F1 ground effect explained: the floor that makes the car

After the chequered flag, a car is pushed into the scrutineering bay and turned onto its side, and an official measures the thickness of a wooden plank running down its belly. If that plank has worn through a single millimetre more than the rules permit, the result is deleted. Nothing about the engine, the wings or the driver is in question. The car simply spent the afternoon a fraction closer to the road than it was allowed to be, and that fraction is worth enough to be policed with a micrometer.

Any account of F1 ground effect has to begin there, because the measurement tells you where the performance lives. The floor of a modern Formula 1 car produces the majority of its downforce, it produces that downforce for very little drag, and the amount it produces depends on how close the underside sits to the tarmac. Everything else in this article follows from those three statements: the venturi tunnels, the sealing problem, the porpoising that embarrassed the sport in 2022, and the reason the whole concept was outlawed for four decades before it came back.

What is actually happening underneath the car

Take a channel, narrow it, and push a fluid through. The fluid has to speed up through the constriction because the same quantity is passing through a smaller opening every second. As it speeds up, its static pressure falls. Widen the channel again and the flow slows and the pressure recovers. That is the venturi effect, and it is the whole physical basis of a modern Formula 1 floor.

Now put one wall of that channel on the ground and the other on the underside of a racing car. The low-pressure region is no longer an abstraction in a pipe. It is a large area of suction acting upwards on the road and downwards on the car, and because atmospheric pressure is still pushing down on the top surface of the bodywork at its normal value, the difference between the two presses the car into the tarmac.

The reason this is so much better than a wing is a question of area and waste. A wing produces its force over the modest area of its own elements, and it does so by hurling a mass of air upwards, which leaves behind trailing vortices and costs induced drag. The floor produces force across almost the entire plan area of the car, and it does so mostly by making air move faster rather than by throwing it somewhere else. Force per unit of drag is the currency that decides which parts get built, as set out in the piece on how downforce is traded against drag in a setup, and by that measure nothing on the car comes close to the floor.

58%24%14%

• Underfloor tunnels and diffuser58
• Rear wing and beam wing24
• Front wing14
• Other bodywork surfaces4

Invented proportions summing to 100, constructed to show why floor development receives most of a team's aerodynamic effort. Not a measured breakdown of any car, and the true split varies with wing level and circuit.

Show the numbers

Worked example: where a ground-effect car finds its downforce

Item	Value
Underfloor tunnels and diffuser	58
Rear wing and beam wing	24
Front wing	14
Other bodywork surfaces	4

The split above is invented, and the ordering is not. It explains a great deal of otherwise puzzling behaviour, including why teams will spend an entire development season on a component the cameras never show and the spectators never see.

The anatomy of a venturi tunnel

A modern floor is not a flat sheet with a diffuser bolted to the back. It is a shaped structure with three distinct regions doing three different jobs.

At the front is the inlet, where air is gathered and directed into the tunnels. The quality of the flow arriving here decides everything downstream, and it is the reason the front wing, the front suspension geometry and the deflectors around the front wheels are all designed with the floor in mind rather than for themselves. Turbulent, badly angled air delivered to the inlet cannot be repaired later.

Behind it is the throat, the narrowest section of the tunnel and the place where the flow reaches its highest speed and lowest pressure. This is where the downforce is generated. The suction peak is concentrated over a relatively small stretch of the car's length, which is one reason the aerodynamic balance of a ground-effect car is so sensitive to how the platform pitches.

Behind that is the diffuser, the expanding section that returns the flow towards ambient pressure before it leaves the car. The diffuser is not an afterthought and it is not merely an outlet. How steeply the channel can be opened without the flow detaching from its surfaces sets the ceiling on how low the pressure at the throat is allowed to go in the first place. A diffuser is a permission slip for everything upstream of it.

Two tunnels run either side of a central section, and the regulations bound their geometry closely. The Technical Regulations describe the floor in terms of permitted volumes, minimum radii, maximum expansion and defined reference surfaces, which leaves the teams competing over a narrow band of legal shapes. That is deliberate. The rules are attempting to specify not just how much downforce is available but how it behaves, and the second of those is far harder to legislate than the first.

• 2Venturi tunnels beneath the car
• 10Plank thickness specified in millimetres
• 9Minimum plank thickness permitted after the race
• 2Separate regulatory eras of ground effect in the sport

Structural and regulatory features of a current ground-effect car rather than performance figures. The plank measurements are the ones taken at post-race scrutineering.

The small parts nobody photographs

Walk past a floor propped against a garage wall and the tunnels are the obvious feature. The details around them are where the season is usually won.

A row of fences stands at the leading edge of each tunnel, dividing the incoming air into separate streams and starting the rotating structures the floor will use further back. Their height, spacing and curvature are among the most tightly regulated dimensions on the car precisely because they set up the flow that everything else depends on.

Along the outer boundary sits the floor edge itself, shaped and often turned upwards, which is where the sealing vortices are generated and where teams found most of their gains once the rule cycle matured. Ahead of the tunnels is the bib, the central structure beneath the survival cell that splits the flow between the underfloor and the outside of the car and sets how much of it goes where.

None of these produce much downforce on their own. They decide whether the parts that do produce downforce receive air worth working with, and a floor with a good throat and a poor inlet is simply a good throat that never gets used.

Why the seal decides everything

Here is the failure that ruins the whole arrangement. Alongside the car, at the level of the floor edge, sits air at ordinary atmospheric pressure. Underneath it sits air at considerably less. Air moves from high pressure to low, so it will attempt to spill sideways under the floor edge and into the tunnels at every point along the length of the car.

If it succeeds, the suction is diluted, the throat cannot hold its low pressure, and the downforce falls away. A ground-effect floor is only as good as its ability to stop the atmosphere getting in from the sides.

The first generation of ground-effect cars solved this with hardware. Sliding skirts ran down each side of the car, held against the road by springs, forming a physical barrier that kept the outside air outside. They worked extraordinarily well. Downforce rose to levels that had no precedent, and cars that had previously depended on wings began cornering at speeds their chassis, their tyres and the circuits around them had never been designed to absorb.

Modern cars are not permitted skirts, so the seal has to be made out of air. Along the outer edge of the floor, geometry is used to generate a strong, coherent vortex that spins along the length of the car. A vortex has low pressure at its core and rotational momentum resisting anything trying to cross it, so a line of them forms a curtain that inflowing air struggles to penetrate. The floor is sealed by a rotating wall of air rather than by a rubbing strip.

This is not merely a workaround. It has a property the skirt never had. A skirt either sealed or it did not, and the transition between those two states took milliseconds. An aerodynamic seal weakens and recovers. Run over a kerb and the edge vortices are disturbed, some downforce goes away, and then most of it comes back as the structures re-establish themselves. The car loses grip for a moment instead of losing it entirely, and a driver can survive the first outcome.

Why ground effect was banned the first time

The tunnels appeared in Formula 1 in the late nineteen seventies, when a Lotus design team including Peter Wright and Colin Chapman worked out that the sidepods of a racing car could be shaped as inverted aerofoil sections running close to the ground, and that sealing their edges would multiply the effect. The Lotus 78 introduced the idea and the Lotus 79 turned it into a dominant car. Within two seasons the entire grid had copied it.

What followed was a period the sport still treats as a warning rather than a golden age, and the reasons are worth separating because they are not all the same reason.

The first was speed itself. Cornering forces climbed to a level that circuits built for slower machinery could not accommodate. Run-off areas designed around one set of assumptions were being approached at speeds those assumptions never contemplated, and the barriers behind them were correspondingly closer than they should have been.

The second was the physical toll. To make the floor work, the cars had to be run extremely low and extremely stiff, because any compliance in the suspension moved the floor and moved the downforce with it. Suspension travel was reduced almost to nothing. Drivers were racing on skateboards with the ride quality to match, over circuit surfaces far rougher than today's.

The third, and the one that actually settled the argument, was the failure mode. When a skirt jammed in its runners, broke, or lifted over a kerb, the seal was gone at once. The car was at that instant travelling as fast as it went all lap and carrying more downforce than at any other point, and all of it left simultaneously. There was no warning, no progressive slide, and nothing the driver could do with the controls, because the forces involved were far larger than anything the steering could generate. Several of the era's most severe accidents carried this signature, and the 1982 season in particular left the sport with casualties it could not argue away. The long arc of how Formula 1 came to take driver safety seriously has this period near its start.

Sliding skirts were prohibited first, in 1981, together with a minimum ground clearance requirement. Teams responded by building suspension that let the car settle onto the road once it was moving and rise again when it stopped, which demonstrated fairly conclusively that a clearance rule alone was unenforceable. For 1983 the governing body took the direct route and mandated a flat floor between the axle lines. With no tunnels, there was nothing to seal.

What the flat-bottom decades did instead

Removing the tunnels did not remove the ambition. It relocated it.

For nearly four decades, teams generated underbody downforce from the only part of the floor still available to shape, which was the diffuser at the rear, and made up the rest with wings and increasingly elaborate upper bodywork. The flat area ahead of the diffuser still worked as a mild ground-effect surface, so the incentive to run low never went away, and the regulations kept having to respond.

A stepped floor arrived in the mid nineteen nineties, raising the central section relative to the reference plane and cutting the effectiveness of the flat area. The plank came in during 1994, after a season the sport would rather not repeat, precisely because ride height needed a policing mechanism that did not rely on measuring a moving car. Rake angle became a development direction in its own right, with cars run nose-down and tail-up to turn the flat floor into a crude expanding channel.

The result was a generation of cars that produced a great deal of downforce from surfaces sitting well above the road, in air arriving from a specific direction at a specific angle. Those are exactly the conditions that a car in front destroys. The sport had solved the sealing problem by making downforce that depended on clean air instead, and it spent twenty years discovering how expensive that was for the racing.

Why the 2022 rules brought the tunnels back

The regulations that took effect in 2022 reversed the choice on purpose, and the reasoning is worth stating precisely because it is often described as a wake-shaping exercise alone.

Moving the downforce back to the floor does two separate things. It shapes the wake, because a car generating its force underneath itself throws less disturbed air outwards and upwards than a car generating it from wings and bargeboards. And, independently, it makes the following car more tolerant, because underbody performance depends primarily on ride height and sealing rather than on the precise angle at which air arrives. A floor-dominated car in a wake loses less than a wing-dominated car in the same wake, before anything at all is done about the wake itself.

The sealing problem came back with the tunnels, as it had to, and this time it was solved aerodynamically. So did a second problem that the skirt era had largely masked, and the sport rediscovered it in public during the first pre-season test.

Porpoising is the floor stalling and recovering, several times a second

Ground effect gets stronger as the car gets lower. That is the whole appeal and it is also the trap, because a force that increases as the gap closes is a force with positive feedback built into it.

1. 1The car gains speed on a straightAerodynamic load rises with the square of speed. The suspension compresses and the floor moves closer to the road surface.
2. 2Suction increases as the gap closesLess clearance means a stronger venturi effect through the tunnel throat. The extra downforce pulls the car down further, which increases the suction again.
3. 3The throat reaches its limitBelow a certain height the flow can no longer negotiate the constriction and the pressure recovery demanded of the diffuser. The boundary layer detaches from the tunnel surfaces.
4. 4The floor stallsAttached flow gives way to a separated, recirculating mess. Most of the underfloor downforce disappears in a fraction of a second, and it does so while the car is at its fastest.
5. 5The springs push the car back upWith the aerodynamic load suddenly removed, the suspension returns towards its static position and the ride height rises.
6. 6Flow reattaches at the higher gapWith more clearance the tunnels work again, the suction returns at full strength, and the car is pulled straight back down towards the road.
7. 7The loop repeatsNothing has damped the cycle, so it runs again immediately. At the frequencies involved the driver experiences continuous violent vertical shaking rather than a series of separate events.

The self-sustaining loop between underfloor aerodynamics and suspension that produced the vertical oscillation seen on the 2022 cars. Each step causes the next, which is why the motion continues without any input from the road surface.

The distinction that matters is between this and ordinary bouncing. A car crashing over a bumpy surface is being shaken by the road, and the cure is suspension. A porpoising car on a billiard-smooth straight is shaking itself, and no amount of damping tuning removes the cause, because the energy is being supplied continuously by the airflow. This is an aeroelastic instability, the same family of phenomena as flutter in an aircraft structure, and it appears wherever an aerodynamic force depends on a displacement that the force itself produces.

Formula 1 suspension makes the problem worse in a way that cannot be avoided. The platform has to be held rigidly for the floor to work at all, so the springs are extremely stiff and the available travel is small. That stiffness is exactly what removes the compliance that might otherwise have absorbed the oscillation. The design requirement and the failure mechanism are the same requirement.

What porpoising actually cost, and what fixed it

The visible cost was physical. Drivers reported blurred vision on straights, spinal and neck loading over race distance, and difficulty reading braking references at the point of the lap where getting them wrong matters most. The governing body eventually intervened on medical grounds, which is not something it does often about a performance phenomenon.

The performance cost was less obvious and arguably larger. A porpoising car is not producing its peak downforce. It is alternating between too much and almost none, and the average across the cycle is worse than a stable car running slightly higher would produce. It is also unpredictable, and a driver cannot commit to a fast corner in a car whose grip arrives in pulses of unknown timing.

Every available remedy required paying for it somewhere.

Raising the ride height removed the car from the part of the map where the stall lived, and cost downforce at every corner on the circuit. Stiffening the platform still further changed the frequency at which the loop could sustain itself, and cost mechanical compliance, kerb usability and driver comfort. Redesigning the tunnel so that separation arrived gradually rather than all at once cost peak downforce and bought a wider window in which the car was usable. Most teams did all three in some proportion, and the ones that recovered fastest were generally those whose floors had been designed with a tolerant stall characteristic from the outset rather than a higher peak.

The regulations moved as well. A mid-season technical intervention introduced a measured metric for permitted vertical oscillation, giving the governing body an objective threshold rather than an argument about how bad a car looked on television. For the following season the floor geometry itself was changed: the floor edges were raised, the diffuser throat was raised, and the structural stiffness requirements around the floor edge were increased so that teams could not recover the lost performance by letting the floor deflect downwards at speed.

Ride height sensitivity is the permanent version of the problem

Porpoising was the dramatic symptom. The underlying condition never went away, and it defines how a ground-effect car has to be operated.

Invented index units in which the downforce at the reference ride height equals 100, plotted against ride height as a percentage of that reference. Constructed to show the steep rise and the collapse below the stall point. Not measured data, and the location of the cliff differs for every floor design.

Show the numbers

Worked example: the shape of a ground-effect ride height curve

Item	Underfloor downforce index
60% of reference height	48
70%	132
80%	122
90%	111
100%	100
115%	87
130%	76

Read the chart from right to left and the appeal is obvious. Lower the car and the downforce climbs steadily, then more steeply. Keep going and it falls off a cliff. The optimum is on the edge of that cliff, which is precisely where no engineer wants to place a car that will be driven over kerbs by a human being at three hundred kilometres an hour.

Everything a team does about ride height flows from this curve. Fuel load changes it, because a car burns off a substantial mass across a race and rises as it does, so the aerodynamic platform at the end of a stint is not the one it started with. Tyre wear changes it, because a worn tyre has a smaller rolling radius. Braking and acceleration change it continuously by pitching the car, which alters front and rear clearance in opposite directions and therefore moves the aerodynamic balance under the driver at the most demanding moments of the lap. Kerbs change it violently for a fraction of a second.

A team's setup work is largely the business of keeping the floor inside the good region of that curve for as much of the lap as it can manage, and accepting the losses everywhere else. Springs, dampers, ride heights and anti-roll bars are aerodynamic components in this era, and treating them as comfort or mechanical grip devices leads to a car that is quick in a simulation and undriveable on a Sunday.

There is a knock-on into rubber that is easy to miss. A floor cycling in and out of its efficient region delivers uneven load into the tyres, and uneven load produces uneven temperature, which shortens the usable life of a set. Some of what looks like a tyre problem in a team's race strategy started life as a ride height problem several hours earlier.

How the rules keep the car off the ground

Since the aerodynamic incentive points relentlessly downwards, the regulations need a mechanism that makes going too low expensive rather than merely risky.

That mechanism is the plank. A skid block of specified material runs along the underside of the car on the reference plane, ten millimetres thick, and after the race it must still measure at least nine millimetres where the rules require it to be measured. Titanium skids set into it produce the sparks visible on television, and their thickness is checked too. Wear beyond the limit is treated as proof that the car ran lower than permitted, and the standard consequence is disqualification.

This turns ride height into a risk calculation rather than an optimisation. A team sets the car as low as it dares with a margin for a bumpy surface, a heavy fuel load, an aggressive kerb or a longer race than expected, and the margin is performance deliberately left on the table. Teams that misjudge it lose results after the flag, in a scrutineering bay, for a reason that had nothing to do with what happened on the circuit.

It also explains why circuit surface quality matters more in this era than it used to. A resurfaced, smooth track lets every car run in the strong part of its curve. A bumpy street circuit with heavy braking zones forces everyone upwards, compresses the field, and produces racing that looks different for reasons no commentator can see.

Where the concept goes next

The regulations do not stand still, and the direction of travel is towards floors that are less extreme in their ride height dependence rather than more.

The sport has now run a complete rule cycle with venturi tunnels and learned what the concept costs as well as what it buys. Peak downforce achieved in a narrow band of ride heights turns out to be worth less than a slightly smaller number available across a wide band, because the second one is downforce the driver can actually use in a race, on a kerb, in a gust, behind another car. The floor changes made after the first season of the cycle were all in that direction, and the thinking behind the next generation of cars continues it, with a less aggressive underbody paired with movable aerodynamic devices that give back some straight-line speed the tunnels take away.

The historical pattern is consistent. Ground effect gets outlawed or curtailed not because it is slow, and not because it is inelegant, but because the versions that produce the largest numbers behave badly at their limits. Every regulation in this area, from the skirt ban to the plank to the oscillation metric, is an attempt to keep the concept and remove the cliff. The broader argument about what a set of aerodynamic rules is trying to achieve runs through how downforce is generated across the whole car, and the floor is where that argument has always been hardest.

What to watch for on a race weekend

The floor is invisible in motion, but its behaviour is not.

Watch the sparks. Titanium skids strike the surface when the car is at its lowest, so a shower of sparks at the end of a straight tells you the car is compressed to the bottom of its range and the team has left very little margin against the plank.

Watch a car over a bumpy braking zone. A car that skips or shimmies under heavy braking on an uneven surface is losing and regaining underfloor seal, and the driver is being asked to trust a grip level that is changing several times per second.

Watch the first laps of a stint. A car heavy with fuel sits lower, and some cars are visibly happier once they have burned some off. A driver who is nowhere for eight laps and then comes alive is often telling you where their floor's window sits.

Watch kerb usage across the field. Teams whose cars can take a kerb without unsettling the floor will use them lap after lap. Teams whose cars cannot will take a wider, slower line through the same corner all afternoon, which is a floor characteristic showing itself as a driving style.

Listen for ride height talk on the radio. A driver reporting bouncing, or a team asking about a bottoming noise, is describing an aerodynamic problem in mechanical language.

Compare team mates over the same kerb. Two cars with identical floors and different ride height choices will use a chicane differently, and the one riding the kerb harder has usually accepted a higher, safer platform in exchange for the freedom to use it.

Watch the scrutineering news after the race. Plank wear disqualifications are rare, and when one happens it means a team pushed its ride height margin further than the surface allowed. That is a setup gamble losing, not a random misfortune.

Judge a ground-effect car by consistency rather than by peak pace. The quickest single lap in this era is usually available to whoever is willing to run lowest, and the race generally goes to whoever built a floor that keeps working when the car is not exactly where the simulation put it. Further writing on the engineering behind the sport is collected in the motorsport section, and the full set of explainers across every sport we cover sits on the blog index.