers are not actively narrating every single thing they are doing in real time - nowhere close. “Brake now. Turn in now. Catch the rear. Unwind lock. Don’t over-slow entry.” Those types of active coherent word-based thoughts are usually not part of the equation. Some do this - and you see this sort of thing a lot in esports / simulator driving, even at the very highest level, but for most it’s different.

The interesting thing is that people sometimes mistake this absence of conscious linguistic self-commentary for the absence of thinking entirely, but that is very obviously not what is happening. In many cases the brain is likely processing more information, not less. It’s just doing it procedurally instead of verbally. The nervous system is still predicting grip, estimating load transfer, tracking yaw behavior, filtering vibration and tire feedback, building expectations for surface conditions, and continuously correcting errors. It’s just not narrating all of it.

I think that distinction matters a lot, because modern culture tends to privilege the kinds of cognition we can consciously explain. If you can articulate it cleanly, quantify it, put it into a diagram or spreadsheet, people tend to treat that as “real” understanding. But motorsport exposes the limits of that assumption very quickly.

There are drivers who can recite tire models, damper curves, and slip angle theory while struggling to execute consistently at the limit. There are also drivers who could not derive a tire force equation if you put a gun to their head, yet somehow immediately understand what a chassis wants underneath them. Neither of those people is “fake”, or “wrong”. They are simply operating with different mixtures of explicit and procedural cognition. And before anybody takes this into anti-engineering territory, that is not where I’m going with this at all.

Physics matters enormously. Data matters enormously. Explicit understanding matters enormously. In fact, some of the best drivers I know are extremely technical people. But what fascinates me is that high-level execution often relies heavily on compressing analytical understanding into lower-latency predictive systems that operate below conscious narration. In engineering terms, it is as though repeated exposure trains a massively parallel predictive controller that eventually stops requiring constant top-level supervisory input. There’s a parallel between this and how AI functions, but I digress, and that’s a story for another day…

Licensing Day

Years ago, when I took my Michigan motorcycle endorsement test, I volunteered to go first during the riding assessment. Very “me” behavior. In front of 30 people no less. On the very first exercise, I put my foot down unnecessarily and lost three points immediately. You could only lose fifteen.

I was so frustrated and irritated with myself over making such a dumb mistake that something interesting happened afterward. I flipped a switch. I basically stopped consciously thinking about the test entirely and just rode the motorcycle. The rest of the assessment became almost automatic, and I ended up with a perfect score beyond the initial penalty, along with compliments from the instructors about my bike control.

What fascinates me looking back is that the skill was obviously already there before I put my foot down. The mistake did not magically create the competence. If anything, it may have removed part of the cognitive overhead surrounding the competence. I stopped auditing myself in real time. I stopped trying to consciously prove I could ride the motorcycle, and simply rode it.

When Thinking Gets in the Way

I think almost anybody who has driven at a high level has experienced some version of this. You have sessions where you are trying incredibly hard to “drive correctly,” consciously monitoring every action, every brake point, every steering input, every tiny mistake. Then you have other sessions where you simply settle into rhythm and suddenly become smoother, calmer, and faster. Ironically, and most often, the harder you try to consciously supervise every detail during execution, the more bandwidth you consume doing so.

This becomes especially interesting from a systems perspective because conscious verbal reasoning is relatively slow and serial in nature. High-speed vehicle control is not. Vehicle dynamics are massively parallel and continuously evolving. Tire force generation, load transfer, transient damping behavior, aero balance migration, yaw acceleration, steering phase lag, brake release timing, and surface friction estimation are all evolving simultaneously and continuously. A driver consciously narrating every one of those processes in real time would be overwhelmed almost immediately. So instead, humans build compressed internal models. That is the part I find fascinating.

Building Internal Models

What intrigues me most is watching elite drivers at unfamiliar tracks or in unfamiliar cars. Some of them seem to almost immediately understand not only where the track goes, but how the car is likely to behave there. Different pavement, different camber, cold grip, standing water, different tire construction, aero balance changes, surface texture, transient understeer on entry, the way a curb loads the chassis, how quickly a rear tire “takes a set.” They often intuitively know. I am not particularly good at this…hence the fascination. I instead ease up to it, and it takes me more repetition and experimentation than is ideal when coming up to speed at most tracks.

I do not think this is a “sixth sense”, magic, or mysticism, or even a “gift” or “talent”. I think it is highly trained predictive cognition operating below constant conscious narration. And the level at which people do this varies by the individual and even by the specific situation in each individual. We all know competent drivers that cannot left foot brake, or heel / toe, or trail brake very well, or who struggle in the wet. The newer driver often experiences driving as isolated instructions in most situations: brake here, turn here, track out there. The elite driver often appears to perceive the entire vehicle as a continuously evolving dynamic system.

That difference matters. The really good ones are not merely memorizing corners. They are recognizing relationships and invariants between systems. How weight transfers longitudinally and diagonally. How tire load sensitivity alters available grip. How quickly the chassis accepts a yaw moment. How certain surface textures correlate with μ variation. How steering release timing changes rear tire workload. How curb geometry excites suspension modes. How aero platform migration changes front response at speed. And importantly, many of them cannot fully explain what they are doing. That part fascinates me too. Humans are capable of becoming deeply competent at things they do not fully understand symbolically. Sometimes the nervous system learns first, and language catches up later.

Why Some Fast Drivers Cannot Coach

This is also why elite athletes are not automatically elite coaches. I was just discussing this with a motorcycle racer transitioning into cars that I was instructing at a track day this past week, and he described a recent coaching experience with a “fastest guy at the track” type who could not successfully coach almost anyone - himself included - a guy that I personally found IMMENSELY coachable - one of the best students I’ve ever had. And as odd as that sounds, it makes perfect sense to me.

Some people learn through equations and telemetry. Some through sensory feel. Some through visuals. Some through rhythm. Some through repetition. Some through metaphor. A good coach is often less about possessing knowledge and more about being able to translate between cognitive frameworks. And honestly, genuinely effective coaching requires a level of empathy that I think often gets overlooked in technical environments. Good coaches find common ground. They figure out how the other person internally models the problem, then build a bridge from there. They are not merely transmitting information. They are translating cognition.

And that also requires humility. I think genuinely effective coaching is fundamentally rooted in the empathy I mentioned above. Not softness, but the ability to understand how another person is perceiving the problem. The ability to see things from their perspective, be “in their shoes”. Good coaches find common ground. They understand that different drivers process information differently, build confidence differently, and internalize vehicle behavior differently.

Toxic or, more generally, ineffective coaches often do the opposite. They leverage ego. They want to impress. They can confuse intimidation with authority and assume that demonstrating superiority is the same thing as teaching. Effective coaches are usually much more humble than that. Their focus is not on displaying expertise. It is on helping another person build understanding and confidence.

The best coaches I have personally encountered are usually not trying to impress the student with how fast they are or how much they know. They are trying to understand where the student currently is, what signals they respond to, what fears or misconceptions are interfering, and what language actually allows the student to progress.

In some cases skilled drivers have never had to consciously deconstruct their own process in the first place, so frustration emerges when another driver cannot simply “feel it” the same way they do. But effective coaching is rarely about making the coach look impressive. It is about making the student feel understood. I think this is also why certain coaching phrases sound almost absurdly vague from the outside.

“Wait for the car.”

“Let the rear come to you.”

“Feel the tire take a set.”

“Don’t rush the platform.”

Those statements sound imprecise, but they are often attempting to interface with embodied cognition rather than symbolic analysis. They are describing highly dynamic systems that operate faster than language comfortably does.

And yet the engineering side still matters enormously. This is where people sometimes oversimplify things into false binaries: feel versus science, intuition versus engineering, natural talent versus technical understanding.

I do not think that is what is happening at all. The strongest drivers I know usually possess both: deep procedural intuition during execution, and analytical frameworks during reflection and refinement. Yet there are a very many people who find themselves in the position of driver coaching, both informally and formally, who cannot let go of that binary. These are the guys who “drive around setup issues”, or alternatively, fixate too much on 0.1deg of camber change as the solution to an issue.

The analysis builds and sharpens the model. The procedural system executes the model. Conscious intervention ramps back upward when anomalies appear, conditions change, or learning stalls.

The Sim Racing Question

This whole topic has also made me think about something else I have noticed in recent years: the increasing number of young drivers who transition from esports and sim racing into autocross and track driving with shocking speed and confidence. As somebody who learned the “old way” first through real vehicles, that progression is fascinating to me.

Historically, many drivers implicitly treated tactile sensation as the foundational language of driving. Tire vibration, steering effort, vestibular loading, chassis movement, fear response, engine noise, and physical consequence were inseparable from the learning process itself.

Then sim racing arrives and younger drivers begin demonstrating something deeply interesting: many high-level competencies appear to transfer extraordinarily well even when large portions of tactile input are absent. That forces an uncomfortable but fascinating question: How much sensory information is actually required to train high-level predictive driving models?

Sims preserve many of the things that may dominate higher-order driving cognition: visual timing, spatial prediction, line optimization, yaw anticipation, steering phase relationships, braking timing, trajectory management, and iterative experimentation. In other words, they preserve much of the predictive architecture of driving even while stripping away many tactile channels.

That may explain why sim-native drivers can adapt so quickly once they enter real cars. The higher-order predictive framework already exists. They more so simply need to “calibrate” additional sensory channels and consequences. From a controls perspective, it is almost like the state estimator is already trained, and the real-world tactile system becomes an additional refinement layer rather than the foundation itself. That possibility is deeply interesting.

Older generations often built intuition through tactile immersion first, then gradually abstracted generalized predictive models from it. Sim-native drivers may be developing generalized predictive frameworks first, then attaching physical sensation onto them later. Those are almost inverted developmental pathways.

At the same time, I do think it is important not to oversimplify this into “sims replace real driving.” Real-world tactile and vestibular cues still matter enormously at the highest levels, especially near adhesion limits, during transient conditions, in endurance scenarios, and under unusual grip environments. The strongest interpretation is probably not that tactile feedback is unnecessary, but that higher-order predictive frameworks may be trainable independently to a much greater degree than many of us previously assumed.

Signal Purity

I have also thought about this in the context of recent comments from Max Verstappen regarding front wheel drive cars and karting. On the surface, statements like “this doesn’t translate” can sound dismissive or overly simplistic. But I suspect there is something deeper underneath them.

Despite what Max said, karting absolutely teaches valuable things: racecraft, momentum preservation, reaction speed, line sensitivity, proximity comfort, and reflexive correction. But karts also operate under a very different dynamic regime than modern race cars. No suspension, very different yaw behavior, different transient loading characteristics, different tire behavior, different compliance effects, and almost no meaningful aero platform management. Similarly, front wheel drive cars train certain predictive assumptions that may diverge substantially from rear-drive or high-downforce vehicle behavior.

A top-level driver may subconsciously categorize certain learned behaviors as highly transferable and others as potentially contaminating to refined predictive models. That does not necessarily mean those disciplines have no value. Quite the opposite. Many elite drivers came from karting and lower-grip platforms. But transferability may be partial and asymmetric rather than universal. The more elite somebody becomes, the more they may optimize not merely for learning volume, but for signal purity. And maybe that is ultimately what expertise really is. Not the absence of thought. But the compression of thought into something faster than language.

Once you start looking at race cars through that lens, a lot of seemingly strange aerodynamic decisions suddenly make sense. Oval cars, dirt cars, lightweight Caterham and Lotus Seven–style cars, and even Pikes Peak hillclimb cars often operate successfully with far less front aerodynamic authority than modern road racing orthodoxy would suggest they “should” need. The reason is not that they somehow care less about grip. It is that the road itself is already doing part of the work.

That sounds abstract at first, but the underlying physics are actually pretty straightforward. A tire does not know whether load came from gravity, aero, banking, compression, pitch, suspension geometry, or weight transfer. It only knows two things: how much normal force is available at the contact patch, and how much combined work it is being asked to do. The entire vehicle dynamics problem ultimately reduces down to that relationship.

The mistake a lot of people make when discussing front aero is treating downforce strictly as a way to “add grip.” In reality, the front tires are continuously balancing multiple competing jobs at once. They are braking, turning, stabilizing yaw, and absorbing transient load transfer simultaneously. The real question is not simply how much front grip a car needs, but rather how much front tire workload has not already been solved elsewhere by the road geometry, chassis behavior, and vehicle state.

The Tire Only Knows Load and Work

At its simplest level, the tire operates on a finite force budget:

F_available ≈ μ × N

Where:

• μ is tire friction coefficient

• N is tire normal force

At the same time, the tire is being asked to perform both braking and turning tasks:

F_demand ≈ √(F_brake² + F_turn²)

The remaining tire margin therefore becomes:

Margin ≈ (μ × N) − √(F_brake² + F_turn²)

This is the important conceptual bridge: front aero improves tire margin by increasing N, but that is only one possible solution. The road geometry itself can also improve tire margin by reducing the amount of work the tire must perform in the first place.

That distinction matters because people often discuss downforce as though it exists independently from the track surface beneath the car. In reality, the track surface is deeply integrated into the tire workload problem. Banking, grade, compressions, crests, and chassis attitude all alter the relationship between demanded force and available force at the contact patch.

The tire itself does not care where the force came from. It only cares about how much work it must perform relative to the load available to support that work.

Banking / Road Camber Changes the Turning Problem

This becomes particularly obvious when you study oval racing. On a flat corner, the tires must generate nearly all of the lateral force required to turn the car. The lateral acceleration requirement is fundamentally:

F_turn ≈ m × v² / r

Where:

• m is vehicle mass

• v is velocity

• r is corner radius

Every pound of that force must ultimately come from somewhere. On a flat surface, nearly all of it comes through tire slip angle and lateral loading at the contact patch.

But on a banked oval or positive camber turn, the road-normal force is rotated INWARD relative to the chassis. Part of what would normally just support the car vertically is now helping generate inward cornering force:

F_bank_assist ≈ N × sin(θ)

Where:

• N is road-normal force

• θ is banking angle

The banking literally redirects part of the support force toward the center of the corner. The front tires no longer need to generate all of the required lateral force through slip angle alone because the track geometry is already contributing to the turning problem.

That changes the aerodynamic requirement of the car. If the front tires are being asked to do less lateral work, the amount of front aerodynamic compensation required also decreases. This is one reason oval cars can often run surprisingly modest front aero compared to road race cars operating at similar speeds. The banking itself has already reduced front tire workload.

There is also a secondary effect that matters enormously in oval racing: the car often exists in a relatively steady-state yaw condition. The chassis is continuously loaded, continuously rotating, and continuously settled into the banking. The front tires are spending less time “rediscovering” direction during transient left-right transitions and more time operating inside a narrower, more stable force envelope.

That means the front axle is not constantly being asked difficult and varied questions. The track geometry and the steady-state nature of the cornering environment have already simplified part of the problem.

Grade Changes the Braking Problem

Hillclimbs like Pikes Peak introduces a closely related phenomenon, although in a different axis. Instead of primarily changing the turning problem, the mountain changes the braking and longitudinal load problem. An uphill road rotates gravity relative to the chassis, meaning gravity itself begins contributing to deceleration during braking zones.

The gravitational component acting along the road surface becomes:

F_gravity ≈ m × g × sin(θ)

Where:

• m is vehicle mass

• g is gravitational acceleration

• θ is road grade angle

During uphill braking, gravity assists the braking process because part of the vehicle’s weight is already acting opposite the direction of travel. The tires therefore do not need to generate as much longitudinal braking force for the same total deceleration. In practical terms, the front tires are simply being asked to do less work.

That is the deeper connection between oval banking and mountain roads. Banking primarily reduces lateral tire workload. Grade primarily alters longitudinal tire workload. But both are fundamentally the same category of physics problem: the road surface rotates the force environment relative to the chassis.

The important nuance is that the road is no longer acting as a passive flat reference plane. The terrain itself begins participating in the load map of the vehicle.

On hillclimbs with steep grades like Pikes Peak this becomes even more significant because the road is rarely doing only one thing at once. Braking zones may simultaneously involve uphill grade, off-camber surfaces, compression loading, and cresting behavior all within the same corner sequence. The force environment becomes dynamic and heavily terrain-coupled.

The Front Tires Rarely Do One Job at a Time

This becomes especially important once the front tires are operating in combined states of braking and turning simultaneously. Real race cars are almost never asking the front tires to do only one job. Corner entry, trail braking, transient rotation, and stabilization all overlap.

The combined force demand becomes:

F_combined ≈ √(F_brake² + F_turn²)

Any reduction in either component increases remaining tire margin. Banking reduces the turning component. Uphill grade reduces the braking component. Both reduce the total workload burden placed on the front axle.

This is where the idea of front downforce as “grip” begins to fall apart. Front aero is not simply adding capability. It is compensating for tire workload that has not already been reduced elsewhere by geometry, terrain, or chassis state.

Once you start thinking about the problem that way, a lot of aero decisions that initially appear strange begin making much more sense.

Aero Balance Is Contextual

Aero balance is often treated as a fixed target number: 50 percent front, 45 percent front, 52 percent front. In reality, the optimal balance depends heavily on how much front tire workload the road geometry has already removed from the equation.

A smooth road course with rapid transitions, heavy braking zones, and large transient yaw states may require substantial front aerodynamic assistance because the front tires are constantly being asked difficult questions. But on a banked oval, the turning problem is already partially solved by the track itself. The car may also exist in a much steadier yaw state with fewer violent directional transitions, further reducing front tire workload.

Pikes Peak becomes even more interesting because the force environment is constantly changing. Grade changes alter braking demand. Compressions alter normal force. Cresting alters pitch sensitivity. Air density changes continuously with altitude. In that environment, maximizing front downforce is not necessarily the same thing as maximizing front confidence. An overly aggressive front aero platform can become nervous, pitch-sensitive, or over-authoritative as the terrain rapidly changes beneath the car.

This is especially true in lightweight cars because terrain-induced load changes represent a larger percentage of total vehicle mass and tire loading. The road itself becomes a far more active participant in the handling balance.

Duncan Cowper’s DAX and the Pikes Peak Problem

A great example is Duncan Cowper’s DAX Rush Pikes Peak car. By modern road-racing aero standards, the car appears heavily rear biased aerodynamically. The rear wing and rear aerodynamic authority substantially outweigh the front aero package, and at first glance that can seem counterintuitive to people accustomed to modern time attack or GT-style aero balancing philosophies.

But the important question is not “why doesn’t the car have more front aero?” The important question is “how much front aerodynamic compensation does the car actually need given the environment it operates in?”

The answer changes dramatically on a mountain like Pikes Peak.

The car is continuously operating in a force environment where uphill grade reduces braking workload, compressions temporarily add substantial normal force, rapid elevation changes alter pitch attitude, and the low mass of the chassis amplifies terrain-induced load changes. At the same time, Seven-type cars already possess extremely low yaw inertia and naturally rapid rotation characteristics. If the front aero platform becomes too authoritative in that environment, the car can very quickly transition from responsive to nervous.

That means the road geometry itself is already solving part of the front tire workload problem.

The rear wing on a car like Cowper’s DAX therefore is not simply “more rear grip.” In many ways, it acts as a stabilizing anchor for a chassis that naturally rotates easily while already experiencing large terrain-induced front load variation. The aero package does not need to solve the entire handling problem alone because the mountain itself is already participating in the load distribution and tire workload map.

What may initially appear to be an “imbalanced” aero package often makes perfect sense once you stop thinking of front downforce as a universal objective and instead view it as compensation for unmet tire workload.

When the Road Starts Sharing the Work

We often think of aerodynamics as something happening “above” the car while the road merely provides grip. In reality, the road surface itself is deeply integrated into the aero and tire workload equation.

Banking, grade, compression, cresting, pitch, and yaw all alter the relationship between available tire force and demanded tire force. The road is not merely the thing the car drives on. In many situations, it becomes an active participant in how the car generates and distributes grip.

Once you understand that, front downforce stops being a universal objective and starts becoming something far more contextual: a compensation system for whatever tire workload the road, gravity, banking, braking, and chassis behavior have not already solved.

On a normal track day, that process works. The first lap isn’t perfect, but it doesn’t need to be. You get a few laps to sort things out, adjust your references, and gradually move toward the limit.

The problem is that first lap isn’t “neutral”. it’s far from it. It teaches you a version of the car that isn’t quite right. Grip is changing faster than you can process it, pressures are lagging behind what the contact patch is actually doing, and your brain starts building a model based on that early, incomplete information. Usually, that’s fine because you have time to correct it.

One Lap of America - The Crucible

At One Lap of America, you don’t get that time. You show up, you grid up, and you get three laps - and one lap that matters most is that first one, which you do from a dead stop, after a quick sighting lap. Cold tires, cold brakes, no real reference. Whatever model you bring with you is the one you’re going to drive, whether it’s right or not, unless you change your thinking.

For me, the issue I’ve had in One Lap first laps hasn’t really been about outright grip. It’s been about commitment in those first few seconds and corners. Off the line, I’m typically spinning tires to get some heat in them more than I should if optimization of acceleration is the goal. I’m not doing anything obviously wrong, but I’m not committing early enough to acceleration either. Then the first real braking zone shows the same pattern. There’s a slight hesitation, a slight delay, and by the time I turn in, everything is clean but conservative.

Each of those moments is small on its own. On a normal day, they disappear over the next few laps as everything comes together. In a One Lap format, they don’t. That early hesitation effectively sets the ceiling for the rest of the run, and there’s no opportunity to claw it back later.

Rewiring Your Driving Brain

What changes in that environment isn’t just the pace. It’s the order of operations. In a typical session, you learn the limit first and then commit to it. Here, you have to commit before you’ve really confirmed anything, and then adjust as you go. If your process depends on feeling things out before you trust them, you’re always going to be a step behind.

That’s where the idea of acting without confirmation comes in, and it’s something that shows up in other forms of driving too. In autocross, you don’t get an out lap either. You get a course walk, maybe a visual reference, and then you’re expected to be near the limit from the first input. Over time, that builds a different habit. You’re not waiting to feel grip and then reacting to it. You’re working from an expected level of grip and operating inside that, adjusting in real time as the run unfolds.

Track driving tends to reward the opposite approach. You build into it, you let the car come to you, and you refine your inputs over multiple laps. That’s a good skill, but it doesn’t translate directly when you remove the ability to iterate.

Cold tires get blamed for a lot in these situations, and there is less grip available early on. The carcass is stiffer, the surface isn’t as compliant, and the peak is lower. But the bigger issue is that the window is tighter and the feedback is less clear, which pushes you toward hesitation. That hesitation ends up costing more time than the reduced grip itself.

There are really two ways to give time away in that first lap. You can overstep the available grip and have to catch the car, or you can never quite get to the available grip in the first place. Most drivers are focused on avoiding the first problem, but in a One Lap format, the second one is usually where the bigger loss is.

What has worked better for me, and what I’m still trying to get more consistent with, is shifting away from the idea of starting cautious and building toward aggression - but in a measured way. Being to confident too early has consequences too if you’re not reading the car or track accurately.

A better way to think about it is committing early, but controlling how quickly you get all the way to the edge. The inputs are still deliberate. You brake with intent, just a little earlier than you would once everything is fully up to temperature. You turn in cleanly without waiting for the front end to “prove” itself. You go to throttle early enough that it matters, but you don’t spike it.

The Goal Isn’t to be Perfect Right away. It’s to Avoid Being Late.

That’s an important distinction, because “being more aggressive” is easy to say and not very useful. What actually helps is making your inputs predictable and repeatable so the car is easier to trust before it’s fully come in. It also helps to use normal track days a little differently. Instead of writing off the first lap, you can treat it as if it matters and practice engaging earlier, even if you’re not pushing all the way to the limit.

It also helps to think in terms of an envelope you can access immediately rather than the absolute limit. If you know you can operate at something like 85 to 90 percent right away, that’s usually enough to be competitive in a One Lap scenario without asking questions the tire isn’t ready to answer yet. From there, you’re making small adjustments rather than trying to discover the car from scratch.

When I think back to my own runs over the two One Lap events I’ve competed in, the places I’d change are pretty straightforward. I’d get off the line with more intent instead of easing into it, or spinning the driven tires too much to generate “safe” heat , and I’d commit more decisively to that first braking zone and turn-in instead of waiting for confirmation that the grip is there. Those aren’t dramatic changes, but they happen early enough that they shape the entire lap. Don’t tip toe - be decisive, and be able to monitor internally how close to the limit you will be in those early moments

On a normal track day, the out lap tends to understate how much grip the car really has, and you correct for that as the session goes on. In something like One Lap of America, your instincts tend to overstate the risk early on, and there’s no time to recalibrate.

The underlying problem is the same in both cases. You’re making decisions based on incomplete early information. The difference is that in one case, you have time to fix it, and in the other, you don’t.

I know what I’m supposed to do in that first lap, and I know the car has more in it than it feels like it does. The part that still takes work is acting on that right away instead of waiting for it to confirm itself. In a format like this, the meaningful time isn’t lost at the limit. It’s lost in the first few seconds, before you ever really get there.

There’s a persistent idea floating around car culture that braking performance scales with hardware. More pistons. Bigger calipers. Larger pads. It feels intuitive. More stuff should equal more stopping power. But brakes don’t work like engines. They’re not power generators. They’re energy converters. And that distinction changes everything.

The First Truth: Your Tires Set the Limit

Before we talk about calipers, pistons, or pad size, we need to establish the hard ceiling:

If you can lock the tire, you already have enough brake torque. That’s it.

Once you’re at threshold braking, the system is no longer limited by the brake hardware. It’s limited by tire grip. So when people say “these brakes stop harder,” what they usually mean is that they’re easier to modulatE, they resist fade better, and they feel more confidence-inspiring. Not that they generate fundamentally more stopping force.

The Two Conversions (Going in Opposite Directions)

Braking systems are interesting because they perform two different conversions, in opposite directions:

1. Mechanical → Hydraulic (Pedal to Line Pressure): You push the pedal → the master cylinder converts force into pressure

2. Hydraulic → Mechanical (Pressure to Clamp Force): That pressure acts on caliper pistons → creates clamp force on the rotor

Then:

3. Mechanical → Thermal (Clamp + Friction → Heat): The rotor converts kinetic energy into heat

Hydraulics Flow:

In an automotive braking system, the different variables work like this simplified flow:

Pedal Force → Lever Ratio → Booster → Master Cylinder → Line Pressure → Caliper Piston Area → Clamp Force

And the key takeaways are :

Smaller master cylinder → higher pressure, more pedal travel, and

Larger caliper piston area → more clamp force, more fluid volume required

So you are always trading force vs travel vs volume - You don’t get all three.

Piston Count vs Piston Area

Hydraulically, piston count doesn’t matter. Period. Total piston area is what determines clamp force. Clamp Force = Pressure × Total Piston Area. So a 4-piston and a 6-piston caliper with the same total piston area will produce the same clamp force. That’s not where the real difference comes from.

Why More Pistons Actually Exist

Multi-piston calipers solve a different problem - they distribute force across the pad. Large pads need even pressure to prevent taper wear, reduce pad flex, and maintain consistent friction behavior. That’s it. More pistons are about control and distribution, not raw braking power.

Pad Area: What It Does (and What It Doesn’t)

Pad area is one of the most misunderstood variables in braking. It does not increase friction force. Friction force is:

Friction = μ × Normal Force

Not area.

So increasing pad area does not increase torque directly. What it does do is it reduces pressure (force spread over larger area), slows temperature rise, improves thermal stability, and increases pad life.

Torque Comes From Somewhere Else

Brake torque, or more simply put - the torque that a brake system produces is:

Torque = Clamp Force × μ × Effective Radius

and there are three independent variables : Clamp force (hydraulics), μ (pad compound), and radius (rotor diameter / pad position)

This is where things get interesting. You can increase torque by increasing clamp force, increasing friction coefficient, or increasing rotor radius. But these don’t behave the same way.

increasing clamping force is the easy one - more force at the pedal means more force on the rotor.

Increasing radius increases leverage. It also reduces required clamp force, and inherently improves thermal distribution because the brake rotor will naturally be larger - all things equal.

Increasing increasing μ generally increases sensitivity of the pad to rotor interface (but not always), introduces changing behavior with temperature (more “racy” pads don’t like cold temperatures, and can sometime introduce instability and inconsistencies because of these things.

looking at it from an outside in perspective, you could create the same torque outcome, completely different system behavior adjusting these variables…

The Real Problem: Heat

Brakes don’t fail because they lack force. They fail because they run out of thermal capacity. Every braking event converts kinetic energy into heat:

Energy = ½ m v²

That energy has to go somewhere. If it isn’t absorbed, distributed, or rejected you get fade.

Two systems can produce identical braking force but one can survive lap after lap ehile the other falls off a cliff. Why? Same force. Different thermal outcome. That’s why “street brakes feel fine” — until they don’t.

Where Pad Size Actually Matters

Pad size plays directly into thermal behavior.

Small pad:

• higher pressure

• faster heat buildup

• quicker response

• less stability over time

Large pad:

• lower pressure

• slower heat buildup

• more consistent behavior

• longer operating window

So the trade-off is response vs stability, not braking power.

The Big Brake Myth

Most “big brake kits” don’t increase stopping distance performance. Because stopping distance is tire-limited. What they do is improve consistency, reduce fade, improve modulation, increase thermal headroom. That matters on track. It barely matters on the street. If you see a car with huge brake rotors on it on the street, it’s because the designers are considering uses of the vehicle off the street.

What Actually Makes a Faster Car

If you’re optimizing a braking system, the order matters - Tire grip, Brake bias, Pedal feel / modulation, Thermal capacity, Weight and inertia….then Everything else. If you think about it in this order, you’ll do far better in life.

oh, and piston count is way down the list.

Where the Internet Gets It Half Right

You’ll see a lot of comments on line that have some partial truths to them, but lack context.

• “More pistons feel better” → yes (stiffness, distribution)

• “Pad area helps heat” → yes

• “Hydraulics only care about area” → correct

• “Friction doesn’t depend on area” → mostly correct

But the missing link in most of these confident sounding comments is this: these variables don’t live in the same layer of the system. That’s why debates go nowhere. But now you know…

Technical Appendix:

Hydraulics:

Pressure = Force / Area (master cylinder)

Clamp Force = Pressure × Caliper Area

Tradeoff:

• Smaller M/C → higher pressure, more travel

• Larger caliper area → more force, more fluid displacement

Torque:

Torque = Clamp × μ × Radius

Thermal:

Heat per stop ≈ ½ m v²

Temperature rise depends on:

• rotor mass

• pad area

• airflow

• material properties

Pad pressure:

Pressure = Clamp / Pad Area

Higher pressure:

• increases local temperature

• increases wear rate

• can destabilize μ

Implementation Notes - an example:

For a track car - let’s use a random example to illustrate:

• If you can already lock the tire → don’t chase clamp force

• Increase rotor diameter before increasing piston count

• Use pad compound as your primary tuning tool

• Add cooling before adding caliper

• Use pad area for longevity and stability, not power

The Simple Way to Think About It

Brakes don’t make you stop. They let the tires do it repeatedly. Everything else is just how cleanly, consistently, and confidently you get there.

Toward the end of the podcast, they got into tire pyrometers. Dave made a point that stuck with me. He’s moved away from relying on them, especially after seeing real-time IR data from tires on track. His takeaway was simple. Tire temperatures vary so much, so quickly, that trying to measure them in the pits can be misleading, regardless of how quick or careful you are. His preference now is to focus on pressure instead.

At first glance, that sounds almost backwards. Tire temperature is the fundamental metric. As discussed here previously, every tire has a temperature window, and getting into that window is what unlocks grip. So walking away from temperature sounds like ignoring the thing that actually matters. But the more you sit with it, the more it makes sense. And also, the more you realize what’s missing.

The key is understanding what problem he’s actually solving. If you look at modern tire behavior, especially with real-time IR, the clean mental model most of us grew up with starts to fall apart. We tend to think of a tire as something that reaches a steady temperature across the tread. Maybe a little hotter on the inside or outside depending on camber, and caster (or toe) but broadly stable. Something you can measure, write down, and use to make a decision. But that’s not what’s happening.

Surface temperatures spike and fall corner by corner. The contact patch isn’t uniform, it’s dynamic. Load moves. Slip moves. Energy input isn’t evenly distributed. You don’t get a smooth gradient, you get localized hot spots that appear and disappear as the car goes around the track. By the time you come into pit lane and stick a probe in the tire (even deeply), the system has already started to decay. Heat is moving. The sharp gradients are softening. What you’re measuring is already a blend of what just happened and what’s currently happening.

So the question becomes uncomfortable.

Are you measuring the tire, or just the last corner you drove? That’s the problem Dave is reacting to.

Pressure, on the other hand, doesn’t care about that level of detail. It’s slow. It’s boring. And that’s exactly why it works. Pressure is an integrated signal. It reflects the bulk carcass temperature over time. It carries the history of the run. Every braking zone, every long corner, every moment of slip energy input gets averaged into a single number. That averaging filters out the chaos. It gives you something stable enough to tune against when everything else is moving too fast to measure cleanly. That’s why pressure works so well in the real world. Not because it’s more precise, but because it’s more usable.

And if you track it the right way, especially delta from ambient, monitoring pressures becomes a powerful tool. Cold pressure sets the initial condition. Hot pressure reflects the system response. The difference between the two tells you how much energy the tire absorbed over the run. And that pressure rise gives you a strong indication of whether you’re operating your tires in the proper temperature window. That’s a clean, repeatable signal.

What Pressure Doesn’t Tell You

Pressure isn’t JUST telling you about temperature. A tire isn’t only a thermal system. It’s a spring and a damper, operating in series with the rest of the chassis. Pressure is the primary adjustable factor that sets that spring behavior trackside. Change pressure, and you’re not just shifting where the tire sits in its thermal window. You’re changing how it supports load, how it deflects, how it transfers force into the chassis. That feeds directly into what the driver feels. Support on entry, compliance over bumps, how quickly the tire takes a set.

Temperature still plays a role. It influences carcass stiffness and internal damping characteristics. But those effects are secondary compared to what pressure is doing mechanically. So when you adjust pressure, you’re doing two things at once. You’re moving the tire thermally. And you’re tuning a spring in the system. That’s part of why pressure becomes such a dominant tool at the track. It doesn’t just describe the system. It actively shapes it. But that still isn’t the whole picture.

The Life and Times of a Racing Tire:

Pressure tells you where the tire lives. It does not tell you how it lives there. And this is where things get real, both from an engineering standpoint and from the driver’s seat.

Take a high horsepower front wheel drive car that struggles to put power down on corner exit. You’ll often see the inside front spin under load. From the driver’s perspective, it feels like the car just won’t hook up. You feed in throttle and the car doesn’t go where you want it to, even though nothing feels dramatically wrong mid-corner. From a pressure standpoint, that tire can look completely normal, or even slightly elevated. It’s working. It’s generating heat. But that heat isn’t coming from productive grip. It’s coming from slip.

If you look at temperature across the tread in these cases, you’ll often see uneven or localized heating patterns. Not always clean, not always perfectly repeatable, but enough to suggest the tire is doing work in the wrong way. Pressure tells you the tire is busy. Temperature helps you see whether that effort is useful.

The same thing shows up in more traditional setup problems. You can have a car with a camber issue that still produces a reasonable pressure rise over a session. On paper, everything looks fine. But the temperature profile starts to show the imbalance. One side of the tread consistently hotter than the other. The contact patch isn’t being used evenly, even though the overall energy in the tire looks right. From the driver’s seat, this often shows up as a car that feels inconsistent. Grip is there, but it isn’t stable or predictable. Again, same integrated signal. Different underlying behavior. What it really comes down to is that within a tight window of parameters, pressure can tell you a lot of things, but without temperature, you’re always guessing at some level. Experience helps here, but don’t through away your pyrometer just yet…

Your Eyes as a Tire Tuning Tool

And then there’s the part that no sensor replaces. You still have to look at the tire. ChatGPT and Claude are getting better at this sort of thing, but once you know what you’re doing, you can always trust what is likely your most important human sense

Wear patterns, surface condition, buildup, pickup, tearing, graining. These are all signals about how the tire is interacting with the track and how it’s being driven. A tire that’s building up rubber in patches is telling you about localized slip. A tire that’s tearing is telling you about energy input relative to compound window. A tire that looks clean and even is often telling you the system is working together. From the driver’s perspective, this is often the difference between a car that feels connected and one that feels like it’s fighting itself. Sometimes the most useful signal is still right in front of you.

So What Should I Be Doing?

The real workflow isn’t pressure versus temperature. It’s layered. You use temperature to understand how the tire is working. You use pressure to keep it in the window. And you use your eyes to sanity check both. Ignore any one of those, and you lose part of the picture.

Let’s Get Philosophical

There’s also a deeper shift underneath all of this. The more chaotic the system, the more valuable integrated signals become over time. Not because they’re more accurate, but because they survive noise. They give you something stable enough to act on when everything else is moving too fast to measure cleanly. Pressure carries memory. It reflects what the tire experienced over the entire run. Accuracy doesn’t win on pit lane. Stability does. But stability has limits.

Integrated signals tell you where you’ve been. They don’t always tell you when something is changing in real time. A tire starting to fall off, a balance shift mid-stint, a change in driving style that adds new energy into the system. Those can lag. So even here, you don’t get to be passive.

If there’s a clean way to frame all of this, it’s this. Pressure is the best tool for controlling a tire. Temperature is the best tool for understanding it. And you only get to rely on pressure alone once the tire has nothing left to teach you. Some people skip that step. They go straight to pressure because it’s clean and repeatable. But pressure only works cleanly when the tire is already behaving the way you think it is. If it isn’t, you’ll still get a number. It just won’t mean what you think it does every time.

What This Looks Like Between Sessions

All of this sounds clean when you write it out. It’s a lot messier when the car comes in hot and you’ve got a couple minutes to decide what to do next. Most of the time, I’m not chasing numbers. I’m trying to answer a few simple questions as quickly as possible. Did the tire live where I expected it to live? That’s pressure. Hot versus cold, and more importantly the change between them. If that number is off, I’m not thinking about anything else yet. The tire didn’t operate in the window I thought it would, so any feel I got from the car is already suspect. If the pressure looks right, the next question is whether the car felt like it was using the tire correctly.

That’s where driver feel starts to matter. Did the car take a set cleanly? Did it hold on consistently, or did it fall away in certain phases of the corner? Did it put power down the way it should? If something feels off, I don’t ignore that just because the pressure looks right.

That’s when temperature becomes useful, if it’s taken quickly enough to mean anything. Not to chase a specific number, but to look for a pattern. Is the tire working evenly across the tread, or is it concentrated somewhere it shouldn’t be?

And then, regardless of what the data says, I look at the tire. Not casually. Intentionally! Is there pickup? Is it building up in patches or evenly? Is the surface clean, or does it look like it’s being overworked? Is there any sign that the tire is sliding more than it should be? Those visual cues tend to confirm or challenge whatever story the numbers are telling.

Most of the time, the answer isn’t “change everything.” It’s a small adjustment with a clear reason behind it. A pressure tweak to move the tire back into the window. A setup change if the contact patch isn’t doing what it should. Or sometimes nothing at all, if everything lines up.

The important part is that each of those tools has a role. Pressure gives you a stable reference point. Temperature gives you context when something doesn’t add up. Your eyes tell you if the story makes sense. Use them in that order, and you avoid chasing noise.

The temptation is always to simplify. To find one number, one tool, one answer that makes the whole system easier to manage. Pressure gets close to that, which is why it’s so valuable. But tires aren’t simple systems. They never were and never will be. The goal isn’t to reduce them to one signal. It’s to understand which signals you can trust, and when. And more importantly, how much to change them when you decide to act.

Because a half psi isn’t just a number. It changes how the tire supports load, how quickly it takes a set, how it transitions from grip to slip. Small changes, but not small effects. That’s where this really starts to matter.

I received a note from respected instructor and analytical mind Mario Korf, who spends time coaching at both low-speed tracks like Pineview and high-speed tracks like Watkins Glen, after my last article.

The theme was simple: Cars feel looser at speed. We all know it. But what is actually changing, and how should we teach around it?

What follows is a Q&A referencing his well thought out and grounded questions.

Q: Skid pads are low speed but constant turning. Does yaw behavior act the same there as it does at speed? Does constant side loading reduce yaw damping and make it feel more like high speed?

A: No. A skid pad does not behave like high speed, and constant side loading does not reduce yaw damping.

A skid pad is a low-speed environment by design. You simply cannot reach the speeds where high-speed yaw behavior shows up, so the system never enters that regime.

At low speed, yaw rate is higher for a given lateral g, and yaw damping is very effective relative to the motion of the car. The system responds quickly and corrects itself quickly.

A skid pad is steady-state. The car is not transitioning in and out of load, which makes the balance easier to feel because the system is continuously expressing slip.

At high speed, yaw rate drops and yaw damping becomes less effective relative to the motion of the car. The system responds more slowly and takes longer to correct, so small imbalances persist.

So even though the tires are always loaded on a skid pad, it actually feels more communicative, not less.

Q: What increases yaw inertia at higher speeds? Would a longer wheelbase or more caster help with feel?

A: Nothing increases yaw inertia with speed. The effects become more dominant, but it does not increase. Strictly speaking, yaw inertia is fixed. It comes from mass distribution and does not change with velocity. I previously lumped a number of terms into the phrase yaw inertia, but to be strictly technically clear here - what changes is how dominant inertia feels in the high speed regime relative to the rest of the system.

At higher speeds, yaw rate drops and yaw damping becomes less effective relative to the motion of the car. The system is slower to respond and slower to correct, so inertia feels more dominant even though it has not changed.

When I’ve used the term yaw inertia, I am using it as a shorthand for that combined effect. Not just the physical inertia of the car, but how resistant the system feels to being rotated and corrected.

A longer wheelbase increases actual yaw inertia and will generally make the car more stable but slower to respond. More caster improves steering feel by increasing aligning torque, which helps the driver sense the front tires better, but it does not change the underlying yaw dynamics.

Some changes improve stability. Others improve feedback. They are not the same thing.

Q: How do different tires and slip angles factor in? Do slicks have less feel at speed than street tires?

A: No, slicks do not have less feel. But they have a narrower and more precise operating window.

Race tires operate at lower slip angles and respond more quickly. Street tires operate over a broader slip range and feel more progressive.

At higher speeds, slip angle is smaller for a given lateral acceleration because forward velocity is higher. That compresses the range over which the tire communicates.

So a slick can feel like it has less feedback at speed, but what is really happening is that the usable range is tighter and builds more quickly. A street tire spreads that response out, which can feel more forgiving.

Both are subject to the same scaling problem at speed. But with slicks, the system is quieter, not less capable.

Q: Tires get grip from slip angle. Does that grip decrease with speed? Is there less time for the tire to deform into the surface?

A: No, grip from slip angle does not simply decrease with speed. The tire still generates force the same way, but the conditions change.

Slip angle is smaller at higher speeds for the same lateral g, because it is tied to lateral motion divided by forward velocity. That makes the deformation harder to perceive, not less present.

It is not that the tire has less time to deform. It is that the deformation is happening over a smaller angular range.

At the same time, load sensitivity and thermal effects become more sustained. The tire is under load longer and small differences persist because the system is slower to correct them.

So the grip is still there, but it becomes harder to read.

Q: Why can changing tires make a neutral car feel “oversteery” when nothing else changes?

A: Because the tire defines how the car generates and balances grip.

Changing tires changes cornering stiffness, slip angle behavior, load sensitivity, and compliance. If the front and rear do not change proportionally, the balance shifts.

A tire with more rear compliance or slower response will delay how the rear builds force. That can make the car feel looser even if overall grip is lower.

Nothing else changed on the car, but the relationship between the front and rear can and does. Staggered setups, and / or large differences in longitudinal mass distribution can magnify small differences in tires.

Q: How does sidewall stiffness affect yaw damping? Is a taller sidewall more damping or more feel?

A: A stiffer sidewall increases how quickly the tire responds. A softer or taller sidewall slows that response down.

Yaw damping comes from how quickly the tire generates corrective forces when slip develops. A stiffer sidewall builds that response faster, which feels like stronger damping.

A taller sidewall is usually more compliant. It delays and spreads the response, which can feel more progressive but less immediate. So a taller sidewall does not mean more damping. It usually means more delay.

It is not about more or less. It is about response speed and clarity.

Q: Given how people sense traction, what is the best way to teach limit driving? Low speed or high speed?

A: Start at low speed. Apply it at high speed. I’m a big fan of starting on the skid pad, and I really wish more track facilities had them on hand for beginner drivers to start with.

Low speed gives you larger, faster signals. The system responds quickly and corrects itself quickly, which makes it easier to understand what the car is doing.

High speed gives you smaller, slower signals. The system is less effective at correcting itself relative to its motion, which makes it harder to interpret.

Trying to learn at high speed is difficult because the system is quieter. But avoiding high speed entirely leaves a gap, because that is where the scaling effects show up. And learning there also has more inherent consequences of course.

Q: How does camber or banking affect yaw damping and feel?

A: Banking changes how load is carried, not the fundamental yaw mechanics. It reduces how much of the lateral load the tire has to generate on its own, which can make the car feel more stable.

But it can also mask front-to-rear differences because the tire is not working in the same way as it would on flat ground.

So it can either help or hide what the car is doing, depending on the situation.

Bringing it all together

At low speed, the system is fast and strongly self-correcting.

At high speed, the system is slower and less effective at correcting itself relative to the motion it carries.

It starts to feel like yaw effects take over, not because inertia increased, but because correction has fallen behind motion.

That is the whole shift.

Closing thought

The car does not stop telling you what it is doing. It just says it more quietly, and takes longer to fix things when they start to go wrong. Low speed teaches you what to listen for. High speed tells you whether you actually learned it.

That was intentional. Not because it is the perfect textbook term, but because it points at something drivers actually feel. As speed rises, the car stops feeling like a simple, obedient object and starts feeling like a system carrying motion, delay, and memory.

That feeling is real.

But what I wrapped into that term is not just one thing. It is a combination of actual yaw inertia, yaw damping, and the way yaw response itself scales with speed. Add to that the fact that small front-to-rear imbalances persist longer as the car covers more ground, and you get the behavior we are trying to describe. This is where we zoom in.

Yaw inertia, literally, is just rotational inertia about the vertical axis. It is the rotational equivalent of mass. A short, light car with centralized mass rotates easily. A longer car, or one with more mass spread out, resists rotation and feels slower to take a set. That part does not change with speed, which raises the obvious question. If inertia itself is fixed, why does the car feel more inertia-dominated in a fast corner? Because everything around it changes.

The first shift is yaw rate. For steady cornering, lateral acceleration is a_y = V² / R and yaw rate is r = V / R, which combine to r = a_y / V. So for the same lateral g, yaw rate actually goes down as speed increases. A tight, slow corner requires the car to rotate quickly. A fast sweeper at the same lateral g rotates more slowly, even though the car is covering far more ground. That alone changes the feel.

Now bring in yaw damping.

Yaw damping is the tire-generated resistance to yaw motion and slip angle differences. It is what helps the car clean itself up when something starts to deviate. It is the stabilizing side of the tire’s job.

The important nuance is that yaw damping does not disappear with speed, and it does not simply drop as a clean function of velocity. What changes is its effectiveness relative to the motion of the car.

Slip angle is tied to lateral motion relative to forward velocity, so as forward speed increases, the same disturbance produces a smaller angular slip. The tire is still generating force, but with less angular leverage in the way the system expresses yaw behavior. At the same time, the car is carrying more energy and covering more distance.

So the system still has damping, but it is less effective at quickly correcting motion relative to the state the car is in - that is the key relationship. And that leads directly to what you feel from the driver’s seat.

At lower speeds, yaw rate is high and yaw damping is relatively effective. The system is quick to respond and quick to clean itself up. Small imbalances come and go before they really develop.

At higher speeds, yaw rate is lower and yaw damping is less effective relative to the motion of the car. The system responds more slowly and takes longer to settle. Small imbalances do not get cleaned up as quickly, so they persist and build. The car starts to feel like it has memory. That is what shows up as looseness.

It is worth being precise here. This does not necessarily mean the car is more oversteer-biased in a steady-state sense. Many cars are designed to become more understeer-stable with speed. But the feel can still move the other way because the yaw system is slower to respond and slower to correct itself. That distinction matters.

Where roll stiffness fits into this

This is also where roll stiffness starts to matter more than it seemed to at low speed. Roll stiffness does not directly create or remove damping. What it does is control how load is distributed across the tires, front to rear. Since tire grip is load-sensitive, that distribution affects cornering stiffness at each axle, which in turn affects how much yaw moment and yaw damping each axle contributes.

At lower speeds, the system is fast and strongly damped, so small differences in roll stiffness balance are partially masked. The car responds quickly enough that those differences do not have much time to express themselves.

At higher speeds, where yaw damping is less effective relative to the motion of the car, those same differences persist longer. A slightly rear-biased load distribution, or a rear axle that builds slip angle more quickly, will now show up more clearly because the system is slower to correct it.

So roll stiffness is not changing the fundamental physics of yaw damping. It is influencing how the front and rear axles contribute to it. At speed, when the system has less ability to clean things up quickly, that influence becomes much more visible.

That is why a setup that feels neutral in a slow corner can feel loose in a fast one without anything actually changing.

Technical Breakdown

For steady-state cornering:

a_y = V² / R

r = V / R

Substitute R:

r = a_y / V

So for a fixed lateral acceleration, yaw rate decreases as speed increases.

Example at roughly 1.2g (≈ 38.6 ft/s²):

At 52 mph (76.3 ft/s): r ≈ 38.6 / 76.3 = 0.506 rad/s ≈ 29.0 deg/s

At 73 mph (107.1 ft/s): r ≈ 38.6 / 107.1 = 0.360 rad/s ≈ 20.6 deg/s

At 90 mph (132.0 ft/s): r ≈ 38.6 / 132.0 = 0.292 rad/s ≈ 16.7 deg/s

Same lateral g, lower yaw rate at higher speed.

Yaw dynamics can be represented in simplified form as:

M_z = I_z × ṙ + C_r × r

Where M_z is the net yaw moment from the tires, I_z is yaw inertia, ṙ is yaw acceleration, and C_r represents yaw damping.

The inertia term resists changes in rotation. The damping term resists ongoing motion and helps stabilize the system.

The important point is not that C_r disappears with speed, but that the effectiveness of yaw damping is tied to slip angle, and slip angle scales with lateral motion divided by forward velocity. As speed increases, the same disturbance produces less angular slip, which reduces the system’s ability to generate strong corrective yaw response per unit motion.

So while I_z remains constant and C_r still exists, the balance between stored motion and corrective capability shifts.

Visual Interpretation

The chart above represents this as a single trend. As speed increases at a fixed lateral g, yaw response slows while yaw damping becomes less effective relative to the motion of the car. The curve represents the resulting reduction in correction authority, normalized for clarity.

At lower speeds, the system is quick to respond and quick to correct. At medium speeds, response slows and correction begins to fall behind. At higher speeds, the system is slow to respond and slower to correct, which is what shows up as looseness. The exact shape of the curve is illustrative, but the trend is the point.

Closing thought

The faster corner is not asking the car to rotate more. It is asking it to rotate more slowly, for longer, while carrying more energy and relying on a system that is less effective at correcting itself relative to that motion. Once you see it that way, the feeling stops being mysterious. It becomes expected.

If you’ve felt it, you know exactly what I mean. The same car, same tires, same lateral g can feel planted and confident in a hairpin, then take on a vague, floaty, almost nervous character in a long sweeper. Nothing changed on paper, but everything changed in how it feels.

Aero is parked nicely in that compact car spot for this conversation. That’s a whole different layer.

What we’re talking about here is purely mechanical grip, and how the balance of influences shifts as speed and corner radius increase, even at the same lateral acceleration

At low speed, steering geometry is doing a lot of the work. Large steering angles, high front slip angles, short duration. The front tires are clearly in charge. You feel weight in the wheel, the front bites, the rear follows along without much to say. The car feels decisive, almost heavy on its nose.

As the corner opens up, steering angle drops away and that strong geometric contribution fades. The car starts to feel lighter, but not necessarily better. That’s where tire load sensitivity starts to lean in.

As speed rises, lateral load transfer becomes more influential, and tires play their favorite trick. The outside tire doesn’t gain as much grip as the inside tire loses.

In a tight corner, this effect is brief and somewhat masked. In a long, fast corner, it has time to fully develop. The rear axle, which was mostly along for the ride in the hairpin, now carries sustained load imbalance.

That subtle shift is enough to move the balance toward oversteer. Not because anything changed mechanically, but because the distribution of usable grip changed.

Carcass deflection is always part of the mix, but it starts to speak up more at speed. Slip angle exists because the tire is deforming. At low speed, it’s transparent. At higher load and longer duration, that deformation builds and persists.

Now the rear tires are no longer quiet passengers. Their slip angles grow closer to the front, and the car starts to feel like it has stored movement in it. Not instability, but potential. You feel that as looseness.

Then there’s heat.

At low speed it’s basically invisible. At track speeds, especially in long corners, it matters. Energy goes into the tire continuously, and the grip you get back starts to evolve even within a single corner.

Whether you’re climbing up the grip curve or falling off it, the key is this. The longer the corner, the more the tire’s behavior drifts away from its initial state. That drift tends to show up more on the rear axle because of the sustained load it’s carrying.

And finally, the big one. What I loosely call yaw inertia.

That’s not strictly the textbook term. It’s a useful umbrella for how the car behaves in yaw at speed. It’s really a combination of a few things working together: the car’s resistance to rotation, yaw damping from the tires, and how yaw response itself scales with speed.

In a quick, tight corner, this whole system barely has time to matter. You turn the wheel, the car rotates, and it’s over before anything can build. But in a long, fast corner, all of it shows up.

For the same lateral acceleration, yaw rate actually decreases as speed increases. The car is covering more ground while rotating less per unit time. At the same time, the tire’s ability to clean up small slip angle differences, what you can think of as yaw damping, becomes less effective in practice.

So now you have a car carrying a lot of energy, rotating more slowly, and correcting itself less aggressively. That’s where the looseness comes from.

The car isn’t just turning anymore. It’s sustaining a state of motion, and any small imbalance between front and rear has time to persist, build, and become something you can feel.

And to be clear, this doesn’t necessarily mean the car is mathematically oversteering more in a steady-state sense. Many cars are actually designed to become more understeer-stable with speed. But it feels like oversteer because the system is slower to respond and slower to correct itself.

Put all of that together, and you get a clear picture. At low speed, the car is dominated by geometry and immediate tire response. At high speed, it’s dominated by load sensitivity, sustained deformation, thermal effects, and slower, less damped yaw behavior. Same lateral g. Completely different system.

I built a chart to visualize this, stacking each of these contributions across corner radius. For reference, I dropped in Gingerman T10b, Gingerman T1, and Mid-Ohio T1 so you can anchor it to something real. Get this model in your head, and the next time the car starts to feel loose in a fast corner, it won’t feel random. It’ll feel inevitable.

If you want to go one layer deeper, the next piece is roll stiffness distribution and how it feeds directly into that load sensitivity term. That’s where setup either fights this behavior or makes it a lot worse. I might do the next article on that topic if there’s enough interest here…

Every car on track is moving through a fluid, and at any meaningful speed, that fluid is not a small part of the problem. Aerodynamic drag is typically the dominant resistive force once you get out of low-speed regimes, and even relatively small changes in flow conditions can materially affect the forces acting on the car. Yet in most grassroots data systems, we don’t measure anything about the air directly. We log the car’s response to it, but not the input itself. So what we end up with is a dataset that is internally consistent, but externally ambiguous.

—

If you start from first principles, the structure is simple enough. Longitudinal force is:

F = m · a

and the corresponding power at the wheels is:

P = F · V

If you expand that into something closer to what’s actually happening on track, you get a balance of forces:

m · a = Fₜ − F_drag − F_roll − F_grade

For most track scenarios, grade (slope of thetrack surface) is negligible, so what you’re really looking at is the interaction between tractive effort, aerodynamic drag, and rolling resistance. The problem is that drag is not just some constant offset. It scales strongly with velocity, and more importantly, with velocity relative to the air:

F_drag = 0.5 · ρ · Cd · A · V_air²

That V_air term is where things start to fall apart in typical data. What we have available is V_ground, usually from GPS. Those are only the same thing if the air is stationary relative to the ground, which is almost never true on track in any consistent way. A headwind effectively increases the apparent velocity term in the drag equation, a tailwind decreases it, and crosswinds introduce additional complexity that shows up as inconsistency rather than something you can directly attribute.

So when you take your logged data and try to build something like a power vs speed curve, what you’re actually plotting is:

P ≈ m · a · V_ground

with all of the aerodynamic effects embedded implicitly in the acceleration term, but modulated by a velocity term that is missing the correct reference frame. That’s why repeated runs often don’t collapse onto a single curve the way you’d expect. Instead, you get scatter that gets attributed to driver variation, or sensor noise, or just “track conditions,” without a clear way to separate those effects.

—

For large changes, this isn’t necessarily a problem. If you make a big enough adjustment - significant power increase, major aero change, something that shifts the force balance substantially - you’ll still see it. The signal is large relative to the uncertainty. But once you start chasing smaller deltas, which is where a lot of meaningful development actually happens, the limitations become obvious. A one-degree change in wing angle. Adding a small gurney. Subtle ride height adjustments. These are changes that might move the force balance by a few pounds at speed. On the power side, that might correspond to fractions of a horsepower at a given velocity. Those effects are absolutely real, but they live in the same magnitude range as the unmeasured variability introduced by the air.

This is where the absence of an air-referenced measurement becomes the dominant issue. Without some way of estimating V_air, or more generally the dynamic pressure:

q = 0.5 · ρ · V_air²

…you don’t have a clean handle on the primary term driving aerodynamic force. You can assume a standard density. You can assume zero wind. And sometimes those approximations are “good enough” for rough comparisons. But they’re not stable across sessions, and they’re not robust when you’re trying to make decisions based on small differences.

What’s interesting is that you don’t necessarily need a perfect measurement to start improving things. Even an approximate or relative measure of dynamic pressure begins to structure the data in a way that ground-referenced signals alone cannot. Once you introduce an air-relative component, you can start to see which portions of your data scale with V², which portions are more consistent with rolling effects, and which differences between runs are actually attributable to external conditions rather than changes to the car.

From a practical standpoint, this changes how you interpret everything downstream. Instead of looking at a single blended signal and trying to reason backwards, you start to separate contributions. You can look at repeated straight-line segments and see whether they collapse when referenced to air-relative terms. You can begin to distinguish between a true increase in tractive effort and a reduction in resistive force. You can start to ask whether a setup change altered the coefficients in your model, rather than just whether the lap time moved.

None of this eliminates the other sources of uncertainty. Tire behavior is still complex. Rolling resistance is not perfectly constant. IMU data has noise and bias. But those are second-order effects compared to completely omitting the air as an input to the system. Right now, in most grassroots workflows, we’re trying to resolve relatively small differences while leaving one of the largest variables unmeasured.

It’s not that the math is particularly complicated, or that the underlying physics are unknown. It’s that the measurement framework is incomplete. We’ve gotten very good at instrumenting the car itself, but we’re still largely blind to the medium it’s operating in. And as soon as you start looking for smaller and smaller gains, that blind spot becomes harder to ignore.

At some point, if the goal is to move beyond “that felt faster” or even “that looked faster in the data,” and toward something closer to cause and effect, you need to bring that missing reference frame into the picture. Not at wind tunnel levels of complexity, but enough to stop treating the air as a constant.

Because it isn’t.

There are ways to add that layer without completely changing how you log or analyze data. We’ve been working through what that looks like in practice - something that fits into the tools people are already using, but fills in that missing piece.

I’ll share more on that soon…

The Zeppelin company tried welded duralumin alloy in their rigid airship frames. They cracked in service, and engineers went back to rivets. Junkers ran corrugated aluminum skins with welded joints in the 1920s. Within hours of the start of vibration testing, cracks spread from the bead lines. NACA put welded tube structures into test rigs in the 1930s and saw the same pattern: premature failures, while riveted joints survived thousands of cycles.

By WWII, no serious manufacturer was welding structural aluminum. They used high-strength alloys similar to 2024 and precursors to 7075, which are unweldable, and fastened them mechanically. The Navy’s flying boats sealed the verdict. Welded hulls cracked and leaked until manuals flat-out banned aluminum welding for repairs.

The physics behind it hasn’t changed. Weld aluminum and you create three distinct zones:

- The bead, solidified and hardened, usually stronger but less ductile than the base metal.

- The heat-affected zone (HAZ), annealed (softened) by thermal cycle of welding, often the weakest point.

- The interface, where bead and base metal meet, creating a classic stress riser.

Fatigue starts there. Microcracks form at ripple valleys or along the weld toe. They’re almost always invisible at first. Cyclic loading and vibration pry them open a fraction more each time they’re cycled. The surrounding aluminum work-hardens as it flexes, which raises strength but destroys ductility. The crack grows faster, sharper, until it runs. The part often doesn’t warn you. It holds until it doesn’t, which can be catastrophic.

A racing wing mount lives in exactly the same world. You might not be in the air, but thin aluminum plate, variable downforce versus speed, and endless vibration can be just as harsh, if not harsher than in the air. A weld that only ties into skin with no spar or backing plate is already on the clock. Even if it looks perfect, the fatigue process has started the moment the joint was made.

If you’re running a aluminum wing with welded upright mounts, keep these things in mind:

- Inspect under bright light at a shallow angle. Run a fingernail along the bead and toes — if it catches, assume it’s a crack.

- Use dye penetrant to make the invisible visible.

- Add Sharpie witness marks or torque stripes across the joint. Movement will show itself before total failure.

- Treat aluminum welds in aero load paths as consumables. They aren’t lifetime parts.

What sort of upright mounts are better, if this type is bad? Generally speaking, mounts should tie into structure like ribs, stringers or spars, not just skins.

If the construction of your wing is such that tying into structure isn’t possible, use backing plates or load spreaders - either internally of externally. Rivets work well, and don’t be shy with the quantity used - but do make sure they’re stainless, or even better, structural aircraft rivets (pull or bucked - either are pretty readily available online through experimental aircraft materials supply companies). Bonding with high strength epoxy adhesives like Hysol is also good - but both rivets and epoxy together are likely the best.

And one more thing to note - if your wing is aluminum, be aware of galvanic corrosion. Any backing plates or load spreaders made out of non-stainless steels (which can still cause issues) or carbon will need to be protected from direct contact with the aluminum to avoid corrosion.

The widest tire isn’t always the fastest. Inertia, weight and aerodynamic drag effects aside, the basic math for friction doesn’t even have an area term, so you don’t magically get more grip just because the footprint got bigger. The real world with tires made of rubber is a bit more complicated than that as the mechanical interface between the tire and the surface it contacts does see some rising Mu initially, while load is increased.

Because of that, when you increase tire width you might actually be giving away grip because you’ve lowered the loading per square inch, which means less heat and less mechanical contribution. What you really gain with more tire width is the ability to carry more load (and heat) before the tread overheats and falls off - more rubber to soak up the heat, makes sense, right? - but that’s not the same thing as going faster. And if the tire never reaches its operating window in the first place (a far more common problem than you might expect), all that extra width, weight, and aero drag from your big meat is just ballast.

Every tire has a comfort zone where load and temperature work together. We’ve reviewed this before. A 200TW street tire usually finds its best grip early and at coolish temps, around 135 to 140°F. A slick built for heavy aero wants more heat, with peak grip closer to 210°F. Get it right and the car feels glued to the track. Miss it and you’re either tip-toeing around on something too cold, or sliding on something that’s cooked.

Scrub is where this gets interesting. A tire will always see some scrub in normal use — a bit of slip angle in the corners (which is required to generate any dynamic friction), a little scrubbing in hard braking (that’s slip angle too, ya know) and that’s fine. It’s productive heat, helping you get the tread into its window.

The problem is excess scrub. Bad toe, poor camber, messy hands on the wheel, unnecessary steering inputs, or even running an undersized tire that’s being overworked on a Subaru with whack ass weight distribution can all pile on heat that doesn’t help you accelerate, brake, or corner. It just makes the “with scrub” temperature curve climb faster. That might sound good if you’re struggling to warm a tire, but once you’re in the window it just means you’ll overshoot and fall off the peak sooner - sometimes a lot sooner.

The chart tells the story. The solid temperate line is clean heat build, the dashed line is normal scrub in a car that is set up well, with tires that are suited to the task, and the dotted line is excessive scrub. Once you see how quickly that dotted line rockets past the optimal range, it’s easy to understand why two cars on the same tire can feel completely different by mid-session. The driver / car who manages scrub stays in the sweet spot longer. The one who doesn’t finds grip falling away, no matter how much tire is under the car.

The fastest tire for you is the one that peaks where, when, and how you actually run it - and the one you can keep in that window without wasting it on heat you didn’t need.

But if you’re not tuning for tire temperature first - as the primary variable - you’re not tuning for grip. You’re guessing. Guessing will get you close, don’t get me wrong. But if you want the most out of your tires - the only thing on your car that should be touching the ground - understanding tire temperatures is fundamental.

Measuring “hot” vs. “cold” pressures isn’t enough. Nor are pyrometer readings after cooldown - measure them in hot pits after a series of 100% laps. You can apologize to your brakes later.

Even better is if you have on-board tire temperature measurement and logging. Then you can start thinking about “what temp, where, and when”. Which corner builds it, which one loses it. Whether your camber’s keeping the contact patch hot and uniform across the face or burning one edge while the other skates under braking or under power coming out of a corner. Ultimately it’s all about whether the tire is in its window, how fast it gets there, and what it does after.

A 200TW tire comes up quick, gives you confidence early, then fades as the lap counts increase. An R-Comp builds slower, takes commitment, but pays it back over the full run. On the same car. Same weather and track conditions. Ultimately, If you don’t know how each brand, size, and spec tire wants to be driven, you’re not testing.

The flip side, and where experience matters is that you can get decent-looking temps with a garbage setup. High speed shock tuning off? Car oversteering through trail? Toe in / out just a touch too aggressive? Diff spinning the inside wheel on power out of a corner? Now you’re making temp, but it’s not translating to grip - it’s cooking sidewalls, eating edges and building pressure you can’t trust. The number lies. The temp lies. The tire tells the truth - you just have to know how to listen, read, and iterate.

And if you’re dreaming of running F1 slicks on your track car after you look at the included graph, stop. They don’t work on road chassis. You’ll never get them hot enough, never load them hard enough, never see the shape of their curve. They’re not just slow when cold. They’re dangerous. Grip you can’t activate isn’t grip. It’s dead weight.

Temperature is the first thing you need to understand, and the last thing most people tune. Get it wrong and everything else is noise. Get it right and suddenly everything makes sense: why your fast lap always happens in lap 3. Why the car feels numb in one corner and alive in another. Why your “better” tire made you slower.

It’s not about finding grip. It’s about creating the conditions that let it find you.

As mentioned in that post, the project began as a brainstorm about using boundary layer suction-based aerodynamic assistance to enhance a car’s downforce without resorting to giant wings or wild bodywork, or mega power fans needing high voltage and lots of space, or sealed skirts and all the hassles that go with them.

The goal was to stay true to the Lotus ethos: simplicity, lightness, and cleverness over brute force. Power is always the easy way to go faster, but adding 3-400 horsepower to a Lotus misses the point in my opinion. Adding thoughtful and lightweight solutions that improve lap time without undue complexity and mega horsepower doesn’t.

From there, things evolved into a full development exercise. I designed a test rig (more in that later), and “built” a range of car configuration models to see how a small, actively managed suction system could influence pressure distribution, flow stability, and lap time - first virtually.

Aerodynamic Evolution:

The aerodynamic evolution of things I looked at followed a clear path during my theoretical optimization exercise:

• The factory splitter extends ~2 inches forward, while an optimized splitter grows to about 5 inches forward and adds dedicated tunnels, shifting the center of pressure rearward and reducing pitch sensitivity.

• The optimized plain diffuser stretches the factory layout and tightens the seal around the wheels.

• All enhanced configurations include wheel spats, helping reduce turbulence and drag around the rotating wheels.

• The drag-optimized rear wing is a true airfoil section, cantilevered slightly past the tail to create downforce to the rear wheels by using leverage while reducing induced drag.

• Ride height decreases progressively in all non-factory configurations, complementing the underbody aero efficiency.

The optimized front tunnels move the high-pressure stagnation region rearward and lower the overall sensitivity to ride height. Together, the splitter and diffuser changes act to stabilize balance through transitions and reduce dependency on static rake, and turn the floor of the car into an aerodynamics “system” as opposed to individual parts and pieces that will develop downforce at a rate greater than the sum of its parts.

Concept and Modeling:

What I’m now affectionately calling (will it stick?) the “Level 0” suction concept represents the next step: a single or dual-fan system designed to stabilize flow attachment across the rear diffuser and maintain consistent downforce regardless of speed. It’s clever rather than complex - low weight, no exotic hydraulics, high voltages, or specialized materials, and manufacturing techniques and a natural fit for a car that values balance over bravado.

All of the lap time projections in the table below come from a physics-based simulation environment built on principles similar to OptimumLap, which I’ve fine-tuned over the years. It uses a vehicle dynamics core that couples aero coefficients, tire friction, and drivetrain particular while integrating energy losses. The coefficients and other useful parameters have been validated against real-world data from hundreds of logged laps I and others have run at GingerMan Raceway on South Haven, MI in different cars, setups, and performance levels that I have data for.

GingerMan makes an ideal baseline for this kind of modeling. It combines a long back straight for assessing power sensitivity, fast aero-loaded sweepers that reward stability and efficiency and where a lot of time can be made with good aero, and a few tight corners that expose the limits of mechanical grip. This balance means that changes in downforce, drag, or tire performance show up clearly in both sector deltas and total lap time - making it the perfect track for comparing how each configuration actually translates into performance.

Comparative Modeling:

I modeled 19 total configurations to explore how each design change - engine, tire, or aerodynamic - contributes to overall lap time.  The following represent the most instructive steps, showing how each modification builds upon the last.

The trend is clear. If it can be made to work, the suction system doesn’t just add downforce - it adds control. It allows the car to maintain consistent aero balance without resorting to heavy or high-drag solutions. Even with relatively modest power levels, the lap time delta is substantial when the system is allowed to breathe at the right rate.

Testing and Validation:

Usefully, and unlike most aero development, which depends on expensive CFD or wind-tunnel work, the suction approach can be characterized by simple pressure monitoring. Measuring pressure in the plenum alone, where the fans are mounted over the duffuser, gives direct insight into system efficiency - something that normally requires F1-style pressure rakes and hours of post-processing to do.

The first phase of testing will use a flat static test panel, built with over 8,400 precision-matched openings, to benchmark suction uniformity and pressure response inside a plenum designed to marry the suction panel to the fan. The plenum pressure, fan speed, and suction bleed flow will be logged continuously, as will a few downstream areas within the diffuser, allowing me to determine how stable the system is under different simulated conditions, including airspeeds and pressure gradients once static testing is done.

From those results, I’ll be able to identify how the fan’s PWM duty cycle scales with airflow velocity, and where diminishing returns begin. This will inform both control strategy and diffuser geometry for the vehicle prototype.  Logging will reveal whether or not the fan is capable of delivery stable pressure and flow, including during transients.

Once the single-fan and static-panel tests are complete, the plan is to build a full vehicle system for validation under real-world conditions. The car will use a modular mounting setup that allows rapid swaps between configurations - allowing swapping a passive diffuser and the suction-diffuser for side-by-side evaluation as the definitive real world test.

The design and control logic of the system will likely evolve based on data. PWM response, suction stability, and pressure recovery rates will all feed into new calibration maps for both single- and dual-fan operation. Those adjustments will be guided by test data rather than speculation - subject to change as we learn what the system prefers.

Broader Context:

Projects like this can help illustrate how much performance lies in understanding and using  airflow rather than overpowering it. The difference between “fast” and “right” is often just how much you let physics do the work for you.  That, and if this system can be made to work, it will demonstrate an effective way to develop straightforward systems that leverage some of that free-of-charge air swirling around your race car using first principles without depending on hours and hours of CFD and wind tunnel modeling.