If you are serious about your all-motor power, then a custom header is going to be mandatory for your project vehicle. By this point your motor is built to the limit or near-limit. You have high compression, a ported head, different intake manifold and perhaps a different combination of bore and stroke. You have camshafts with an aggressive profile. The last of these modifications, the camshafts, have the largest affect on header design. A camshaft's duration and lobe centers have a direct effect on overlap and the optimal primary length, diameter and collector type of the header. If you have a highly modified engine then regular pre-made headers probably won't cut the mustard.
Generally, the off-the-shelf headers on the market have been designed around stock camshafts. The worst headers are going to be copies of something or just simply made using the designer's 'experience' in making headers. This can be as bad as having some fabricator stuff some tubes into the engine compartment with no consideration toward tuning. Motors with stock camshafts are less sensitive to variations in header design but once you start upping the duration and increasing overlap, the engine becomes increasingly sensitive to header and exhaust design and thus the more important optimizing the header becomes. Creating harmony between a built engine and the header's tuning is one of the main reasons why custom headers can extract surprising amounts of power from heavily modified motors over off-the-shelf offerings. I have personally seen a custom header make 30 hp on a heavily modified street motor over the best off-the-shelf header design. The next time you go to a drag racing event and walk through the pits, note that none of the top all-motor cars are using off-the-shelf headers.
Design your own header
If you want to try your own hand at header design, I have reduced down some of the math involved to these simple equations and tables that should get you in the ballpark. The actual formulas are much more complex, but the complex stuff is reduced down to constants here. These simple formulas are not the end all solution for the ultimate in header design, but they are much better than a 'WAG' and they will get you in the ballpark. This information has helped me many times in my career to date, and it has never resulted in a header that sucked. Even if you don't want to design your header, you can use this information to sort through the design specs of a bunch of off-the-shelf headers to help pick one that is the most likely to work well on your motor.
The first step is to calculate the length of the primary tube. The formula for Primary length is:
|P = ||Primary Length |
|ED = ||180 degrees plus the amount of degrees before bottom dead center that the exhaust valve opens |
|RPM = ||the RPM that the header is tuned to work best at. |
|You can roughly calculate primary internal diameter with this formula: |
|ID = ||(The square root of cc/(P+3) x 25) x 2.1 |
|ID = ||Inside Diameter |
|cc = ||Cylinder Volume in cc |
|P = ||Primary Length |
Having a tube with a slightly larger cross sectional area than the exhaust port is a decent starting point as well.
If you wanted to design a Tri Y or an interference branch style header, you first determine the best overall primary length by using the above equation or my handy dandy table. Make the length to the first Y junction from 13-16 inches. Subtract this from the overall primary length to determine how long to make the tube from the first junction to the main collector.
To find the inside diameter of the first junction use the equation we last used to determine the ID of the primary pipe. From this diameter we can determine the diameter of the next branch using this equation:
|ID2 = ||the inside diameter of the secondary primary |
|ID = ||the inside diameter of the first part of the primary |
|The collector should ideally be a merged collector with an included merging angle of 14-20 degrees. |
|To find the diameter of the collector, this formula can be used to get you in the ballpark: |
|Collector ID = ||(the square root of cc x 2/ (P+3) x 25) x 2 |
|cc = ||cylinder volume in cc |
|P = ||primary length in inches |
When designing a header for low-end power and street use, you typically want to tune the header for the rpm of the estimated torque peak. For forms of racing that need a useful powerband like road racing or short circle tracks or perhaps a serious streetcar, you may want to tune the header for somewhere between the torque and power peak. For all out drag racing with a close ratio gear box in a light car you might tune for the rpm of the power peak.
For a fun mental exercise, try calculating what should work with a stock engine with a stock cam at various streetable rpm ranges, and then compare your findings with typical off-the-shelf headers. Afterwards, "design" some headers for the same engine with available performance and racing camshafts at higher but realistic RPM ranges. See the big differences? Now do you wonder why you rarely see any market headers on the cars in the all-motor class?
Although these equations are what many engineers use when designing a header, they don't take into account many of the recent design trends for header design that are being proven to work quite well. Many of these latest trends cost a lot more to make and are not likely to be found in an off-the-shelf production header. These trends are proven power adders or powerband wideners which makes designing a custom header incorporating these features more and more worthwhile.
Some of the latest design trends are: stepped primary tube diameters, anti-reversion chambers, merged collectors and venturi collectors. A stepped primary diameter steps up in primary diameter two to three times over the length of the primary. Measurements like 1.75 inches to 1.875 inches to 2 inches are common in high revving import motors. Usually these steps go in lengths of 7 inches or so. By making the propagation of waves and refractions as discussed last month more gradual, stepped primaries generally give a wider powerband with no loss of top end power. Most engines with larger camshafts respond well to stepped primaries.
Anti-reversion chambers are controversial. These chambers are areas in the primary tube with a larger ID over a short distance usually about 5-7 inches away from the head flange. The chambers sort of look like goiter bulges in the primary pipe. They are supposed to damp out the return of the reflected acoustic wave to prevent the short-term spike in primary tube pressure around the exhaust valve on overlap. Whether they actually do anything is a fierce source of debate among header designers.
As discussed last month, merged collectors are the best for power production and width of powerband. They are exceeding difficult to fabricate, however. Burns Stainless sells many variations of merged collectors of exquisite quality, which can greatly aid in fabrication of your custom header. The majority of fast all-motor racers in this country use Burns Collectors. Additionally, many of the best fabricators use Burns collectors as a labor-reducing component in there own custom headers since no one does it better.
A venturi collector has a necked down area just past where the primary tubes merge. Generally this is a cone shaped neck down with a 7-10 degree taper with a megaphone with a similar taper stepping the collector back up to the full diameter of the exhaust pipe. For most compact cars, the venturi goes from 3 inches in diameter at the merge down to a 2 3/8-inch venturi, back up to 3 or more inches to the exhaust pipe. Sometimes a short reverse cone is added to the end of this megaphone before the exhaust pipe starts to add yet another back pulse to help broaden the powerband further. The purpose of this venturi is to speed velocity and create a stronger low-pressure rarefaction at the exhaust valve without reducing flow much. Some header builders use a short primary tube for good top-end and use the venturi collector to help maintain a broader powerband. Burns Stainless offers prefabricated venturi collectors, some with removable and tunable venture sections.
One of the best ways to design a custom header is to let a professional engineer do it. Burns Stainless offers such a service. If you can give them some detailed information about your engine, they can do all of the calculations for a nominal fee to spec out a header's primary length, diameter, step sizes, collector taper, megaphone length and diameter. The fee is applied towards your purchase of a merged collector for your project. It's a little known fact that a majority of the successful naturally aspirated racing motor headers in this country have had design input from Burns. Even if you don't go custom, you can take Burns' suggestions and use them when shopping to buy a shelf header that comes closest to what Burns's calculations ended up being.
Even with the best equations and calculations, the header created usually still is not the optimal for your engine. Even a change in cam timing done on the dyno can change the optimal tuned length for the header's primaries. If you have the budget, dyno testing is the best way to fully optimize your header to your combination. Burns Stainless sells slip fit collectors that are held to the header primaries with springs. This enables you to make a test header where you can alter the tuned length of a header during testing in short order to find what works best. Burns also makes a megaphone merged collector with slip in venturis which can be exchanged with different sizes to determine which works the best during dyno testing. This also allows the header to be tuned for different track conditions as well.
Log turbo manifold
Many debate the usefulness of simple log turbo manifolds versus a header combined with a turbo. It is true that a turbo motor can make lots of power with a log-type manifold. It is also true that a log manifold can spool faster than a tuned manifold. This is because a turbo is driven partly by heat energy and expansion of the hot exhaust gasses. The longer tubes in a tuned header tend to dissipate a lot of this heat energy before the turbo, which can result in more turbo lag. However, turbos can use the pulse energy for better breathing and to help spool the turbo faster as well.
In my experience, log manifolds and a properly engineered tuned manifold will have nearly the same boost onset rpm. The log manifold spools the turbo faster and more violently while the tuned manifold has a smoother, more gradual onset of boost that is more manageable and controllable with the throttle. The engine with a tuned manifold will be snappier off boost. A properly designed, tuned turbo manifold will have from 30-100 more horsepower than an untuned, or log manifold, at the same boost level. It's usually a good trade off. The lag and heat loss of a longer runner tuned turbo manifold can be minimized through the use of stainless steel in the manifolds construction. Stainless has poor thermal conductivity and keeps the heat in the pipes and thus transfers more heat to the turbine. By adding tricks to the manifold design like pulse conversion where the runners 180 degrees out of phase are matched with each other to let the turbine have evenly timed pulses hitting it, the lag can be reduced or even improved over a log manifold while retaining the advantages of a tuned system's good breathing. As an example in a 4-cylinder: you would pair cylinders 1 and 4 with 2 and 3, side-by-side right before the turbine. Doing this with a twin scroll turbine housing makes for a huge improvement in spool time in 4-cylinder and rotary engines.
When taking the effort to build a custom header, it's a wise idea to use premium materials. Stainless steel is the material of choice; and a 304 alloy is the minimal grade you would want to use, while 321 is preferred because it is tougher and more heat resistant. For a turbo engine, 321 is the only choice. If you are going to the expense of a custom header, the cost to upgrade to 321 is not too much of an issue. If you are going to be running your turbo motor at wide open throttle for long periods of time, as you would in road racing, then the aerospace super alloy Inconel should be considered. It's a little known fact that stainless steel headers make more power than mild steel headers. This is because stainless steel has much poorer thermal conductivity than mild steel. Mild steel conducts heat 220 times better that your typical stainless alloy. Stainless keeps the heat in the tubes, subsequently keeping the exhaust energy higher. Higher energy equates to higher velocity and better scavenging. Sure you can do this with thermal barrier coatings or thermal wraps, but coating cost extra money and wraps can cause the headers to crack. Again, Burns Stainless is one of the best places to buy tight radius bends and tubes of 304,321 and Inconel to make your header.
Typically a custom stainless header might set you back from $1500 to $3000. Although this seems pricey, when you have a fully built engine, it's relatively easy free horsepower on the table and is still relatively low on the bang-for-your-buck scale.
Hopefully this has made you an expert on headers, enabling you to optimize your setup or make wise decisions when buying a header for your ride. Until next time...
|RPM ||50 ||55 ||60 ||65 ||70 ||75 ||80 ||85 ||90 |
|4000 ||46.0 ||47.0 ||48.0 ||49 ||50.2 ||51.2 ||52.0 ||53.3 ||54.4 |
|4500 ||40.5 ||41.3 ||42.3 ||43.2 ||44.3 ||45.1 ||46.0 ||47.0 ||48.0 |
|5000 ||36.2 ||37.0 ||37.7 ||38.6 ||39.5 ||40.4 ||41.0 ||42.0 ||43.0 |
|5500 ||32.6 ||33.2 ||34.0 ||34.8 ||35.7 ||36.4 ||37.0 ||38.0 ||38.8 |
|6000 ||29.6 ||30.3 ||32.0 ||31.7 ||32.5 ||33.1 ||33.8 ||34.5 ||35.4 |
|6500 ||27.2 ||27.6 ||28.4 ||29.0 ||29.8 ||30.4 ||31.0 ||31.7 ||32.4 |
|7000 ||25.0 ||25.5 ||26.0 ||26.7 ||27.5 ||28.0 ||28.6 ||29.2 ||29.8 |
|7500 ||23.0 ||23.6 ||24.2 ||24.7 ||25.4 ||26.0 ||26.5 ||27.0 ||27.6 |
|8000 ||21.5 ||22.0 ||22.5 ||23.0 ||23.6 ||24.0 ||24.6 ||25.2 ||25.8 |
|8500 ||20.0 ||20.5 ||21.0 ||21.5 ||22.1 ||22.5 ||23.0 ||23.5 ||24.0 |
|9000 ||18.7 ||19.2 ||19.6 ||20.1 ||20.6 ||21.0 ||21.5 ||22.0 ||22.5 |
|9500 ||17.6 ||18.0 ||18.4 ||19.0 ||19.4 ||19.8 ||20.2 ||20.7 ||21.2 |
|10000 ||16.5 ||17.0 ||17.4 ||17.8 ||18.3 ||18.6 ||19.0 ||19.6 ||20.0 |
|10500 ||15.6 ||16.0 ||16.4 ||16.8 ||17.2 ||17.6 ||18.0 ||18.4 ||18.8 |
|11000 ||14.8 ||15.0 ||15.5 ||16.0 ||16.3 ||16.7 ||17.0 ||17.4 ||17.8 |
This table is handy to roughly estimate primary length. The numbers in the table are primary length in inches.