It has been said many times that an internal combustion engine is nothing more than an air pump, and the more air it can pump, the more power it will make. As with any simplification, it is not all true, but it's a heck of a good place to start. Here, we are going to take a serious look at not only what it takes to get air through an engine, but also how to use it as effectively as possible on the way through.

The number-one flow restriction in any engine is within the cylinder head. Get air to effectively pass these restrictions and you are good to go. It's a big subject to cover, so to chop it down to more manageable bites, we are going to apply five basic rules:

1.Locate the point of greatest restriction, and work on that first.
2. Try to let the air move the way it wants to, not the way you think it should.
3. Air is heavier than you think. Keep the port velocity up and avoid redundant cross-sectional areas.
4. Mixture motion (swirl or tumble, or a combination of both) is important--do not ignore it.
5. Shape is all-important--a shiny finish is not.

Locate the Point of Greatest Restriction
Like it or not, the valves are part of the ports, and when they are closed their ability to flow is exactly zero. This means until they have opened quite a way, the valve is the main restriction to the engine's airflow. Even when the valve is at a decent lift, it still presents a tortuous path for the air to travel on its way into or out of the cylinder. Making a comparison of the valve's ability to flow compared to the rest of the port on a stock production small-block Chevy head demonstrates just how much of an impediment the valve is even when it is wide open.

First the main body, Section A. This is a highly flow efficient, rectangular cross-section straight tube and flows 300 cfm. Section B, the turn into the seat area, flows some 200 cfm. Now let's consider section C. This is the only part of a port that has a moving component--the valve. When the valve is seated, flow is zero. If the valve is lifted high enough, it moves beyond the seat's influence, but to do so takes a lot of lift. In practice, half the valve's typical full lift is about representative of its average flow potential. In our example, this is about 140 cfm. From this we can see that at some small amount of valve lift, flow restriction is 99 percent at the valve seat and one percent in the port. At very high valve lift (.75 inch or more) this situation is pretty much reversed.

Our number-one priority then is to make the valve capable of passing as much air as possible--whatever the lift involved. To do this we need to address both size and efficiency. What we are searching for is the biggest effective valve.

Having identified the most restrictive point along the engine's gas flow path, let's see what can be done to improve it. Although it may be the last operation during a porting exercise, the valve seat design is the first priority toward effectively filling a cylinder. So let's look specifically at valve seat design.

Forms for Flow
Let's consider the valve seat isolated from the rest of the port. Fig 2 shows a progression from a very basic form "A" where the throat of the port is exactly the same size as the valve, to a relatively well-developed form at D. At first sight, it would seem the hole under the valve head needs to be as large as possible so as to flow the most air. Before flow benches became popular, it was often common practice (and often written up as such) to make the valve seat as thin as possible so as to achieve the maximum throat diameter. The flow bench shows this to be a bad move. In reality, maximum flow is always a combination of size and form around the valve before and after the seat. As you can see from Fig 2, the form for the most effective valve/seat/throat combination progresses toward that shown in Fig 3.

Applying the basics outlined for seats will get good results, but to get the max requires a lot of time on a flow bench, as subtle changes can make measurable differences. Often some changes in seat angles are used by top pros to enhance some particular feature. For instance, most Cup car heads have 50- or even 55-degree valve seats. These give up a little low-lift flow, but start to pay back above .5-inch lift. Also, the steeper valve angle acts as an impact damper, reducing the valve's tendency to bounce off the seat at closure time. For big-inch under-valved engines, 30-degree seats can be advantageous, as they make the valve appear bigger than it really is and flow more air during the initial opening phase.

Seats and Port Approach Angles
Air has mass and does not like to hug a port wall around a short-side turn. That is why purpose-built race heads have steeply down-drafted ports. But when heads have to fit under low hoods, port angles have to come down. With low-angle ports, the air (at mid and high valve lifts) does not make it around the short-side turn very well. As a result, most of the air goes out of the long-side turn. This is a situation that becomes more exaggerated the higher the lift becomes. As a result, the streamlining of the port on the long side needs to cater for low, medium and high lifts, while the seat approach on the short side needs only to deal with the requirements of low-lift flow. Fig 4 illustrates what is going on with a low angle of attack port (i.e. less than about 30 degrees).

Valve Shrouding
Just so you know, we are still working on rule number one, and the subject is now valve shrouding. Remember, the goal was to have as large and efficient valves as possible. For a typical parallel- or nearly parallel-valve head, valve shrouding is about the biggest impediment to achieving that goal. Worse yet, the bigger the valve gets, the bigger the valve shrouding problem becomes. Valve shrouding seriously impedes flow and starts having a negative impact at about 75 thousandths lift. If we are talking two virtually parallel valves, then the negative effect of shrouding increases until the valve gets to a lift value equal to about a quarter of its diameter, or as it is more often know, 0.25D (Fig 5) If nothing impedes the flow and the air exits uniformly all the way around its circumference, then the valve can be said to be un-shrouded. Unfortunately, few cylinder heads have totally un-shrouded valves. Fig 6 shows the situation for most parallel-valve, two-valve engines. As you can see, the presence of the cylinder wall, and (in this case) the proximity of part of the combustion chamber wall, cuts the area that the air can use to escape from around the valve's circumference. The circles around each valve show how far the cylinder and chamber walls would have to be clear of the valve's periphery to produce an un-shrouded situation. In essence, this would be like having a valve at the bottom of a cone with the walls 38 degrees off being flat (Fig 7).

Be it the intake or exhaust, the worst part of the port for air to negotiate is the short-side turn. If the air fails to make it around the short-side turn, there obviously won't be much air exiting the valve in that area. Since the valve is not being fully utilized around the short side, there is no point in de-shrouding the valve to the extent needed on the long side. A modern high-performance head has little shrouding from the combustion chamber, but be aware, it is possible to cut the chamber wall too much while un-shrouding the valve, especially the exhaust, and pay a performance penalty. This is just one reason why a flow bench is a great asset.

If we were to plot out the flow efficiency of either valve for a parallel two-valve engine, we would find that above 0.25D of valve lift the flow efficiency starts to climb back up from a low. The increase in efficiency starts at a point when the flow stops trying to exit all around the valve and instead "windows" predominantly out one side of the valve (Fig 8). This is why two-valve engines with well-developed ports thrive on ultra-high valve lift. Since I have brought in the subject of "windowing," let's see what the implications are.

Mods to Minimize Shrouding Effects
At this point, we are easing into the application of rule number two. If shrouding is to be minimized, we must take steps to establish what alternative route the air could have to find its way into the chamber in a relatively unimpeded fashion.

A study of the flow into and out of the cylinder of a two-valve engine shows that the principle direction of entry or exit is either toward (for the intake), or away (for the exhaust) from the center of the cylinder. This being the case (and applying rule number two) we need to find a way to let the air flow more nearly along its preferred route. This is done by "port biasing" in the direction the flow wants to "window." This helps reduce the negative effect of shrouding. Applying this biasing technique can make substantial improvements in high-lift flow. Although port bias pays off from mid-lift values, it really comes into its own when lift exceeds 0.25D.