Roll center: the most abused point in vehicle dynamics

It has no bearing, it moves constantly, and half of what's said about it on forums is wrong. It also decides how hard your springs have to work — so here's what it actually is, and the one number you get from it.

You have a point that isn’t there (the instant center) and a length derived from it (the swing arm). Now the third construction, and the one people argue about hardest.

The roll center is where the car’s body is considered to roll about, at one axle, at one instant. Find it the same way you found everything else in this module — with a straightedge:

Draw a line from the instant center to the contact patch. Where that line crosses the car’s centerline is the roll center. Do it at both axles, join the two points, and you have the roll axis the whole car leans about.

That’s the entire construction. No algebra.

What it actually does for you

The roll center matters because of the distance between it and the center of gravity. That gap is the lever the car’s mass gets to pull on when you turn.

M_{\text{roll}} = m \cdot a_y \cdot (h_{\text{cg}} - h_{\text{rc}})

$m$ is mass in kg, $a_y$ is lateral acceleration in m/s², and the heights are in meters. The result is the roll moment in newton-meters — the twisting effort your springs and anti-roll bars have to hold up.

Worked example, on the Mini. 1200 kg, CG at 470 mm. Its roll centers are 90 mm at the front and 162 mm at the rear, so the roll axis sits around 126 mm at the middle. At 0.8 g:

M_{\text{roll}} = 1200 \times (0.8 \times 9.81) \times (0.470 - 0.126) \approx 3240\ \text{N·m}

Now raise the roll centers to 250 mm — a change of about 12 cm, which sounds modest:

1200 \times 7.85 \times (0.470 - 0.250) \approx 2070\ \text{N·m}

A 36% reduction in roll moment, for free. No stiffer springs, no bigger bar, no worse ride. That is the seductive part, and it is why “just raise the roll center” is such common advice.

Why it isn’t free

Because the same geometry that resists roll also pushes the car upward. When the tire generates lateral force, the suspension link geometry converts part of it into a vertical force at the body — jacking. It scales with roll-center height:

Take that Mini front axle. At 0.8 g the front tires are making roughly 6000 N of lateral force, the track is 1683 mm, and the roll center is 90 mm up:

F_{\text{jack}} \approx 6000 \times \frac{0.090}{1.683} \approx 320\ \text{N}

Modest — about 32 kg trying to lift the front of the car. Now run the same sum with the roll center at 250 mm and it becomes 890 N. The car starts standing up on its outside wheel mid-corner, which raises the CG, which increases weight transfer, which is the opposite of what you wanted. In its extreme form this is what people mean when a car “trips over itself.”

So the trade is: higher roll center means less body roll and more jacking. Lower means more roll and less jacking. Neither end is free, and the numbers above are how you decide where to sit.

And below ground?

A roll center under the ground plane is not an error — plenty of lowered cars have one, and the strut layout above goes there at full compression. It reverses the jacking force, pulling the body down under cornering, which some designers actively want. What it usually signals, though, is that the geometry has been pushed far from where it was designed to sit, and chapter 5 is about how far it then wanders.

Go look at it

Open the strut layout.

In the front view you’ll see the roll center marked directly on the drawing as RC h=90 mm, with a construction line running from the contact patch toward the instant center. The GEOMETRY METRICS panel gives the same figure as RC — Roll Center 90.1 mm.

Grab the Lower Bushing — the inner end of the lower arm — and drag it downward about 15 mm. Watch RC — Roll Center fall. Keep going and it will pass through zero and go negative: the roll center is now underground.

Reload, and this time drag that same point upward. The roll center climbs. You are trading roll stiffness against jacking, live, with one hardpoint.

Now check what else you changed. FVSA — Swing Arm moved too, so your camber curve is different from the one you designed in chapter 3. This is the recurring lesson of the module: there are three or four numbers and only one set of pivots.

Compare axles. Switch to the Rear Axle: that wishbone layout’s roll center sits at 162 mm, far higher than this strut’s 90 mm. That difference is deliberate — it is one of the levers that sets whether a car understeers, which is Module 3.

Draw a line from the instant center to the contact patch. Where it crosses the car's centerline is the roll center, and its height above ground is what sets your roll moment.

Watch the construction live — Mini front →

What it costs you

Chasing a roll center height in isolation is the classic beginner error, because the pivots that set it also set the swing arm. Every millimeter of roll-center height you buy by moving an inner pivot is paid for in camber curve, and often in track change too.

The other cost is subtler and catches experienced people: the construction above is only valid at one position. Everything on this page describes the car standing still. The moment it rolls, the two sides no longer share a geometry, and the roll center you designed is not the one the car is using. That is chapter 5, and it is the reason roll-center arguments on forums never resolve — the participants are usually describing different positions of the same car.

Rules of thumb

Starting points by class. All of them assume you have also read chapter 5, because a static number without a migration figure is half a specification.

Road car, front: 30–100 mm. The strut layout above sits at 90 mm.
Road car, rear: usually higher than the front by 30–100 mm, which is part of how manufacturers build in stabilizing understeer.
Circuit car: often close to ground level, sometimes slightly below, accepting more roll in exchange for predictable, low-jacking behavior and a geometry that barely moves over small travel.
Rally-raid and desert trucks: roll centers are usually well above ground and the roll axis is high, because ground clearance forces the arms upward — but with 350 mm or more of travel the migration dominates the static value entirely. Design the envelope, not the number.
Roll moment arm ( $h_{\text{cg}} - h_{\text{rc}}$ ): 250–400 mm is typical for a road car. Under about 150 mm and the car resists roll so well that you lose the roll cue drivers use to sense grip.

Try this

Find the pair of lower-arm positions that give the Mini the same roll-center height but noticeably different swing-arm lengths. What does that tell you about designing for one number at a time?
Put the front roll center at exactly 0 mm. Then run ▲▼ Bump. Does it stay there? (Chapter 5 in one experiment.)
Work out the roll moment for your own car at 0.7 g, then again with the roll center 50 mm higher. Is that worth what it does to your camber curve?

Next: everything above was true for one position of one wheel. Now we let the car actually move, and watch the roll center refuse to stay where you put it.

Preset hardpoints are illustrative sketches, not measured factory specs. The relationships are real; treat the absolute numbers as a starting point for your own model.