legacycentral bbs

WRX Headers vs. Modified Stock Manifold - reading time about 7 minutes



OK I have some results, which might lead to some assumptions, which could be wrong or right. I wish I could supply some hard empirical test numbers which related apples to apples, but I’m afraid the best I can offer is less, but still probably worth something to somebody.



Here’s what I’m working with, and where I went with it.



There is a magazine called “Speed Style & Sound”, which published a WRX Header Comparison. http://www.s3mag.com/headers/



Seven headers of various configurations, as well as the stock system, were tested on two different WRX vehicles at two different “stages” of tune. Charts were generated for Boost, Power and Torque.



Except for some of the most widely spaced lines, because of the small size of the charts, only some generalizations were discernable, and that along with the article narrative, provided an insight into the conclusions.



Because of the way I like to digest data, I wanted more discrete information, so I undertook to enlarge and print the charts, and overlay them with a grid providing enough discrimination for me to extract approximate quantitative figures, which could then be translated into differential percentage ratios.



The published test was quite surprising to me because I’m an advocate of equal length, 4-into-1 headers as the optimized ideal, yet there were designs ranging all over the place, and no configuration stood out as the hands down winner.



The four designs which seemed to place higher in the pack for either the Stage 2 or Stage 4 or Boost charts are of wildly different configurations, though the variables introduced during the testing procedure also caused some separation.



Of great interest to me was the superiority of the stock system over all the headers at low RPM, validating the claims that headers may not add much to performance unless other changes are in place, and in fact may actually cause low and midrange RPM performance to suffer.



Another pitfall is the off-the-shelf challenge of finding a header which just happens to be a step up rather than a step down, depending on what the rest of the engine has going on.



The consolation, at least from the data made available by the article, is that for moderately modified engines, probably any header out there will be better than stock at midrange and higher RPM’s, with the gap decreasing as the engine speed approaches redline.



However, if absolute power or torque is a goal, that is where a truly optimized custom system will probably prevail at producing the best gains at a predetermined point or range of engine speeds. Also I am not convinced my ideal design was exactly represented here, but cannot say whether it would have proven significantly superior, and the evidence seems to suggest otherwise.



Speaking with other people and taking note of the conclusion expressed by the article, led me to a little test of my own. The article indicated a modified stock system might present the best of both worlds, and that certainly seems reasonable, since some of the headers tested did little more than essentially reproduce the stock system configuration in tubing!!! So modify the stock system is exactly what I did.



I bought a spare exhaust manifold, took advantage of my personal friendship with Dan Paramore and put it on the flow bench at DPR Racing, sent it to Extrudehone, and flow tested it again after Extrudehone did their thing.



So after all this long winded stuff, just what were the differences in the equipment tested ?



And that’s where I can offer you something less than apples to apples but maybe more than apples to oranges. I have interpolated figures for boost, power, torque, and airflow then reduced all to a percentage change from the minimum. Not to say that a 10 percent increase in airflow will equate to a 10 percent increase in power, but at least we can look at everything from the perspective of absolute numbers and relative changes, instead of just lines on a chart. I believe when we crunch the numbers some correlations emerge which will be valuable to some in aiding in the decision whether to spend the money one way or another.



So finally here we go;



STAGE 2 had some weird stuff going on with some of the systems where both boost and torque experienced peaking around 3 grand then falling off quickly to a more stable and representative curve, where I chose to select my measurement points.



STAGE 2 BOOST (PSI)

Max = 15.33

Min = 14.13

Range = 1.20

Max - Min Delta = 8.5%



STAGE 2 POWER (SAE HP)

Max = 233

Min = 215

Range = 18

Max - Min Delta = 8.4%



STAGE 2 TORQUE (Lbs-Ft)

Max = 226

Min = 206

Range = 20

Max - Min Delta = 9.7%





STAGE 4 BOOST (PSI)

Max = 16.13

Min = 15.20

Range = 0.93

Max - Min Delta = 6.1%





STAGE 4 POWER (SAE HP)

Max = 270

Min = 255

Range = 15

Max - Min Delta = 5.9%



STAGE 4 TORQUE (Lbs-Ft)

Max = 234

Min = 223

Range = 11

Max - Min Delta = 4.9%





STOCK MANIFOLD AIRFLOW (CFM)

Before Extrudehone = 312

After Extrudehone = 351

Range = 39

Max - Min Delta = 12.5%





I was pleasantly surprised that the stock manifold improvement figure after processing was significantly greater than any of the other deltas. The figures are encouraging to me because when I disassembled the stock manifold and examined the interior of the crosspipe I was horrified to see the end of a smaller piece of tubing floating around in the middle, under the expansion joint!



I had not expected there would be any significant flow improvement because of this obstruction, but apparently the Extrudehone process cleaned up the cast end pieces so that quite a diff was realized, though it could not do anything for the crosspipe. So in fact I now believe even greater gains can be found by replacing the stock crosspipe with unobstructed tubing, and the result should yield that best of both worlds suggested by the article.



Then maybe I’ll have a PART 2 of this if anybody expresses interest in what I’ve done so far, and maybe even with an actual power and torque gain if I can find a volunteer with access to a dyno to borrow it run some numbers.



ANY TAKERS? Meanwhile please somebody give me feedback as to whether this article was of any interest, or if I should not bother you with such trivia

reading time? took me about.. six seconds.


then again, I just scanned on down the page.

Someone did this sort of test quite recently: http://forums.nasioc.com/forums/showthread.php?t=974977

Weed through a few pages. Tested a stocker, a ported/polished stocker, a ported/polish/coated stock with a larger crosspipe, and an equal length setup. Each was better than the last with no real disadvantages.

I'm using a 4-into-1 unequal header. I noticed a slight reduction in spool time (although that was with a modification in final drive, and spool was mostly back the way it was when the final drive returned to stock).

With the 3.90 final drive, and the headers, I noticed about the same spool, but a very smooth roll-onto-boost. Car didn't feel like it surged as much, and the line of off-to-on boost seemed blured.

An improvement? At 10psi, the car feels quicker than it did, though it feels smoother.

I'd say it was an improvement without really any downfall. But a significant one? Likely not. I'd say it was worth it for the sound, and I'm betting that at 16psi, it'll feel a lot stronger up top. Espically after tuning.

One note:

Going to GP-Moto-modified 4-into-1 headers on a customers car, with only a mild tune change, there was a 50whp gain over stock headers. Of course, this was in the realm of 425whp before and 478whp after. There was no real loss in low-end and there was no real loss in spool time.

The gains were noticeable and definite. I'm a BIG header fan if power is the name of the game.

Thawa thank you for the link to that other test. There was good info there plus I provided my article to the NASIOC site.

BAC5.2 Yes, as HP increases, the stock system will present more of a restriction and overall performance will likely benefit more from a change.

I cannot stress strongly enough that headers are very sensitive to the engine to which they are mated. before loading up on Off-The-Shelf headers, one should inquire as to the GENERAL target performance level they are intended to be used for. Headers that jam for a 450 HP engine may cause a DECREASE in performance on a stock engine. Equal vs UNequal length is only part of the equation, with other important factors such as tube diameter and collector design often providing more effect on performance either good or bad. Also TRADITIONAL header design, as originated for NORMALLY ASPIRATED engines, made power for a different reason and in a different manner than headers designed for a turbo application. That's one reason it can be disappointing to bolt headers onto a stock engine that has a pretty good stock exhaust manifold.

Header design is one of my fave subjects, and one thing that sticks out for me is the criticality of balancing all the variables to arrive at an optimum system, admittedly a more expensive proposition, but which is arguably preferential to less, which may not be worth it at all.

One major "problem" is the boxer design. What works for an inline engine isn't going to necessarily work for a Boxer. Same goes for a V configuration engine.

I think for most people, stock headers are fine.

Another problem is wastegate integration. Hard to do well. We've found a good way to do it well, but I see a LOT of externally gated Subaru's running around with sub-par setups.

The basic layout of a Subaru doesn't really follow any norms, and there isn't exactly a "perfect" setup. If you want to make a LOT of useable power, your setup will be less efficient than someone just trying to make a lot of power. The person making lots of power is not going to make a lot of low-end power where the person with a less efficient setup will beat them out.

It's all a catch-22.

BAC5.2 wrote:One major "problem" is the boxer design. What works for an inline engine isn't going to necessarily work for a Boxer. Same goes for a V configuration engine.

I think for most people, stock headers are fine.

Another problem is wastegate integration. Hard to do well. We've found a good way to do it well, but I see a LOT of externally gated Subaru's running around with sub-par setups.

The basic layout of a Subaru doesn't really follow any norms, and there isn't exactly a "perfect" setup. If you want to make a LOT of useable power, your setup will be less efficient than someone just trying to make a lot of power. The person making lots of power is not going to make a lot of low-end power where the person with a less efficient setup will beat them out.

It's all a catch-22.

Right I agree. I am after lots of useable low and midrange torque. My car is a daily driver. I want the torque curve that climbs right away and stays flat across the rpm range rather than the one that peaks higher but only in a 200 RPM band

As long as I reach full boost by 50% of my redline, I'm happy. My car isn't a daily anymore though. Though I would be happy to drive it daily, it's not economical and my girlfriend refuses to ride in it.

heheh. im running some 4-1 unequal length headers as well. for me, i have a fairly stock setup. other than the headers, i have underdrive crank pulley, 3 inch downpipe 3 inch hi flow cat and cat back. boost is set to about 10 psi.

my car seems to spool slower than with a stock header. maybe i need to wrap them?

Project_Legacy wrote:heheh. im running some 4-1 unequal length headers as well. for me, i have a fairly stock setup. other than the headers, i have underdrive crank pulley, 3 inch downpipe 3 inch hi flow cat and cat back. boost is set to about 10 psi.

my car seems to spool slower than with a stock header. maybe i need to wrap them?

The stock exhaust manifold beats just about everything at quick spool up

dont wrap ss headers. they will crack sooner than later. i used 4-1 tuned length gtspec euro gen 3 headers for a few weeks. did exactly what i thought they would do. moved my power band to the right and put 2 seconds on to my times. im a big perponent of stock cast iron headers with a cobb coated inconel uppipe instead of headers. if i were to recomend a header it would have to be the aps which i would like to get but a number of things deter me. one, the header like all headers is stainless, secondly UR doesnt make a kit that goes with these headers, third tuning to a header imo is verry diffcult. they get too cold too quickly making the 400 whp you made on the dyno after sitting on it for a few hours only 320 on the streets wiht cold headers. this was really noticeable with my friends wrx with headers and a holeset. after 8 hours of tuning we finaly got the maps just right and the next day it wouldnt run right until we floged it till the header got hot enough. im with larry in that useable power is the name of the game and more importanly consistency.

i talked to christan at cobb at length about headers for the subies and he strongly dislikes them and doesnt like to tune for them. that thought alone from him cements in my mind that headers are a bad idea.

for thoes who dont know christan tunes all the accessport maps that cobb puts out. he is very good at what he does even though among subie crowds isnt as famous as the phil @ element and gadiel's out there.

hm... i c. azn2nr, it made ur quarter mile times 2 sec's longer?? holy crap. think im gonna find my stock pieces this weekend and try it on.

ive never had the stock header on my car, cuz when i did the swap, i just straight up used the Stainless steel header.

thanks for the insight, maybe my car will be faster, cuz as of right now, its pretty slow.

see thats the thing though you have a manual trans. when dragracing you can rev inorder to spool adn heat your turbo but really only when your on the track. that power becomes unconsistent on the streets adn imo really anoying. sometimes having a manual you wont suffer down low as much because you have a launching advantage but me having an auto all it did was mimick the powerband. i went from 14's back to 16's when i ran 17's stock. thats was what i atributed to the headers as nothing had changed between my two runs.

really? holy crap. thats the thing though, most my runs are in the streets. im gonna try out the stock headers i have lying around somewhere and see how i like it.

thanks for your input!