41 replies to this topic

[JollyRoger]

Anybody care to explain the way an exhaust system aids flow, alters power/torque etc? Or know of a good site for that has this stuff?

Also curious how can having multiple pipes going into a one that is much smaller than the sum of the others not ****** flow?

Any takers?

[bobdar]

I'm certainly no expert, but will try to give you the info as I know it. First of all, the cylinders aren't all exhausting simultaneously, so the secondary does not have to equal the sum of the primaries. As a slug of exhaust gas travels down its primary to the collector, you want to maintain an optimum velocity to assist the scavenging of exhaust gases from that cylinder. When that "slug" enters the collector, its velocity will generate a negative pressure pulse in the other primaries. In a certain RPM range, and with the correct primary length, that negative pressure pulse can assist in "pulling" the exhaust from the next cylinder. As you can see, timing is everything in this situation, and becomes more critical as the RPM increases.

It's really a science of wave theory, pressure pulses and velocities, and undoubtedly requires a lot of dyno work to get real results.

[Jezztor]

Just skimming through that last post, I would consider it accurate. Well said.

Exhaust gas is obviously very hot. In basic principles, high temperature = low pressure. So if you run your exhaust gas over a wing like in F1, it assists both in forward 'propulsion' (i hate that word) and helps with a higher change in pressure through and under the wing elements. Like McLaren tried last year, this can also be applied in the diffuser. (they had exhaust outlets coming out the diffuser plane). The above is extremely basic, there are many variables and exceptions.

With regard to your multiple pipes question, although exhaust gas is not like normal 'air', it is compressible. So basically the multiple pipes 'squeeze' air into the final tail piece and the air is also accelerated. You can get the idea of this by putting your finger on the front of a syringe tube, squeezing the air until uncompressible, then remove your finger. Something like that.

Jezz

[MRC]

Bobdar's explanation of the all cylinders not exhausting simulatneously, seems spot on. The flow is anything but steady state, and to assume steady state flow is a bad assumption.

Quote: "It's really a science of wave theory, pressure pulses and velocities, and undoubtedly requires a lot of dyno work to get real results." (bobdar)

I don't think you can say it any better than that. A very good engine simulation package might also get you good results, but it still needs to be compared to some baseline dyno work.

One thing though, as a pressure wave hits the collector, there will be a pressure wave traveling up the adjacent pipe(s), not a rarefration. A rarefraction wave will be reflected up the pipe that the original pressure wave came from. I also am not getting how the high temperature exhaust is compressible, yet air in the intake, is not. Also, if the exhaust is accelerated in the tailpipe, would this not mean a drop in average pressure (in the tailpipe)?

The subject of how exhaust system variables and options will alter the torque curve, of a single cylinder engine, could fill a volume of books, let alone multiple pipe manifolds. My biased opinion is stay away from the web sites as much as possible, and go visit your local library instead. Better yet, go to a university library. May I also recommend you check out books by Gordon Blair, and "Design Techniques for Engine Manifolds" by DE Winterborne & RJ Pearson for some good information on wave theory and the like. SAE papers are also excellent sources.

[12.9:1]

look here for "tech articles"

[Jezztor]

Intake air is compressible - turbo's work by compressing air and 'injecting' it. I am no expert on this, but I don't think there is a pressure loss within the pipes if the air is accelerated. The reason for this is, it is not expelled in regular periods, it is expelled all the time. The gas particles being squeezed together upon ajoining of the pipes, are also squeezed out the exhaust, to cause more pressure in the tail end of the pipes. This is what expells the air out.

someone please correct me if I am wrong. I'm going to bed :

Jezz

[MRC]

Like I said before, you simply can not assume anything close to steady state flow, even in the tailpipe. If the flow velocity is increased in the tailpipe, there will be a pressure drop. More likely the pressure will increase, partly due to a decrease in gas speed. Of course the stagnation pressure will always drop along the length of the pipe. While a larger tailpipe, will tend to decrease the backpressure, this effect falls off, long before it's area reaches the area of all the pipes joining together at the collector. Having the same size tailpipe, and changing the collector volume, will also have an effect on the backpressure.

[DOHC]

Originally posted by MRC
I also am not getting how the high temperature exhaust is compressible, yet air in the intake, is not. Also, if the exhaust is accelerated in the tailpipe, would this not mean a drop in average pressure (in the tailpipe)?

I have no experience with exhaust pipes, but from the point of view of physics, temperature can't be neglected in the exhaust pipes. After combustion, the gas is very hot and at high pressure, as the valve opens it further expands and loses temperature traveling down the pipe, which I would think should be compatible with losing pressure, according to the gas law. (This, assuming of course that no further combustion takes place in the pipe; sometimes a questionable assumption I think.) At the open end of the exhaust pipe the pressure is atmospheric, so the exhaust gas should lose pressure. Or is this thought fundementally wrong somewhere?

As the noise is different with different exhaust pipes, there must be acoustics involved, so compressibility certainly must play a role.

Again, I'm absolutely not an expert on exhaust, so please enlighten me if I'm off in these speculations.

[MN]

Originally posted by Jezztor ....... Exhaust gas is obviously very hot. ......

Indeed.
'83 Sprit-Honda.

[MRC]

DOHC, you are completely right that you have to take the high temperature of the exhuast into account. I don't think I said anything counter to this. With exhaust you have flow with wave propogation included right? Take a single cylinder. Imagine that the pipe has a step increase in pipe diameter somewhere along it's length. The average pressure will increase immediately after the step. Any pressure wave will be partially reflected as a rarefraction, with the main pressure wave still travelling the same direction. This was my point. Then you have heat transfer all on top of this. Not to mention friction, which will reduce actual and stagnation pressure, and accelerate the gas.

Let me ask you this, if you have a pipe connected to a tank with a very high pressure, would you assume that the pressure at the end of the pipe is that of atmospheric.

I can't say that I exactly know what is all happening in the collector and/or the tailpipe, but I highly doubt anyone could give a good synopsis in a page of text.

[desmo]

Might the compressability of the exhaust be influenced by it's velocity? I understand that exhaust port velocities are designed to be locally supersonic.

[Jezztor]

Read this .

Jezz.

[AS110]

Originally posted by Jezztor
Read this .

Jezz.

This is more like it!

People always get confused with exhaust flow - we have the physical exhaust gas (the smoke) and a sonic wave (the noise) one travels at about 350ft/sec and the other at about the speed of sound.This artical explains it well.

Think of the physical exhaust gases as a river of water,a very wide river has a slow current,but as we narrow the banks,the water speed increases,until the water can't all get through and a lake forms.We have to have a flow between these extremes.Too big a dia pipe slows gas speed,too narrow is resrictive.

GAS FLOW is controled by diameter.

SONIC PRESSURE WAVE is controled by length.

[DOHC]

Jezz -- nice article

[Jezztor]

AS110 - Precicely

[DOHC]

AS110, thanks for putting a few things straight. Yes, the distinction that has to be made is between gas transport (subsonic) and acoustics (sonic waves) in the exhaust header.

Desmo, I doubt that there's much supersonic going on in a combustion engine, neither supersonic gas transport nor supersonic flame fronts. Plenty of turbulence though. Combustion engines were often in the old days incorrectly called "explosion motors"; this is inocrrect because at a detonation the flame front propagates at supersonic speed; in combustion engines it is at best a deflagration taking place. In a deflagration the front or interface between the area where chemical reactions are taking place and the area where the charge has yet to ignite moves slower than the local speed of sound. It is by no means slow, but there's a considerable gap between sonic speeds (>300 m/s) and typical piston speeds (25 m/s) in high-revving engines. (We often think of the pistons as moving very fast, but it's not true, really---the complete car is much faster!)

The reason for compressibility in the exhaust, while compressibility is not an issue in the intake header, is because of the very high pressures built by the exothermal chemical reaction in a confined volume. This high pressure is causing acoustic waves, hence noise. If you think if it, it's rather clear, crank the engine without switching the ignition on, and all you hear is mechanical noise and hissing sounds from gas transport. When the ignition is on, though, all hell breaks loose, as compressibility and acoustics come into play.

MRC, yes, I'd say that at the open end of the pipe the pressure is "atmospheric". That's not entirely true of course in this case, and again I think you have to make a distinction between gas transport and acoustic phenomena. If you only have acoustics (cf. an organ pipe, where air transport is negligible) the pressure is atmospheric at the open end. If you have significant gas transport with acoustics, as is the case with exhausts, it's most likely not a very accurate approximation. But the physical domain outside the pipe where the pressure deviates from atmospheric is rather small. That's inside the flames in that picture above.

Really interesting topic, this!

[H. Eckener]

DOHC,

I don't think the relationship between whether or not a gas is compressible and the NVH of the intake or exhuast system holds up as you describe it.

Cause on my hopped up 350, 0.030" over, when I take the air cleaner/snorkel off the intake and drive around town it makes more noise then if I had the air cleaner on the engine. Hell I got me a fancy high breathen tricked out dual exhust with twin glasspacks, and that them sucking sound from the engine is still noiser then the exhuast when I give it the gas.

Me thinks those Chevy engineers designed the 350's induction system with compressible effects in mind, because my intake sure can be loud. I don't know about the Ford boys, I ain't never heard of a Ford as loud as my Chevy.

[DOHC]

Mr Eckener, sounds like your Chevy makes a whole lot of noise! To be honest, I have no idea how that suction noise originates, but suction can be loud too without compressibility (e.g. vacuum cleaners). Maybe turbulence plays a role? I would be very surprised though if one would reach air velocities in the intake header in excess of 100 m/s. Does anyone know?

[testarosa]

A large part of intake noise (at least, in my experience) comes from the valve overlap period. I don't think it comes from compressibility, because the intake system should be lower pressure than atmospheric, except when resonant.

DOHC- the movement of the valves causes a very powerful resonance in the exhaust and intake system, so even when cranking the engine with no fuel, there would still be resonance at the appropriate rpm.

[juanjo24]

not super-technical but still intersting

http://www.insiderac....com/drgas.html

[12.9:1]

Am I mising somthing, perhaps some protocol -- or ?
When I posted the BURNS STAINLESS link ......................ok nothing, that's ok.
Later Jezztor posts the BURNS STANLESS link and

_______________________

This is more like it!
_______________________

Jezz -- nice article
________________________________

and indirectly
_______________________________

thanks for putting a few things straight
_______________________________



I suppose it's nothing ?

Incidentaly good stuff on Titanium, Inconel applications, and fabrication info on this site.
And yes Desmo -- racing engines tuned for a very broad power band will, almost by definition-- { if properly designed} Go Supersonic at the EX valve - @ WOT/maxBMP

[DOHC]

Originally posted by 12.9:1
racing engines tuned for a very broad power band will, almost by definition-- { if properly designed} Go Supersonic at the EX valve

So there are shock waves in the exhaust header?

The article mentions average gas speeds of 350 ft/s or approx 100 m/s. That's fast but far from supersonic. Locally speeds could of course be considerably higher, and the highest is clearly at the EX valve, given the geometry and the state of the gas. Is there a reference to supersonics in this application, in particular shock waves?

[12.9:1]

Only at the EX valve.

[12.9:1]

I should say "standing shock wave" --- at the EX valve.

[desmo]

I put up that link a year or two ago! Where's my props?

I frankly don't understand why, but a source I cannot quote said that local supersonic flow velocities around the EX valve during blowdown could be exploited for improved VE (Volumetric Efficiency). Any thoughts on how this might work?

[12.9:1]

Desmo, This notion is contrary to all I know, But would be power-full if true.
I'll ask a round.

[AS110]

Originally posted by 12.9:1
Am I mising somthing, perhaps some protocol -- or ?
When I posted the BURNS STAINLESS link ......................ok nothing, that's ok.
Later Jezztor posts the BURNS STANLESS link and

_______________________

This is more like it!
_______________________

Jezz -- nice article
________________________________

and indirectly
_______________________________

thanks for putting a few things straight
_______________________________



I suppose it's nothing ?

Incidentaly good stuff on Titanium, Inconel applications, and fabrication info on this site.
And yes Desmo -- racing engines tuned for a very broad power band will, almost by definition-- { if properly designed} Go Supersonic at the EX valve - @ WOT/maxBMP


regards, Eric

Sorry 12.9:1,but I was responding to the latest post,and Jezztor was linked right to the important page.

[12.9:1]

Quite all right AS110, and by the way - "It was hell" - One of My all time favorits

[JollyRoger]

Thanks guys, brilliant stuff.

One other question that comes to mind. Is clearing either part - particle or resonance more important than the other?

[12.9:1]

Posted by Desmo
______________________________________________________________________________

I frankly don't understand why, but a source I cannot quote said that local supersonic flow velocities around the EX valve during blowdown could be exploited for improved VE (Volumetric Efficiency). Any thoughts on how this might work?
______________________________________________________________________________


I've asked around and, I'll admit My question was "have you heard any crakpot theorys regarding
useful supersonics".
No one had, all agreed - supersonic flow in a pipe leads only to blockage.

perhaps your source was referring to " near supersonic " or " transonic " ?

[desmo]

Could be, I am out of my depth here. Nothing in open literature I've read suggests that supersonic velocities in exhaust flow would be desirable.

"racing engines tuned for a very broad power band will, almost by definition-- { if properly designed} Go Supersonic at the EX valve - @ WOT/maxBMP"

By this do you mean that in a properly designed race engine that local exhaust velocities should be tuned so that they reach or exceed Mach 1 under those conditions? Also, how accurately is it possible to experimentally measure or calculate instantaneous local velocities in the intermittant flow and extremely hostile environment found there? I am wondering how good a picture of what is actually transpiring as the EX valves crack open and blowdown occurs we really have.

[DOHC]

Originally posted by 12.9:1
all agreed - supersonic flow in a pipe leads only to blockage.

Happy to hear it, that's what I thought too, even if my point of reference here is only physics and not knowledge of exhaust pipes. Supersonic channel flow usually only gets you into trouble. Rocket engines do have partly supersonic flow, but then the exhaust opening is a very carefully designed bell-shape and not a pipe. And my guess is that this is what the exhaust ports of an engine would also look like if the flow were supersonic---I don't think you would dream of putting a pipe and collector in the way of such flow, unless we're talking of mufflers for everyday cars.

I believe that you could have transonic flow around the edge of the EX valve, but transonic flow is a different animal. That would indeed lead to the (presumably small and local) "standing shock wave" you mention.

I've read somewhere in physics literature that measurements in combustion engines are done using laser light, but IIRC it was a very complicated technology. Light is otherwise a possible way to locate shock waves due to the change of refraction index across a shock. So this type of measurements can be done, but I guess only in the lab and with prepared engines.

PS You can see shocks in transonic flow with your naked eye, quite fun! Commercial jetliners operate in the transonic regime, and then there's a standing shock wave on the upper wing surface located at approx 70% of the wing chord. If you have a window seat around there, and the sunlight shines at a certain angle, you can see the shock line quite clearly as a thin, light shadow along the wing surface. Don't expect to find it on your first try though, you really have to know what your looking for...

[12.9:1]

APOLOGY to desmo -----

Yes exactly so, one would know wen either port has gone super by the drastic drop in performance.
Your confusion {in this} is completely my fault, I meant to say, yet some how did not -
that the port should be "on the verge of going supersonic"
A substantial slip - sorry


The idea is - to broaden the power band one must sacrifice some peak torque !
it's commonly assumed that peak gas velocitys occur at peak rpm.
However it is at peak torque or bmp-max that the greatest amount of air/fuel is ingested
per intake cycle.


Peak torque is but a zone in the rpm range of every engine, where maximum synergy
occurs. Moving away from this zone - higher or lower rpm, will degrade the tuning of the
engine{musical instrument is a very good description}. Going to higher rpm's has the
advantage of the multiplying effect of rpm's i.e.- horsepower, so while instantaneous
torque may be falling horsepower can still rise. Going to lower speeds however is always
towards lower torque and horsepower. So to stretch the power band it's the lower end
that needs the most help.


And the best tonic for a lazy low end is high gas velocity, i. e. smaller crossection inlet and
exhaust, this will improve both filling and combustion efficiency, The cost - at peak torque,
under sized valves/ports/pipes, leading to supersonic blockage. but if this compromise is
skillfully executed the natural degradation of the tuning {on rising rpm} will arrive just as
the ports/valves are about to go super !

[DOHC]

Enlightening!

[desmo]

Thanks for clarifying, that's what I took it to mean. I did a Google search to see if I could turn up any references to the topic of trans/supersonic exhaust velocities, and found this, I believe, dealing with Pro Stock drag engines:

"Port Support -Exhaust Ports

The exhaust port is a proper place to start. For years we’ve tweaked exhaust seat configurations and used theoretical radiuses to create ports that flow well without turbulence. In the late '70's we found that if you decreased the cross section of the exhaust port to the point where the flow exceeded mach 1, the amount of work necessary to properly scavenge the cylinder was almost nil. By placing the port's smallest cross section well below the seat insert, the flows maximum velocity was now in an area where we could control the sonic shock and use it’s placement to allow the port to literally "pull" the flow out. On the flow bench, these ports tend to become extremely turbulent and noisy until approximately 7% net lift is obtained and then the port smoothes out, making a smooth "scream" that certainly requires ear protection. The early "noise" was created when the flow was in the trans-sonic range, and once past mach 1, it’s a screamer. To relate to typical exhaust port operation on a flow bench, one normally needs to adjust the pressure drop with each incremental valve opening to maintain the correct test pressure. With the sonic configuration, once port velocity is past mach 1, little or no additional adjustment to the pressure drop is necessary and the flow rate will continue to increase dramatically with each sequential valve opening. Since we’re not using additional pressure to increase flow as the valve opens, we have, not only a very efficient port, but also a port that acts as a vacuum, literally sucking the exhaust out, and eliminating the pumping losses normally incurred. An additional benefit is that since the point of maximum velocity is "down in" the port, the seat area is no longer sensitive at all, and simple three angle seats now flow as well as the radiused seats previously necessary. Torque is also greatly enhanced by the small area ports, and the shapes and areas used to control and place the shock wave do not allow the ports to flow well backwards at all, so there is little inert gas left to contaminate the fresh intake mixture. "

This from the article here.

Anyone care to run this through their bullshit detector? It does sound a bit dodgy given what I've read, but I am emphatically not qualified to pass any judgement on it.

[H. Eckener]

DOHC,

I was being facetious. How do you define when a gas is compressible and when it is non-compressible, perhaps that is where you are confusing me?

[MRC]

Gas flow speeds are influenced by the diameter, the length (total length and distance along pipe), and the friction factor, not just the diameter. Although the diameter is the largest factor, of gas speed.

Length will have an effect on the traveling of a WAVE along with the diameter.

Air and exhaust are both compressible. There are compressible effects in the exhaust and intake, even under motored conditions. Even if there were not compressible EFFECTS going on in the exhaust and intake, that does NOT make the exhaust or intake gases, non-compressible. When people speak of compressible air flow, they talk about considering air being compressible above about mach 0.3. That means that they can neglect the compressible EFFECTS below that speed, that is not to say that the gas is not compressible, nor to say that there are no compressible effects below that speed. With an intake and exhaust system, this will not be a good assumption below mach 0.3, due to the waves, and the waves often being decidely non-linear. So, velocity might determine when you should take compressibilty effects into account, for any calculations you might be doing, but it certainly will not negate them. Any half ass or better engine simulation code, will take into compressibilty effects.

I agree the particle speed will be based on the both the local sound speed and the transport velocity. However I disagree with the Burn's article that the sound speed will be based just on temperature (in the exhaust). The local sound speed will be based on the gas temperature and to a smaller degree any change in pressure (ie- slope of pressure rise). This can lead to a traveling shock in the exhaust. A pessure wave in the exhaust will tend to steepen. This doesn't mean it will form a shock, but will tend towards it. An expansion wave will tend to spread; opposite of the compression wave. This is described by the Earnshaw equations if you would like to research this.

Can the gases go supersonic after the exhaust valve? Like DOHC referred to, it would required that the proper area ratio along the length of the duct, to be followed. Changes in pressure ratio, and the likely improper area ratios, would lead me to think that if you did manage to get the flow to go supersonic, that it would likely not last long, and would just evetually shock somewhere farther down the pipe, obvioulsy giving you a sub-sonic speed after the shock.


quote from desmo's article:
"To relate to typical exhaust port operation on a flow bench, one normally needs to adjust the pressure drop with each incremental valve opening to maintain the correct test pressure. With the sonic configuration, once port velocity is past mach 1, little or no additional adjustment to the pressure drop is necessary and the flow rate will continue to increase dramatically with each sequential valve opening. "

Once the port velocity is at mach 1 anywhere, the mass flow is constrained, for a given front side pressure (ie- the pressure in the cylinder). If the back pressure is decreased (ie -higher pressure ratio) with mach 1 occuring in the port somewhere, you simply will not get an increase in mass flow. An increase in the pressure the cylinder side will give an increase in density, and will give an increase in mass flow. So an increase in pressure ratio after mach 1 is reached, via increasing the cylinder pressure, will give an increase in mass flow.

So I would have to say that I would be interested in seeing what these guys' test setup looks like and how they are correcting their volumetric flow readings. Right now, this is registering about a 7 or an 8 out of 10, on my bullshit-o-meter. I guess if you really want to get into semantics, you could see an increase in volumetric flow rate, but mass flow is what I care about.

As for supersonic flow, I can't see this helping you out in any way.

[12.9:1]

Desmo Asks
_____________________________________________________________________________
Anyone care to run this through their bullshit detector? It does sound a bit dodgy given what I've read, but I am emphatically not qualified to pass any judgement on it.
______________________________________________________________________________


Two points 1. A whistling sound or even a "screaming" sound are just noise ! - oscillating
pressure waves, and in no way constitute a scientific measuremet. Besides whistling ports,
{at high flow rates} is rather the norm.

2. Ever-since Hotroders in the 50's reasoned that biger is better, grossly oversized ports
have been the norm for virtually all stock block racers.
Only in the 90's had the revelation of high gas velocity, become more or less common knowledge.
The exhaust, resisted resizing for years after intakes had started to "scream",partially because
many as cast ports were already big enuf {but couldn't be left like that !} some actually to big
for racing. The old boys just couldn't believe, that for all those years their exhausts often had
double the area requird.

SO it may be that, wen one indipendetly "discovers" proper sizing, it may well seem that something
magical is happening i.e. supersonic assist.

[DOHC]

Originally posted by H Eckener
How do you define when a gas is compressible and when it is non-compressible, perhaps that is where you are confusing me?

Sorry for causing that confusion. MRC has the answer right above.

[AS110]

Originally posted by 12.9:1
Desmo Asks
_____________________________________________________________________________
Anyone care to run this through their bullshit detector? It does sound a bit dodgy given what I've read, but I am emphatically not qualified to pass any judgement on it.
______________________________________________________________________________


Two points 1. A whistling sound or even a "screaming" sound are just noise ! - oscillating
pressure waves, and in no way constitute a scientific measuremet. Besides whistling ports,
{at high flow rates} is rather the norm.

2. Ever-since Hotroders in the 50's reasoned that biger is better, grossly oversized ports
have been the norm for virtually all stock block racers.
Only in the 90's had the revelation of high gas velocity, become more or less common knowledge.
The exhaust, resisted resizing for years after intakes had started to "scream",partially because
many as cast ports were already big enuf {but couldn't be left like that !} some actually to big
for racing. The old boys just couldn't believe, that for all those years their exhausts often had
double the area requird.

SO it may be that, wen one indipendetly "discovers" proper sizing, it may well seem that something
magical is happening i.e. supersonic assist.

An easy visual history of this is motorcycle front pipe diameters over the years.Pre war bikes had huge pipes,and the 50's Triumph twins had fat pipes.But in the sixties,after much tuning work(mainly flatrack racing) it was found Triumph valves and ports were TOO large.The later bikes had narrow dia front pipes,and you can see the step down from the original size at the cyl head.They speeded up gas flow with smaller pipes.

[Christiaan]

Originally posted by desmo
Might the compressability of the exhaust be influenced by it's velocity? I understand that exhaust port velocities are designed to be locally supersonic.

Yes it does. There is a link between sonic velocities, bulk modulus (which relates to compressibility) and density. Its something I can model in Control Theory but not quite explain physically.

[desmo]

"When the exhaust valves open (approx. 95 deg ATDC)
the in-cylinder pressure is above 10bar, sonic speed
occurs inside the seat gap for about 50 deg CRA at
rated speed. Therefore, the development work for
exhaust ports is performed under sub-sonic and supersonic
conditions."

Mario Ilien- SAE Paper 2002-01-3359