LME Dyno Cathedral Port Versus Rectangle Port Heads

LME Dyno Tests Cathedral Port Versus Rectangle Port Heads

In pretty much any subject which is discussed on the internet, there are going to be two constants. The first is that there will be two distinct approaches to solving a problem, with a “rule of thumb” explaining when one is better than the other. The second is that someone will disagree with that rule of thumb.

In this particular case, we are discussing intake port shapes on Gen III/IV LS engines, and the widely held internet rule of thumb is that cathedral port heads produce more torque, while LS3-style rectangular ports will make a higher overall horsepower number. “There’s this online debate constantly occurring, as well as us constantly getting phone calls, asking what should customers run, aftermarket cathedral port heads or if they would be better off with aftermarket or even production LS3-style heads,” shares Bryan Neelan, owner of Late Model Engines .

Late Model Engines just happened to have two identical shortblocks on their dyno back to back, one topped with a set of cathedral port heads, and one with LS3-style rectangle port heads. Neelan decided that was too good of an opportunity to let slip by, and made a video comparing the results of the two dynos.

Cathedral port vs. rectangle dyno results. What heads would you put on your engine?

Posted by Late Model Engines on Friday, October 12, 2018


Both engines utilize 427 cubic-inch shortblocks with 9.8:1 static compression ratios. Beyond the shortblock, both engines utilize a Holley Hi-Ram intake manifold, and two-inch primary headers from American Racing Headers. “The tests were run on two different back-to-back days, but on the same dyno, and with the same shortblock,” Neelan says. “While not identical, the camshafts are very, very similar. Really, there were as many similarities as was possible.”

Where the two engines differ are in the cylinder head choice. One engine utilized a set of TEA-ported Trick Flow 245 cathedral port castings. On the second engine were a set of Late Model Engines CNC-Ported 280cc Brodix BR3 rectangular port heads. Both top performers in their respective applications, but distinctly different from one another.

The results of the test were fairly significant. “We just happened to run identical engines back-to-back with the same shortblock, same holley Hi Ram intake, and the same headers. Both engines were run on pump gas, and the Trick Flow-headed engine wanted about 0.5-degree more timing than the LS3-headed engine, with very similar timing curves,” says Neelan. “At the end of the tests, [the cathedral port engine] made 645 horsepower and 545 lb-ft of torque, and the LS3 heads made 671 horsepower and 560 lb-ft. That’s a 26 horsepower gain and 15 lb-ft gain with the LS3 head.”

While it might seem only logical that the engine with the larger heads made more power, that’s not necessarily what the internet “rule of thumb” says it should be. “Those are real results there,” Neelan says. “The LS3 style made both power and torque over what many would think would be a higher torque number combination [the cathedral port] because of the higher velocities and smaller port. This test shows that contrary to that line of thought, [LS3-style heads] basically annihilate the cathedral port.”

While Neelan’s statements are based on real world testing, we’re sure that there is someone who will disagree, because after all, it is the internet, and a subject with this many opinions will never be settled by one video and one test. However, it definitely appears to be an additional feather in the cap of the rectangular port camp.

The red trace is the rectangular port LS3-style heads, while the black trace is the cathedral port heads. While there is a clear difference between the two above 5,000 rpm, we have to admit it looks like there might be some face-saving difference in the power and torque below 5,000 rpm for the cathedral port crowd. This one simple test that happened because of sheer coincidence of dyno scheduling certainly isn’t going to settle the debate that has been going on since the LS3 was released.

Post a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.