Extractors/ Headers for NA EJ25

[Gannon]

I have had my Outback a little over 3 weeks and the modifying bug has already set in.

My first issue is with the stock headers. Although a little better than some of the cast manifolds on inline fours, there is still room for improvement.

They no longer sell NA EJ20/25 headers on ebay aus (there are some on ebay us, and although they are $150, freight doubles their cost)


If i were to make my own, i would buy 4 mandrel 1-3/4" bends and make my own, and replaced the silly shaped factory cat.
Something similar to this design, but without the silly hump, in 4-2-1 (over the 4 into 1) design cos its more suited to everyday driving (so i was told)


But then a little voice in my heads says, shouldnt you join cylinders 1 & 2, and 3 & 4 in the primaries like they do on other 4 cylinder engines
Some EJ turbo manifolds do this, but no NA ones. Probably cos its too hard to package.


So what im asking is your opinions.
Ive looked at RSLC and OzLib, the NA guys say they get gains from headers, but it was just from the type featured in the top picture. I rekon there is more to be gained from joining the pipes in the right order and getting the right pipe sizes and lengths, but i cant find any info on designing them

Let me know what you think

[Xtreme_RX]

Gannon,

I have a set of the first pic of the headers on my EJ22. They work very well they give a massive torque increase.
That design is proably the best setup. Rember you motor is now NA not turbo charged. The exhaust headers dont need to be 'fancy'

More later......

[AlpineRaven]

Gannon,

I have a set of the first pic of the headers on my EJ22. They work very well they give a massive torque increase.
That design is proably the best setup. Rember you motor is now NA not turbo charged. The exhaust headers dont need to be 'fancy'

More later......

In your opinion - is it worth the upgrade?? Do you think that it has given you more torque?
Cheers
AP

[Gannon]

Im thinking of making my own based on the design of the 2nd pic. If im gonna do the headers, i wanna also get rid of that funny shaped cat converter

I just need to get hold of some EJ exhaust/head flanges and 4x 1-3/4" mandrel bends

[AlpineRaven]

The "Hump" in the 2nd picture would have been that idea to make the same length for left and right bank to give it even power output...

In the 3rd Picture - wouldnt make any difference from picture one and 2, because of the way the engine works - while the cylinder is under compression the other cylinder will exhaust so it works both ways as close and open in both cylinders on each bank - that what i think?
Cheers
AP

[Gannon]

I think this is where Discospud03 can enlighten us on the grouping of cylinders in manifolds.

It has something to do with the firing order, and the pause between pulses of exhaust gas. Whether or not i will make much difference, i dont know.

I also need to calculate the desired length of the primary tubes to get them to work best in the required torque range.

I will also be freeing up the rest of the exhaust too.

[AlpineRaven]

I think this is where Discospud03 can enlighten us on the grouping of cylinders in manifolds.

It has something to do with the firing order, and the pause between pulses of exhaust gas. Whether or not i will make much difference, i dont know.

I also need to calculate the desired length of the primary tubes to get them to work best in the required torque range.

I will also be freeing up the rest of the exhaust too.

In my opinion, dont forget, Subaru has made exhaust in that type of way and there is a reason behind that for long term.
Ive seen cars with high flow exhaust or after market types that isn't really suited for the engine have become to have wear and tear rapidly, Like my brother in law had 2.5" high flow exhaust and he had it since the car was new, its now 80,000 kms and needs new valves as they're badly damaged due high flow...
Cheers
AP

[AndrewT]

lol...

...they give a massive torque increase....

....Do you think that it has given you more torque?...

Yeah I agree with the point that Subaru know what they're doing with headers most of the time...I don't think the expense of replacing the headers is really going to be worth it in a "bang for buck" scenario.

[discopotato03]

Yeah , it can be hard to have a bang for buck win with non stand exhaust headers . The thing is that the state of tune of the engine is designed around the exhaust flow characteristics of the std exhaust - from heads to chrome tip . One you start to change things you get into a grey area unless you know exactly what the factory got from their std setup .
In a perfect world you'd be able to mimic their systems positives and improve upon any negative compromises .

Probably the hardest thing to do is to decide on tube diameter and primary length . Both of these affect how well the engine breathes out and how well the header system allows the cylinders to scavenge at the valve overlap phase .
Really to get an insight you'd need to dyno the engine so you have an idea of its torque characteristics - as in where the factory intends it to be .
Its pointless trying to enhance upper rev range breathing if the engine is designed to pull from low revs which most are nowdays .
This is often why less than adequate , small enough , tube sizes are used because people don't understand how to make torque at user friendly revs .

As for how to get the correct cylinders linked .
In a nutshell you want to have all the exhausts putts or exhaust events evenly spaced and Subaru don't normally do it like that . What they do is link the exhaust ports on each head together because its cost effective and packaging wise the way to go .
Now because Fuji's flat fours fire the adjacent cylinders in each side or bank before crossing to the other pair two exhaust events are close together followed by a pause then over to the other side or pair .
This not to say that you don't get evenly spaced power strokes , it is to say that these engines operate more like two in line twins sharing a common crankshaft .
From a strictly functional engineering point of view the front two cylinders and the rear two cylinders should be linked , but Subaru only do this on their twin scroll turbocharged flat fours . The single scroll turbo engines get the same crap system as the NA fours .

I had a sticky at this thread this afternoon , yesterday afternoon now , and went looking for pics I have somewhere of Subies twin scroll header pipe system - couldn't find them .
At least they are designed by the factory to fit under an EJ , provided you use the matching sump pan and oil pick up .
Gannon you may have to search around on some of the American sites such Legacy GT /Legacy Central and NASIOC for pics of the factory TS headers .

I don't know what your cars cat layout is like but usually the best solution flow wise is to use a larger capacity one and neck the exhaust tube up before and down again south of the cat . My 185 Kw Skyline and my "8" Kw Rx L Turbo use the exact same 3" cat though the Rx has 2.25" exhaust where the Nissan is probably 3" - Nizmo thing . The Subie is not noisy at all , neither are .
Occasionally engines with two banks/heads have a separate converter before they join to the single tail pipe . Four cam Nissan VG30 V6's in Z32's are like that .

I know it's not what you want to hear but think briefly before you dive in if it's a better bet to trade up to the 3 liter 6 if there is one in that model range . Plenty of people have proved over time that factory engineering can be easier and cheaper than taking stabs in the dark . At least it's a known quantity if you can find and try a six for yourself .
There is a risk in trying lots of plausible things out , if they don't do much or you go backwards the money's already spent .

Food for thought , nite all .

[Gannon]

Ive been playing with a header calculator and have come up with these figures

EJ25 with 165Hp at 5600RPM

Primary Pipe Inside Diameter 1.39"
Primary Pipe Length 35.4" or 90cm
Collector Inside Diameter 2.12"
Collector Length 13.5" or 34.3cm

It estimated my engine displacement to be 149.9cid, and the EJ25 actually has a displacement 152.5cid

This was calculated with a performance factor of 4
Maximize lower midrange and midrange torque, with good top-end horsepower. Part throttle operation important.

Now if i go for a performance factor of 6
Maximize midrange torque, with excellent top-end horsepower. Throttle response important.

Primary Pipe Inside Diameter 1.41"
Primary Pipe Length 34.4" or 86cm
Collector Inside Diameter 2.21"
Collector Length 14" or 35cm

Doesnt seem to be much difference between the two 2 different performance factors in pipe sizes. They seem very small though. If i were to use 1.5" pipe and make each primary 90cm long, its quite a bit smaller and longer that the pipes shown in the pictures of premade headers shown in the images in my first post

To get an idea of what the performance factor means
Point 5 should be around 4000rpm and point 10 will be 5600rpm (for an EJ25)

[discopotato03]

By all means do the maths but I'd be curious as to what the mob was using as a generic test engine because there really isn't any such thing .

Big fours , I think the Porsche 3L was among the largest , are getting towards not being nice things mainly because the piston weight and therefore inertia starts to work against them .
They often don't have wide rev ranges so the manufacturers try to make them work or have torque low in the rev range and tie it all in with appropriate gearing .

If you search you should be able to find performance data , torque vs revs , and people sometimes want to equate peak torque revs to tallest gear cruising revs at around 100-110 km/h .
As I said torque peak is usually in the same region as the engines volumetric efficiency peak but it all ties in with dynamic or effective compression ratio .

Julian Edgar was the original editor of FF/R and then Zoom , he now does it for I think Autospeed which is an online thing .
Very early in the peace he played with an RS Turbo and found some gains just from taking restrictions out of the inlet system . Good filter elements , free flowing ducting , cold air pickups , maybe oil/air separators if the engine is fumy .
The real thing with non forced induced engines is that you only have atmospheric pressure with which to charge or fill your cylinders so you have to take every step possible to ease the flow path from atmosphere to cylinder - if you want best performance .
Turbos make it easier because they pump the air in and you can have more performance without using very high revs .
Its reasonably straightforward to raise an OEM turbo engines performance by 20-50 percent , not nearly so with a factory NA engine .
Broadly speaking only two things markedly increase an NA engines torque output over the normal rev range , increased capacity and or higher compression ratios . So either higher cylinder pressure on the std capacity or std pressure on the increased capacity .

Best of luck , cheers A .

[Gannon]

Small update

Im going to get 4 x 1-5/8" bends, 2 x 2inch bends and 3 collectors from the zorst shop, make up some flanges and see how i go.

Basic design will be 4 x 1-5/8 inch primaries about 400-500mm long(the exhaust ports in the heads are 1-1/2 inch) that feed into 2 x 2inch secondaries that feed into a 2-1/4" collector just before a high flow cat. It will then join to the existing exhaust system just before the second cat with a length of 2-1/4" pipe.

I'll take a picture of a mock up of how im gonna do it.

I will be concentrating on good flowing bends and merges rather than equal length for now.

[steptoe]

and while you are there I wonder if an old turbo from your pile of spares might just team up with your mandrel bends, check out where the oil feed will come from.....

[Gannon]

This is the design im aiming for



The primaries will be equal length, but the secondaries wont be. This isnt a concern because im not matching the right ports to make full use of scavenging anyway.

[Venom]

Thats basically what i had made up for the EJ25 when i put it in my L series.

[Gannon]

Got pics?

Id love to see how you did it

[discopotato03]

I found the pic of the twin scroll Subaru factory header but its on the other computer , will E mail it to you in the near future .

Cheers A .

[Gannon]

Would that be these?


Here are an aftermarket set that are of the same concept (but are made with mandrel bends and no pre-turbo cat)