Automotive Horsepower Guide: Exhaust Headers

British engineer and author Philip H. Smith wrote a 250-page book entirely on intake and exhaust systems because the subject warranted it. We don’t have space here to do that, so you are going to get the hands-on basics of an important and complex subject.

 


This Tech Tip is From the Full Book, DAVID VIZARD’S HOW TO BUILD HORSEPOWER. For a comprehensive guide on this entire subject you can visit this link:
LEARN MORE ABOUT THIS BOOK HERE

 

SHARE THIS ARTICLE: Please feel free to share this article on Facebook, in Forums, or with any Clubs you participate in. You can copy and paste this link to share: http://www.musclecardiy.com/performance/automotive-horsepower-guide-exhaust-headers/


 

The energy available to scavenge the combustion chamber is far higher than is often supposed, and harnessing it can produce a power increase equal to 3 to 4 pounds of supercharger boost. Such a power gain is there for the taking and requires little more than getting the exhaust manifold (header) dimensions right.

The principal and most critical dimensions we have to take care of are shown in Illustration 13-1. What we see here is about the simplest form of exhaust manifold: the 4- into-1 system. There are others, such as the 4-into-2-into-1 or 6-into-3- into-1 systems. As you may expect, due to the number of cylinders involved and the engine’s layout, variations abound.

To get a handle on how the system “tunes,” we look at a simple onecylinder system. In Illustration 13-1, we have a single exhaust pipe with the exhaust-valve end to the right and the open end to the left. When the exhaust valve initially opens, high-pressure gas from the cylinder is rapidly dumped into the pipe. The high-pressure wave then travels along the pipe at the speed of sound (about 1,300 ft/sec in hot exhaust gases).

 

In a high-performance engine, the exhaust has a greater effect on the induction system than the piston does going down the bore on the induction stroke. To make use of this valuable performance asset, it is necessary to understand what is going on. The fundamentals of such are dealt with in this chapter.

In a high-performance engine, the exhaust has a greater effect on the induction system than the piston does going down the bore on the induction stroke. To make use of this valuable performance asset, it is necessary to understand what is going on. The fundamentals of such are dealt with in this chapter.

 

13-1. Dimension A (yellow line) is the primary length. Dimension B is the secondary (collector) length. Dimension C is the primary diameter and Dimension D is the secondary/collector diameter.

13-1. Dimension A (yellow line) is the primary length. Dimension B is the secondary (collector) length. Dimension C is the primary diameter and Dimension D is the secondary/collector diameter.

 

When the wave reaches the end of the exhaust pipe, two things happen: It is reflected, and it changes from a positive-pressure wave (above atmospheric) to a negative-pressure wave (below atmospheric). This negative-pressure wave, which can be as much as 7 psi below atmospheric pressure, then travels back along the exhaust pipe to the exhaust valve. If this negative-pressure wave arrives at the exhaust valve while it is open and in the overlap period, it not only sucks out the exhaust residuals in the combustion chamber but also lowers the chamber pressure to something well below the intake-port pressure. This results in the charge flowing into the cylinder, even though the piston may not have quite reached TDC. By the time the piston is about to start the intake stroke, the intake charge can already be moving at as much as 100 ft/sec into the cylinder.

For the events to happen at the intended RPM, the exhaust-pipes need to be a certain length. There are many formulas out there that give a good working result. These formulas take into account the valve opening and closing events as well as the RPM involved. Rather than bog you down with yet more formulas, I feel a simple graph (see Chart 13-2) gets the job done more easily.

The chart is easy to use, but it does make an assumption that you need to take into account. The assumption here is that the cam being used will get longer, but the duration of a cam and the characteristic it produces are very much intertwined with the valve-size-to-displacement ratio. For instance, a 280-degree cam in a fourvalve engine looks like a small race cam, and power can peak as high as 8,000 rpm. That same 280-degree duration, in a 500-inch big-block Chevy, may peak at no more than 6,200 rpm and run like a very civilized street cam. The company that manufactured your cam should provide the RPM for peak power with the right exhaust. As for applications, this chart is for one-, two-, four-, and six-cylinder engines and in-line or flat-cylinder configurations. It is not for V-8s.

 

13-2. Shown here in much simplified form is the process by which the exhaust can be made to scavenge the combustion chamber and make the intake flow start much sooner.

13-2. Shown here in much simplified form is the process by which the exhaust can be made to scavenge the combustion chamber and make the intake flow start much sooner.

 

In this graph, the intake-port velocity is computed from the measured intake pressures. Of interest here is that the intake velocity, when the piston is at TDC and the intake stroke about to commence, exceeds 80 ft/sec (nearly 60 mph).

In this graph, the intake-port velocity is computed from the measured intake pressures. Of interest here is that the intake velocity, when the piston is at TDC and the intake stroke about to commence, exceeds 80 ft/sec (nearly 60 mph).


 

When it comes to primary length selection of a two-plane crank V-8, all those length calculations go out the window. What we have is actually the most complex situation, in terms of function, paired with what must be the simplest solution.

To see how this works out, it must first be understood that a two-plane crank V-8 is (unlike Cosworth’s old DFV F1 V-8) not two inline-fours put together, but two V-4s. The firing pulses down each bank do not occur at regular intervals but are, starting at cylinder number-1, spaced 0-270-180- 90-180-270 and so on. Chart 13-4 shows two pressure pulses seen at the entrance into the collector. Note that cylinders number-5 and -7 (being only 90 degrees apart) appear as if they are one single, much-larger cylinder.

But that’s not all; that pair of cylinders also appears to the collector as if it is revving more slowly than the other pair. The result here is that a two-plane crank V-8 simply does not conform to the primarylength predictions (shown in Chart 13-3). Now that may sound a little disappointing but, by a quirk of fate, it all works in our favor.

What we find in practice is that a V-8 is very much primary-length insensitive. If the primaries are somewhere between about 28 and 40 inches, things work just fine. This is also a reason why getting all the pipes the exact same length for a V-8 is something of a waste of time. It’s best to make sure that all the bends are a big radius and the route to the collector is the least tortuous.

 

Pipe Diameter

The diameter chosen is equally important, and there are abundant methods that show you how to calculate it. However, none that I know of actually takes the valve lift and the flow capability of the port into account.

 
MuscleCarB


 

Primary Pipe Diameter

Sorting out the length needed to tune to the desired RPM really is only half the way toward achieving an effective primary pipe spec. Over the years I developed some curves explicitly based on dyno test results that appear, in most instances, to deliver very near optimal results. These curves, as seen in Chart 13-5, rely solely on the flow capability of the port at the full valve lift figure delivered by the cam.

 

13-3. This chart assumes a progressively longer cam as RPM across the bottom scale increases. This being so, it predicts the primary length for a short street cam at the low end of the RPM scale and a race cam for the high end. This means you must choose a realistic tuning RPM that also represents about where the cam allows peak power to occur, given a properly tuned exhaust. To make the most use of the tuning effect, choose an RPM that is about 5 percent less than the expected peak power RPM. Note: All these lengths are from the valve, so the length of the exhaust port in the head casting must be subtracted for the length computed here.

13-3. This chart assumes a progressively longer cam as RPM across the bottom scale increases. This being so, it predicts the primary length for a short street cam at the low end of the RPM scale and a race cam for the high end. This means you must choose a realistic tuning RPM that also represents about where the cam allows peak power to occur, given a properly tuned exhaust. To make the most use of the tuning effect, choose an RPM that is about 5 percent less than the expected peak power RPM. Note: All these lengths are from the valve, so the length of the exhaust port in the head casting must be subtracted for the length computed here.


 

Secondary Dimensions

At this point, we have sorted out the primaries. To get a workable secondary diameter, multiply the primary by 1.68 for a hot street application, 1.73 for street/strip, and 1.78 for all-out race. If we are dealing with a one-cylinder engine, better results may be seen with a smaller secondary of about 1.6 times the primary. Determining the secondary/collector diameter was straightforward enough. The same applies to the collector but, because it is easy to adjust, we don’t need to take theory as the last word. For a V-8, the secondary, unlike the primary, is very sensitive to length tuning. The test in Chart 13-6 shows how just 12 inches of collector length can affect power.

 

13-4. Here are the exhaust pulses, as seen at the collector, from one bank of a small-block Chevy V-8. Note that, although cylinders No. 1 and No. 3 produce a distinctive pulse of their own, the pulse as seen at the collector from cylinders No. 5 and No. 7 appears to be one much bigger cylinder. This causes problems when attempts are made to tune in the same fashion as an in-line four.

13-4. Here are the exhaust pulses, as seen at the collector, from one bank of a small-block Chevy V-8. Note that, although cylinders No. 1 and No. 3 produce a distinctive pulse of their own, the pulse as seen at the collector from cylinders No. 5 and No. 7 appears to be one much bigger cylinder. This causes problems when attempts are made to tune in the same fashion as an in-line four.

 

13-5. From these tests on a 350 small-block Chevy, equipped with a hot street cam, we can see that primary lengths from a 29-inch average to a 38-inch average showed only minimal changes in output.

13-5. From these tests on a 350 small-block Chevy, equipped with a hot street cam, we can see that primary lengths from a 29-inch average to a 38-inch average showed only minimal changes in output.

 

To use this chart, you need to know (or need a good estimate of) the exhaust flow at full valve lift. Using this figure, go up from the bottom of the chart to one of the curves (green for street, purple for street/strip, blue for race). An example looks like this: port flow at max valve lift on a 130- cfm street/strip application. From the 130-cfm mark on the bottom scale, go up to where you intersect the purple line, and then go left to read the pipe ID—in this case 1.375.

To use this chart, you need to know (or need a good estimate of) the exhaust flow at full valve lift. Using this figure, go up from the bottom of the chart to one of the curves (green for street, purple for street/strip, blue for race). An example looks like this: port flow at max valve lift on a 130- cfm street/strip application. From the 130-cfm mark on the bottom scale, go up to where you intersect the purple line, and then go left to read the pipe ID—in this case 1.375.

 

The effect of an even moderately-toolarge collector diameter can be seen from this dyno test. The largest diameter delivered less output all the way through the RPM range. Even at peak power, it was marginally less than either of the smaller-diameter collectors.

The effect of an even moderately-toolarge collector diameter can be seen from this dyno test. The largest diameter delivered less output all the way through the RPM range. Even at peak power, it was marginally less than either of the smaller-diameter collectors.

 

13-6. As this dyno test on a V-8 shows, secondary length is critical.

13-6. As this dyno test on a V-8 shows, secondary length is critical.


 

The advantage of engine or chassis dyno testing exhaust systems (to optimize secondary lengths) is that adjusting the length is easy. That means there is almost no excuse for not getting on the dyno and doing just that. In the collector length tests, the blue curve was with about a 1.5-inch collector length after the 4- pipe merge section. By adding 12 inches of collector length, the output rose to that shown by the red curve. The gains are far from inconsequential: 40 ft-lbs at 3,700 rpm, and 12 hp at 6,200 rpm.

On occasion, I have built V-8 exhaust systems for street use with a secondary as long as 60 inches. This proved to be capable of boosting low-speed (2,000 to 3,000 rpm) torque as much as 20 ft-lbs over a more usual 20- to 30-inch collector. At this juncture, I need to point out that the secondary length for a V-8 application is, at lower RPM, somewhat longer than for a one-, two-,or four-cylinder in-line engine. If you are working with a non-V-8 application, you can get a good starting point for the secondary length by making it 15 inches less than the primary length.

 

Alternative Configurations

So far, we have discussed basic manifold configurations based on all the primaries dumping into a secondary. This gives us the commonly seen 4-into-1 system used on V-8s and high-RPM four-cylinder engines. If the engine is a four-cylinder unit, and is not likely to see the top side of 7,800 to 8,000 rpm, then a 4-into- 2-into-1 system is the way to go. Such a system has the advantage of fitting the confines of the engine bay easier and produces more torque up to about 6,800 to 7,000 rpm. At the lower speeds (2,000 to 4,000), such a system can be up by as much as 12 ft-lbs per liter (0.2 ft-lbs per cubic inch). Above 7,800 to 8,000 rpm, it’s a clear-cut advantage to go with a 4-into-1 system.

 

13-7. For four-cylinder applications below 7,800 rpm, a 4-2-1 system (as seen here) gives better results than a 4-1 system. The primary and secondary lengths need to be about the same and generally run around the 13- to 18-inch mark.

13-7. For four-cylinder applications below 7,800 rpm, a 4-2-1 system (as seen here) gives better results than a 4-1 system. The primary and secondary lengths need to be about the same and generally run around the 13- to 18-inch mark.


 

The V-8 world has what can best be described as its own version of a four-cylinder 4-2-1 system. In this instance, though, it is more of a 4-2-1 collector design. As can be seen in Illustration 13-7, the system still has long primary pipes but, instead of merging into a 4-into-1 collector, each pair of primaries merges first into one short secondary and then makes another merge into the collector.

As of 2010, these are the systems of choice for some of the most powerful, naturally aspirated, single-4- barrel-carbed V-8s. Just how much they are worth on an 890-hp 350, I cannot say with any certainty. But on a 740-hp 350, they showed some 10 hp more at the mid-to-top end, but maybe a little less horsepower at 500 rpm under the peak torque. For a street application, they represent a big additional dollar investment over a conventional 4-into-1, and so may not represent a cost-effective alternative.

 
MuscleCarB

 
 
 

Using This Header Tech

Because of space limitations here, I have taken some short cuts concerning exhaust manifolds. However, most hot rodders do not make their own manifolds and, instead, buy them either as off-the-shelf components or as custom-built items. This being the case, the intent of this chapter was more to illustrate what goes into the selection of a manifold. This chapter should leave you, as a prospective customer, with a good idea of the implications involved if you make a poor choice. Heed what I say here and listen to the advice of someone who has expertise with exhaust systems for the particular engine type you are working with. Above all, given the choices, I advise you err on the conservative side. The one common theme all through the dyno tests shown in this chapter is that, past a certain point, bigger is not better—a little too small does not hurt to any measurable extent.

 

This four-long-tube V-8 manifold (header) utilizes a 4-2-1 merge collector. Carefully conducted dyno testing indicates about a 1.2-percent increase in output above the peak torque RPM.

This four-long-tube V-8 manifold (header) utilizes a 4-2-1 merge collector. Carefully conducted dyno testing indicates about a 1.2-percent increase in output above the peak torque RPM.

 

This LS6 Chevy’s power output climbed from 739 hp with the original 4-1 collector to 750 hp with the change to the 4-2-1 merge collector shown here.

This LS6 Chevy’s power output climbed from 739 hp with the original 4-1 collector to 750 hp with the change to the 4-2-1 merge collector shown here.


 

Written by David Vizard and Posted with Permission of CarTechBooks

GET A DEAL ON THIS BOOK!

If you liked this article you will LOVE the full book. Click the button below and we will send you an exclusive deal on this book. Deal-Button