Exhaust Physics

[clint78z]

Thought I would start a thread on the science of exhaust systems, and try an take away some of the mystery behind the subject. Basically when you look at an engine it is an air pump. Fluid dynamics and thermodynamics can be used to understand what is goin on.

 

First off I would like to make the statement of people using horsepower and torque as two different entities. HP is just a math number that is made up of torque at a certain rpm.

 

I will start with the combustion chamber and go from there, the exhaust gas hot (600f-1800f)and is waiting to get out of the chamber. It has lot's of volume the higher temp or engine load you get. As the cam opens the valve it has a certain speed as it exits through the port it as a pulse into a manifold. What we try to do is tune the pulses so they fall one behind the other. This creates a scavenging effect and actually sucks the gases out of the chambers. This tuning usually happens at certain load on the engine (a sweet spot if you will). It collects to a main pipe of given diameter 2-3" usually. The exhuast gases cool as the reach the muffler the velocity goes down.

 

Things to note, and exhaust system is a package and all components must be designed to work together to optimize performance for a given powerband.

 

Another common thing I hear is "Backpressure creates more torque" This is not directly true, and I will explain why. Backpressure due to friction loss is an evil thing, and will always steal the amount of power and engine can make. It is true that a 2.5" exhaust will usually create a bigger torque number at a lower rpm than a 3". It is not the backpressure which gives you the more torque it is the velocity of the gas at that rpm that gives you the bigger number.

 

So for designing an exhaust system you must take into account the powerband of the motor you want to use. The camshaft you are using, size of motor, head design. Fortunately the legwork has been done by most other people.

 

Things to note spend more time at the manifold side, making sure exhaust flow has a smooth path of flow. Since gas is so hot this is where major restrictions can happen. For example if you can use two 2.5" 45 elbow or two 3" 90degree elbows which one would have less restristion ??

 

The answer is the two 2.5" 45 elbow has 3 time less restriction than the two 3" 90's.

 

It might be neat if some of us took a typical 6 cylinder and a 350 chev and did some of the calculations on this.

[johnc]

For example if you can use two 2.5" 45 elbow or two 3" 90degree elbows which one would have less restristion ??

 

That's an odd comparison - two 45 degree bends to two 90 degree bends? The first is a total direction change of 90 degrees and the second is a total direction change of 180 degrees. Of course the second would be more restrictive then the first.

 

Also, having built a few racing exhausts, I can't see where an exhaust builder would be faced with that kind of a choice. You either need to change directions 90 degrees or 180 degrees. Normally you don't have an option of one or the other.

[Drax240z]

Er, you can have 2 45 degree or 2 90 degree bends and end up with a 0 degree direction change John.

[Tim240z]

My assumption was that the statement was intended to say: 2 x 45*, or 1 x 90*.......in which case I would chose the 2 x 45* with as long of a straight pup inbetween.

[Bartman]

I think the point he was making is that "the smooth path of flow" is more important than the size of the pipe. The 45 degree difference in direction is more restrictive than the difference in the diameter of the pipe.

[johnc]

Er, you can have 2 45 degree or 2 90 degree bends and end up with a 0 degree direction change John

 

That's "outside the exhaust" thinking, not exhaust flow thinking. In both examples the exhaust flow itself has made two direction changes. In the first, one 45 degree change and then another 45 degree change totaling 90 degrees. The second totals 180 degrees. Where the exhaust ends up in relation to the world outside the exhaust pipe is irrelevant to flow, system pressure, velocity, etc.

 

My assumption was that the statement was intended to say: 2 x 45*, or 1 x 90*.......in which case I would chose the 2 x 45* with as long of a straight pup inbetween.

 

Would have been my asusmption too but the pipe OD was added as another variable to the comparison that makes it less valuable.

 

It also assumes a lot regarding the volume of flow in the exhaust itself. A 1L stock Suzuki 3 cylinder engine will see absolutely no difference in performance using a 2.5" OD exhaust with two 45 degree bends vs. a 3" OD exhaust with two 90 degree bends.

 

Clint's basic premise is correct - reduce the total direction change in the exhaust system and you will increase the total flow capabilites. This is true to the point that an exhaust with a larger tube OD but greater direction change MIGHT not be as effective and a smaller OD tube with fewer direction changes. But the 3 times number quoted depends on so many other variables that it can't be taken as a trueism.

[Drax240z]

That's "outside the exhaust" thinking, not exhaust flow thinking. In both examples the exhaust flow itself has made two direction changes. In the first, one 45 degree change and then another 45 degree change totaling 90 degrees. The second totals 180 degrees. Where the exhaust ends up in relation to the world outside the exhaust pipe is irrelevant to flow, system pressure, velocity, etc.

 

Fair enough.

 

Clint, what are you basing your pressure loss calculations on for bends? I can see it in industrial piping, even crimp bends than your "3x" number can be accurrate. I can't see the same being true for 2 mandrel bends being compared however. Maybe I am wrong, but my gut tells me otherwise.

[Drax240z]

Another note on exhaust gas pulse theory. Clints basic explaination is true, but a real world model becomes increasingly more difficult to predict. Clint's above example is complete for a single cylinder engine, but once you start throwing in collectors and multiple pipes merging, you start to get pulse reversions through the non-pulsing cylinders as well. On top of that there is more than just 1 wave created off each pulse, each wave created can reflect a few times and each reflection acts as another pulse. This is a bit more commonly used in intake tuning, as the pulses travel a shorter distance, but significant gains can be made with 2nd pulse tuning on top of tuning for first pulse. Of course you can't have everything, so first pulse tuning is more important overall, but if you can make use of 2nd or 3rd pulses as well you are doing well. (and most of the time the frequencies of these additional pulses fall into place in your usuable RPM range)

 

Exhaust collector design and routing as I said becomes increasingly difficult with more tubes at the collector. the goal is to stack the pulses to have the first cylinder aid the flow of the second, and so on down the line. So knowing your firing order you can figure out what cylinder needs to collect where. (specifically on the more common 4-2-1 or 6-2-1 design) The seconday pipe length and diameter also plays a significant role.

 

Applied to most exhaust manufacturing this is beyond the scope most people need to reach. It is becoming increasingly important in new cars, and high RPM cars/motorbikes can really benefit greatly from this wave theory and tailoring their designs to maximize their results. In the end it is all about increasing efficiency at the rpm the engine is used at.

[clint78z]

Yes I did leave a little out, I did that so the main point I was trying to get across was not lost. Yes I was hoping to spark even more disscusion from all the talented fabricators, engineers, and tuners on this board. Thanks for the great discussion. It's already heating up.

 

I guess I will start in order here with johnc, I was talking of the flow restrictions of just the elbows not the entire system being 3 times. Sorry to muddy the water on that one. I used a fluids chart nomograph to roughly compare the two long radius elbows. The 2 1/2" 45 degree has an equivalent length of pipe of 3 feet of pipe, the 3" 90 degree has an equivalent length of 10 feet of pipe. More than likely there would mabey be a 0.5psi difference. I was thinking about on a 280zxt turbo, lot's of people want that 3" downpipe on their car and have to use two 3" 90 degreee bends. When more than likely the better performance would be gained by using two 45 degree 2 1/2"bends like ScottyGNZ used to make. You are wonderfully right in stating air hates to turn directions. My point was that just because you have a big fart pipe on the car doesn't make it fast, it has to be designed correctly. JohnC you are correct again that on 1 liter Suzuki, no difference would be made between the 3" and the 2 1/2" system. It simply does not flow enough air to make any change. It would have something sized to keep the velocity of gases moving out of the cylinders.

 

Drax yes I was talking about multi cylinder engines, the pulses from each cylinder can produce a scavenge effect. I didn't say predicting these principle were easy, thermodymics is a beast I deal with every day.

 

I know a fair bit about this stuff, but can be wrong about certain things and though it would be fun to discuss. Hey you never know when you can learn something. Too bad it's not as often as we forget things

[datsunan]

Maybe you read the thread posted earlier today about 2.5 and 3" down pipe, I have been curious about the best way to go as far a a new down pipe. I fabed my own exhaust to save money but did not replace my DP. Now I plan on doing that and am thinking about redoing the entire exhaust but I am not a designer or engineer of exhaust systems so I will most likly be building it myself again. It seems to me the money that would need to be spent on R and D for a custom exhaust that out flows an exhaust system that is well planned and built would not be worth it for a street/race track car. I have often wondered what power I might be missing out on using my exhaust but who does this kind of tuning and what does it cost? Seems to me if your talking about the manifold the tuning would be very important dew to the heat and velocity but on a turbo charge engine the pulses are all blended by the turbine. So after the tubine I would think it all comes down to, the restiction of flow = less hp and there would not be much need for tuning the rest of the pipe. Yes or no?

[Guest tony78_280z]

Interesting discussion. I was just thinking about this subject because someone had mentioned that they do not think my exhuast has enough back pressure.

clint78z says

It is not the backpressure which gives you the more torque it is the velocity of the gas at that rpm that gives you the bigger number.

According to this statement, I should now worry about back pressure but scavenging. And this brings me to my question which no one has touched yet. How does one design a good scavanging exhaust system? Or is this designing know how beyond the average fabricator/builder (like me). And, if one does not go with a prefab set up is he just gambling on his results?

[DaleMX]

Ok, Question. Does the pulse travel at a faster speed in a smaller pipe? Say for design purposes @ 3000 rpm you need a 30" 1 ¾" pipe to scavenge. So does it seem logical to assume that a 1 ½" pipe would need to be shorter?

 

[Guest Aaron]

Here is a link Grumpyvette once posted on tri-Y headers. http://www.thorleyheaders.com/tech_basics_triy.html I think he posted more, but I do not have them in my favorites. I will see what else I can find.

 

Ok, found this thread. Header Math

[clint78z]

Hey Drax thanks for the tidbit on second and third pulse, it makes sense that would happen but something I have never thought about.

 

So how can a person go about designing their own exhaust system without doing a ton of hit and miss calculations ??

 

Keeping a smooth path, smooth welds reduce turbulence and resonance. Try to avoid transitioning from different pipe sizes. The exit on the pipe should be a bellmouth to avoid exit loss. Mandrel bent is a good option, if your going to make a turn the less the angle and longer the radius the better. Crush bent is definately out, it's noisier and robs power.

[Drax240z]

Also if you have to change diameters, take a look at getting a "mandrel bent" reducer and use as shallow an angle as you can. (ie: stretch the size transition over a large length, rather than having it abruptly change diameter)

 

http://www.burnsstainless.com sells the reducers I am talking about, I am sure they are available at other places as well. Burns collectors are VERY VERY nice quality as is all their stuff in my experience. (when only the best will do!)

[datsunan]

What about this as an option for the tight fit spots? Would these flow as good as a round 3"? They aint cheep but then speed costs!

http://www.burnsstainless.com/OvalTubing/ovaltubing.html

[Drax240z]

It's a great option for increased ground clearance, just pricy. They'll flow very well.

[Nismo280zEd]

Hmm... so what would happen if you threw in a one way valve in the exhaust, this woudl eliminate reverse scavenging, but create tremendous initial backpressure I would think.

-Ed

[Guest aryanramsey]

my mechanic recently told me that he could install dual side by side exhaust for my car a hell of alot cheaper than dual exhaust out the back, what do you guys think?

[proxlamus©]

I have a quick question...

 

As air compresses it generates heat.. hence why turbo cars *should* have an intercooler.

 

Now as air expands.. it cools. technically.. as the exhuast gas exits the engine... wouldn't it be more efficient to go from let's say a 2.25" to a gradual 3" piping in a relatively small distance? The exhuast gases would "expand" as the piping got larger, resulting in cooling, which means a faster flow??

 

I could be wrong.. but a crazy assumption?!