This point is emphasized so that the results that follow are not misconstrued. A good operator will be able to use a flow bench as a tool with which to optimize cylinder head ports in order to increase engine power. But a flow bench is not a dyno, and the dyno is the only true test as to whether one cylinder head works better than another.

That being said, a flow bench is still a very useful tool to use in extracting more power from an engine. The key is that the user must be experienced and knowledgeable. One can often make a port flow better on a flow bench by simply making it bigger, yet this does not necessarily yield more power on the dyno. That is mainly because a flow bench does not give very beneficial information about flow velocity. And flow velocity is critical to inertia supercharging, which along with intake and exhaust pulse tuning are the main reasons why it is possible to achieve greater than 100% volumetric efficiency in a naturally aspirated internal combustion engine. The way that air is caused to enter the combustion chamber can also have a strong impact on power production. For instance, it may be desirable to cause the incoming air to "swirl" into the chamber. This swirl is initially induced within the ports, but cannot be measured on a flow bench.

As a general rule of thumb one would like the smallest port cross section that will yield good flow. Yet there is much more to it than that. Factors such as port profile and valve seat angles must also be addressed. Through careful attention to these parameters a good cylinder head technician might well be able to increase the flow of a given port without substantially modifying its cross-sectional area, thereby increasing power over a broad range of the useful rpm band.

Another important fact to keep in mind when examining flow bench results is that a flow bench operates with a constant pressure drop across the port(s). In reality however, the pressure drop across the ports is constantly changing (along with valve lift) as the piston moves down the bore on the intake stroke (and beyond), up the bore on the exhaust stroke, and as any pulse tuning effects play their part. And the manner in which the pressure delta across the ports changes with piston position in the bore is also dependent on RPM as well as rod ratio (an often overlooked contribution). So in theory a flow bench test session should be run with the test pressure varying according to valve lift, with the pressure vs. lift data having been determined from measurements on an actual running engine. And even that data would only be applicable to one throttle setting, presumably full throttle. This could present quite a challenge. However, flow bench experts have found over the years that quite good results can be obtained by running the flow bench somewhere close to the "average" pressure drop that the port is likely to see in actual application.