This is pretty open to debate so I'll throw in my .02. I have a decent amount of experience flowing heads on a bench, porting heads and manifolds, talking to a handful of guys who have been developing cylinder heads for OE and racing applications for longer than I've been alive, and I've spent the past 4 years getting my degree in automotive engineering, so I like to think I have a decent grasp on engine airflow dynamics.

To start with what was brought up first, oversized valves. IMO, 99.9% of people do not need oversized valves in their evo. Valve sizing has a lot of determining factors but the largest two are port size, and cylinder bore size. I haven't measured the port size in the evo head yet but just from eyeballing it, it doesn't look large enough that the valves are over restricting it.

The larger factor though is bore size, more specifically, the valve to bore ratio. Pretty easy to measure, just the area of the valve (multiplied by 2 for a 4 valve) over the area of the bore. There is a certain point of diminishing returns on increasing valve size and this is usually pretty easy to find by looking at the valve/bore ratio. The reason for this point of diminishing returns is because the larger your valves get, the more shrouded they become. 4 valves are especially bad in this sense because the larger the valve gets, not only does it become more shrouded by the cylinder wall but it also becomes more shrouded by the valve next to it as well.

I haven't run the numbers for an evo 4G63 but the evo has 0.5mm larger intake valves, 2.5mm larger exhaust valves, and a 2.5mm smaller bore than the last engine I did the math on and that engine was already just past the point of diminishing returns on the intake side and was getting close on the exhaust side, IIRC I found that +1mm valves were about as big as needed to go on the exhaust and the intake was fine at stock size.

Another thing that relates to valves, though doesn't determine sizing for them is their relation to cams. Cam selection in terms of lift can often be determined by valve size. By looking at the lift/diameter ratio you can find where the point of diminishing returns is on lift. L/D ratio is simple, just divide the lift of the cam by the diameter of the valve. The magic number for this is often around 0.35. Going beyond means you'll pick up fractions of a percent of flow. If you look at most grind options for evos you'll see that most fall below this 0.35 number. The reason for the point of diminishing returns is because this 0.35 number is often right around where the curtain area (curtain area = (valve head area - valve stem area) * valve lift) becomes larger than the intake port so the point of greatest restriction in the port is no longer at the valve. Often times porting the head to move the restriction back to the valve means making the ports so large that you lose tons of velocity in the air charge which means you'll lose tons of driveability. Bigger isn't always better.

So the disclaimer here is yes, larger valves will flow more air, but it doesn't mean you'll make much more power. So it's up to you to decide if paying for a set of valves, seats, getting them ground, new guides (guides must be replaced when seats are ground) and all the other necessary machine work to pick up maybe a couple percent in power. Lastly, larger valves don't last as long as smaller valves. The reason being that for a valve guide clearance X, a larger valve will move off center relative to its seat more than a smaller valve because the radius is larger. Because of this larger off center, it is more likely to wear the seat and face of the valve unevenly which will result in a poorer seal over time.

So with that mass of text out of the way, I'd say that the biggest reason to move away from stock valves would be to move to a superior material that can handle more heat, or has a coating that carbon won't adhere to easily or rejects heat, weighs less, etc.

The absolute most important parts to a cylinder head are the valve seats and the port bowls. A properly shaped bowl and valve seat profile are key to making power on an engine. Some cylinder heads are blessed with good bowls from the factory, others require extensive work to keep up, once again I haven't spent much time examining an evo head to look at the bowls but I was pretty satisfied with how the short turn radius looks when I last had my head off. The other key part is the valve seat profile. You'll see much better gains from a good 5 angle valve job than you will from going to larger valves. Furthermore, larger valves usually only pick up power on the top end, whereas an improved seat profile will pick up power across the entire powerband. And likewise, the valve face is equally important. You'll pick up more power on back cut valves because their profile has a better flow promoting shape and there's usually isn't much of a downside to a back cut.

Regarding other components in the head, valve springs are obviously important. The largest factor to selection is how high the engine will be revved. Lighter valvetrain components are always better because lighter parts put less stress on valve springs. Valves are probably the best place to save weight, usually done by changing material (titanium $$$), or moving to hollow stems, or moving to tulip shaped valves (or machining them to be tulip shaped). Other places that weight can be saved are in the spring itself, and the retainer.

One other thing I'd like to briefly cover about springs is that boost level does not affect spring pressure (in most cases and I'll explain the exception). I know plenty of people will disagree with me on this so I'm going to explain this and then not talk about it again because I've gotten into long debates about this and I'm sick of it honestly. So the reason that boost pressure doesn't affect valve spring pressure is because there's a pressure differential across the valve and it decreases to be negligible. As the cylinder fills and the pressurized air flows across the valve, the cylinder will pressurize and if your system is efficient enough, those two pressures on either side of the valve will equalize and it would be as if there is no pressure on the valve at all. Furthermore, the cylinder pressure will exceed the intake tract pressure as the piston rises, which would actually help to close the valve. The exception to this is on inefficient systems (most of the time supercharged applications or turbos that have extremely restrictive turbine housings). How do you know a system is inefficient? High back pressure in the exhaust. If the back pressure is high, it can hold exhaust valves open longer than they're supposed to be. The reasoning for this is exactly the same as the intake side. As back pressure rises, it will attempt to push the exhaust back into the lower pressure cylinder and this positive pressure differential holds the exhaust valve open. (so much for keeping that brief)

Valve guides are somewhat important. For the most part, OE ones do just fine unless you're doing something crazy and have super high EGTs or cylinder temps and you're melting valves. Valves dissipate their heat through the face when it's in contact with the seat and through their stems. Different material guides transfer heat better. Most OE guides are powdered metal steel so they do an okay job of moving heat but materials like bronze manganese do a much better job and they also have better wear properties. The same goes for valve seats. They're often powdered metal steel in OE applications but moving to a beryllium copper seat will dissipate heat better and keep your valves cooler. Stuff like this is extremely important for titanium valves because titanium transfers heat very poorly so these better materials help keep titanium valves alive.


That's kind of mostly what I've got to say about headstuff, hopefully some of you find it useful information. Damn, I suck at keeping **** short.