The reason all tunnel-ram intake manifolds have V-shaped bottoms is that it presents a str
Kyle Buesing; via CarCraft.com: I have seen several cylinder head flow bench tests in books and magazines, including Car Craft, in which there is a flow reading on the exhaust, with and without a short section of pipe attached, and there are higher flow numbers with the pipe attached. Since an assembled, running engine is not used with either setup, what flow number is relevant and which one is useless gibberish? The two different numbers could mean the difference between needing a single-pattern or dual-pattern cam. Tricks on a flow bench to change the numbers mean nothing if it does not reflect something that would actually be used on a running engine, such as attaching an intake manifold to the cylinder head on the flow bench. This has made me curious for quite a while, and any light you could shed on the subject would be greatly appreciated.
Jeff Smith: I've noticed the quality of questions over the past couple of years has been steadily improving in that they are more in-depth and reveal our readers are putting some genuine thought into them; great job, guys! This is especially true with this question, Kyle. Not long after we began testing cylinder heads on the flow bench, I learned that adding a short length of pipe to the exhaust port tended to improve the flow numbers by roughly 5 percent (slightly higher at higher valve lift numbers and less for very low lift numbers, such as 0.100 inch, where the volume and flow velocities were very low). This was not intended as a trick as such but rather an attempt to replicate the real-world situation in which the exhaust port is mated to a header tube. The reason the pipe is short is because the flow effect of the pipe occurs mainly in the first 6 or 8 inches, so the pipe can be short and still see the same net results as if it were 32 to 36 inches in length. If your question has to do with which number is more accurate, I would say the flow numbers with the pipe.
For this same reason, if you were really interested in accurate flow numbers on the inlet side, I would flow the head with the intake manifold you were planning on using so you would have a more accurate idea of the difference in flow with and without the manifold. Keep in mind that any additional ducting on the inlet side will be a restriction, but a better intake port is still going to flow more than a poor one, regardless of the manifold choice.
[A] [B] Any bend in tubing will create a short side radius (A) and a long side radius (B).
Your question about camshaft choice based on the exhaust side is also well taken. The standard seems to be if the exhaust port is capable of 70 to 75 percent (or more) of intake flow, you should consider choosing a single-pattern cam (the same exhaust duration as the intake) or a cam with only a small amount of additional exhaust duration—perhaps along the lines of 4 to 6 degrees. Conversely, a cylinder head with poor exhaust flow to the intake (like the LS-series rectangle-port L92 heads) can really benefit from a dual-pattern camshaft with 6 or more additional degrees of exhaust duration.
Your question also reminded me of a flow bench test Jim McFarland had me do probably 25 years ago when he was the engineering vice president at Edelbrock and I was first learning how to run a flow bench. He made up four exhaust stubs that were about 12 inches in length using 1 5⁄8-tubing and welded to exhaust flanges to fit a small-block Chevy head. The first pipe was straight, the second had a 30-degree bend, the third a 60-degree bend, and the last had a very tight 90-degree radius bend just downstream of the mounting flange. He had me flow-test each pipe across the entire valve-lift curve and then compare the flow numbers. Not surprisingly, the straight and the 30-degree bend pipes flowed the best, the 60-degree pipe offered less flow, and the 90-degree pipe flowed the least. Then he had me measure the lengths of the inside and outside turns on the pipe. This represented roughly 4 inches of difference, with the inside radius obviously being shorter. We marked the outside radius point where the lengths were the same and trimmed the pipe with a belt sander until the pipe exit was angle-cut to equalize the short and long side radii. Then he had me flow-test the pipe again. Amazingly, the flow increased to the point where this modified 90- degree-bend pipe flowed almost as well as the one with the minor 30-degree bend.
The point of this exercise, as I learned from McFarland, is that you can fool air into acting like the pipe is not bent if the effective lengths of the pipe on the short and long side radius are as similar as possible. In a practical application, the idea is to minimize how tight the radius bends are with header or inlet pipes, or, if that's not possible, to attempt to equalize the lengths of the walls of the port or flow pipe. In the case of an exhaust header, this might not be possible, but at least you should attempt to make the headers exit the exhaust port as straight and as long as possible before the first bend. That's why the dyno headers at Westech are shaped the way they are. I call them Sprint Car headers because that's what they look like. This is not an accident and the difference in exhaust flow between these Sprint Car headers and a tight set of chassis headers is measurable. Another excellent example of this is the tunnel-ram intake manifold. If you look at the original design of the very early tunnel-rams, they were all flat-bottom boxes with curved runners leading to the port inlets. Then somebody realized that if they created a V-shaped floor, the ports could exit the manifold with a straight shot directly into the ports, and power improved. Development ideas like this are why horsepower numbers keep improving by reducing restrictions to airflow.