Suck Squish Bang Blow: All You Wanted to Know About How Headers Work (Part 1)

by Mike Kojima

Headers are one of the most common and basic of your bolt on mods, however very little is understood about how they actually work. How and why headers work is an arena full of old hot rodders and internet keyboard guru’s tall tales and myths.  We can easily understand why this is true as there are not many accurate stories on the net on how headers actually work.  Most of us know that headers can produce substantial amounts of power on a motor with very few negative compromises.  Properly designed headers work well in producing extra power with without many of the typical negative side effects of power mods such as loss of low end drivability and mileage reduction. Headers are a rare, win-win modification with hardly any negative trade-offs.  This makes them a mod that is almost essential for any engine build-up from serious to mild bolt on.

More on Mike's continuing series on how engine stuff works.


This is a stock exhaust manifold off of a typical late model car.  It was primarily designed for good cold start emissions as the catalytic converter is very close to the exhaust ports.  Most of a modern car's emissions occur during the first few seconds after cold start so speeding the amount of time it takes to light off the cat makes a huge difference in overall emissions.  You can almost bet that a header will help this car.  Some cars, like recent Hondas take this to an even greater extreme where the exhaust runners are cast into the head and are extremely short.  Good for cat light off but impossible to modify.  


For space, cost and catalytic converter light-off reasons, most cars come from the factory with crude, cast iron, log exhaust manifolds. A log manifold is simply a cast tube with short, stubby legs connecting the exhaust ports to the main tube.  Once the tooling is designed they can be made by the thousands for very little. Besides being cheap, a log manifold is also good for conserving exhaust heat to quickly light off a catalytic converter during cold starts. A cast manifold can be compact, hugging close to the cylinder head, preserving valuable under hood space within today’s crowded engine compartments.  All of these attributes are important to an OEM engineer; however a log manifold is detrimental for power production which is often a secondary concern to the OEM designer. A few of today's newer cars come with more efficient tubular exhaust manifolds most of which have short runners and crude stamped transitions. Not the greatest for performance either but much better than a simple log. Most OEM tubular manifolds are double walled and have insulation sandwiched in-between the walls to speed cat light off.  In short, an OEM manifold has many design parameters to fulfill with power production perhaps not as high a ranked priority.


This is a log manifold where very short runners dump into a larger tube that collects all of the exhaust gasses.  The stock manifold is on top and an Extrude Honed ported version is on the bottom.  As you can see a log manifold is hurting from both a flow and an overall tuning perspective.


What exactly is a header? Let’s first start off with the most basic of basics.  By the enthusiast definition, a header is an exhaust manifold fabricated from tubular sections of pipe.  Full radius mandrel bends are preferred so the pipe’s tight radiuses will not be crushed down promoting better exhaust flow and more equal flow from cylinder to cylinder.  Each individual exhaust port is treated to its own separate primary runner of close to equal length instead of merely dumping into the shared main pipe of a log manifold.  The equal length, or close to equal length primary pipes converge at a single, larger diameter point or collector.  The collector then leads to the main exhaust pipe.


You can see how the stock manifold on top is really corked up compared to the Extrude Honed part.


An old hot rodders tale is that headers produce more power by reducing backpressure.  Intuitively, when looking at a kinked up stock log manifold this seems correct but backpressure reduction alone is not the main reason why headers work so well. Case in point, there are a few cars like a D15/D16 Honda engine or the DOHC Ford/Mazda V-6 engine where adding headers to an otherwise stock engine produces very little additional power despite really crappy looking stock exhaust manifolds.  You might wonder why this is so.


More comparison.  These manifolds were on a turbo engine so increasing the flow really helps power even though it has no effect on pulse tuning.


Page 1 of 5 Next Page
Bookmark and Share
Wednesday, January 30, 2013 6:47 AM
Fantastic first article. Can't wait for part 2. Many of these things can be translated into intake manifold design as well I suppose. Want to tackle that after this series Mr. Kojima? I do remember an article in SCC years ago that had some info on intake manifolds in it.
Wednesday, January 30, 2013 6:53 AM
So this article brought three questions to mind.

1. Is there design software that can model proper header design so that there is less trial and error (and math...) besides propietary ones?

2. Wouldn't the lack of backpressure (especially since they are often NA) on hot rods with open headers cause a power loss or is there so much pressure coming out of the primary's that any bottleneck causes a loss?

3. Is the M3 in the lead photo going to be a project car :D
Mike Kojima
Mike Kojimalink
Wednesday, January 30, 2013 7:00 AM
I have not been able to design an intake manifold for an NA car that worked well. Intake manifold design is tricky.Even guys with good analytical tools like CFD and complex engine modeling like GT Pro have trouble designing them. I am not an expert enough to write about the subject.

I think I wrote somewhere in there that less backpressure never causes a power loss, just bad tuning due to hacking something off the exhaust system.

The M3 was either Verena Mei or Billy Johnsons beater car that just happend to be there.
Wednesday, January 30, 2013 7:07 AM
I'm going to take issue with one of the examples that Mike uses in this article: The Honda D15/15. These are pretty much exclusively what I've worked on tuning and now even building headers for for the past 8 years. In my experience with these engines and the cars that they are powering, the reason why headers don't make too much of an impact on power when slapping them on is tuning, not the header itself. A typical suite of I/H/E will lean out the engine, especially under 3500RPM, which is why many people will report a power loss at lower RPM when going through your basic bolt on phase. However, once the engine is properly tuned, you regain all the power "lost" and can typically add back a healthy 3-5WHP all over (or more, depending on the header, engine and if the moon is in the Second House of Aquarius).

Basically, my point is that with the Honda D15/D16, insensitivity towards a header comes down to the ECU.

As far as making more power with those engines, it is certainly possible to do so with a custom header. For example, this header I built:


Made this flat of a torque curve:

on a D16 from 3500-7500 (names and RPM withheld at the time to keep the racing team's info confidential), which is unusual, at least in my experience.

Anyway, my point is that at least in this example, I wouldn't chalk up header insensitivity to the headers themselves. The tune of the engine plays more of a part in that than most suspect, as the stock ECU is very optimized for the restrictive manifolds they come with from the factory and that once the engine is allowed to breath, the ECU runs the engine lean, loosing power.
Mike Kojima
Mike Kojimalink
Wednesday, January 30, 2013 8:47 AM
My experience was trying to get more power from a D16 in a ITA car which was restricted to a stock cam. A super duper custom header really didn't do a whole lot better than a DC shelf header. We tried different venturi sizes and primary lengths but still could not make an improvement. In my opinion header insensitivity is usually due to a lack of overlap in the cam. That's been my experience with several different motors as well. Maybe the D motor was not the best example because a header does help but the super custom header did not improve much even with a bunch of tuning, or at least did not justify the expense.
Wednesday, January 30, 2013 8:51 AM
Ah! That makes more sense. Also, if it was a D16A6 (CRX/Civic), the intake manifold really sucks, and that could be another reason why you had issues. For the IT crowd, it sure is a pain to get hampered by stock parts when even moving to later model stock parts would make a huge improvement.

Thanks for the reply!
Mike Kojima
Mike Kojimalink
Wednesday, January 30, 2013 8:59 AM
Now can you teach me on how to make a proper intake manifold? I have been able to work with a stock intake for various cars to make more power doing things like extrude honing and porting but every single time I have tried to do a ground up manifold on a NA motor it sucked. If it made more power, it lost a bunch under the curve. Turbo cars are a different story, give them plenum volume and a decent runner and they zing but NA motors are beyond me, at least I am gun shy because I have wasted a lot of time and money trying to do them.
Wednesday, January 30, 2013 9:26 AM
I'm also in the near-clueless category with this one.

My experience is VERY limited to pretty much Hondas (I've had my CRX since 1998) and I've only recently started dorking around with other vehicles. That being said, I've noticed a few things along the way.

If you take a look at the Edelbrock PerformerX intake manifold for the D-series, you'll find a lot of really good clues as to what works well, at least for I4s. The throttle body inlet is bent. The runners are tapered and very even as far as length goes (much more so than most of the OEM manifolds) and the plenum size isn't too big or too small. This was originally part of their turbo kit, but has proven to work VERY well on NA engines as well, even stock blocks. (Definite area under the curve improvements nearly everywhere!) I had mine flow tested out of curiosity (I had a friend at the time who bought a used Superflow bench), and the results were pretty effing amazing. With no throttle body or even radius on the TB flange, the manifold flowed a nominal 250CFM @ 28inHg, with the variance of only +/-3CFM between the cylinders. I don't have the velocity figures at the runner outlets, but I remember the comment was "very good." Overall, it works.

Other things I've noticed that make good power for NA engines is most manifolds that make good power have tapered runners, radiused inlets or even radiused inlets raised off the floor (usually following some formula that depends on who is making the manifold), a plenum designed to keep as much energy inside (as oppose to being lost out the intake tube) via non-parallel surfaces, a plenum end that extends far enough past the farthest inlet (on a side-inlet manifold) to prevent cylinder packing (I think the angling of the throttle body inlet helps with this, specifically), and enough plenum volume for the purpose at hand.

I mostly dork around with turbos, but help out a bunch of guys/teams who do track racing now a days. Most of the time, if you can get away with it, starting with the best OEM manifold you can and improving it (cutting the plenum, reshaping inlet redii, creating anti-reversion steps where you can, porting it for more plenum volume, tapering the runner inlets if there is enough meat to port the runners much, etc.) has been a much better use of time and resources than going from scratch.

NA tuning is something I do to learn more so I can make better turbo setups. LOL! The header I post above was certainly an eye opener to me, as it flies in the face of a lot of "conventional Honda thought," but works well enough for me to keep pursuing those ideas and trying to incorporate them into my turbo endeavors, as I like smaller, efficient turbo setups that minimize lag, more mimicking a larger NA engine than having a typical "lag lag lag OMG SHIFT" feel.
Mike Kojima
Mike Kojimalink
Wednesday, January 30, 2013 9:38 AM
I never messed with tapered runners because I could not model them (at least I don't know how to do it) They are not buildable by normal fabricators either! If I had access to some SLA stuff it would be cut and try heaven. Tried most of the the other tricks but positioning the throttle to keep the waves in the intake manifold is one that never occurred to me. My intake manifolds were simplistic. tried to keep it to a one degree of freedom system with the tuning in the runners with radiused inlets. Then use a plenum with non parallel walls, 4" from the end of the runners and at least 1.5 times engine displacement for overall volume. This is all that I am smart enough to try to figure out. This sort of thinking has always worked with a turbo motor but with NA has usually done something like more peak power with a substantial loss everywhere else. We even made one manifold with an adjustable plenum to try and tune the volume and stand off height. That didn't work either. I bet tapered runners to improve velocity over a wider range and throttle body placement might hold some keys.
Wednesday, January 30, 2013 10:00 AM
You can build tapered runners from thin sheet and careful rolling. It is not fun, but entirely doable and time consuming. Another way is to start out with thin-wall tubing and cut a triangle out of the side with an angle grinder, then carefully smash and tack, then massage the rest of it into shape. I've seen this sort of thing done with dies . . . but that is $$$$$

I do suspect that the tapered runners, with nice radiused inlets, make the engine less sensitive to runner length, do the the nozzle effect of the taper and the efficiency gain (by reduction of some "pumping losses" at the mouth of the runner), which is usually a good thing because with OEM packaging, there nearly always has to be a compromise on runner length (usually too short, unless there is some variable length mechanism).

I do think the inlet angle is a huge thing that most guys just can't "see" well enough to wrap their heads around. I get a fuzzy picture that just shows me how little I know of what is actually going on inside. Looking at the "best performing" manifolds I see results from, nearly all of them (for I4 engines, anyway) employ the bent throttle inlet, where the centerline of the inlet tends to "hit" near the middle of the back wall of the plenum.

Some things I have wondered about but not either seen or had the chance to mess with are offset plenums (inlets shifted towards one of the sidewalls), alternate throttle body placement (there are some OEMs that do this, the first one that comes to mind is BMW, and I guess Mercedes, too, with their inline engines), directional fins (I've seen this done in turbo setups, but never NA) and encapsulated secondary injectors (a fuel rain supported line of injectors above the runner inlets, completely encapsulated in the plenum, as most of the time I see this either "open" on an ITB setup, or simply capped in what I consider a crude plenum over an ITB setup for rules compliance about blocking flamables).

There is a whole world of things to try. I just don't have the time, money, or clientèle (yet) to build this stuff.
Wednesday, January 30, 2013 10:02 AM
Ech. I meant "usually too long" in the above. There are some pretty short runners from OEMs, but I think that is much more unusual.
Wednesday, January 30, 2013 12:27 PM
Jeez Beave, way to rule the comments :P

I've found that anything you can do to increase the velocity into and out of the combustion chamber will have positive results in torque and overall response. Low-overlap motors can have issues seeing these gains, but adding a cam with a little overlap can really wake a motor up!
Wednesday, January 30, 2013 12:32 PM
Hey, I know you. LOL! I just happen to love this subject and can really ramble. Which you know. heh

It isn't always a matter of straight up increasing velocity, but when you are increasing that velocity, since nearly nothing works at all times.

You are certainly right about the cam, and that is easily one of the best things to do, but, in some cases, the cam is restricted, which is more than a little annoying to me, but I don't make the rules.
Bob Holmes
Bob Holmeslink
Wednesday, January 30, 2013 12:44 PM
According to an aticle in Racetech or Racecar Engineering, the BTCC did a test on intake manifolds and found that none, that's right not one, of the bespoke intake manifolds in the series produced more power than the stock manifold.

Clearly its hard to do better.
Wednesday, January 30, 2013 12:48 PM

Do you have a link to that? I don't know much about BTCC engines, but are they limited or restrictorized?
Bob Holmes
Bob Holmeslink
Wednesday, January 30, 2013 12:58 PM
Ok, I'm getting frustrated, twice I've added lengthy comments, hit the add comment button and the damn thing disappeared. Arrrrghhhh.
Bob Holmes
Bob Holmeslink
Wednesday, January 30, 2013 12:59 PM
Over the last three years, I’ve participated in the Engine Masters Challenge with an early Hemi. We’ve tested at least three headers each year, as well as collector length and muffler type. We’ve used Larry Meaux’s Pipemax (a great buy at $45), Dynomation, and team members’ proprietary formulas to try and determine the best header, as well as run off the shelf products. The only way we’ve been able to pick the best header is with testing. There have been times when the “all science-out” header has been beat soundly by the off-the-shelf product. The next year, with a change in rules and engine specifications, the roles will reverse. The largest bolt-on horsepower increase (everything else staying constant other than fuel/air tuning) we saw was 50 hp, with that trend continuing from the hit to redline. This was on an engine that peaked in the mid-600 horsepower range.
One item that clearly made a difference was collector length. We test in Chico, Ca at John Beck’s Promachine. His dyno room is different than the competition room at UNOH in Ohio. This last year, we were down on power at the completion, but we managed to make the final three. One sharp eyed team member noted that the effective collector length at the competition was shorter than what we used in Chico. That evening we crunch the numbers and determined that the collector was 8” too short. We had the kids at UNOH make us up some collector extensions and the resulting increase in power vaulted us from third to second in class.
Although unrelated, we tested the engine with and without exhaust driven PCV. We place the PCV exit at the apex of the merge collector. When operating, the installation provided 10 hp/tq throughout the power curve.
There is no doubt that there is substantial power to be found in exhaust tuning.
Bob Holmes
Bob Holmeslink
Wednesday, January 30, 2013 1:02 PM
BTCC engines are "performance equalized." They have been moving from a NA formula to the "world racing engine" 1.6T. Its through the testing that they attempt to equalize the competitors. I'll see if I can dig it up, I think I gave the magazine to a buddy as we are fiddling with our 2.3t roadracing engine and I wanted his take on the article.

It would have been over the last 4 months.
Wednesday, January 30, 2013 1:05 PM
I had the same issue earlier, Horsewidower. I've been editing comments out of the browser, then pasting them and trying a few times. Isn't all the stuff supposed to just work by now?

I definitely like long collectors! There doesn't seem to be much, if any, downside to them. I've also been experimenting with megaphones with some interesting results, but I am not at all sure how they work yet.
Bob Holmes
Bob Holmeslink
Wednesday, January 30, 2013 1:20 PM
Me too, I've gone to drafting in Word and pasting.

Megaphones are interesting, we haven't tried them yet. I have seen them on some of the "BIG" shop competitors entries. Obviously something there.

This year I'm working with a different team, and we'll be trying sets of headers spec'd by three name designers. Of course we say that now and we'll see who actually performs. LOL.
Wednesday, January 30, 2013 1:25 PM
If you've invested more than 15 seconds into a comment I HIGHLY suggest using the copy/paste method. Going into the detail would bore you to tears...
Mike Kojima
Mike Kojimalink
Wednesday, January 30, 2013 1:30 PM
My comment just vanished....
Mike Kojima
Mike Kojimalink
Wednesday, January 30, 2013 1:30 PM
I wrote some stuff about my megaphone observations but I guess that can wait till part 2.
Wednesday, January 30, 2013 1:53 PM
We will certainly be waiting, Mike!


I've had to delve deep into automotive historic literature to get much info I can understand on megaphones. It seems that they were forgotten except by a few for a very long time, at least outside of the 2-stroke world. I'd love to gather some more real world experiences with them to figure out what is going on. I don't have access to a dyno (that I can afford) and mostly only have anecdotal evidence to go on so far. Most of what I've done is public (but I'm not in the business of self promotion, usually) as I love it when I do something that makes people think!
Dan DeRosia
Dan DeRosialink
Wednesday, January 30, 2013 10:22 PM
What's interesting is that I know there's some people using megaphones on 4-strokes in weird ways... wandering around the paddock at Road America, I ran into some frontrunning Formula Atlantic cars with something that looks a lot like a 2-stroke expansion chamber in their exhaust, though I have no good idea why. And never let it be said that Loyning and the like don't know how to build a naturally aspirated 4-banger, so it's got to be there for some reason...

Hm, think I have pictures of that around somewhere. Will have to see if I can dig them up tomorrow.
Wednesday, January 30, 2013 11:15 PM
I've got some experiments I would like to finish this year:


The collector on the left in the above picture is my attempt to see how big is too big. The header on the right started off as one of the most hyped honda D-series headers on the market, that promptly proved to be nothing but headaches for owners by falling apart at the welds, but aside from that, the collector is HUGE IMO, so when it fell apart (for the third time), I got a hold of it and fixed a few issues with it and the owner is shocked at the difference it made in the engine. (In a good way.) I know it is anecdotal, but, that's most of what I've got so far. It is leading me to believe long primary, low angle collector (both 4-1 and 4-2-1) headers with a proper nozzle and post-nozzle megaphone is a pretty potent combo.

I'm hoping things work out this year with some more empirical testing of my first "sratch built" design versus one of the above headers on a track car.
Dusty Duster
Dusty Dusterlink
Thursday, January 31, 2013 4:43 AM
Awesome article. It's a shame that most OEMs employ log manifolds instead of headers; usually the bean-counters are to blame. Interestingly, the Ford Coyote 5.0 has "shorty" headers from the factory. An article in *5.0 Magazine* detailed the arguments that engineers had to make with the accounting department to justify the added expense.
Der Bruce
Der Brucelink
Thursday, January 31, 2013 8:02 AM
Aaron - Haha, "Boar Me" or "Bore Me"!?

Mike - Great info! I'm hoping that one of the parts will detail Turbo headers a little more and the difficulty one runs into with massive diesel turbos where really it's tough to beat a cast OEM unit for strength .
Thursday, January 31, 2013 9:13 AM
@ - Bruce Yesss, tired not pig ;0) heh
Post Comment Login or register to post a comment.

MotoIQ Proudly Presents Our Partners:

© 2018 MotoIQ.com