Thom

If I understand correctly the design of the S4/GTS intake manifold, the flap closes as vacuum decreases to simulate shorter runners by reducing the resonance effect through two separated plenums?
In twin plenum mode the runners still look excessively long to me to provide "enough" ram effect past 5k rpm.

ptuomov

Not to go off topic on this thread, here are two threads on the 928 S4/GT/GTS intake manifold:

https://rennlist.com/forums/928-foru...and-ideas.html

http://speedtalk.com/forum/viewtopic.php?f=1&t=35634

On this thread, I'd love to hear more from Turbo Tim how he created that "blonde's brainwave" torque curve.

Duke

Part of the reason is probably the engine dyno. E.g if you do a static test you can reach full boost in a rpm zone you would need a 8th gear to do on the street

ptuomov

Valid point. However, my car is right now on an eddy current dyno with Klam retarders and we can have it cycle thru all the load cells on the fuel map at whatever speed until the retarders overheat. The torque curve still doesn't look like a flat line. I don't think it's just loading up the turbos.

Thom

It seems to me this far that the stock S4/GT/GTS intake is a particularly complex part that, coupled with the remarkably restrictive cams, was designed to provide the widest possible torque band while maximising fuel efficiency over peak power... in a similar way to the 944T design. I can relate to the interest in understanding how the stock set up works but if I were to make a parallel with how things go with the typical highly-strung 944T engine, I see many more drawbacks to advantages with the stock intake/cam combination.
In your shoes I would simplify the intake and make something similar to the Cayenne intake, try much sharper cams and size down your turbine wheels. This is a good way to flatten the torque band and relieve the engine from working against the high rpm restriction caused by the stock intake+cams combination.
How much back pressure are you seeing with your GT35 turbines?

ptuomov

The S4 fuel maps run too rich at WOT and high rpms, so I don't think they cared about fuel efficiency there. At cruise loads and rpms, emissions and fuel efficiency were obviously a concern. Yes, I do own some better camshafts and will try those down the road - it's just that my project is moving at a snail's pace over Marco Polo's distance. It's fun though, and car's already very fast. I don't want to pollute this thread with my stuff, if someone is interested in commenting here's the right thread for that: https://rennlist.com/forums/928-foru...e-streets.html

You write that the S4 intake manifold "...was designed to provide the widest possible torque band while maximising fuel efficiency over peak power... in a similar way to the 944T design." But if the 928 S4 and 944T intake manifold were designed to give the broadest possible torque band, isn't it a remarkable accomplishment that TurboTim has been able to create a manifold that has a significantly broader and flatter normally aspirated torque curve?

I really think that TurboTim's manifold is fantastic and I would like to very much know the plenum volume and runner lengths. I likely can't ever recreate that manifold on a V8 because of the firing order and distance between banks (can't cross the valley without longer runners), but I'd still like to know those dimensions to try to understand how he did it. What might be going on that there are two manifold torque peaks, one on each side of the camshaft torque peak, resulting in this perfectly flat line. But that's just speculation.

Which cams is TurboTim's build using? I assume something like stock 3.0 16V N/A cams, right? Don't recall anything on that in this thread, except they are reprofiled.

Would these be the 3.0 16V N/A valve events?
Intake opens @1mm 3° after TDC
Intake closes @1mm 47° after BDC
Exhaust opens @1mm 39° before BDC
Exhaust closes @1mm 7° before TDC

I programmed the 944 S2 to a cheap simulator this morning, and all I get form going from assumed stock 15" runners to 4" runners is the torque peak moving up from 4000-4250 rpm to 5000-5250 rpm. But that proves nothing, especially since I had to assume a lot about the stock manifold and Turbo Tim's manifold and used stock S2 cams.

Thom

I suppose stock engines will always run a bit rich under full loads with the OEM mapping, to maximise reliability. Stock Mitsubishi EVOs run pig rich on full load.

The intake manifold is a recurring hot topic on here, and if you look at most 4 modern cylinder turbocharged engines they all feature "short" runners intakes.
Whether all units available as "upgrades" on the aftermarket were thoroughly engineered remains to be proven, but as far as I have seen long runners replacement intakes are non-existent. Some may rightfully argue that the 944T 2V head is an old design that may not work best with excessively short runners, but a few have shown otherwise, if a lot of work is needed to see those gains. On the much better 4V head it is IMO a non-brainer.

Tim mentioned earlier he used an aftermarket Supra MkIV intake plenum that he welded to 944 S2 runner flange cut from a stock 944 S2 intake manifold.

If you look at Lindsey Racing's intake for the 8V 944T head it uses a similar runner length, Refresh951's too.

The (basic) calculations on this page allowed me to get a good picture on how runner length works with pulsations. They correlate pretty well with I have exerience between the stock 944T intake and a short runner aftermarket unit in terms of moving the ram effect through the RPM band.

ETA : here is a custom intake we did for an E39 M5. Shorter runners than on a 928 intake. 500hp and 540 ft.lbs at 7psi of boost / no real problem with lack of low down torque.

ptuomov

I have no doubt in my mind that "not all upgrades on the aftermarket were thoroughly engineered!" ;-)

However, Turbo Tim's intake manifold and turbo selection are nothing short of amazing. The flattest NA torque curve I've ever seen, full boost for 3500-7000, and a flat-line boosted torque curve without RPM specific boost profile.

I've been trying to correlate simple formulas with results, and I have grown skeptical of any simple formulas that can be implemented in an Excel spreadsheet. I did the two chamber, two pipe Helmholtz resonator model, and it's not very predictive of reality. The reality is a combination of something like a finite wave with the valve open, something like an acoustic wave with valve open, and something completely different during the overlap. I don't think any spreadsheet is going to get you closer than the bread ballpark.

One thing that I am kind of interested in whether you see these short-runner manifold engines respond to the plenum volume or not. If not, that would simplify the problem a bit.

ptuomov

Here's the flattest curve we've been able to run with a constant boost profile and with the stock intake manifold with the 2x944 engine:

https://rennlist.com/forums/11315969-post814.html



Not making much more than Turbo Tim yet, despite 40% more displacement, and the torque curve is not as flat. I hope we get there one day...

Sorry for the hijack, I'd really like to hear more from Turbo Tim how he got that flat torque curve and what the intake manifold dimensions are.

EliteThink

wow impressive.

turbotim3

Well I'm getting there slowly. She's finally up and running and hopefully up the road at the weekend. Hoping for some dry weather.

Cheers Tim

ptuomov

Any updates?

ptuomov

Any updates, especially on the manifold? What runner length, runner diameter, runner taper, and plenum volume?

turbotim3

Sorry but I've been having problems with mounting a centre push clutch slave cylinder. Should be sorted shortly.
As for the manifold, I haven't got any measurements. As I mentioned in the thread it was made to fit the engine bay, to clear the bonnet and turbo. it does work well though.

944crazy

Wow, awesome build. Keen to see where this goes.