Direct Connection ( Mullen LA ) heads

K.O. SWINGER said

This information confirm my knowledge type of casting or design it is very strong at low lift numbers and benefit from more duration. I realizing compression is a consideration but it would help you extend the RPM range while decreasing un desirable angles and possibly unstable valve train. If we were talking about Apache heads or ls'cathedrals I would say yes and jam those valves to the Moon but we're not? 6,500 or 7,000 is one thing 8,500 or more it's night and day.

Click to expand...

Who said 8500? .600 lift is nothing. I do it on the street with heads that break over at .500 without an intake manifold and at 28 inches. I know the limits of the flow bench which is why I ignored the flow numbers.

I guess this won’t make sense to you until you get a flow bench and start testing. It appears you are still in love with flow numbers and that is not a good thing.

Look at the shape of the curve. Forget about much under .150 lift because it doesn’t matter. Ever. If you stop and plot out the lift curve against piston movement you’ll see why I can say that.

Then spend some time looking very close at overlap and the heads we are discussing. Overlap is looked at as a bad thing, but IMO if overlap is done correctly, and you shape the valve seat correctly you can lose quite a few CFM at .100 or even .200 lift and make way more power.

And then there is reverse flow...as a general rule when I see big low lift numbers (compared to valve size) I can say with almost 100% accuracy that the head will flow in reverse almost what it flows forward and that is not only pissing away power in the middle it kills driveability.

Edit: seat to seat duration is a bad way to make up for not running enough lift. That right there is a sure fire way to kill the middle and make driveability go to hell. Lift averaging proves this out. Add to that guys love to drop the intake duration (or add exhaust duration depending on your philosophy) and then blow the LSA out to get the RPM they want is another sure fire way to kill the middle of the power curve. Again, this relates to overlap, the overlap triangle and header design and function. If you use junk headers the engine becomes almost oblivious to changes in LSA.