You guys are very technical and it's fascinating to see the level of analysis applied to any given challenge.

My approach is far more rudimentary, I understand that the longer headers will produce better cylinder filling by simply creating a bigger vacuum effect which is probably helped with a longer tract which affects the front most cylinders most - which happens to be the cylinders with the biggest inlet manifold challenges. This will be helped by the change in cams also.

The inlet manifold 'runners' to me look terrible for the front most cylinders and inside the plenum chambers the bell mouth design / set up is a mess (sorry Porsche but you could have done better). I will make limited changes to the inlet manifold to clean up castings, match components and help the mess inside the plenum chambers - I can't employ too much science as I don't have the time to measure the individual changes or replicate the engine flow properties but I can have a best guess based on everything I see here which is the next best thing!

Many thanks, I had not seen this thread before and it will be used when I start to shark tune in a few weeks

I've made 2 of these for testing on a 6.5 liter.

No gains over stock, for me.

Just put in some plenum spacers....even better, since people can see them.

Boring out the housing and installing an oversize 80mm throttle blade increase the air flow by only 4,2% which hardly will make a noticeable gain in power. Adding a proper porting job to the housing, the air flow increase is measured to 16,5% which is the best that can be achieved and should bring with it a good power gain.
Åke

Did that years ago and love the look. Bling Bling.

Alright, since you're set on modifying the existing manifold, I have an idea for you, and it's slightly counter intuitive.

The biggest problem with the stock intake is its inability to equally fill the cylinders. Because of this, the tune must be compromised to accomdate the manifold's best-flowing cylinders, 2 and 6. As we know, for the most part, 2 and 6 will be 'perfect' as far as air/fuel goes while the remaining cylinders will be too rich to varying degrees with 5 and 8 the richest. Also, timing will be set to accommodate 2 and 6.

Since it is impossible to make 5 and 8 flow as well as 2 and 6, how can we help equalize flow to all cylinders?

What we can do then, is to decrease the flow thru 2 and 6, thus bringing them closer to 5 and 8.

But how can we do that without worrying about the new restrictor breaking-off and being ingested into the engine?

How about simply cutting the trumpet-ends off of runners 2 and 6?

The flow difference between a straight-ended pipe versus a pipe with a trumpet is quite dramatic depending upon which paper you read.

So lets go conservative and say there will be a 5 to 7 percent difference, thus, 2 and 6 will now flow 5 to 7 percent less air because their trumpets are now missing.

This will bring 2 and 6's flow much closer to 5 and 8. What one can do next is retune, and the new tune can be more aggressive and safer, now that overall cylinder filling is more equalized amongst them.

Of course I haven't performed the surgery nor dyno'd the result since I thought that up while programming today. However, theoretically, it does make sense.

The best part about the idea, it's a relatively easy fix/compromise to help get around the short-comings of the stock 32v intake.

What evidence do you have that, say, cylinder 6 fills better than say cylinder 3 at 6000rpm? My data says that, with stock exhaust manifolds, cylinder 6 actually fills relatively poorly at and above 6000rpm compared to other long-runner cylinders. Cylinder 6 does knock relatively easily at high rpms with exhaust manifolds, but that's not because of the better filling. Both poor filling of cylinder 6 and its high likelihood of knock are both caused by the cylinder 5 blowing into the exhaust manifold right after cylinder 6 and causing a pressure spike at high rpms in the exhaust port of cylinder 6 during cylinder 6's overlap. That prevents the cylinder six from filling well with fresh charge and contaminates the charge with hot exhaust gas. In my opinion, how each cylinder runs at each rpm changes pretty dramatically if you switch from stock exhaust manifolds to headers.

Since the original poster is planning to run headers, I would recommend the following. Up the compression as much as the cams and available fuel allow. 11:1 at least? Get the experimental JDS Sharktuner code version that allows for cylinder-specific ignition trim if there's any indication that the long-intake-runner cylinders knock easier at mid range rpms, and selectively back off timing from those cylinders. When porting the intake, don't intentionally make any runner to flow worse than it could, just get the most flow out of each cylinder without increasing the cross-sectional area any more than necessary. What this means in practice is improving the bell mouths and somehow smoothing over the head-to-intake transition without expanding the cross-sectional area. Finally, don't worry too much about fueling, as all indications are that the power that 928 S4 makes isn't very sensitive to small variation in fueling in the neighborhood of low-thirteens AFR. That's what I'd do, anyway. Goes without saying, this is just my opinion.

You can and should resurface the heads and continue to use the thinner head gasket. That'll up the compression a bit.

Any normal long-tube 4-1 headers will achieve the pulse separation you need.

Some interesting comments there. I found that 6 and 2 control the amount of advance that can be applied at higher rpms. I also assumed that those cylinders were flowing more air thus running leaner so your comment is interesting. What I did find through experimentation with the patch that Ken wrote for me was that 6 and 2 need to be pulled by about 3 degrees to get similar knock characteristics in the higher rpm range but then the patch was an across the board type of thing, not cell specific. This led me to wonder whether one can increase the timing across the board and rely on the knock sensors to pull timing on the "noisy cylinders" as needed. This also told me that the motor needs 98 RON and that 95 RON is a bit limiting. That I had to pull a bit of timing compared to stock was a bit disappointing thus but then I later learned that having a faster exhaust compared to stock also suggests less advance needed.

I also found that AFR did not impact advance in the value range 12 to 13.5. I was also advise dthat the 928 inlet manifold is quite efficient and that best power is typically made in the low 13's.

Hopefully Marti will advise what he finds when Sharktuning.