An opening does not move air. A difference in pressure across a restriction moves air, and the opening only decides how much moves for a given difference. Enlarge the inlet and the difference across the core stays the same, so the flow rises only in proportion to the reduced inlet loss. If the inlet was never the problem, enlarging it changes almost nothing, and that is the most common result of front-end work.
The half of the system that gets designed is the half you can see from outside the car. The other half, the exit, is usually whatever space is left between components. It is the exit that sets the pressure on the back face of the stack, and therefore it is the exit that sets the difference the whole system runs on.
Treat the bay as a plenum. Air enters through the front, passes through one or more heat exchangers, and has to leave somewhere. If it leaves as easily as it arrives, the pressure inside stays low and the difference across the cores stays high. If it leaves badly, the bay pressurises, and the pressure behind the cores rises until the flow through them falls to match what can escape.
This is a self-correcting system in the worst possible way. The engine does not warn anyone; it simply gets less cooling air. Everything downstream of the cores, from the gearbox tunnel to the gap above the subframe, is part of the exit whether or not it was designed as one, and the smallest of those paths becomes the limiting factor.
An exit wants three things: a region of low pressure to discharge into, a path with as few changes of direction as possible, and enough area that its own loss is small compared with the rest of the system. The first is the one that decides everything else, because discharging into a high-pressure region gives no benefit no matter how large the opening.
Regions of low pressure on a moving car are specific and local. The area under the nose ahead of the front wheels is not one; the area behind the front axle generally is. The base of the windscreen is one only where the flow stays attached over the bonnet. Behind a wheel is one, but the air there is turbulent and carries debris. An exit placed by appearance rather than by pressure does nothing useful.
The path matters as much as the destination. A large exit fed by a tortuous route is a small exit in practice, because every bend and every gap the air has to squeeze through adds loss. The best exits look boring: a straight, expanding path from the back of the stack to a low-pressure region, with nothing in the way.
System loss is the sum of inlet, core and exit losses, and the exit term is easy to underestimate because it is usually a shape rather than a component. A sharp-edged opening in a panel has a substantial loss coefficient; the same area with a radius or a short diffuser has far less. That means a smaller, well-shaped exit can outperform a larger, badly shaped one, which is exactly the opposite of how most people approach it.
There is also a distribution problem. An exit that draws from one corner of the bay leaves the rest of the bay stagnant, so the cores see uneven flow across their face even though the total might be adequate. Spreading the exit across the width of the stack keeps the face velocities closer together, and a core with an even face velocity uses its whole area.
At speed, the exit discharges into fast external flow. At low speed there is no such flow, and the pressure difference across the stack collapses to whatever the fan can generate. In that condition an exit near the front can reverse: hot air that has just left comes back in, and the stack is fed with air it has already used.
Recirculation is the reason some cars overheat in traffic despite looking well vented. The vent works when moving and becomes an inlet when stationary. Preventing it needs a one-way arrangement, a duct that discharges away from the inlet, or a fan and shroud combination that keeps the bay at a lower pressure than the outside air in front of the car.
Underbody geometry decides the pressure the bay discharges into, which makes it part of the cooling system whether anyone intended it. A smooth floor accelerates the air beneath the car and lowers its static pressure, which helps extraction. A rough floor with exposed subframes and a big diffuser throat does not.
But a flat floor also reduces the amount of air that reaches the bay from below, so the bay has to be fed entirely from the front. That is a reasonable arrangement and it is how most modern cars work, but it means the front inlet has to be sized for the whole demand rather than for part of it. Improving the underbody without revisiting the inlet is a way to make the bay hotter while making the car faster.
The usual sequence of front-end development is to open the grille, add a duct, fit a larger core and then find the temperature unchanged. Each step reduced the inlet loss, which was already a small part of the total. The exit loss, which was the large part, was not touched because it was not a part anyone bought.
The diagnostic is straightforward once the idea is in place: measure the pressure in the bay and compare it with the pressure in front of the car. If the bay is close to the free-stream pressure while the car is moving, the exit is the problem, and no inlet change will help until it is addressed.
The plenum model assumes the bay is a single volume with roughly uniform pressure. In reality, components divide it into regions with quite different pressures, and a path that looks open on a drawing can be blocked by a wiring loom or an air pipe that nobody considered structural. Pressure varies from one corner of the bay to another by enough to change where the air goes.
It also stops being predictable when the exit is close to a rotating wheel or a tyre. Those regions are low pressure and highly unsteady, and the flow there alternates between extraction and ingestion depending on speed and steering angle. A design that relies on that region will behave differently on each side of the car.
Exit design is decided early, and it is expensive to revisit later.
The cores and their ducting are on our intercooler airflow pages, with the pressure side under intercooler and the test method under intercooler pressure testing. The inlet half of the same argument is under air intake hoses, and heat that should never reach the bay in the first place under exhaust heat shield and exhaust heat blanket.
