The Importance of Header Design.



HEADER DESIGN COMMENTS - The Importance Of Header Design.

Under Construction GIF


This page will be added to and updated considerably over the next several months
and probably won't be acceptable to us until Summer 2012.
Once deemed somewhat acceptable, the Under Construction sign above will be removed.


One of the many things that I find disturbing as I read car magazines, various automotive books and Internet discussions and watch various automobile-oriented TV programs, is the almost complete lack of discussion as as to how important header design really is (most discussion is centered more on motivating you to BUY a set of headers).  This almost total lack LACK of objective discussion unfortunately seems to be interpreted by many as meaning that GOOD Header Design is of minimal importance in the overall "performance" picture.

The following discussion will, I hope, convey to you not only how important GOOD Header Design is but even how often critical it can be in many situations.

In any given header there are a number of design parameters that affect performance.  Their presence (or absence) or to what degree they are achieved ALL have an effect on the total amount of performance gained (or lost) by the addition of a header or a change of header design.  These design parameters (essentially listed by importance) are:


1.  Tube Size

a.  "Restriction" based Header Design Theory

b.  "Velocity" based Header Design Theory

c.  Two Tube Size Error (when selecting tube sizes)

2.  Tube Length
     3.  Equal Length Header Design

4.  Collector Size (open exhaust)

5.  Collector Transition Shaping
     6.  Collector Length

7.  Exhaust Port Matching

8.  Construction Problems


Tube Size

Of all the Header Design parameters that affect performance, tube size, by far, is the most important.  While I still hear comments about BIGGER IS BETTER when it comes to header design, many, many times it does not prove out to work as expected.  In fact, when I talk to guys that are complaining about the performance (actually, lack of) of headers they've bought elsewhere, one of the questions I find myself asking way too often is "Why did you buy headers that are so big?", (along with "Why are they so short? and "Why are they so UNequal in length?).


"Restriction" based Header Design Theory

One of the most basic MISunderstandings of how headers work to improve performance is based on an assumption that headers basically improve performance by reducing restriction to exhaust flow.  While true to a certain degree, this point of view - which I call the "restriction" based point of view of header design theory - not only dictates that BIGGER IS BETTER but also includes other opinions like "turns in header tubes are bad", "welds are bad", etc.  In other words, anything that can be interpreted as being restrictive to flow in some manner is considered bad and must be avoided if maximum performance is to be achieved.

  When I first began to build headers many years ago I actually had that point of view too - for example, thinking that if I built a header simply bigger than what someone else was making, I was assured of an automatic power gain.  Well, that point of view disappeared quite quickly as by 1964 I began to realize that there was no consistency in the "restriction" based header design theory.  While I had been building headers for only about two years, I had already seen bigger headers slow cars down, headers that others said were "too small" work great (compared to larger headers), big headers that helped top end power but blew away bottom end and mid-range power - totally messing up street performance, etc., etc.  Plus, at times I found myself quite confused as I was often watching the exact opposite of what I believed true working better!  Because of the lack of consistency of my "restriction" based Header Design theories I began to realize there had to be a better way (a theory) of looking at exhaust system design - one that would not only work CONSISTENTLY when applied but also one that would deal with the "goofy" things I was seeing as well .


"Velocity" based Header Design Theory


     To the right is the set of headers that basically changed my mind about header design.  While the car itself was nothing special - a street-driven 1958 Pontiac with a STOCK (unmodified) 370" 2 barrel engine that ran L/SA at the local drag strip - it was what these headers DID to the performance of the car that forced me to really question my "restriction" point of view of header design.
     Through factory exhaust the car ran consistent 15.80s.  With the new headers and open exhaust, the car ran 15.0s the first Sunday out at the drag strip and, with jetting and ignition timing optimized for the headers, ran 14.80s after that - a full second ET reduction!
     At that time this was the greatest ET reduction I had ever seen solely due to a set of headers that I had ever built PLUS, even better, the gain was obtained on a totally STOCK engine! 
Header Picture
1958 370" Pontiac Header (Left Side)
Header Picture
1958 370" Pontiac Header (Right Side)
HEADER DESIGN:  1 5/8" OD Header Tubes on end cylinders, 2" OD Header Tubes on
the center cylinders.  Tube Lengths were 40" long.  Collector is 3" OD x 14" long.
(Headers have an unusual appearance as they do NOT go straight back but,
because the car had an "X" shaped frame (much like 58-64 Chevs), they
followed the routing of the factory exhaust system that went around the
OUTSIDE of the frame and then turned back toward the rear of the car).
     Keeping in mind that this car probably weighed in excess of 4000 pounds with factory rated horsepower in the 220-230 range, the improvement in Elapsed Times probably represents a gain of 50 or more horsepower!  PLUS, looking at the header design some forty five years later and factoring in what has been learned about designing BETTER headers over those many years, I think an even BETTER header could be made now.  Having header tube sizes of 1 1/2" on the end tubes and 2 1/8" for the center tubes seems more appropriate to me now plus the the collector size of 3" is simply way too big.  While not obvious in the photos, the collectors' transition shape and length is nowhere as good as what we're providing in our Custom Collectors now so another gain could be achieved there as well.  Setting up the headers with collector flanges at the end of the collector tapers and then experimenting with bolt-on collector extensions of different lengths as well as smaller diameters would improve performance even more!  I think that these header design changes could easily reduce ETs at least another .3s (another 10-15 horsepower in this particular instance)!  When one also factors in how much better tires are now than they were in the mid-60s, this same car with updated headers and new slicks would probably run mid to high 13s.  WOW! 13s with this heavy a car and a STOCK engine.  WOW again!
     One last comment deals with header parts availability.  The header flange kit I bought then from a big company had starter tubes made out of 1 3/4" tubing.  Had I made the headers based on their parts (assuming that the starter tubes were the right diameters), the end tubes would have been even bigger (worse than 1 5/8") yet the center tube would have been much smaller (an even bigger error!) and the performance gain would not have been much - probably better than a manifold but nowhere near what I proved was possible.  Also - I don't remember that 2 1/8" OD x 3" radius mandrel bent U-Bends were even available in 1964 (manufactured 2 1/8" headers didn't even start to be offered until the early 1970s!).  I also remember that the collectors were for two 1 3/4" tubes and one 2" tube so they didn't even match the tube sizes correctly - yet they were the ONLY collectors I could find at that time even close to fitting the selected tube sizes.  Another problem with the collectors I had to use had a transition length of less than 2" - a terribly short taper length.  PLUS, the use of the one tube for the two center cylinders is due to the fact that 1955-1958 Pontiac cylinder heads only have a single exhaust port in the center (shared by the two middle cylinders).  In 1959 Pontiac split the previously shared center port into two ports so, from then on, each cylinder did have its own exhaust port.  Had these been '59 or later cylinder heads, I would have DEFINITELY built a four tube header as using a single header tube for two cylinders is a poor, Poor, POOR way of designing a header (cheaper - yes, better - no way).







Two Tube Size Error (when selecting tube sizes) and its horrendous effect on performance - street OR racing.


Tube Area
Area Increase
over next
smaller tube
Area Increase
over second
smaller tube
This is a portion of one of several charts found in our
Header Design InfoPAK.
































Tube Length





Equal Length







Collector Size (open exhaust)






Collector Transition Shaping





Collector Length







Exhaust Port Matching





Construction Problems







Notes for future comments:


Importance of Header Design on street driven vehicles.


"Goofy" things being done.  Example:  Sherman and Professor Peabody getting into their "Wayback Machine" (time travel) and visiting a Chevy dealer in 1966 to look at two different 350HP Chevrolets (a SB Nova and a BB Chevelle) where, with two engines of the same power rating, the headers available (then and now) differed by 3 tube sizes!


Many header companies copy one another leading to fellows thinking there are "standard" tube sizes for certain engines.  This "standardization" ignores all the differences that can be there.  Mention sending for catalogs around 1970 and almost total lack of design difference in BB Chev headers - almost all were 2" even though tests and experiences we were having showed many, many times that 2" was too large - even slowing cars down at the drag strip!


 Vehicle Performance "Dynamics" - NOT being paid attention to in the :  (1) Optimization of street vehicle performance relative to rpms, (2) Optimization of SCCA race car performance s (include references to data acquisition devices to develop better - more performance effective - headerr designs;  (3) Optimization of Pulling Header Designs by analyzing what is ACTUALLY going on in a 300' pull and how header design is NOT even considered in most pulling vehicles (relate local puller with a 572" Big Block Chev engine in his truck that consistently beats engines as large as 640" - 70 cubic inches larger - because he pays attention to the importance of the design of his headers!)


Critical - local group of guys built a Bonneville streamliner and paid NO attention to the header design they constructed - AND, unless the car runs in a class with a "soft" record AND the reason for the car to be constructed was to set a new class record - that design error will prevent the car from EVER performing to its maximum performance potential - literally negating the thousands of dollars and thousand of hours of work spent and denying these guys the very goal they have set.


Also:  E-mail us with suggestions you would like us to comment on?