# Instant centre & roll centre

> The front-view instant centre and the roll centre decide how much lateral load transfer goes through the links rather than the springs, how much the body rolls, and whether the car jacks up in corners.

Area: Suspension Geometry & Kinematics · Starting level: L1 Foundation · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/sg-roll-centre

## L1 · Foundation

**Instant centre (IC).** In front view, a double-wishbone upright moves – for small motions – as if it rotated about a single point: the intersection of the lines extended through the upper and lower wishbone pivots. That point is the **front-view instant centre**. The distance from the tyre to the IC is the **front-view swing-arm length (FVSA)**.

**Roll centre (RC).** Draw a line from the centre of each tyre contact patch to that side's IC. Where the left and right lines cross (on the centreline for a symmetric car at static) is the **kinematic roll centre**. The roll axis is the line joining the front and rear RCs.

**What it does.** The lateral force at the tyres can be thought of as acting at the RC. Part of the lateral load transfer then goes straight through the links into the chassis – **geometric load transfer** – without compressing springs:

$$\Delta F_{z,geo} = \frac{m_{s,axle}\,a_y\,h_{RC}}{t}$$

with $m_{s,axle}$ the sprung mass carried by that axle and $t$ its track. The rest acts through the moment arm between the CG and the roll axis, rolls the body and goes through springs and bars – **elastic load transfer**. Raising the RC:
- reduces body roll (shorter moment arm),
- moves load transfer from springs to links (faster, undamped),
- increases **jacking** (the car lifts in corners).

**Typical static RC heights**
| Car | Front | Rear |
|---|---|---|
| Single seater (F4–F2) | ~0–30 mm | ~20–60 mm |
| GT3 | ~20–70 mm | ~40–100 mm |
| Road car (for contrast) | 30–100 mm | 50–150 mm |

**The big misconception:** the body does *not* physically pivot about the roll centre like a hinge. The RC is a convenient point for describing how forces enter the chassis. It moves as soon as the car rolls or heaves.

## L2 · Practitioner

**What the race engineer can actually change.** On most cars the RC is set by inner pickup positions (wishbone shims, alternative pickup holes) and by ride height. In practice you meet it via:
- **Ride height changes** – lowering the car usually lowers the RC *more* than the ride-height change (on short-FVSA geometry, often 1.5–3 mm per mm). A 5 mm drop at the front can take 10–15 mm off front RC height and change front geometric load transfer and roll.
- **Manufacturer pickup options** – GT3 and prototype manuals often offer two or three inner wishbone positions labelled "RC high/low". These are homologated options in GT3; in junior formulae geometry is usually fixed.

**Balance effect.** Higher RC at one axle puts more of that axle's lateral load transfer into the geometric (link) path and reduces roll. Raising the front RC **increases front lateral load transfer in steady state** – by $(m_{s,f}\,a_y/t_f)(1 - K_{\phi f}/K_\phi)$ per metre of RC rise – and reduces rear transfer, because the roll axis rises at the CG and less roll moment is left for the springs to share. It also makes the front share arrive faster, because the link path bypasses the dampers. What is car-specific is the *balance outcome*: front RC up often sharpens initial response but adds front tyre load transfer (mid-corner understeer tendency), and it comes with camber-gain, jacking and roll changes. Check with a LLTD/roll model rather than folklore.

**Jacking.** At the outside tyre the lateral force points toward the turn centre and, with the IC above ground, has a component pushing the chassis up along the line contact patch→IC. The inside tyre pulls down but carries less lateral force. Net result: the body rises in proportion to RC height and the outside/inside force difference. You see it in data as rear or front ride height *increasing* mid-corner on the laser sensors, together with a sum of damper positions moving to droop.

**Measuring and modelling**
- From pickup coordinates (CAD, manual, or a FARO arm) with a 2D/3D kinematics tool (SusProg, Lotus SHARK, OptimumK, a spreadsheet).
- Validate against **roll gradient** from damper pots (Segers §9.2): if the calculated roll moment arm is wrong, the measured roll gradient will not match the predicted one.

**Pitfall:** never move the RC to fix a balance problem you can fix with springs/bars – geometry changes also move camber gain, bump steer and anti-geometry. Use RC changes when you need what *only* RC height gives: less roll without more spring stiffness (e.g. to protect aero platform in roll), or reduced jacking.

## L3 · Advanced

**Construction, properly.** Kinematic RC = intersection of the two contact-patch→IC lines. Its height for a symmetric car at static:

$$h_{RC} = \frac{t}{2}\cdot\frac{h_{IC}}{d_{IC}}$$

where $h_{IC}$ and $d_{IC}$ are IC height and lateral distance from the contact patch. The camber gain is $\approx 1/d_{IC}$ rad per unit of wheel travel. **RC height and camber gain are therefore coupled**: for a given IC height, a short swing arm (more camber gain) also raises the RC. The geometry designer trades them off by moving the IC up/down and in/out.

**Migration in roll.** As the body rolls, the left and right ICs move asymmetrically and the kinematic RC moves laterally – on many race cars by 100–500 mm for one degree of roll, sometimes outside the track width on very low RC designs. Its height can rise or fall. A large lateral migration means the car's force-balance "description" changes enormously while the physical state changed little – a sign the kinematic RC is losing meaning.

**Force-based RC (RCVD Ch. 17).** Milliken prefers to define the RC from forces: apply lateral forces at each contact patch, compute the vertical *jacking* forces from the link geometry (lateral force × tangent of the contact-patch-to-IC angle), and find the point where a lateral force produces no roll. With unequal left/right lateral forces (the real cornering case) this gives a different – and more useful – point than the kinematic construction. Per wheel:

$$F_{jack} = F_y\,\frac{h_{IC}}{d_{IC}}$$

and the net body jacking is the sum of the outside (lifting) and inside (pulling down) terms.

**Interaction with LLTD.** Total lateral load transfer at the front axle (the rear is analogous):

$$\Delta F_{z,f} = \frac{a_y}{t_f}\left[m_{s,f}\,h_{RC,f} + \frac{K_{\phi,f}}{K_{\phi,f}+K_{\phi,r}}\,m_s\,(h_s - h_{RA}) + m_{u,f}\,h_{u,f}\right]$$

with $m_{s,f}$ the sprung mass carried by the front axle, $m_s$ the *total* sprung mass (the whole roll moment is shared by the roll-stiffness ratio), $h_s - h_{RA}$ the roll moment arm at the sprung CG, and $m_{u,f}$, $h_{u,f}$ the front unsprung mass and its CG height. The geometric part reacts instantly through the links; the elastic part through springs, bars and dampers. That's why RC height also changes **transient** balance even when steady-state LLTD is preserved with compensating ARB changes.

**Below-ground RCs** (common at the front of single seaters with nearly parallel arms at low ride height) reverse the jacking sign: the car is pulled *down* in corners – helpful for aero platform – but the RC can cross ground level mid-corner, flipping the jacking sign and making the car's ride-height response non-linear.

## L4 · Expert

**Limits of the concept for aero cars.** On a downforce car the RC is evaluated at the *dynamic* ride height, which varies with speed by 10–30 mm. A front RC at +25 mm static may be at −5 mm at 250 km/h. Any balance analysis that uses static RC heights for high-speed corners is wrong in sign as well as magnitude. Build RC height and jacking coefficient into the vehicle model as functions of heave and roll (lookup from kinematics tool), not constants.

**Jacking as a platform tool.** Lateral jacking adds a vertical force that is proportional to lateral force – i.e. it acts like a lateral-g-dependent heave spring. Teams exploit or minimise it depending on how the aero map responds: a car whose floor loses load when the front lifts wants low front jacking; one that is stall-sensitive at low rear ride height may benefit from rear jacking under high lateral g. This is a level of tuning that needs a full multi-body model with tyre lateral force distribution, because the jacking depends on *individual* tyre forces, not axle totals.

**Validation.** Compare simulated and measured: ride height (laser) vs lateral g at constant speed; sum of damper positions vs lateral g; roll gradient. A model with the wrong jacking coefficient will show the correct roll gradient but the wrong heave in corners – easy to miss if you only check roll.

## Key formulas

**Roll centre height (symmetric, static)**

$$h_{RC} = \frac{t}{2}\cdot\frac{h_{IC}}{d_{IC}}$$

- t: track (mm)
- h_IC: instant centre height (mm)
- d_IC: lateral distance contact patch → IC (mm)

**Geometric lateral load transfer**

$$\Delta F_{z,geo} = \frac{m_s\,a_y\,h_{RC}}{t}$$

- m_s: sprung mass carried by the axle (kg)
- a_y: lateral acceleration (m/s²)

**Jacking force per wheel**

$$F_{jack} = F_y\,\frac{h_{IC}}{d_{IC}}$$

- F_y: tyre lateral force (N)

## Trackside takeaways

- RC = intersection of contact-patch→IC lines; it is a force concept, not a physical pivot.
- Higher RC: less roll, more link (geometric) load transfer, more jacking.
- RC height moves more than ride height – lowering the car usually drops the RC substantially.
- Evaluate RC at dynamic ride height on aero cars; static numbers can be wrong in sign.
- Do not use RC changes to fix what springs/bars can fix; geometry changes have side effects.

## By car category

### GT3

Manuals often offer homologated alternative inner pickup positions (RC high/low). Front RC changes are a strong but slow-to-tune tool – typically decided in testing, not between sessions.

### FIA F3

Fixed spec geometry; the engineer influences RC only through ride height. Front RC sits close to the ground and moves strongly with front ride height – another reason front ride height is so sensitive.

### LMP2

Low RCs to protect the aero platform; jacking in long high-speed corners is visible on the laser ride-height channels and feeds directly into floor load.

### TCR

MacPherson front: IC is found from the strut-top perpendicular and the lower arm; RC is relatively high and migrates a lot with ride height. Lowering a TCR car too much can drive the front RC near ground and increase roll and camber loss sharply.

### IndyCar

Oval and road-course geometries differ; on ovals asymmetric setups make left/right IC positions intentionally different, so the RC is off-centre even at static.

## References

- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 17 Suspension Geometry (instant centres, kinematic vs force-based roll centre, jacking)
- Milliken & Milliken, Race Car Vehicle Dynamics (SAE 1995): Ch. 18 Wheel Loads (geometric and elastic lateral load transfer)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §9.1–9.2 Roll angle and roll gradient (validating roll moment arm with data)
