When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Thanks, so the intake flows about what a stock head inlet runner flows at peak lift.
Before spending any resource on a new intake it may be far more useful to design tuned exhaust manifolds, though this may not happen considering the lack of room. It seems only the flappy is acting on pulse waves but it's working alone against excessively long (for a flatter torque curve up top) inlet runners and against the untuned exhaust manifolds.
Thanks, so the intake flows about what a stock head inlet runner flows at peak lift. Before spending any resource on a new intake it may be far more useful to design tuned exhaust manifolds, though this may not happen considering the lack of room. It seems only the flappy is acting on pulse waves but it's working alone against excessively long (for a flatter torque curve up top) inlet runners and against the untuned exhaust manifolds.
For a street turbo engine, one is better off with a log manifold than long-tube headers, all things considered. It's not really a choice anyway, because nothing but a log-style manifold fits without chassis modifications anyway. Take the exhaust side as a given.
The intake manifold can be modified to distribute evenly on its own. Given the exhaust, the actual mass flow distribution between cylinders won't ever be perfect anyway, even if every runner on the intake side would be exactly identical.
The ported S4 intake manifold is pretty much the perfect intake manifold if one wants a flat torque curve (without boost) 2500-6500 rpm. It can get you 300 rw lbf-ft in that range while peaking at 350 rw lbf-ft, normally aspirated and with a higher compression. If one matches the turbos to that rpm range, that's a good street engine right there, in my opinion. Turbos currently in my car are a little big for that, but hey it's a personal science experiment.
I may be wrong but the latest dyno results in post #2077 suggest to me you might be able to flatten the torque curve by advancing cam timing by a few degrees. The slightly too large turbos coupled with a little too much valve overlap perhaps may be why you do not see 300 ft.lbs no sooner than 3200 rpm. As you know this should move peak torque rpm earlier thus flatten the top end of the torque curve. Perhaps then adjust the flappy opening and closing points to smoothen the edges. You may have tried all of this already?
Yes this is a great project and one of the most interesting I have been able to read about in the last years, all platforms included. Many thanks to you and John Kuhn for openly sharing about your progress and hats off for the huge work.
The lack of room for tuned headers suggests to me that a supercharger may remain the most straightforward way to improve significantly and for more or less reasonable costs the performance of a stock 87+ engine.
The lack of room for tuned headers suggests to me that a supercharger may remain the most straightforward way to improve significantly and for more or less reasonable costs the performance of a stock 87+ engine.
Ok. Hang on....
You are not new here so this should be a reminder and not come as a surprise:
On the 928 board, for historical reasons, any comparative discussion of turbo vs. supercharger is a bad idea and more-or-less verboten. I’m not the Thread Police so do as you will. But threads that do mix these topics go sideways quickly and end up in The Abyss with, usually, multiple bannings. If you want to talk about supercharging there are multiple threads you can necropost back to the front page.
The lack of room for tuned headers suggests to me that a supercharger may remain the most straightforward way to improve significantly and for more or less reasonable costs the performance of a stock 87+ engine.
Originally Posted by worf928
Ok. Hang on....
You are not new here so this should be a reminder and not come as a surprise:
On the 928 board, for historical reasons, any comparative discussion of turbo vs. supercharger is a bad idea and more-or-less verboten. I’m not the Thread Police so do as you will. But threads that do mix these topics go sideways quickly and end up in The Abyss with, usually, multiple bannings. If you want to talk about supercharging there are multiple threads you can necropost back to the front page.
Funny.
Supercharging makes no sense compared to turbo, until you consider the cost! ;-)
I may be wrong but the latest dyno results in post #2077 suggest to me you might be able to flatten the torque curve by advancing cam timing by a few degrees. The slightly too large turbos coupled with a little too much valve overlap perhaps may be why you do not see 300 ft.lbs no sooner than 3200 rpm. As you know this should move peak torque rpm earlier thus flatten the top end of the torque curve. Perhaps then adjust the flappy opening and closing points to smoothen the edges. You may have tried all of this already?
The heads flow so well that advancing the cams don't gain much. These cams are on S3 cores, which the factory installs significantly retarded in S3's. The cams in the blue car are already significantly advanced relative to that (installed about straight up).
There's not much valve overlap to speak of with these cams, these are cams that would pass emissions and idle like a champ. I don't see any chance of the current cams having too much overlap for the current combination in the 3500-6000 rpm range. What more or less overlap would do outside that rev range is complicated.
The reason why we don't see 300 rw lbf-ft that earlier, like a normally aspirated engine would, is mostly the lower compression and secondarily due to no long-tube headers.
The only remaining way to move increase low end torque with the current system is basically installing various intake manifold spacers to reduce the first, flappy closed resonance peak rpm.
On the 928 board, for historical reasons, any comparative discussion of turbo vs. supercharger is a bad idea and more-or-less verboten. I’m not the Thread Police so do as you will. But threads that do mix these topics go sideways quickly and end up in The Abyss with, usually, multiple bannings. If you want to talk about supercharging there are multiple threads you can necropost back to the front page.
In a personal context and since I have already tuned several German turbo cars I quite like the idea of making my teeth on a possibly supercharged 928 in the future, and I enjoy the open approach and exchanges with ptuomov in this thread, as even if we have worked on different engines we seem to have observed and learned some similar things. I would agree that turbocharging is generally speaking a superior solution, but bring the costs in and the comparison with supercharging becomes apples vs oranges IMO.
The heads flow so well that advancing the cams don't gain much [...] There's not much valve overlap to speak of with these cams, these are cams that would pass emissions and idle like a champ. I don't see any chance of the current cams having too much overlap for the current combination in the 3500-6000 rpm range. What more or less overlap would do outside that rev range is complicated [...] The reason why we don't see 300 rw lbf-ft that earlier, like a normally aspirated engine would, is mostly the lower compression and secondarily due to no long-tube headers.
There has to be an amount of cam advance past which you should see a noticeable difference down low, and if, when reached, it messes up the top end badly by getting your knock counts to explode then this may mean there may be an imbalance between your static compression ratio and your amount of valve overlap, like too much overlap and/or too little static compression ratio. Perhaps the slot for the stock cam gear key is just not wide enough to allow advancing the cams enough to reach this point?
What is the back pressure to boost ratio with your current cams? I think you said it was close to 1:1 but have you tried to increase boost and lower ignition timing to see at which boost level the ratio increases more or less sharply?
There has to be an amount of cam advance past which you should see a noticeable difference down low, and if, when reached, it messes up the top end badly by getting your knock counts to explode then this may mean there may be an imbalance between your static compression ratio and your amount of valve overlap, like too much overlap and/or too little static compression ratio. Perhaps the slot for the stock cam gear key is just not wide enough to allow advancing the cams enough to reach this point?
What is the back pressure to boost ratio with your current cams? I think you said it was close to 1:1 but have you tried to increase boost and lower ignition timing to see at which boost level the ratio increases more or less sharply?
The S4 heads flow a lot at 75 degrees ATDC regardless of the cam timing, ports are very large, and the cams are very short duration anyway. I don't think you're going to see any additional low-end torque from advancing cams past straight up in a 928 S4. PorKen has published a number of cam-timing graphs here, maybe search for those?
The wastegate is fully closed at these rpms, so the boost control is not doing much.
Supercharging makes no sense compared to turbo, until you consider the cost! ;-)
And this is where it always starts and quickly goes down hill. It all ends in banning of folks and abyss'd threads. The person who enjoys your patronage was a casualty of one of the past conflagurations.
Originally Posted by Thom
In a personal context and since I have already tuned several German turbo cars I quite like the idea of making my teeth on a possibly supercharged 928 in the future, and I enjoy the open approach and exchanges with ptuomov in this thread, as even if we have worked on different engines we seem to have observed and learned some similar things. I would agree that turbocharging is generally speaking a superior solution, but bring the costs in and the comparison with supercharging becomes apples vs oranges IMO.
Some folks cannot participate dispassionately in a discussion and take great umbrage if one or the other is designated as superior for any reason whatsoever. Name calling and personal attacks follow. And then consequences.
I've got no dog in this hunt other than that I love this thread and don't want to see it disappear again, and this time for good.
And this is where it always starts and quickly goes down hill. It all ends in banning of folks and abyss'd threads. The person who enjoys your patronage was a casualty of one of the past conflagurations.
Some folks cannot participate dispassionately in a discussion and take great umbrage if one or the other is designated as superior for any reason whatsoever. Name calling and personal attacks follow. And then consequences.
I've got no dog in this hunt other than that I love this thread and don't want to see it disappear again, and this time for good.
So few even left, here, that understand the context of which you speak. You are now talking, literally, of over a decade and a half. In some cases, nearly two.
This is a pump gas run with 16 psi boost. The peak power was 715 rwhp, which indicates good scaling with boost. Recall that my SWAG for the engine without boost is 340 rwhp and (16+14.7)/(0+14.7)*340 = 710 so it’s scaling nicely. With my simple formula for crank ho, I’d say we’re close to 820 at the crank. John’s goal on pump gas is 750 rwhp / 860 crank up, so boost still needs to go up.
Originally Posted by ptuomov
Finally some progress in finding a new bottleneck.
Spec 3+ clutch gives up the ghost at 720 rw lbf-ft of torque. (Rennlist mobile editor on iPhone gives up the ghost when trying to type lbf-ft...). That’s 820 or so at the flywheel and given that the Spec’s spec (puntentional) is 770 at the flywheel, the clutch did what it was supposed to do.
Spec 5 disc will go in as temporary solution.
The preliminary plan is to tune the car with Spec 5 clutch to see what the next bottleneck is, detune it for street 93 E10 and drive it with Spec 3+, and investigate other clutch solutions.
Can you please remind us what was modified in the setup and/or tune between the 2 runs above?
Can you please remind us what was modified in the setup and/or tune between the 2 runs above?
The main differences are:
- The first run was fully loaded at a lower rpm than the second run
- The first run didn't have a slipping clutch
- The first run was at lower octane, lower stoichiometric AFR gasoline than the second run
- The first run had a different fuel map than the second run
- The first run was produced to check the current state while the second run was intended to be a specific tuning exercise
- The first run had a lower maximum boost set point than the second run
The second run in there was just work in progress that I wouldn't have posted if the clutch hadn't slipped and burned. Pay no attention to that graph, other than to indicate the clutch failure.
I see, thanks for the clarification.
With a torque plateau as wide as 3K rpm and holding up beyond 6k rpm, this is in my opinion excellent and may be hard to be improved.
What are the flappy switch points?