Subaru Intake Manifold Development
 

So I decided to develop some KSTech intake manifold designs for the newer turbo Subaru WRX, STI, etc. These will be for very-high horsepower and higher RPM applications - mainly racing, but it would not surprise me if they work well for serious street cars too. Front-mount intercoolers only, and I want a version designed specifically for reversed manifold configurations, as well as the normal back-facing version too. A/C retention is a possibility, but that is to be determined.

The commercially available options all have compromises I don't like:

• Stock WRX/STI intake manifold has virtually no plenum volume, the runners are too long and undersized. All fine for a stock production street car, but that is about it. The flow variation from runner to runner likely contributes to many cracked pistons on high-boost/power cars. Basically this manifold was designed to work with a TMIC, and all other considerations were secondary.
• The newer N.A. manifolds are pretty nice with better and more-even flow and a bigger plenum. But the runners are too long and a little small, the plenum is still not big enough for forced induction.
• Magnus and its ebay copies have been around for a while, but they are basically a bigger-plenum version of the stock design. They have all the compromises requred for the TMIC configuration, small runners, and a reputation for cracking.
• Cosworth's manifold is pretty nice with a good-sized plenum, big runners, and (I assume) even flow. But they still used the twisted and squished TMIC design, and then they didn't make it work with any normal TMIC. WTF mate?
• Forsa has an expensive manifold in forward and back configurations. The plenum volume is too small and the runners are longer than stock. There are also some flow variation issues from runner to runner, especially with the rear-facing version.
• AMS has been working on a dual-plenum manifold forever, and still have not released it. Pairing cylinders #1 and #3 in one plenum, and #2 and #4 in another plenum is NOT a good idea with Subaru's 1-3-2-4 firing order. #1 gets a good gulp of air, then #3 gets starved because it intakes next before the plenum has had a chance to recharge, followed by a long pause while #2 and #4 do the same thing on the other side. This is unacceptable in my view. It looks kewl as hell, but is a bad idea IMO. Maybe that is why it has never been released.
• Corsa Veloce intake manifolds are works of art, but they are $4k+! Google it for pics and you will be impressed. The plenum might also be bigger than necessary, and there are no radii on the runner inlet.
• Tomei has an ITB manifold with a huge plenum (which is okay with ITB's). It is expensive too. The ITB's require a cable-type throttle and there is no idle air control. The runners are very short but would work great in the 8000+rpm range. OBX has copied it (or is just buying it from the same Chinese supplier) for a ridiculously low price. I may get one to see if the quality is good and to see if it worth pursuing.

So my design goals for a KSTech Intake Manifold are:

• Configuration for 450+whp in the 6000-8000 rpm range.
• Even flow between runners (both continuous flow and pulsed)
• Minimize or eliminate the compromises mentioned above without creating new ones.
• Forward-facing versions for stock alternator location and relocated alternator.
• Possible A/C retention on forward-facing version.
• Rearward-facing version that retains the A/C.
• No TMIC version (too many compromises in performance).
• Manifolds to bolt to TGV's for now, and TGV-less versions later.
• Easily built-to order to customer specs.
• Retail price in the $1000 to $1500 range.

I also intend to make individual components available for shops and DIY'ers to fabricate their own intake manifolds. There is nothing currently commercially available to my knowledge. Even if the complete manifolds do not end up being commercially viable (for whatever reason), these parts alone would make the project worthwhile.

I have been sketching intake manifold designs on napkins for years now, and I think I have a concept that I will develop finally. Not going to disclose it yet, but there are some things I can share for now.

First order of business was getting some flanges designed and machined for bolting to the TGV housing. To do this I needed to decide on a runner size. I am going with 2.5" diameter x .080" wall tubing for the following reasons:

• .080" wall thickness for a stronger cracking-resistant runner compared to typical .065".
• Theoretical calculations on runner size point to 2.25" to 2.5" being the optimum size.
• The TGV opening is 2.2" stock, and many people open it up a little when porting. This is a better port match to the 2.5" tubing. 2.38" tube bends are not readily available in .080".

The only problem with the big 2.5" runner size is it puts the welds too close to the bolt locations. So I designed a flange that will use countersunk socket-head screws. Below are some Solidworks renderings produced by SECCS member AtomicLabMonkey, Austin. It also features counterbores for the runner tubing to aid fabrication. I am pretty happy with the way it looks and the cost to CNC machine it is pretty low. My welder, Ray is also happy with the design from a fabrication standpoint. I have a dozen of these on order already.

I think some DIY'ers will want a less-expensive flange and/or one for 2.25" runner size, so I had some cut out of 1/2" plate by a local waterjet shop. The port hole is sized for 2.25" runners, but it can be opened up for 2.5" and/or port-matched to the TGV. I also made the flange perimeter a little bigger to allow for more weld area, as not everybody is as good a welder as Ray Cunningham is.

I will post pics later.

Next up are velocity stacks for inside the plenum. Update coming within a week or so.

Friggin' finally. :lol:

That's what I'm gonna say when you get all your shit out of my shop... or we throw it away. :P

Have you given any thought to making ITBs work on DBW manifolds? I converted the wagon to '05 RS specs or I would just buy the OBX setup.

I probably don't know what I am talking about but couldn't you make the intake manifold and TGV into a single piece?

Are you designing for 450hp+ because the stock manifold is adequate for less hp?

Yes. I will likely get a set and make an adapter. The cable throttle is annoying, though.

Those ITB's should work with an 05 RS motor. Is there some info I'm missing?

• Everybody hates the TGV's, but once they are ported, they are a nice gradual, smooth transition from a round port in the manifold to the heart-shaped port in the head. The engineer that designed them did a nice job really. It's a good port shape, but unfortunately interrupted by the TGV flapper and divider.
• A manifold that bolts to the TGV is simpler, quicker to develop, and less expensive, so people who already have ported TGV's (and there are lots of them) will be arguably better off.
• When I do a TGV-less manifold it will be for top feed injectors only, so people with 04-06 STI and 05-09 Legacy GT would be left out.

I may do a "street" version with smaller/longer runners. But under 400-450whp I don't think the intake manifold is holding people back much. But still, the stock WRX/STI intake manifold is a poor design at any power level if there is an FMIC.

Drivetrain-wise, I have an '05 RS. Drive-by-wire, so no cable throttle, so no easy adaptation of the OBX ITBs.

You could get a pedal and cable assembly from a 2002-2005 WRX or 2002-2004??? Impreza.

It would be interesting to try to adapt the DBW motor to the ITB bell crank. I bet it would not be strong enough to work though. Somebody already tried that unsuccessfully, but he was also running bigger barrel ITB's.

I was considering trying to split the throttle motor signal to both TGV motors, which are identical to it, but I don;t know how that would work electronically. Going to a cable throttle setup causes some issues with DBW ECUs as well, doesn't it?

The two motors might draw too much current, and they would not turn at exactly the same rate.

A DBW ECU might work with cable throttlle. Some rewiring of the the harness would be necessary, and there would be no idle air control. And of course all the DBW CEL's would need to be disabled.

So it is time to figure out the runner length. The runner length is the main factor that determines what RPM the intake manifold is tuned for. I do not want to go into theory, but here is a handy basic calculator:
http://www.velocity-of-sound.com/vel...alculator3.htm

And here is a more detailed one:
http://dairally.net/daihard/chas/Mis...s/DaiPipes.htm

This is all theory and this is just useful for providing a starting point. But from experience from back in my V8 days, and looking at other intake manifolds out there, it should be close.

There are some basic runner-length constraints for this intake manifold on the Subaru engine. Starting at the valve seat where the runner theoretically begins:

1. There are approximately 4" of port length in the head casting.
2. The TGV Riser is approximately 3" tall.
3. There is about .94" of manifold flange, etc. on top of the TGV (in this particular case).
4. There needs to be a 90-degree bend (or more maybe) coming off the top of the TGV to provide hood clearance. This will become more apparent once you see a picture of the whole deal (fairly soon). The 2.5" tube bends I am using have a 4" centerline radius, so that will make for about 6.28" runner length in the bend (2 x 4" x 3.14 (pi) / 4 = 6.28").
5. Beyond the bend I will need about 1.75" of runner extending into the plenum at a minimum.

Add that all up and the nominal runner length is about 16" long. That length is tuned to about 6000rpm, which is about right for an engine with peak power in the 5000-8000 rpm range.

For a street engine I could make the runners longer and tune them to a lower RPM. This will be an option on this manifold.

For higher RPM applications the whole plenum would have to be redesigned to accomodate shorter runners. I already have ideas for this, but that is for the distant future.

So these runners extend inside the plenum. For optimum airflow they should have a nice radius at the entrance, and a velocity stack or "trumpet" is even better. Due to space constraints I had to to keep the velocity stack relatively small, but there are diminishing returns beyond this anyway. Pics below are some Solidworks renderings of what I came up with and what it looks like on the tube intake runner (courtesy of AtomicLabMonkey of course).

Next up is to put the runners on the flanges and position them on top of the engine. Coming in a week or so.

Make sure to use Mil-Spec materials.

No variable length runners? I kid. I like this stuff. Keep posting.

http://upload.wikimedia.org/wikipedi...ol-hacksaw.jpg

http://www.artsjournal.com/jumper/wp.../duct-tape.jpg

Feel free to vary the intake runner length as much as you'd like!

Duh, you need a flex element, too.

http://www.ez-groom.com/catalog/imag...yer%20Hose.jpg

We would also accept a Slinky.....and more duct tape.

I just figured you could hack out a section of the pipes for a track day, then later, tape it back together for the street.

And, I'm pretty sure Samco has a patent on duct-taped wrapped slinkys.

That picture hurts my eyes to look at. It's like it was resized with an extra-aliasing algorithm.

So much for my technical thread... ;)

I have a concept in mind for a plenum with a removable top so that the runners can be easily changed out in a matter of minutes. This could be useful for tuning purposes, or for changing them out for different needs - short runner for drag, medium-short for road race, medium length for aggressive street, and long for everyday driving. Diameters could be changed too. That is for way later though, and it may be cost-prohibitive.