Torque Peak vs. Flow Passage Size
Essentially, this is a subject almost unto itself. However, since we're discussing port size fundamentals and their relative effect on engine performance (torque output), a brief but general discussion seems appropriate.
Here's the deal. Even though intake and exhaust flow is considered "unsteady" and "bi-directional" (at times), it is possible to determine what we'll call "mean flow velocity" (m.f.v.) at a given time (rpm). Simply stated, there is a mean flow velocity that's measurable at any rpm point, specifically at peak torque. Further, it turns out this m.f.v. is virtually the same for any engine at its peak torque rpm. It's a combination of conditions involving fluid dynamics and acoustics.
From a parts selection or component design standpoint, peak torque rpm can be influenced (sometimes significantly) by passage size or cross-sectional area. If you subscribe to the concept that a common m.f.v. will exist among engines at their peak torque rpm, then increasing passage area raises this rpm point and decreasing area lowers the point. While it's true that passage length affects torque output (long vs. short), this dimensional change typically tends to "rock" or see-saw a given torque curve about the rpm point that we'll call the "critical" flow velocity at peak torque. You've seen this characteristic when lengthening or shortening header pipe lengths. Short ones tend to add torque above the peak, longer ones increase torque below peak.
Rather than detract from the core text of this month's Series segment, suffice to say that the "critical" m.f.v. is on the order of 245 ft/sec and is largely governed by rpm, piston displacement and passage cross-section area (intake or exhaust). If you want to test this theory from a practical standpoint, do this: Study a dyno sheet to determine the peak rpm torque point. Even though this point is influenced by factors other than (for example) header pipe section area, the exhaust and intake passages are each major contributors to the determination of peak torque rpm.
Calculate the primary header pipe section area, using pipe i.d. in the calculation for area. For the sake of discussion, if you happen to be using 1.75-inch primary pipes of 0.040-inch wall thickness, this computes to an area of 2.19 square inches.
Next, determine the volume of one cylinder. If the engine displaces 350 cubic inches, one cylinder's volume (for purposes of this calculation) would be 43.75 cubic inches. Now, multiply 2.19 x 88,200 (a mathematical constant) and divide the answer by 43.75, and you'll have determined that these headers should reach their "critical" flow velocity at 4,416 rpm. Installing headers of 2.0 inches o.d. (still of 0.040-inch wall thickness) and making no other changes would yield a cross-section area of 2.895 and, using the same method of calculation, shifts the torque peak rpm upward to 5,837. Again, lengthening or shortening the primary pipe length simply adds (or removes) torque below or above this rpm point.
Now, those of you mathematically inclined have already seen that these dimensional and results relationships can be transformed into a rather simple algebraic equation that looks like the following:
Torque Peak (rpm) = (88,200) x (Passage Area) / Displacement of One Cylinder
Although not as accurate as some computer models that take more functional elements into account (including valve motion characteristics, mechanical compression ratio, cylinder pressure histories, etc.), this simplistic approach can reveal why an engine's torque output may vary (in both quantity and placement) throughout a given rpm range, when contemplating or changing exhaust systems. Given this equation and through simple transposition of terms, you can select an rpm at which you'd like a torque (volumetric efficiency) boost and solve for the passage area required to produce that increase (at that point). It's a pretty useful tool.
And, interestingly, you'll find that this approach can also be applied to intake systems (manifolds in particular) when you're either selecting parts or evaluating dyno (or track) performance of a particular engine package. If personal experience has any measure of value, I've included this method in both the initial and in-progress design of intake and exhaust systems for a number of years. It's a pretty good way to perform evaluative and directional steps, quickly.