10.5-inch Runners: 638 ft-lbs @ 5,400 rpm
7.25-inch Runners: 620 ft-lbs @ 5,400 rpm
Largest Gain: 62 ft-lbs @ 3,900 rpm
The 16½-inch runners were the clear winner in torque production up to the crossover point of 5,400 rpm. The torque gains were as high as 62 ft-lbs at 3,600 rpm over the shortest runner length. The 10⅕-inch runners offered more torque than the 7¼-inch runners all the way up to 6,500 rpm. Only above that did the short-runner combination excel.
Working with the intake and camshaft, the cylinder heads are part of the trio of performance components that dictate the power output of the engine. In the case of the LS3 and (especially) LS7, the factory heads offer exceptional airflow. Unlike cathedral-port heads (706, 317, 243, etc.), it is difficult to improve upon the power output of the already impressive factory heads.
I have seen power gains eclipsing 70 hp when upgrading cathedral-port heads (on a 408 stroker), but the gains were nearly half that (or less) when replacing factory LS3 heads (on a larger 468 stroker). The reason for this is not that the aftermarket doesn’t know how to produce a good LS3 head, but that the factory LS3 heads already flow enough to support such high power levels. A stock LS3 head flows near 318 cfm. This compares to the very best cathedral-port head (317 or 243) that flows 244 cfm. The difference between an LS3 and the best factory cathedral-port is more than 70 cfm.
I have exceeded 690 hp using stock LS3 castings on a 468 stroker, and the stock LS7 heads are even more impressive. Run on a 495-inch stroker, the stock LS7 heads produced 773 hp. The impressive head flow offered by the stock heads is both a blessing and a curse. On the plus side, they offer impressive power right out of the box, but just don’t expect huge power gains when upgrading the heads on your LS3 or LS7.
Bolting on the right set of LS3 or LS7 cylinder heads can yield impressive power gains.
Factory LS3 and LS7 heads offer exceptional flow and power potential, but CNC-ported, aftermarket heads offer even more.
To understand the reason for this, you need to first understand the correlation between airflow and power potential. There is, of course, an equation to calculate the “potential” horsepower offered by cylinder heads using the airflow data. This formula is:
Using an LS3 (with 318 cfm) as an example, the formula indicates (.257 × 234 cfm × 8) that the stock heads will support 653 hp (though have made more than the formula suggests). Were you to upgrade the heads on a stock or mild LS3 (making less than 560 hp), the gains offered by the head swap might be minimal because the stock heads already flow more than enough to support the current power level. Test 1 in this chapter illustrates what happens when you add cylinder head flow to a mild combination.
Using a flow bench is one way to determine the power “potential” of a set of cylinder heads, but the only way to know for sure is to run them on a dyno.
LS3 and LS7 heads offered by the aftermarket are usually purchased based on flow numbers. The problem with purchasing cylinder heads based on airflow is that the airflow numbers represent only a potential power output. As in the example in Test 1, just because you have 800-hp head flow doesn’t mean your combination is in a position to take full advantage of the available flow. This is especially the case in LS7 applications, where aftermarket head flow can exceed 400 cfm (or more). It took a 495-inch super stroker (see Chapter 8) to tax the flow limits of the best LS7 heads, and the stock LS7 heads produced 773 hp.
Even the best heads were only up by 25 hp or so, the gains offered by ported LS7 heads would be even less. On a stock LS7, there may be no gain at all. This is especially the case if the maximum flow rate given for the heads you plan to purchase exceed the lift of the cam you plan to run. Big flow at .700, .750, or .800 lift is useless if you plan to use a .600-lift cam. Besides, you should be more concerned with the mid-lift flow numbers because the valve spends much more of its time sweeping through the mid-lift (opening and closing) than it does at peak lift.
For the ultimate in valve control and RPM potential, a dual-spring upgrade is the way to go on a high-performance LS3 or LS7.
The great thing about the LS engine family is the interchangeability. Cylinder heads from an LS7 physically bolt onto an original LS1, but the small bore size does not allow them to actually run without valve interference. This interchange allows the later LS3 heads to serve as inexpensive upgrades to the earlier cathedral-port engines. The most popular upgrade is to replace the stock 317 cathedral-port heads on a 6.0 truck (LQ4 or LQ9) or LS2 (243 castings) with the rectangular-port LS3 heads. The large valves in the LS3 heads require the 4.0-inch bore of the 6.0 (and do not work on smaller 4.8, 5.3, or 5.7 blocks), but the results are impressive.
The LS3 head upgrade also requires the corresponding offset (intake) rockers and intake manifold, but the stock LS3 heads offer an additional 70 cfm per runner. Tested on a 408 stroker, the LS3 head upgrade was worth almost 40 hp over the stock 317 truck heads. The smaller cathedral-port 317 heads offered more power up to 4,000 rpm, but the LS3 heads pulled away up to 6,500 rpm.
Test 1: Stock LS3 vs Chevy Performance CNC L92 on a Stock LS3
Several tests in this book were designed to illustrate what happens when you install the right part on the wrong application. Unlike factory cathedral-port applications, LS3 engines were blessed with high-flow cylinder heads. In terms of head flow, there was a substantial step up from the cathedral-port LS6/LS2 heads to the rectangular-port LS3 heads. That is why adding LS3 heads to a 6.0 is such a popular swap.
As you learn in this chapter, factory heads can support nearly 700 hp on the right application, but that doesn’t mean ported heads don’t offer any power. Just don’t expect the huge gains normally seen with cathedral-port head testing; the stock LS3 heads flow nearly 315 cfm. This test shows that, especially on a stock application, cylinder head flow was not the limiting factor in terms of performance.
This test was run on the LS3 crate engine from Gandrud Chevrolet in near-stock trim. The engine was equipped with a set of long-tube headers, manual FAST throttle body, and Holley HP management system. Everything else on the engine was left stock, including the camshaft, displacement, and compression ratio. This test was run to illustrate what happens when you increase the head flow on an engine that already has enough cylinder head.
Run with the stock LS3 heads, the stock LS3 produced 493 hp at 5,700 rpm and 484 ft-lbs of torque at 4,800 rpm. I then installed a set of CNC-ported L92 heads from GM Performance (supplied by Gandrud) that flowed nearly 350 cfm