Exhaust Design, Turbo and Naturally Aspirated
by Josh Tenny
Exhaust design is one of the areas of most confusion and can be one of the areas of greatest gain if done properly. There are many factors to consider when designing an exhaust system. Complicating things is the fact that what works perfect for one application may not be even remotely close to optimal for another. Making it even harder to choose the proper exhaust components is the fact that different "tuners" say different things. I have heard so much info myself it is hard to sort through and determine what is marketing and what is correct and truly useful and information. But keep in mind that perception and marketing do play a roll when looking at it from a company's point of view. What good is any product if it does not have enough value for people to buy, or if they do not understand its true advantages. We want to offer a technically superior product and letting our marketing explain why.
One thing I do want to explain before I get too far into this is backpressure. I hear "You need backpressure to make torque." all the time. And it even comes from "tuners" as well as customers. This is flat out not true. Before I arrived into the company of true tuners I fell into the same trap myself. The reality is that backpressure is the enemy. You want to keep it as low as possible. What you do want is to keep velocity up. However doing the things that keep velocity high involves slightly more backpressure under some conditions. You want to keep the gasses moving as quickly as possible to make both good torque and top end power. The perfect exhaust system would keep the gasses moving as fast as they did coming out of the cylinder and have zero backpressure. However this is impossible to achieve in the real world.
Turbo System Exhaust
- Is bigger better all throughout the powerband or does it loose bottom end torque?
- Should I use 2.5", 3.0", or 3.5" for XYZ power?
- Does reducing the pipe diameter towards the end of the exhaust help power?
- What is the best turbo outlet design?
- How does a cat-less exhaust perform compared to a high flow cat?
We get these types of questions all day every day. We get many questions, but most center on the size of the turbo-back system alone and disregard any other parameters. I wish it were that simple, but unfortunately it is not. Post turbo pipe diameter is only one small parameter in many that determine how well the exhaust performs.
Here is a pictorial break down of the WRX exhaust system. Thanks to mattr672 on our forum for a clear image of what the WRX exhaust looks like.
The exhaust system before the turbo and the turbo itself have a greater effect on backpressure than the exhaust behind it. You want the least restriction after the turbo as possible for both top end power and quick spool-up. Careful attention has to be paid to keep velocity high before the turbo and in the exhaust housing of the turbo to spool the turbo up as quickly as possible while not choking off the exhaust gasses on the top end.
The header can be simpler in some ways than a non-turbo header. Bigger dividends can be had by getting the exhaust gasses to the turbo with the least amount of restriction, highest velocity, and the most heat rather than worrying about a tuned equal length design. It would be optimal to make an equal length header, but the packaging of the WRX make a tuned equal length header difficult to design. This helps explain why we usually get near identical results from a factory header when compared to the aftermarket ones we have tested thus far. The factory header gets the gasses to the turbo as quickly as possible and goes a good job of keeping the heat in. Aftermarket headers tend to take a longer path and loose quite a bit of heat in the process. Also, most have a poor collector design that is s a byproduct of the unique packaging of the WRX. In this case a good collector does play a more important role than the length of the pipes. If all of the gasses ram together at a steep angle it causes a lot of turbulence, creates backpressure, slows velocity, and tends to make a mess of things before the turbo, which is the worst spot for inefficiency on a turbo charged car. A good header design would be something like a 4 into 1 design that uses castings or good thermal coatings as much as possible to keep heat in and would also get the gasses to the turbo as quickly as possible. The collector would have to be longer than the ones I have seen and the transitions nice and smooth. The main problem comes from trying to make it package well into the constraints of the turbo Subaru. A well designed header would likely require the header to up-pipe connection to have different flange points than factory, so it would not be as easy to sell in header and up-pipe pieces. The WRX is a hard car to design a proper header for, to say the least. It is very hard to improve what the factory has already done.
The up-pipes duty is to get the collected gasses from the header up to the turbo. The best size is the smallest that does not create excessive backpressure for the intended use. Again, the goal is to keep the gasses moving as quickly as possible while flowing enough gasses to make the desired power. There also needs to be a small amount of flex in the system to avoid cracking, warping, and blown out gaskets. The exhaust before the turbo has a lot of heat differential from one point to the next and adding to that is the fact that different metals have different expansion. This leads to a system that wants to twist, pull and push quite a bit. Without some give, something has to go. The gaskets and welds are usually the first victims. One problem lies in getting flex without having a flex section that is prone to cracking, splitting, and leaking it's self. It is not wise to cure a problem with a part that causes the exact same problem. That would be like sun screen that causes skin cancer.
Turbo Exhaust Housing
With turbos there even more factors than just the design of the exhaust side of the turbo that go into a good turbo for your application. However since we are talking about exhaust theory here I will only talk about the exhaust section of the turbo.
The size and design of the exhaust housing plays a major roll in the spool-up characteristics of the turbo and its ultimate power potential. There has to be a balance met if you want to have the quickest spooling turbo for your power goals. If you go with too large of an exhaust housing you greatly increase lag. Too small of exhaust housing and you severely limit the amount of boost and top end power you can make. You can only push so much gas volume through a small housing without having negative side effects. Adding to the complication is that each pound of boost created makes a ratio of backpressure before the turbo. It is different for each turbo, the amount of boost you are running, the size of the motor, RPM, and load on the motor. Once you start trying to push too much through the exhaust section of a turbo (running too much boost for the turbo) you start making a huge ratio of backpressure, and it only gets higher the more boost you run. This not only limits the amount of power you can make, but makes EGT go up, hinders the motor's ability to get the burnt gasses out of the motor, and makes the car more prone to detonation. This is also a big cause for failed pre-cats in the up-pipe. Choose too large of an exhaust housing for the application and it takes the turbo too long to spool, effecting torque production. The best way to make good torque on a turbo motor is to spool up the turbo as quickly as possible. Also, who cares how big your turbo is, or what power it can theoretically produce if you can never spool it up or if it falls out of the powerband every time you shift. I have heard of WRXs that theoretically make enough power to run in the 11's in the quarter mile actually run a 14 in real life because of mismatched parts. Bigger is not always better.
Adding another factor is the design of the exhaust wheel. It has to have good aerodynamic properties or it is inefficient. A more efficient wheel design means that you will make more power and/or less lag.
Post Turbo Exhaust
The main performance goal of a post turbo exhaust is to create the least amount of backpressure possible. There are a lot of factors that affect this.
Turbulence is one main factor. If the gasses are all stagnating and/or running into protrusions or running into each other it creates more backpressure than a well designed system. The more laminar (smooth and straight) the gas flow, the more the system can flow for a given pipe diameter. Steep angles and abrupt pipe diameter chances should be avoided.
The methods of collecting the outlet gasses and the wastegate gasses add another part of the equation to change. It would be optimal not to join the outlet from the turbo and the wastegate together, but the real world messes with our fun. Just dumping the wastegate to atmosphere is great for a racecar, but not a street car. So a street exhaust should combine them to get all of the gasses through the same cat and muffler system.
Some of the turbo outlet designs include: flanges with a simple pipe, bell mouths, divorced wastegate, and split bell mouths You also have castings and formed piping to choose from. Which one works best is also determined by quite a few different factors and how well they are designed and manufactured.
Flange w/Simple Pipe - The only advantages to this design are cost and simplicity. The pipe does not have to be formed and the flange is simple therefore reducing cost. The labor to weld the pipe to the flange is easy and therefore less costly as well. That is the main factor that make it desirable to the factory and why it is used on the stock exhaust. The wastegate gasses joining the turbo gasses right at the turbo outlet does create turbulence in the worst spot post turbo and reduces flow, thus not making it as desirable for performance as other designs.
BellMouth - This method is much closer to optimal for joining the gasses from the outlets. There is more room for them to join and if the transition is done properly it can flow very well into the main piping. It packages very well and does not have a lot of complexity, making for less to break. We have gotten the best results from this type of downpipe so far. Boost response has been the best out of the outlet designs we have tuned on, it is easy to put a wideband oxygen sensor bung into. We have also had the fewest problems with this design.
Split Bell Mouth - This design separates the gasses in the beginning of the turbo outlet and joins them at the rear of the bell mouth section. It works well and has some of the advantages of the bell mouth and some of the advantages of the divorced wastegate designs. The main deterrent for this is the cost and complexity of adding the splitter. I am a fan of keeping things as simple as possible while still making the product work well.
Divorced Wastegate - Keeping the gasses from the turbo outlet and wastegate separate until farther back in the system is an attempt to combine the advantages of not collecting the gasses and the real world. Combining them far back is closer to optimal than collecting them closer to the outlets. It is also critical to power production and spool-up to join the pipes smoothly and avoid turbulence. The disadvantages are that you add a lot of cost and complexity. You have big temperature differences on each pipe and that makes for a system that can crack. Putting in flex or expansion joints helps, but adds even further complexity and yet another part to fail. With all of the exhaust systems we have tuned with on the dyno we have seen that it is generally harder to bring boost on as quickly with these types of systems as compared to the bell mouth type systems. Perhaps it helps the wastegate function too well. Also, we have had a few situations where the splitter caused problems allowing the wastegate to function properly by not allowing it to open to its full extent, or even open at all. That caused either boost spiking, or no control over boost what so ever. Since the wastegate could not function the turbo ran as if it did even not have one, and the poor turbo just ran whatever boost it could make uncontrolled. The fix was not hard, but the least amount of stuff to go wrong the better. I know that I would not be happy having to pay for someone to install the exhaust only to have another place diagnose the problem, remove the exhaust, repair the part, and re-install the exhaust.
Cast Outlets - Castings have the advantage of keeping a lot of heat in the exhaust as well as freedom with design. You can basically make it almost any shape you want. The disadvantages are more weight and cost. Cast iron pieces can weigh a ton and that is a valid concern for many people. The casting form that the piece is made in is also very expensive and depending on complexity can range from a couple of thousand dollars to well up in the tens of thousands.
Formed Piping -Forming pipe has almost as much design freedom as a casting with less expense and less weight. The only disadvantage lies in if it is not done properly. Poor forming can look bad and effect flow by having creases and crimped spots. You can also get the piping too thin if you try to stretch the metal too far. If done improperly you can also make the metal brittle and it will usually happen where the metal is the thinnest.
Remember, you will only flow as well as the greatest restriction. If you have a poor cat or muffler design then it will choke the flow no matter how good the rest of the system is designed. Fortunately straight through mufflers and newer high flow cats flow very well. Having a cat is not only good for the environment, but we have seen very little power difference in levels in excess of over 350 h.p. Why be dirty when you can make just as much power while keeping tree hugging hippies happy? Also, a cat tends to quiet things down a little.
Pipe diameter does have an effect on flow rates as well, but again it is not the major factor in most cases. 2.5" may flow enough for 300-350 h.p. without being a restriction. 3" is usually capable of flowing 500-600 h.p. before becoming a restriction. This is assuming that you have designed the rest of the system up to par. There are also full 3.5" systems and those that start out at 4" and taper down. Unless you are making over 500-600 h.p. anything over 3" is a case of diminishing returns and in most cases has no advantage. There is more to gain going from 2.5" up to 3" than there is going from 3" to 3.5". A 3" system will not loose torque compared to a 2.5" system if designed properly. In fact if designed properly 3" may be capable of making better low end torque than 2.5". Again, since the way to make the most torque with a turbo exhaust is to get the turbo to spool-up as quickly as possible, it should be the main goal of the entire exhaust system and good flow after the turbo is one way to achieve it. We use 3" as we want our system to flow enough to be capable of coping with a customer's changing goals. Properly designed we can offer it to the big power crowd while still appeasing the low end torque club.
The only reason to reduce the size towards the end of the pipe is for packaging, cost, and noise reasons. Tapering the diameter does not make more power, torque, or bring on boost faster. However having smaller pipe towards the end has less effect that having smaller piping at the beginning. In other words a system that has 3" pipe for the majority, and necks down to 2.5" at the end will flow enough for more power than a complete 2.5" system. The further downstream you neck down the exhaust the better……..if you decide to neck it down.
Attracting unwanted attention and not hearing your stereo or you passenger would make for an exhaust system great for a racecar, but poor for the average Joe. I like hearing the exhaust myself, but there are times I want to listen to the radio or go on a date without screaming at my passenger. Law enforcement and your neighbors do not appreciate loud exhausts either, even if you do.
Designing a non-turbo exhaust system is quite a bit different, most noticeably in the header section. The primary goal of getting the exhaust out with the most velocity and with the least amount of backpressure is still the same, but that is about where the similarities end. The real world also steps in and throws in the same requirements like noise, environmental concerns, and packaging into the mix, which can also compromise power production.
The header has the greatest effect on the power band and ultimate power production of a non-turbo car. There are MANY factors that go into a properly designed header. One factor is the way you join the pipes together. The two possible configurations for a 4 cylinder are 4-2-1 and 4-1. Basically a 4-2-1 design joins two primaries together into a secondary pipe, and then joins the two secondaries together. A 4-1 design joins all four at the same time. Both have advantages, but the 4-1 design allows the gas pulses to interact in a way that makes the best torque.
Primary Pipe Diameter -Smaller diameters keep velocity higher with smaller exhaust volumes. The more exhaust you are trying to push out the larger the primaries need be. The volume of gasses that you need to flow depends on displacement, RPM, and load. The more displacement you have per cylinder the larger the primaries need to be. The same is true for RPM, the more RPM you will be turning, the more diameter you will need as you will be pushing out a lot of volume over time. Higher loads on the motor also create a higher volume of gasses. As with every other variable there is a balance to be kept. If you are not flowing enough gasses for the pipe diameter (pipes are too big) the gasses will loose their velocity If the gasses get too slow you loose torque, and if you go way to large you can even loose top end power as well. Get it right and you get the best of both worlds, good low end torque and good top end power.
Primary Pipe Length -This has a huge effect on the powerband. Generally longer primaries make better low end while shorter lengths move the powerband up in the RPM range. The length affects the powerband by timing when pressure waves reach the cylinder. To put it as simply as possible, the pressure wave comes out of the cylinder and travels down the primary pipe until it hits the collector. There it gets reflected back down the primary pipe as a negative wave. When it hits the cylinder it helps pull more exhaust gasses out of the cylinder and pull more air in to the cylinder. Since power is made by mixing air and fuel and then exploding it, more air and fuel make more power. This effect is known as scavenging and is one of the main goals of a well designed header. Equal length primaries help each exhaust pulse pull the one behind it. This helps create a suction in a sense. Instead of just relying on the pressure of the exhaust stroke of the motor to get the spent gasses out, the suction of the pulse in front of it helps pull it out. One factor some header designers forget when trying to design an equal length header for the Subaru is that the length of the exhaust port is effectively part of the header and needs to be accounted for. Complicating this is the fact that the exhaust ports on the Subaru are not the same lengths. Not accounting for this effects power production.
Collector Type -The collector merges all of the primary pipes together. There are designs ranging from cheap and simple to incredibly complex and costly. If you just joined the pipes in the simplest possible way you would have something that resembled the picture on the right.
The dead space in the middle of all of the pipes would cause a lot of turbulence and hinder flow. Eliminating the dead space is the main advantage of the merge collector. This is a more cost effective way to make the pipes join smoothly. Not quite as elegant as the merge collector, but still very good.
The bad daddy of all collectors is the merge collector. It is from Burns Stainless and is one of the finest collectors you can buy.
Collector Length -The length of the collector also plays a role in determining the powerband of the motor. Generally the longer the collector the more the powerband is shifted up. You also want enough length in the collector to smoothly join the gasses coming from the primary pipes. If the junction is too abrupt they do not interact very well causing turbulence, and again hindering flow. This is also another area of a lot of testing. The volume of the collector has a fairly big effect on the powerband of the motor.
Collector Width -The width of the collector helps control how well the exhaust pulses interact with each other. Make it too big and one pulse cannot help pull the next very well and the gasses can stagnate hurting flow. Make it too small and you hinder flow by causing too much backpressure. Yet another area to test.
Taper Angles - Basically you want the least amount of abrupt changes as possible. This mostly applies to the collector where it necks down to the diameter the exhaust will be. You do not want an abrupt angle as it will hinder flow.
The entries into the primary pipes from the head also have to be as close to the diameter of the exhaust ports as possible. This is so that you do not get yet another area for turbulence to get in the way of things. Protrusions into the gas flow should be avoided here most of all, as they have a much larger effect than in any other point in the system. According to many experts that do not play the marketing game, the stepped header designs are an attempt to cure other problems inherent in the design. The steps also add complexity and cost.
The lay-out of the car dictates a lot of how the header is made. The ports being on opposite sides of the Subaru engine do not make things easy when designing a header for our cars. Each change in length during testing requires almost making a new header on a Subaru boxer motor, thus the rather lengthy design process of our header. Getting lengths equal is definitely a big task given the packaging, and any variance within .5-1" is considered the mark of a top notch header designer.
We have actually tested cat and catless on non-turbo cars as well. If you use a well designed cat there is very little power to be gained by not having a cat. The cat is a place where abrupt angles make a huge difference. Since inside the cat you are making drastic changes going from the diameter of the pipe, into a large diameter area inside the cat, and back to the diameter of the pipe having abrupt angles can really slow things down. This is as true for turbo as it is for non-turbo cars. You also want the gasses spreading out to flow across the complete area of the catalyst bricks of the cat. If the gasses are too concentrated on one part of the cat you will not be able to flow to the full potential of the catalyst bricks. That is why you see the gentle angle at the beginning of the "good" cat rather than at the end of the less optimal "better" cat.
Designing a good cat-back is fairly simple compared to the header. Keeping velocity high is still the goal. Pipe that is too large will loose low end torque as the gas starts moving slower. Pipe that is too small will loose top end power. So again there is a balance to be reached. The same rule applies to keeping the piping smooth and using proper bending techniques. The muffler needs to be as free flowing as possible without being too loud. Besides that there is not a lot of complexity in a cat-back. It is definitely easier to design than a header.
I hope that you learned some new and useful information. My goal was not only basic design education, but to convey why we do some of the things we do. It is not meant to be a full guide, as a full detailed explanation of exhaust design would take a huge book to discuss completely. We do have the resources to design components that would be 100% optimal, but there has to be a balance reached with cost and complexity just as much as any other factors to consider. We try to keep a balance in all of our products.