Exhaust Science Demystified

The fact is most cars are leaving horsepower on the table. We show you how to get it back.

Photography by Various Manufacturers

For me the first really serious look at how to muffle a high-performance race engine without loosing a significant amount of power started in 1980 when I built a 400lb-ft, 404hp 350 to replace the very lame 158hp 305 in my California-spec Pontiac Trans Am. Having worked very hard to build a pump gas fueled engine (gas was really bad in those days), that would cross the 400 hp barrier, I was very disappointed to find that, regardless of what mufflers were used, the output dropped by some 20 lb-ft and 25 hp. Having had some experience designing a no-loss system for the original style British Mini Coopers, I felt confident I could pull off the same stunt for significantly bigger V-8 engines. The result, aided by an acoustics expert friend, was the Sonic Turbo. This design went on to be manufactured by Cyclone (now a division of Walker/ Dynomax). After the smoke cleared from a big muffler shootout (done at Gale Banks facility and published by Hot Rod magazine), a pair of 2.25-inch Sonic Turbos (the 2.5-inch ones were still a couple of months off) sunk everybody else's 2.5-inch items. This, it seemed, was just what the hot rod fraternity wanted and they sold by the hundreds of thousandths. That was good, but more importantly, it appeared to spark the industry into aggressively pursuing significantly more functional mufflers and exhaust systems. The result is that 20-some-years later, all the necessary components to build a highly effective, no-loss system are at hand, and not necessarily that much money either. All that appears to be lacking is widespread know-how as to what is needed to achieve this happy state of affairs. As of now, we are going to make a start on putting that right.

Simple Steps to Success
Although the mode of function of an exhaust system is complex, it is not (as so often is believed, even by many pro engine builders) a black art. To help appreciate the way to get the job done I will go through the process of selecting exhaust system components for a typical high-performance V-8 in a logical manner from header to tail pipe. Although the entire exhaust functions as a system, we can, for all practical purposes, break down many of the requirements that need to be met into single entities. Fig. 1 details the order of business. But before making a start, it is a good idea to establish just why getting the exhaust correctly spec'd out is so important. This will allow realistic goals, improved component choice, and a more functional installation.

The V-8 engines we typically modify for increased output are normally categorized as four-cycle units. Although pretty much the case for a regular street machine, this is far from being the case for a high-performance race engine. If we consider a well-developed race engine, the usual induction, compression, expansion (power stroke) and exhaust cycles have a fifth element added (Fig. 2). With a race cam and a tuned-length exhaust system, negative pressure waves traveling back from the collector will scavenge the combustion chamber during the exhaust/intake valve overlap period (angle 5 in Fig. 2). To understand the extent to which this can increase an engine's ability to breathe, let's consider the cylinder and chamber volumes of a typical high-performance 350 cubic-inch V-8.

Assuming for a moment no flow losses, the piston traveling down the bore will pull in one-eighth of 350 cubic inches. That's 43.75 cubic-inch, or in metric, 717cc. If the compression ratio is say 11:1, the total combustion chamber volume above this 717cc will be 71.7cc. If a negative pressure wave sucks out the residual exhaust gases remaining in the combustion chamber at TDC, then the cylinder, when the piston reached BDC, will contain not just 717 cc but 717 + 71.7 cc = 788.7 cc. The result is that this engine now runs like a 385 cubic-inch motor instead of a 350. That scavenging process is, in effect, a fifth cycle contributing to total output.

But there are more exhaust-derived benefits than just chamber scavenging. Just as fish don't feel the weight of water, we don't readily appreciate the weight of air. Just to set the record straight, a cube of air 100 feet square will weigh 38 tons! If enough port velocity is put into the incoming charge by the exhaust scavenging action, it becomes possible to build a higher velocity throughout the rest of the piston-initiated induction cycle. The increased port velocity then drives the cylinder filling above atmospheric pressure just prior to the point of intake valve closure. Compared with intake, exhaust tuning is far more potent and can operate over ten times as wide an rpm band. When it comes to our discussion of exhaust pipe lengths it will be important to remember this.

At this time a few numbers will put the value of exhaust pressure wave tuning into perspective. Air flows from point A to point B by virtue of the pressure difference between those two points. The piston traveling down the bore on the intake stroke causes the pressure difference we normally associate with induction. The better the head flows the less suction it takes to fill (or nearly fill) the cylinder. For a highly developed two-valve race engine the pressure difference between the intake port and the cylinder caused by the piston motion down the bore, should not exceed about 10-12 inches of water (about 0.5 psi). Anything much higher than this indicates inadequate flowing heads. For more cost-conscious motors, such as most of us would be building, about 20-25 inches of water (about 1 psi) is about the limit if decent power (relative to the budget available) is to be achieved. From this we can say that, at most, the piston traveling down the bore exerts a suction of 1 psi on the intake port Fig. 3.

The exhaust system on a well-tuned race engine can exert a partial vacuum as high as 6-7 psi at the exhaust valve at and around TDC. Because this occurs during the overlap period, as much as 4-5 psi of this partial vacuum is communicated via the open intake valve to the intake port. Given these numbers you can see the exhaust system draws on the intake port as much as 500 percent harder than the piston going down the bore. The only conclusion we can draw from this is that the exhaust is the principal means of induction, not the piston moving down the bore. The result of these exhaust-induced pressure differences are that the intake port velocity can be as much as 100 ft./sec. (almost 70 mph) even though the piston is parked at TDC! In practice then, you can see the exhaust phenomena makes a race engine a five-cycle unit with two consecutive induction events.

With the exhaust system's vital role toward power production established, it will be easy to see that understanding how to select and position the right combination of headers, resonators, routing pipes, crossovers and mufflers will be a winning factor. This will be especially so if mufflers are involved in the equation. I first started putting out the word on how to build no-loss systems as much as 20 years ago and I am somewhat surprised that it is still commonly believed that building power and reducing noise are mutually exclusive. Historically, this has largely been so, but building a quiet system that allows the engine to develop within 1 percent of its open exhaust power is entirely practical. Be aware that knowing what it takes in this department can easily deliver a 40-plus hp advantage over your less-informed competition.

Headers -- Primary Pipe Diameters
Big pipes flow more, so is bigger better? Answer: absolutely not. Primary pipes that are too big defeat our quest for the all-important velocity-enhanced scavenging effect. Without knowledge to the contrary, the biggest fear is that the selected tube diameters could be too small, thereby constricting flow and dropping power. Sure, if they are way under what is needed, lack of flow will cause power to suffer. In practice though it is better, especially for a street-driven machine, to have pipes a little too small rather than a little too big. If the pipes are too large a fair chunk of torque can be lost without actually gaining much in the way of top-end power.

At this point determining primary tube diameters is starting to look like a tight wire act only avoidable by trial and error on the dyno. Fortunately, a little insight into what it is we are attempting to achieve brings about some big-time simplification. Our goal is to size the primary pipes to produce optimum output over the rpm range of most interest. The rate exhaust is dispensed with, and consequently, the primary pipe velocity, is strongly influenced by the port's flow capability at the peak valve lift used. From this premise it has been possible to develop a simple correlation between exhaust port-flow bench tests and dyno tests involving pipe diameter changes. This has brought about the curves shown in the graph Fig. 4 which allow primary sizing close enough to almost eliminate the need for trial-and-error dyno testing.

Primaries For Nitrous UseSince nitrous injection is so popular, it's worth throwing in the changes needed to optimize with the nitrous on. For a typical race V-8 the area of the primary pipe needs to increase about 6-7 percent for every 50hp worth of nitrous injected. For street applications, where mileage and performance when the nitrous is not in use is the most important, pipe size should not be changed to suit the nitrous.

Headers -- Primary Pipe Lengths
Misconceptions concerning exhaust pipe lengths are widespread. Take for instance the much-overworked phrase "equal-length headers." More than the odd engine builder/racer, or two, have made a big deal about headers with the primary pipes uniform within 0.5 inch. The first point this raises is whether or not what was needed was known within 0.5 inch! If not, the system could have all the pipes equally wrong within 0.5 inch! Trying to build a race header for a two-planed crank V-8 with lengths to such precision is close to a waste of valuable time. Under ideal conditions it is entirely practical for an exhaust system to scavenge at or near maximum intensity over a 4,000 rpm bandwidth. Most race engines use an rpm bandwidth of 3,000 or less rpm. If the primary pipe scavenging effect overlaps by 3,000 rpm then it matters little that one pipe tunes as much as 1,000 rpm different to another. Since this is the case, then all other things being equal, pipe lengths varying by as much as 9 inches have little effect on performance. A positive power-increasing attribute of differing primary lengths is that it allows larger-radius, higher-flowing bends and more convenient pipe routing to the collector in often confined engine bays.

Apart from the reasons just mentioned, there is also another sound reason why we should not unduly concern ourselves about equal primary lengths. In practice, the two-plane cranks that typically equip V-8 race engines render the exhaust insensitive to quite substantial primary length changes. Experience indicates inline four-cylinder engines are more sensitive to primary pipe length, but a two-plane cranked V-8 is not two inline fours lumped together. It is two V-4s and, as such, does not have even exhaust pulses along each bank. With a conventional, as opposed to a 180-degree header, exhaust pulses are spaced 90, 180, 270, 180, 90 and so on. The two cylinders discharging only 90 degrees apart are seen, by the collector, as one larger cylinder and accounts for the typical rumble a V-8 is known for. This means the primaries act like they do on a four-cylinder engine, but the collector acts as if it were on a 3-cylinder engine having different sized cylinders turning at less revs. (Doesn't life get complicated?) This, plus the varied spacing between the pulses appears to be the cause of the system's reduced sensitivity to primary length.

These uneven firing pulses on each bank seem to work in our favor. Evidence to date suggests that single-plane cranked V-8s, which have the same exhaust discharge pattern as an in-line four-cylinder engine, make less horsepower and are more length sensitive. Dyno tests with headers having primary lengths adjustable in three-inch increments show that lengths between 24 and 36 inches have only a minor effect on the power curve of V-8s that you and I can typically afford, although the longer pipes do marginally favor the low end.

Secondaries -- Diameters and Lengths
Well, so much for primary pipe dimensions and their effect on output. Let us now consider the collector/secondary pipe dimensions and configurations. The first point to make here is that the secondary diameter is as critical as the primary. A good starting point for the collector/secondary pipe size of a simple 4-into-1 header is to multiple the primary diameter by 1.75. Fortunately, the collector can be changed relatively easily and it is often best optimized at the track rather than the dyno.

As for the secondary length-that is from about the middle of the collector to the end of the secondary (or the first large change in cross-sectional area), we find a great deal more sensitivity than is seen with the primary. Ironically, few racers pay heed to collector length even though it is easy to adjust. In practice, collector length and diameter can have more effect on the power curve than the primary length. A basic rule on collectors is that shorter, larger diameters favor top end while longer, smaller diameters favor the low end. Except for the most highly developed engines, many collectors I see at the track are too large in diameter and either too short, or of excessive length. For a motor peaking at around 6,000-8,500 rpm, a collector length of 10-20 inches is effective.

Getting secondary lengths nearer optimal can be worth a sizable amount of extra power as Fig. 5 shows. If you want to bump up torque at the point a stock converter starts to hook up the engine, you may want a secondary as long as 50 inches but something between about 10 and 24 is more normal. The shorter of these two lengths would be appropriate for an engine peaking at about 8,500 rpm whereas the longer length would be best for an engine that peaked at about 4,800-5,000 rpm.

Mufflers -- Two Golden Rules To Avoid Power Loss
Inappropriate muffler selection and installation (which appears so for better than 90 percent of cases) will, in a very effective manner, negate most of the advantages of system length/diameter tuning. The question at this point is what does it take to get it right and how much power are we likely to loose if the system is optimal? The quick and dirty answers to these questions are "not much" and "zero." This next sentence is the key to the whole issue here, so pay attention. To achieve a zero-loss muffled high-performance race system we need to work with the two key exhaust system factors in total isolation from each other. These two factors are: the pressure wave tuning from length/diameter selection, and minimizing backpressure by selecting mufflers of suitable flow capacity for the application. If we do this then a quiet (street-legal noise levels) zero-loss system on a race car is totally achievable without a great deal of effort on anybody's part. Ultimately, it boils down to nothing more than knowledgeable component selection and installation, so let's look at what it takes in detail.

Muffler Flow Basics
We select carbs based on flow capacity rather than size because engines are flow sensitive, not size sensitive. This being so, why should the same not apply to the selection of mufflers? The answer (and here I'd like muffler manufactures to please note) is that it should, as the engine's output is influenced minimally by size but dramatically by flow capability. Buying a muffler based on pipe diameter has no performance merit. The only reason you need to know the muffler pipe size is for fitment purposes. The engine cares little what size the muffler pipe diameters are but it certainly does care what the muffler flows and muffler flow is largely dictated by the design of the innards. What this means is that the informed hot rodder/engine builder should select mufflers based on flow, not pipe size.

A study of Fig. 6 will help to give a better understanding as to how the design of the muffler's core, not the pipe size, dictates flow.

Let's start by viewing a muffler installation as three distinct parts. In Fig. 6, drawing number 1, these are the in-going pipe, the muffler core and the exit pipe. Drawing number 2 shows a typical muffler which has, due to a design process apparently unaided by a flow bench, core flow significantly less than an equivalent length of pipe the size of the entry and exit pipe. Because the core flow is less than the entry and exit pipe then the engine "sees" the muffler as if it were a smaller and consequently more restrictive pipe as per drawing number 4. If the core has more flow than the equivalent pipe size, as in drawing number 5, it appears larger than the entry and exit pipe. Result: the muffler is seen by the engine as a near zero restriction. A section of straight pipe the length of a typical muffler, rated at the same test pressure as a carb (10.5 inches of mercury), flows about 115 cfm per square inch. Given this flow rating, we will see about 560 cfm from a 2.5-inch pipe. If we have a 2.5-inch muffler that flows 400 cfm, the engine reacts to this just the same as it would a piece of straight pipe flowing 400 cfm.

At 115 cfm per square inch, that's the equivalent to a pipe only 2.1 inches in diameter. This is an important concept to appreciate. Why? Because so many racers worry about having a large-diameter pipe in and out of the muffler. This concern is totally misplaced, as in almost all but a few cases, the muffler is the point of restriction, not the pipe. The fact that muffler core flow is normally lower than the connecting pipe can be off set by installing something with higher flow, such as a 4-inch muffler into an otherwise 2.75-inch syste