unequal length tri-y secondary merge points.

[PrettyCoolWagon]

Ok, this question / topic relates to inline 4's specifically and the 1,3,4,2 firing order I'm familiar with as found in 4A & 7A engines..

Like this:
https://www.youtube.com/watch?v=MBkqOXaJXco

I've been scouring the internet and coming up with a ridiculous amount of hits for for the subaru wrx guys who either hate unequal or same length primaries OR V-tech guys who just can't wait to get that wrx sound
Plus the American V8 performance tech threads as well.

But I could find little reading on unequal length secondary tri-y merge points for inline 4 cylinder engines for the specific reason of broad range low to mid level torque torque. Basically trying to get rid of some of the resonant flat spots in that area of the rpm range.

(before you say super charger or small turbo, high comp etc. I'll just say F-off in advance - This is the tech thread not any old opinion thread )
No offence,
I just want to keep the topic specific to headers. Of course people will be correct by saying rpms, gear ratios, cam size and forced induction vs. N/A will all have a direct impact on torque usage and header design. That's fair enough.. maybe we can get into that later..

As a basis - Let's just say were talking about an N/A 7AFE in my favourite wagon - the AE95 that weighs about 1200kg's driven on the street ONLY.

So, given that there's flat spots and power bands in the rev range, I'm thinking:
Long narrow primaries straight out of the head to begin with, nice flowing bends whilst avoiding any tight corners and nice 10 degree taper merge points. (BTW - Does any one know the internal port diameters of 4A's & 7A's?)

Here's how I see it:1 & 3 pulses into a secondary, while 4 & 2 pulses continue on a bit longer as individual primaries before also merging into a secondary at a calculated/tuned length that considers both the firing order of all cylinders as well as the first merge point with the goal of keeping the pulses tuned so they still merge evenly at the final 2 into i point.

[PrettyCoolWagon]

Sorry about the font colour guys - I was trying to highlight the background in yellow.
Last paragraph reads:

Here's how I see it:1 & 3 pulses into a secondary, while 4 & 2 pulses continue on a bit longer as individual primaries before also merging into a secondary at a calculated/tuned length that considers both the firing order of all cylinders as well as the first merge point with the goal of keeping the pulses tuned so they still merge evenly at the final 2 into i point.

[madmont]

Traditionally you merge 1 and 4 together and 2 and 3 together and merge them together or have equal length primaries merging 4 into 1. Why do you want to do it differently by merging 1 and 3 and 2 and 4 ? Is this for some different sound or something?

[ToyTA22]

I'm with madmont in the merge points are usually opposing firing points, i.e 1 and 4 etc. It is the positive pulse in 1 which causes a negative pressure in 4,hence extraction effect,then the same with 3 and 2. Again as madmont's comment, could you please give a insight why changing merge points?

[knightrous]

The only reason you would join 1 to 3 and 2 to 4 is if you had a 90 degree throw crank where Cyl1 @ 0/360 deg, Cyl2 @ 90 deg, Cyl3 @ 180 deg, Cyl4 @ 270 deg.
However, all A series motors have 180 degree throw crank.

[Duk]

Your thinking is very lateral and I doubt that there would be many, if any, that have tried such an approach. Ultimately, I think you'll have to make something and see.
Beware the dreaded placebo effect when doing these things. To quantify your results, you will need accurate before and after.
Personally, I'd just build some longer secondaries in either 1 7/8" or 2" pipe, with twin flex joints.

[PrettyCoolWagon]

Cheers guys,
It appears I threw some of you off track by suggesting the incorrect pairing of cylinders. I wasn't aware of the finer details, regarding merge points being the opposing firing points, ..So please disregard.

They key point of the discussions was to focus on reducing flat spots in the torque curve at low to mid range revs by staging the 2 into 1 merge points at different intervals.


(late night post - I rambled a bit and didn't have a look at the pairing on my headers)

[PrettyCoolWagon]

You guys are right.

Just checked a bunch of 4AFE/4AGE tri-Y headers on Google images and the vast majority are as you said: "..1 and 4 together and 2 and 3 together.."

...but then oddly enough, I did find this pic:

http://www.todoautos.com.pe/attachme...age-6ec8_1.jpg

Seems to be an anomaly..

That's a bit off track though... Since I can't find a photo or much reference to what I'm talking about, I'll have a crack at sketching what I mean.

[Xauterus]

I'm not sure you are going to find any miracles in exhaust manifold design. Are you sure the flat spots are related to the exhaust at all?

[Sam_Q]

I know who ended up with the wrong cylinders joined together due to the incompatence of a workshop, it ran rough and didn't make the power it should. It was quite amazing how much difference it made.

[knightrous]

I'm not sure you are going to find any miracles in exhaust manifold design. Are you sure the flat spots are related to the exhaust at all?

There are all sorts of flat spots that can be introduced by the intake due to resonance and in the exhaust due to reversion.
Lot of maths to work out the possibilities and even more hands on to see if they are true.

[PrettyCoolWagon]

Yep. pretty much.
What knightrous said.
There's resonant flat spots in both intake & exhaust due to shck waves from the valves opening and closing ( and also sound waves - but lt's keep it basic)

Calculating the merge points for best torque at a chosen point in the rpm range is quite a skill.

There's all sorts of online calculators from really basic, like this one: www.bgsoflex.com/bestheader.html
to much more complex ones.

Even the racing nerds will tell you there's only so much simulation and calculation you can do, ..and it's still only going to get you in the ball park.

Some people advise, to be prepared to make 2 sets of headers.
The first set you do all your trial and error on, (test fits and test runs) ..The second set you make as perfect as possible.

So I did a quick sketch anyway, to illustrate what I have in mind.
(..and it IS a 'sketchy' sketch ) 20150310_204024.jpg

[knightrous]

With your unequal mergers, you going to have issues with AFRs .
Cylinders 1 & 4 will have a different tuned length, so they will be running richer or or leaner then cylinders 2 & 3 at the same RPM.
To make it work, you will need an aftermarket ecu that can handle dual wideband 02 (one for each merger point) and injector cylinder trims (to be able to handle the different AFRs between cylinders)

IMO, this is not a sensible way of smoothing out a flat spot.

[PrettyCoolWagon]

With your unequal mergers, you going to have issues with AFRs .
Cylinders 1 & 4 will have a different tuned length, so they will be running richer or or leaner then cylinders 2 & 3 at the same RPM.
To make it work, you will need an aftermarket ecu that can handle dual wideband 02 (one for each merger point) and injector cylinder trims (to be able to handle the different AFRs between cylinders)

IMO, this is not a sensible way of smoothing out a flat spot.

There's probably some truth to that...
If it was such a good idea, someone would already be doing it, ..but maybe it's too time consuming calculating a complimentary balance of unequal merge points. So manufacturers don't bother..
Bear in mind, it's just a sketch and the merge points wouldn't be so far apart.

My understanding is - Of the following 3..
Intake design,
02 sensor position
and header design..

It's the position of the 02 sensor and intake design that have the greatest effect on air fuel ratios at mid to low rpms.

For example, factory intake designs like 7afe aren't perfect at feeding the air in evenly anyway.
Meanwhile injectors feed equal amounts of fuel regardless. So air fuel ratios aren't perfect to begin with.
Add to this the fact that oil trickles into cylinders 2 & 3 from the cam cover breather line - further affecting the air fuel ratio of those 2 cylinders.

What's more..
Take a look at both the factory 7afe header and aftermarket replacement in this picture.
Neither of them has equal length merge points. In fact number 1 primary on both headers is way longer than all the others.
http://www.4nmuscle.com/pictures/Cor...arison%201.jpg
I can fabricate better headers than those aftermarket ones in the picture..

The factory shortcomings suggest there's an acceptable range of difference between lean & rich across all 4 cylinders at mid to low rpms.
The 7afe produces peak torque at 2800rpms and peak power at 5600rpms.
The aim of this header design is to improve torque in the low to mid rpm range.

The narrow valve angle of a 7afe is useless for high rpms.
At low to mid rpms there's not enough velocity in the exhaust pulses or overlap in the valves to create the kind of vacuum that would cause issues with AFR's. So an aftermarket ECU & wide-band might not be necessary at all..

Knightrous makes a really good point though...
Considerations around header design and the affect on AFR's is better applied to the GE sports heads, high revving street-performance and flat-out racing purposes where vacuum can be anywhere from roughly 4 - 7psi ..sucking in huge amounts of air from the intake when the cams and valves overlap...
In this case an aftermarket ECU & wide-band would definitely help tune across the whole rpm range, especially when forced induction is involved.. For these purposes, I wouldn't build a tri-y header at all. I'd simply build a 4 into 1 with long primaries.


This low to mid rpm header idea would place the 02 sensor down near the final collector, to get an overall reading.
The unequal merge points and final collector would have to be set at 3 complimentary resonant distances.
The secondaries would be slightly oval too.

[oldeskewltoy]

You want to begin with a OEM 7AFE header after 1996...



The above header is basically a copy of the Toyota 20V headers, but with a flange to fit the "F" head.

The issue becomes one of making something custom so this header can fit your AE95. Problems are it will hit the radiator because of how far away it sits, and the downpipe isn't designed to fit around the all-trac transaxle





In the end I didn't use it, I went with the cast iron P.O.S.

This is the power my 4AFE made @ the wheels... remember this is an All-Trac drivetrain so a 25%+ drivetrain loss is typical