# Combined slip and the friction ellipse

> A tyre has one friction budget shared between lateral and longitudinal force. The friction ellipse describes the trade, and its non-linearity explains trail braking, exit traction limits and many balance complaints.

Area: Tyres · Starting level: L1 Foundation · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/tyres-combined-slip-friction-ellipse

## L1 · Foundation

A tyre cannot give its full lateral grip and its full braking grip at the same time. Plotting the maximum available $F_x$ against $F_y$ gives roughly an **ellipse**:

$$\left(\frac{F_x}{\mu_x F_z}\right)^2 + \left(\frac{F_y}{\mu_y F_z}\right)^2 \le 1$$

For racing slicks $\mu_x$ and $\mu_y$ are similar (often within ~10%), so it is close to a circle; hence the common name "friction circle".

The key engineering insight is the **non-linearity**. Because it is a circle and not a diamond, the first bit of longitudinal force costs very little lateral force, and the last bit costs a lot:

| Lateral force used | Longitudinal force still available |
|---|---|
| 50% | 87% |
| 80% | 60% |
| 90% | 44% |
| 95% | 31% |

So a rear tyre at 90% lateral utilisation mid-corner has under half its traction available. That is why exits are limited by how fast the driver can unwind steering while adding throttle, and why a car that is "loose on throttle" is often simply asking the rear for more than the ellipse allows.

The same logic explains **trail braking**: the driver blends brake pressure out as steering builds, keeping the tyre on the edge of the ellipse rather than inside it. A driver who brakes in a straight line then turns, with a gap between, leaves the tyre under-used in that gap. In g-g data this shows as a "notch" or diamond shape between the braking and cornering lobes.

## L2 · Practitioner

**Seeing combined slip in data.**
- **g-g plot** (lat vs long G). An experienced driver's trace hugs a rounded envelope. A less experienced driver shows a "cross" or diamond shape, with braking and cornering done separately. Segers §7.2 uses exactly this to compare two drivers through a chicane.
- **Combined G** $\sqrt{a_x^2 + a_y^2}$ against distance. Dips between braking and apex show lost time in the transition.
- **Brake release vs steering.** Plot brake pressure and steering angle together. A good trail brake has brake decaying roughly as steering rises, with release rate matched to the car's rotation needs.

**Setup levers that act through the ellipse:**
- **Brake bias** sets which axle reaches the ellipse edge first on entry. Rearward bias with trail braking puts the rear closer to the edge and gives rotation, or instability.
- **Differential** coast/power locking decides how longitudinal force is shared left/right, which moves the inside and outside tyres to different points on their ellipses.
- **Engine braking / coast maps** add rear longitudinal force on entry, using rear budget.
- **TC** limits rear longitudinal demand when lateral demand is high. Modern GT TC uses lateral G or steer as an input for this reason.

**Pitfall:** the vehicle g-g envelope is **not** a single tyre ellipse scaled up. Load transfer moves four tyres to different loads, aero changes the size with speed, and the longitudinal-positive side is power-limited. "We are not using the circle" sometimes means the driver is not committed, and sometimes means the car's envelope really is diamond-shaped in that phase because of balance (e.g. entry oversteer stopping the driver trailing the brake).

## L3 · Advanced

**Slip-vector view.** In a brush model with isotropic friction, combined behaviour is best understood through the **combined slip vector** $\sigma = (\sigma_x, \sigma_y)$, with $\sigma_x = \kappa/(1+\kappa)$ and $\sigma_y = \tan\alpha/(1+\kappa)$. Once sliding dominates, the resultant force points opposite to the sliding-velocity direction. Two consequences:

1. Adding longitudinal slip at a fixed slip angle *rotates* the force vector and reduces $F_y$ **even below the peak**. In the linear region the reduction is small; near saturation it is large. This is why a small amount of brake on corner entry can rotate the car (rear $F_y$ falls with rear $F_x$, which is a yaw moment).
2. Different longitudinal and lateral stiffness ($C_\kappa \ne C_\alpha$) means equal-force contours are not circles at low utilisation. The ellipse is the **envelope** of peak forces, not a constant-slip contour.

**Pacejka MF combined slip** uses weighting functions $G_{x\alpha}(\alpha,\kappa)$ and $G_{y\kappa}(\kappa,\alpha)$ multiplying the pure-slip forces, plus a $\kappa$-induced lateral force $S_{Vy\kappa}$, a function of $\kappa$ whose amplitude depends on load and camber and is reduced as slip angle grows. When you fit tyre data (RCVD Ch. 14), combined-slip coefficients are the least well-measured and the most important for corner entry and exit.

**Ellipse anisotropy.** $\mu_x > \mu_y$ by a few percent is common on slicks. Tread pattern (wets) and construction can reverse it. Load sensitivity is not identical in x and y either, so the ellipse shape changes with load. At high downforce the envelope gets bigger but slightly more elliptical.

**Vehicle-level synthesis.** With four tyres each on its own ellipse, the vehicle's available $(a_x, a_y)$ at each moment is also limited by **yaw moment balance**. On corner entry you may have spare tyre capacity but no way of using it without exceeding the yaw moment the driver can control. RCVD Ch. 8 (the MRA moment method) and Ch. 9 (the g-g diagram) formalise this. The g-g diagram is the projection of a force–moment problem, which is why stability (entry oversteer) can cap the envelope as surely as grip.

## L4 · Expert

**Limits of the ellipse idea.**
- **Transient path.** The ellipse is a quasi-steady boundary. Under fast brake release the longitudinal and lateral relaxation lengths differ, so the actual force path cuts inside the steady ellipse. Fast steering-plus-release inputs lose a few percent of grip that a smooth driver keeps.
- **Thermal coupling.** Operating on the ellipse edge in combined slip generates more heat per unit force than pure slip at the same force, because slip velocity rises. A driver who trails heavily into every corner can be quicker on lap 1 and slower by lap 10 through front surface overheating. The ellipse shrinks with temperature, so lap-time-optimal and stint-optimal driving differ.
- **Asymmetric ellipses per wheel.** Inside and outside tyres have different camber, load and temperature, so their ellipses differ in size and shape. Diff locking couples the two. Vehicle-level "available grip" is a constrained optimisation, which is why quasi-steady lap sims (QSS) and transient lap sims differ in corner-entry predictions.
- **Driver modelling.** A lap sim optimal-control solution uses the full combined-slip model and often finds trail-braking profiles that real drivers cannot hold. Drivers trade 1–2% of envelope for margin. When correlating sim and data, gate the comparison by phase (entry, mid, exit) and compare the *shape* of the g-g trace, not only the peaks.

Expert trackside heuristic: if the driver's g-g trace shows good braking and good mid-corner peaks but a pronounced notch in between, check the car for entry instability (rear rotation under combined slip: bias, coast diff, engine braking, rear toe) before blaming commitment.

## Key formulas

**Friction ellipse**

$$\left(\dfrac{F_x}{\mu_x F_z}\right)^2 + \left(\dfrac{F_y}{\mu_y F_z}\right)^2 \le 1$$

- F_x: longitudinal force (N)
- F_y: lateral force (N)
- \mu_x,\mu_y: peak friction coefficients
- F_z: vertical load (N)

**Remaining longitudinal capacity**

$$\dfrac{F_x}{\mu_x F_z} = \sqrt{1 - \left(\dfrac{F_y}{\mu_y F_z}\right)^2}$$


**Combined G (vehicle)**

$$a_{comb} = \sqrt{a_x^2 + a_y^2}$$


## Trackside takeaways

- One friction budget per tyre. At 90% lateral use only ~44% of longitudinal grip remains.
- Trail braking keeps the tyre on the ellipse edge through the entry; a gap shows as a notch in the g-g plot.
- Adding longitudinal slip reduces lateral force even below the peak, which is why brake, coast diff and engine braking rotate the car.
- The vehicle g-g envelope is limited by yaw stability as well as tyre grip. A "diamond" g-g trace can be a car problem, not a driver one.
- Combined-slip driving generates more heat per unit force, so lap-1 technique can cost you by lap 10.

## By car category

### GT3

ABS lets drivers brake deep into the corner with a lot of trail; TC trims rear demand on exit. Check that the ABS map is not cycling at moderate pressures mid-trail, because that shows as a ragged envelope on the g-g plot.

### One-make Cup

With no ABS (992 Cup), the trail-brake phase is entirely in the driver’s feet. Brake-pressure release rate is the most coachable variable, and bias/coast settings change how much rear capacity remains for rotation.

### TCR

The front tyre carries steering, braking and drive. On exit the front ellipse is shared between lateral force and traction, so throttle mid-corner directly causes understeer. Drivers use lift and rear rotation to manage this.

### FIA F2

Big aero makes the envelope speed dependent. A high-speed trail into a fast corner works on a much larger ellipse than a hairpin trail, and the change in ride height during the trail moves aero balance as well.

### Super Formula

High downforce and light weight give very large combined-G envelopes. Drivers trail deep. With no tyre warmers, out-lap ellipses are much smaller and entry rotation is less predictable until temperature builds.

## References

- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 2 Tire Behavior (Combined slip, friction ellipse)
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 8 Force-Moment Analysis (Yaw-moment limits on the envelope)
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 9 "g-g" Diagram
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 14 Tire Data Treatment (Combined-slip data)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §7.2 The Traction Circle
