The study of engine airflow and ways to improve on-track power are ongoing challenges among engine builders. Although airflow benches are useful tools for accomplishing this goal, it's first worthwhile to have a fundamental grasp of how air and fuel move into and out of the combustion space as exhaust residue.
For purposes of discussion, let's assume we're working on improving the mass flow of a cylinder head. Even if a stock head is required, much of the following discussion applies to it in terms of how valve timing can be used to increase volumetric efficiency when stock or modified porting is involved.
Suppose we begin on the intake side of the combustion space. If we plotted mass airflow as a function of crankshaft angle, we'd discover that about 30 percent into the opening side of valve motion, the piston's descent is sufficient to create a significant pressure drop, enabling the potential for valve motion or port modification to be effective. Depending upon a given engine's stroke/rod-length ratio, at the point of greatest piston speed, you don't want insufficient port area or valve lift to create a restriction to flow. Actually, at peak valve lift, the pressure drop across the inlet path tends to stabilize as flow rate diminishes during intake valve closing.
However, effort to increase low-lift airflow can be particularly beneficial to net power. Of the methods applied here, blending the valve seat with the bowl is one that works. The best approach is to use a radius valve tool with a live pilot as opposed to traditional valve seat cutting machines. Back-cutting intake valves also tends to aid low-lift flow. As an example, for Street Stock engines, you can back-cut with a 30-degree tool so it narrows the 45-degree seat to equal the width of the valve seat. But we digress.
So, how can we apply all this seemingly-confusing language to our airflow analysis? Zora Duntov once told me that if you're going to develop an intake path (cylinder head or cylinder head and intake manifold combination), it's beneficial to "park" the intake valve at about 65 percent of maximum lift and optimize flow at this point. He emphasized that the port would experience this flow rate (intended to be at or near peak piston speed) twice in a complete inlet cycle, so take advantage of optimizing flow at this point. From experience, I share with you that it works. In addition, since flow rate tends to slow down at peak valve lift, creating higher lift points doesn't appear to be a good way to increase flow.
There is also the issue of dealing with some degree of reversion during early intake valve opening. Of course, as engine speed increases there is less time for this event to occur, thereby diminishing the ability of residual exhaust gas to contaminate fresh air/fuel charges. One practice that relates to reversion containment involves flowing intake paths both forward and backward. Some will argue that this is of no value since improvements in airflow toward the combustion space generally improve flow in the opposite direction. This isn't always the case, however. As a rule, since cylinder pressure is typically higher than manifold pressure, early in the inlet cycle, reverse flow measurements at valve lifts that occur when the piston is at TDC and beginning its descent are usually sufficient. You will also likely find that it is in the areas of valve seat widths and bowl angles that reverse flow reductions can be made.
As a potential point of interest, whether along the intake or exhaust paths, changes in passage section area will cause a corresponding change in flow rate. Even wet-flow benches have difficulty enabling evaluation of flow rate changes. Energy is required to create a flow rate change; e.g. increasing or decreasing energy robs the engine of some degree of volumetric efficiency. This goes to the issue of a change in flow momentum.
What is momentum? It's a mathematical product of a mass and its linear velocity. On the intake side, air/fuel charges basically consist of two different masses (air and fuel) trying to move at the same velocity. In a steady-state world, approaching that condition is more doable than during times when either the mass of either is changed or the velocity of both varies. This typically occurs during changes in flow path direction, changes in path section area, or both.
However, as stated in other "Enginology" columns, the two masses experience unsteady, sometimes bi-directional flow, primarily because of the interrupted flow conditions caused by valve events. It's during such changes in momentum that air and fuel can mechanically separate or net volumetric efficiency is decreased; either or both. It's obviously beneficial to at least keep these conditions minimized since we know they are not completely unavoidable.
On an airflow bench, it's obviously difficult to replicate the flow conditions of a running engine. And it's equally difficult to find ways to measure (or at least characterize) problem areas in a flow path. Some experienced bench operators can tell by the airflow sound. Others choose to combine two types of pitot tubes connected to a manometer. One is used with the open end pointing in the direction of flow, indicating port flow velocity. The other is U-shaped with the open end pointing against the direction of flow, indicating instability in the flow stream.
I have found that using them both simultaneously comes closer to identifying problematic flow, particularly on the intake side, than either by itself. In fact, when used with the open end moved along flow passage walls (especially in area when flow directional changes are made); you can spot locations of boundary layer separation (or turbulence) where air and fuel tend to diverge. Correcting these problems with passage shape, surface condition, or other such changes tends to improve volumetric efficiency and aid fuel suspension in the inlet air stream.
While there are similarities in how exhaust passages (ports) are treated, by comparison to intake passages, you will find the short-side (short turn or floor) of most exhaust ports are very sensitive to change. Creating or preserving a generous short-side radius is desirable, along with back-cutting these valves. Once again, you can use the U-shaped pitot tube to locate and (to a degree) quantify any boundary layer issues.
By the way, Duntov had a suggestion about flowing exhaust ports as well. If memory serves, he said to set the exhaust valve at about 80 percent of maximum lift and perform modifications and tests at this lift, suggesting this is where exhaust ports appear to be the most sensitive to change.
Finally, and when using the two types of pitot tubes, don't become confused by the fact each will influence a U-tube manometer in opposite ways. The straight tube will register pressures less than atmospheric (flow rate). The U-tube version won't read anything at all unless it sees a disturbance, at which point it will register a manometric reading in an opposite direction. After all, a typical airflow bench creates a pressure condition less than atmospheric. It's the force of atmospheric pressure attempting to fill that void that pushes air into the bench, as it does on a running engine. The word "suction" does not apply.
In the final analysis, at least on a flow bench (wet or dry), the rate at which atmospheric pressure can push air into the bench is largely a function of a given flow passage's configuration. Aside from any applicable environmental correction factors, atmospheric pressure is a relative constant when comparing it to the influence of passage design.
As a side note, and beyond the scope of this particular column to address in detail, it's possible to use airflow benches to examine flow conditions during the valve overlap period. To do so requires that both intake and exhaust valve lifts can be adjusted to values associated with a given overlap period. Flow is established through the intake track, passing through the combustion space, and out through the exhaust. This requires hooking the exhaust side to the bench's inlet and allowing atmospheric pressure to act on the entire system.
It's known that during the overlap period, a condition of "scavenging" (sonic tuning, wave tuning, or whatever term is used for the process), helps make it possible to achieve levels of volumetric efficiency beyond 100 percent, a common occurrence in racing engines.