# Braking physics, longitudinal load transfer & ideal brake balance

> Braking moves load from the rear to the front axle in proportion to deceleration × CG height / wheelbase. The ideal brake force split follows the axle loads, so it changes with deceleration, grip level and downforce — and it is not the same as the pressure bias shown on the dash.

Area: Brakes · Starting level: L1 Foundation · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/brakes-physics-ideal-balance

## L1 · Foundation

When the car decelerates, the inertial force $m a_x$ acts at the centre of gravity, a height $h$ above the ground, while the braking forces act at the ground. The resulting pitching moment moves vertical load from the rear to the front axle:

$$\Delta F_z = \frac{m\,a_x\,h}{WB}$$

This is **longitudinal load transfer**. It happens regardless of springs or dampers (they only change how much the car *pitches*, not how much load transfers in steady state).

**Example (GT3, no aero):** $m = 1350$ kg, 46 % static front weight, $h = 0.43$ m, $WB = 2.70$ m, braking at 1.3 g. Transfer $= 1350\times1.3\times9.81\times0.43/2.70 \approx 2740$ N. The front axle load rises from 6090 N to ~8830 N: **66.7 % of the total** is now on the front.

**Ideal brake balance.** Maximum deceleration needs every tyre at its friction limit at the same time. With the same friction coefficient front and rear, each axle's brake force must be proportional to its load:

$$\frac{F_{xF}}{F_{x}} = \frac{F_{zF}}{F_z} = w_F + \frac{a_x}{g}\frac{h}{WB}$$

So the ideal **front force share** for our GT3 at 1.3 g is ~67 %. At lower grip (wet, 0.8 g) it would be $0.46 + 0.8\times0.159 \approx 59$ %. **The ideal balance moves forward with deceleration.**

- Too much front brake → fronts lock first: car goes straight, understeer on entry, flat-spotted fronts — but stable.
- Too much rear brake → rears lock first: the car loses rear lateral grip, rotates, possibly spins. **Rear lock-up is the dangerous one.**

**Brake balance vs brake bias on the dash.** Teams usually quote the *hydraulic pressure* split $p_F/(p_F + p_R)$ — typically 52–60 % on GT3 or formula cars. That is not the force split: front calipers/discs are bigger (more torque per bar) than the rears, so ~52 % pressure can mean ~67 % force (worked example at level 2). Always know which number you are talking about.

Typical peak decelerations (high-speed zones, where downforce adds load): F4 ~1.6–1.9 g; GT3 ~1.6–2.0 g; F3 ~2.5–3 g; LMP2 and Hypercar ~2.5–3 g (Hypercar downforce is capped by regulation); F2, Super Formula and IndyCar (road course) roughly 3–4 g. These are indicative only: they vary strongly with track, speed and downforce level, so check logged data for your car.

## L2 · Practitioner

**From force split to pressure bias.** The brake force produced by one wheel is

$$F_x = \frac{2\,\mu_{pad}\,p\,A_{pist}\,r_{eff}}{R_{tyre}}$$

($A_{pist}$ = piston area on *one* side of an opposed caliper; two friction faces). Define the axle **brake gain** $G$ in N per bar. Then the pressure ratio needed for a force ratio is $p_F/p_R = (F_{xF}/F_{xR})\,(G_R/G_F)$.

Worked GT3-style example:
- Front: 6-piston, per side 30/34/38 mm pistons → 2749 mm²; $\mu_{pad} = 0.50$, $r_{eff} = 0.165$ m, $R = 0.34$ m → **267 N/bar per axle**.
- Rear: 4-piston, per side 30/34 mm → 1615 mm²; $r_{eff} = 0.155$ m, $R = 0.35$ m → **143 N/bar per axle**.
- Ideal force split at 1.3 g (level 1): 66.7 % / 33.3 % (ratio 2.0) → pressure ratio $2.0\times143/267 = 1.07$ → **pressure bias ≈ 51.7 %**.

At 1.3 g ($F_x = 1.3\times13\,240 \approx 17.2$ kN) that is ~43 bar front, ~40 bar rear — realistic GT3 numbers.

**With downforce.** The axle loads also include aero:

$$F_{zF} = w_F\,mg + AB\cdot L + \frac{F_{x,tyre}\,h}{WB}$$

Use the **tyre braking force** (total deceleration minus drag) for the transfer term: drag acts on the body roughly at CG height, so the moment of the drag force about the ground cancels the inertial moment of the deceleration it causes; and a wind-tunnel aero balance quoted at the axles usually includes the drag moment already. With $F_{x,tyre} = \mu(mg + L)$ and $\lambda = L/mg$, the ideal front share becomes

$$\frac{F_{zF}}{F_z} = \frac{w_F + AB\,\lambda}{1 + \lambda} + \mu\chi$$

— a load-weighted mix of static and aero balance, plus a transfer term that does not change with speed.

Same GT3 at 250 km/h: $L \approx 4600$ N at 40 % front ($\lambda = 0.347$), μ = 1.3 → tyre braking ≈ 1.75 g, plus ~0.2 g of drag ≈ 1.95 g total. Front share $= (0.46 + 0.40\times0.347)/1.347 + 1.3\times0.159 \approx 0.445 + 0.207 =$ **65 %**, slightly rearward of the 66.7 % no-aero value: the rear-biased downforce ($AB < w_F$) dilutes the front share. That 1.5-point difference is several adjuster clicks. On a high-downforce formula car the picture is stronger: $\lambda$ can exceed 1, so with rear-biased aero the ideal front share is clearly *lower* at the start of the braking zone than at its end. (Counting drag in the transfer would wrongly give ~67.5 %.)

**Practical consequences:**
- A fixed hydraulic bias is a compromise across the braking zone and across corners (fast vs slow, uphill vs downhill).
- **Wet:** lower decel → ideal moves rearward. Teams typically move bias 2–6 % rearward in the wet (Segers Fig. 5.9 shows ~47 % wet vs ~53 % dry on a GT car), limited by entry stability.
- **Fuel load / tyre wear / track rubber** change the achievable decel and therefore the ideal, usually by fractions of a percent to ~1 %.
- Use the brake-balance calculator to explore $w_F$, $h$, deceleration and aero.

**Misconception:** "brake balance should match the static weight distribution." Static weight is the *starting point* of the ideal; at 1.5 g the transfer adds 20–25 % of front share on a GT car.

## L3 · Advanced

**Lock-up boundaries for a fixed distribution.** Let $\phi$ be the fixed front *force* fraction, $z = a_x/g$, $\chi = h/WB$, and $\mu$ the tyre friction coefficient (no aero). The front locks when $\phi\,z\,mg = \mu\,mg(w_F + z\chi)$, the rear when $(1-\phi)z\,mg = \mu\,mg(1 - w_F - z\chi)$:

$$z_{F,lock} = \frac{\mu\,w_F}{\phi - \mu\chi}, \qquad z_{R,lock} = \frac{\mu(1-w_F)}{1 - \phi + \mu\chi}$$

The achievable deceleration is the smaller of the two; **braking efficiency** $\eta_b = z_{max}/\mu$. With $w_F = 0.46$, $\chi = 0.159$, $\phi = 0.667$:
- $\mu = 1.3$: both limits ≈ 1.30 g → $\eta_b = 100\%$ (the bias was set for this grip).
- $\mu = 0.8$ (wet): front limit 0.68 g, rear 0.94 g → fronts lock at 0.68 g, $\eta_b = 85\%$. The ideal $\phi$ for 0.8 is 0.587 — a large move in force terms.

There is one crossing grip level $\mu^* = (\phi - w_F)/\chi$ for which a fixed distribution is ideal. Below it, the front locks first (safe); above it, the rear locks first (unstable). Road cars are designed with $\mu^*$ above any realistic road μ for exactly this reason; race cars are set near the racing μ and accept the risk, which is why bias is adjustable.

**Combined braking and cornering.** During trail braking the inside wheels are unloaded by lateral transfer and part of each tyre's friction ellipse is used laterally. With rear braking force applied, the rear's remaining lateral capacity falls; a bias that is ideal in a straight line can feel unstable at turn-in. Drivers who trail-brake deep often want slightly more front bias (or less rear braking via migration) at low pedal pressure.

**Refinements.**
- **Drag moment:** aero drag acts above the ground (near CG height on many cars), so strictly the load transfer comes from $(m a_x h - D\,h_D)/WB$. With $h_D \approx h$, or if your aero balance convention already includes the drag moment (loads at the contact patch), this is simply the tyre braking force × $h/WB$ — the convention used in the level-2 aero example.
- **Pad μ** varies with temperature and pressure (iron pads μ 0.40–0.55, carbon rising strongly with temperature), so the force split drifts even at constant hydraulic bias.
- **Engine braking** adds a rear (or all-wheel) brake force that changes with gear — see the entry-stability lesson.
- **Anti-dive geometry** changes pitch, not the steady-state load transfer.

## L4 · Expert

**The ideal distribution curve.** Plot front braking force $F_{xF}$ vs rear $F_{xR}$ (normalised by weight). The ideal curve, from $F_{xF}/F_{zF} = F_{xR}/F_{zR}$, is a parabola, most easily written parametrically in $z$:

$$F_{xF} = z\,mg\,(w_F + z\chi), \qquad F_{xR} = z\,mg\,(1 - w_F - z\chi)$$ A fixed bias is a straight line through the origin; it crosses the parabola at one $z$. Proportioning valves (above a knee pressure, rear pressure keeps rising but at a reduced ratio, e.g. 0.3–0.5:1; a *limiting* valve would cap it) create a kinked line that follows the parabola better. With downforce, the parabola becomes a family of curves indexed by speed — the brake-by-wire "migration" concept is essentially a controller tracing that family.

**Dynamic effects.** At brake application, load transfer builds with the pitch response (heave/pitch frequency 2–5 Hz on race cars; damper-controlled). During the first ~0.1–0.3 s the front is less loaded than the steady-state value: a rapid pressure rise with forward bias risks a front lock *before* the transfer arrives, while a slow rise wastes distance. This is why braking "quickness" and front low-speed bump damping interact — stiffer front bump damping accelerates load transfer onto the front tyres.

**Tyre side.** Optimum slip ratio under braking is typically 8–15 % for race slicks, with a flat-ish peak. The front and rear tyres may peak at different slip ratios and have different load sensitivity (narrower fronts on formula cars), so "ideal" is really "both axles at their respective peaks" — μ is not equal front/rear. Load sensitivity also means the heavily loaded front has lower μ than the rear, which moves the true ideal slightly *rearward* of the pure load-share value.

**Limits of the model.** The parabola assumes steady state, straight-line braking, equal μ and a rigid suspension. In practice: bumps (wheel load fluctuation ±20–30 %), kerbs, camber change under dive, brake-temperature-dependent pad μ, aero pitch sensitivity (front RH drops under braking → more front downforce on many cars, which moves the ideal *forward* at the start of the zone) and engine braking all move the target during the event. An expert engineer uses data (lock-up statistics, ABS activity, decel achieved vs. the g-g envelope) to set the bias empirically, using the model to understand directions and magnitudes.

## Key formulas

**Longitudinal load transfer**

$$\Delta F_z = \frac{m\,a_x\,h}{WB}$$

- m: mass (kg)
- a_x: deceleration (m/s²)
- h: CG height (m)
- WB: wheelbase (m)

**Ideal front force share (no aero)**

$$\frac{F_{xF}}{F_x} = w_F + \frac{a_x}{g}\frac{h}{WB}$$

- w_F: static front weight fraction

**Front axle load with aero**

$$F_{zF} = w_F\,mg + AB\cdot L + \frac{F_{x,tyre}\,h}{WB}$$

- AB: aero balance (front fraction)
- L: total downforce (N)
- F_{x,tyre}: tyre braking force = total decel force minus drag (N)

**Ideal front share with aero (at the friction limit)**

$$\frac{F_{zF}}{F_z} = \frac{w_F + AB\,\lambda}{1+\lambda} + \mu\,\frac{h}{WB}$$

- \lambda: downforce / weight, L/mg
- \mu: tyre friction coefficient

**Brake force per wheel**

$$F_x = \frac{2\,\mu_{pad}\,p\,A_{pist}\,r_{eff}}{R_{tyre}}$$

- \mu_{pad}: pad friction coefficient
- p: line pressure (Pa)
- A_{pist}: piston area one side (m²)
- r_{eff}: effective disc radius (m)
- R_{tyre}: loaded tyre radius (m)

**Front lock-up deceleration (fixed split φ)**

$$z_{F,lock} = \frac{\mu\,w_F}{\phi - \mu\,h/WB}$$


## Trackside takeaways

- Load transfer = m·a·h/WB: a GT3 at 1.3 g moves ~20 % of its weight onto the front axle.
- Ideal front force share = static front fraction + (a/g)(h/WB): it moves forward as deceleration rises.
- Pressure bias ≠ force split: ~52 % pressure can be ~67 % force because front brakes have more gain.
- Lower grip (wet) → ideal moves rearward → move bias rearward, carefully (rear lock is the unstable one).
- With aero, ideal front share = (w_F + AB·λ)/(1+λ) + μ·h/WB (tyre force, not total decel, drives the transfer); a fixed bias is always a compromise.

## By car category

### GT3

Typical hydraulic bias 52–58 % front, adjusted in-car. ABS masks the lock-up boundary but not the balance consequences (see ABS lesson). Peak decel ~1.6–2.0 g.

### One-make Cup

In Cup cars without ABS (e.g. the 992.1 Porsche 911 GT3 Cup) bias setting and driver adjustment are critical; the 992.2 911 Cup (2026+) has Bosch M5 ABS fitted ex-works, and Ferrari Challenge and Super Trofeo cars also have ABS, changing how the driver uses the bias.

### Formula 4

Low downforce and iron brakes: ideal balance is close to the no-aero formula. Bias adjuster in the cockpit; lock-ups are a common learning issue for young drivers.

### FIA F3

Significant downforce: the ideal force share varies through the braking zone. Drivers adjust bias between corners in qualifying and as fuel and tyres change.

### Hypercar (LMH/LMDh)

Brake-by-wire on the hybrid axle means bias is a controlled quantity with maps and migration; the driver changes target bias and migration settings from the wheel.

### IndyCar

Driver-adjustable bias with carbon brakes; on road and street courses drivers commonly adjust per corner and through a stint.

## References

- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 20 Driving and Braking
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 18 Wheel Loads
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §5.5 Brake Balance
