Backpressure and low end hp.
07-30-2004, 05:37 PM
#1
Nordschleife Master
Thread Starter
Backpressure and low end hp.
I've been told several times that backpressure is good for low end torque. (which you multiply by the rpms, to get hp.)
That makes no sence to me, and apperently, no sence to anyone else eighter. However, I haven't read a really good explanation of what actualy happens. Install a big bore exauste, low end torque goes down, but high end torque goes up.
Well, I got to looking into it doing fluid dynamic calculations to try and figure out why the heck reality is like this. I have a proposed theory, which I can't find enough real world information to prove conclusively, but have enough theory so as to belive it. There is a way to prove it, but I don't have the reasources to do so.
One of the most counterintuative phenominon of fluid flow is the diffrence between turrbulant flow, and laminar flow. You'd think that tubulant flow would be more restrictive, but the truth is the oppsite in internal flows. Turbulant flow disrupts the shearing of the fluid in question, sets up a diffrent style boundry layer, and reduces the restriction. (strange uh?) Therefor, for best exaust flow, we want a turbulant flow.
A key decideing factor for if a flow is turbulant for laminar is the Renold's number. Turbulant flow happens at high Renolds numbers, laminer flow at low. Renold's number is a dimitionless facotor that is determined from flow stream velosity(v), pipe dia(D), fluid viscosity(u), and fluid density(r).
Re = (v*D*r)/u
And since, in our case, the velosity of the fluid is a fuction of the diamiter of the pipe, it gets even similer, to somthing like this: (c number of other factors, like pie, and constant numbers. Q is the volumetric flow rate.)
Re = (c*Q*r)/(D*u)
Here you can see, the smaller the diamiter of the exauste pipe, the higher the Renold's number. Thus we get into turbulant flow at a lower flow rate.
OK, in an exaust system, a small bore exuaste pipe will get into turbulant flow at a lower flow rate than a big bore system. The flow rate through the exaust, at WOT, is just about dirictly related to the rpm of the engine. If the exaust pipe is small enough, it may be in turbulant flow from the instant the the throtal is opened up.
This should be true for the entire exaust system. The biggest problem is with the baffals, of which there are so many varations, I didn't even bother to worry about them.
Thus, from working work with the above theory, it would seem that installing a big bore exaust system, and then a varrable backpressure muffler at the end, your gaining nothing from a proformance stand point. With a varable back pressure muffler, you can't reduce the size of the exaust pipeing, or even as I understand it, the internal dimintions of muffler. Therefor, it won't gain you back that low end torque.
The correct solution would be to design a system that had a varable size exaust plumbing. Maybe a system that would work with dual pipes at high rpms, and a single pipe at low rpms, allowing the exaust to be reoptamized at will with a valve.
Thoughts?
That makes no sence to me, and apperently, no sence to anyone else eighter. However, I haven't read a really good explanation of what actualy happens. Install a big bore exauste, low end torque goes down, but high end torque goes up.
Well, I got to looking into it doing fluid dynamic calculations to try and figure out why the heck reality is like this. I have a proposed theory, which I can't find enough real world information to prove conclusively, but have enough theory so as to belive it. There is a way to prove it, but I don't have the reasources to do so.
One of the most counterintuative phenominon of fluid flow is the diffrence between turrbulant flow, and laminar flow. You'd think that tubulant flow would be more restrictive, but the truth is the oppsite in internal flows. Turbulant flow disrupts the shearing of the fluid in question, sets up a diffrent style boundry layer, and reduces the restriction. (strange uh?) Therefor, for best exaust flow, we want a turbulant flow.
A key decideing factor for if a flow is turbulant for laminar is the Renold's number. Turbulant flow happens at high Renolds numbers, laminer flow at low. Renold's number is a dimitionless facotor that is determined from flow stream velosity(v), pipe dia(D), fluid viscosity(u), and fluid density(r).
Re = (v*D*r)/u
And since, in our case, the velosity of the fluid is a fuction of the diamiter of the pipe, it gets even similer, to somthing like this: (c number of other factors, like pie, and constant numbers. Q is the volumetric flow rate.)
Re = (c*Q*r)/(D*u)
Here you can see, the smaller the diamiter of the exauste pipe, the higher the Renold's number. Thus we get into turbulant flow at a lower flow rate.
OK, in an exaust system, a small bore exuaste pipe will get into turbulant flow at a lower flow rate than a big bore system. The flow rate through the exaust, at WOT, is just about dirictly related to the rpm of the engine. If the exaust pipe is small enough, it may be in turbulant flow from the instant the the throtal is opened up.
This should be true for the entire exaust system. The biggest problem is with the baffals, of which there are so many varations, I didn't even bother to worry about them.
Thus, from working work with the above theory, it would seem that installing a big bore exaust system, and then a varrable backpressure muffler at the end, your gaining nothing from a proformance stand point. With a varable back pressure muffler, you can't reduce the size of the exaust pipeing, or even as I understand it, the internal dimintions of muffler. Therefor, it won't gain you back that low end torque.
The correct solution would be to design a system that had a varable size exaust plumbing. Maybe a system that would work with dual pipes at high rpms, and a single pipe at low rpms, allowing the exaust to be reoptamized at will with a valve.
Thoughts?
07-30-2004, 06:14 PM
#2
928 Collector
Rennlist Member
Rennlist Member
If you have 8 cylinders pumping exhaust gases into the same tube, the exhaust gases flow at a particular rate. If you split those cylinders' outlets into two groups, you have less synergical flow, which means when one fires and the other closes, the rhythm of the gas flow is not maintained. So, you need scavenging, ie when one closes, another is there to backfill the direction of exit flow, so that the rhythm of exit flow is maintained.
If you raise the exhaust pipe diameter above a point where the gases flow in a linear direction, this will result in turbulence. And this causes flow to be slower than in a smaller pipe.
None of this has anything to do with backpressure.
If you raise the exhaust pipe diameter above a point where the gases flow in a linear direction, this will result in turbulence. And this causes flow to be slower than in a smaller pipe.
None of this has anything to do with backpressure.
07-30-2004, 08:01 PM
#3
Rennlist Member
Four explosions puking into a series of pipes every revolution is a very complex thing to model: The Formula 1 guys likely have it down better than most, now dealing with 10 cyls @19K RPM.
The first issue to contend with is the point(s) Heinrich made. A resonant phenomenon with multiple harmonics is established as in acoustic theory: Scavenging is the beautiful orchestration of these exhaust pulses such that the momentum of the gas column is maintained in a fashion most beneficial to the continued optimal filling of the cylinders with fresh mixture - and the 'exhausting' of the combusted gasses. The sounds we hear are very much influenced by managment of pipe sizes, lengths, crossovers, etc - like a musical instrument.
Calculation of flow characteristics is severely influenced by the rapid quenching of exhaust gas temps, and so on - but regards the point of torque - my slight understanding is that the breathing cycle of a cylinder is largely determined by the camshaft(s) design: While they must be reasonably optimized for HP at higher RPMs, their intake/exhaust overlap combined with scavenging at 'driving RPMs' can lead to perfectly good, unburned mixture being sucked out of the cylinders prior to yielding the torque potential - unless there is adequate backpressure to prevent the great escape. Enthalpy to entropy?
The first issue to contend with is the point(s) Heinrich made. A resonant phenomenon with multiple harmonics is established as in acoustic theory: Scavenging is the beautiful orchestration of these exhaust pulses such that the momentum of the gas column is maintained in a fashion most beneficial to the continued optimal filling of the cylinders with fresh mixture - and the 'exhausting' of the combusted gasses. The sounds we hear are very much influenced by managment of pipe sizes, lengths, crossovers, etc - like a musical instrument.
Calculation of flow characteristics is severely influenced by the rapid quenching of exhaust gas temps, and so on - but regards the point of torque - my slight understanding is that the breathing cycle of a cylinder is largely determined by the camshaft(s) design: While they must be reasonably optimized for HP at higher RPMs, their intake/exhaust overlap combined with scavenging at 'driving RPMs' can lead to perfectly good, unburned mixture being sucked out of the cylinders prior to yielding the torque potential - unless there is adequate backpressure to prevent the great escape. Enthalpy to entropy?
07-30-2004, 09:30 PM
#4
Rennlist Member
You will make your head hurt if you dig too deep with the calculations. Basically, for our cars, less back pressure the better. by removing the cat and Mufflers, we got 10ft lbs improvement from 3k to 6k. yes HP is torque x speed , but at the lower rpms you dont have much rpms to make HP. (even if you did get slightly more torque at the max torque area) as I always say, its all about HP. (or area under the hp curve that is being used)
k
k
07-31-2004, 06:56 PM
#5
Nordschleife Master
Thread Starter
Garth:
My understanding has been that the scavageing ability of the engine is based on a great many things working together. Things like head design, cam profiles, header designs, exauste designs, etc. The end goal being that when they all work together, it's almost a mutiplicitive effect. If any of the issues are not taken care of, the whole system falls apart like a house of cards.
Apperently, part of of the issue is the idea of diminishing returns. You can install a fairly open exaust, and get good gains. Then you install one that results in even less restriction, and get nearly no gains. You've hit the point of diminishing returns. Instead of working on the exauste, you need to work on the biggest bottle neck, be that the cam profile, the head design, etc. Or maybe, the exauste system is scavageing the cylinder exelently, but your just not getting enough air in.
All of the pre 32v engines have very little valve overlap, I dought sweeping the mixture out would be an issue at all. I don't think the 32v cars have much overlap eighter. Compared to the cams the 944 guys, and the aftermarket chevy guys run, the 928 cams are mild as can be.
As I understand it, most production cam shafts have very little overlap. Cams with much overlap apperently work great at high rpm race applications, but are also apperenlty very peaky, and very much effected by the way the exauste and intake pulses line up. As a result, the engine may run like a scalded rabbit at 6,000 rpm, but may not idle at all.
heinrich:
I was understanding you very well, untill you said "If you raise the exhaust pipe diameter above a point where the gases flow in a linear direction, this will result in turbulence. And this causes flow to be slower than in a smaller pipe." Which makes no sence at all to me. Care to explain further?
Mark:
I'm an engineering student. All I do is ask questions that I can't answer, yet, and then try to find the answer. I agree with you that hp is a good thing, and pretty much the savor of all that is holy and good. However, I have been told the above many many many times, and it makes no sence to me. Thus the question, "Why?" If I can't find a good answer, the question gets shelved for future reference.
To accuretly model this stuff, you'd end up with an enormous amount of calculations. Baicly, as I understand it, you've got multipul varables that are all interlinked. You'd have to go through, and set up defferential equations for the heat flow, set up differential equations to model the whole exaust pulse stuff. Then you'd have to set up a system of equations, and solve the thing. The equations aren't simple, and would probably take years of reasearch to find, and then you have to solve the whole deal. That may just about take a Cray to do right. The F1 guys can pull that off, I can't!
I'm not really looking for calculations to accuretly model an exaust system. I'm looking for an explination of one smallish phenominon. Why is the statement "backpressure is good for torque" true. Or is it true? Is it the backpressure itself that is creates additional torque, or something else normaly linked with backrpessure, but behaves in a counterintuitive way?
My understanding has been that the scavageing ability of the engine is based on a great many things working together. Things like head design, cam profiles, header designs, exauste designs, etc. The end goal being that when they all work together, it's almost a mutiplicitive effect. If any of the issues are not taken care of, the whole system falls apart like a house of cards.
Apperently, part of of the issue is the idea of diminishing returns. You can install a fairly open exaust, and get good gains. Then you install one that results in even less restriction, and get nearly no gains. You've hit the point of diminishing returns. Instead of working on the exauste, you need to work on the biggest bottle neck, be that the cam profile, the head design, etc. Or maybe, the exauste system is scavageing the cylinder exelently, but your just not getting enough air in.
All of the pre 32v engines have very little valve overlap, I dought sweeping the mixture out would be an issue at all. I don't think the 32v cars have much overlap eighter. Compared to the cams the 944 guys, and the aftermarket chevy guys run, the 928 cams are mild as can be.
As I understand it, most production cam shafts have very little overlap. Cams with much overlap apperently work great at high rpm race applications, but are also apperenlty very peaky, and very much effected by the way the exauste and intake pulses line up. As a result, the engine may run like a scalded rabbit at 6,000 rpm, but may not idle at all.
heinrich:
I was understanding you very well, untill you said "If you raise the exhaust pipe diameter above a point where the gases flow in a linear direction, this will result in turbulence. And this causes flow to be slower than in a smaller pipe." Which makes no sence at all to me. Care to explain further?
Mark:
I'm an engineering student. All I do is ask questions that I can't answer, yet, and then try to find the answer. I agree with you that hp is a good thing, and pretty much the savor of all that is holy and good. However, I have been told the above many many many times, and it makes no sence to me. Thus the question, "Why?" If I can't find a good answer, the question gets shelved for future reference.
To accuretly model this stuff, you'd end up with an enormous amount of calculations. Baicly, as I understand it, you've got multipul varables that are all interlinked. You'd have to go through, and set up defferential equations for the heat flow, set up differential equations to model the whole exaust pulse stuff. Then you'd have to set up a system of equations, and solve the thing. The equations aren't simple, and would probably take years of reasearch to find, and then you have to solve the whole deal. That may just about take a Cray to do right. The F1 guys can pull that off, I can't!
I'm not really looking for calculations to accuretly model an exaust system. I'm looking for an explination of one smallish phenominon. Why is the statement "backpressure is good for torque" true. Or is it true? Is it the backpressure itself that is creates additional torque, or something else normaly linked with backrpessure, but behaves in a counterintuitive way?
07-31-2004, 07:14 PM
#6
Burning Brakes
Join Date: Aug 2001
Posts: 1,117
Likes: 0
Received 0 Likes
on
0 Posts
VU,
Brush up on your spelling and get a job at one of the 'Big 3' - they have the computing power and the software to FEA an entire engine in it's running state.
The computer will even suggest areas to add/remove material to reduce NVH, and to manage thermal transfer. Not perfect yet, but it is *quite* good stuff.
Greg
Brush up on your spelling and get a job at one of the 'Big 3' - they have the computing power and the software to FEA an entire engine in it's running state.
The computer will even suggest areas to add/remove material to reduce NVH, and to manage thermal transfer. Not perfect yet, but it is *quite* good stuff.
Greg
07-31-2004, 07:29 PM
#7
Nordschleife Master
Thread Starter
I know they do. I also know severl other folks that have similer ability. The company I work for part time runs really complex flow calculations for all sorts of stuff. Pretty much anything you want to model, they can. Of course, I'm just a part time emploeye, so I don't have access to that sort of stuff.
The big 3, and other sources, just don't publish it to outsiders. It's propriatray stuff you see. I'm not a member of the SAE, so I can't access their papers on the subject.
I'm not asking for that kind of knowlage. Just a general theory that explains it under most circumstances.
The big 3, and other sources, just don't publish it to outsiders. It's propriatray stuff you see. I'm not a member of the SAE, so I can't access their papers on the subject.
I'm not asking for that kind of knowlage. Just a general theory that explains it under most circumstances.
Trending Topics
07-31-2004, 07:51 PM
#8
Burning Brakes
Join Date: Aug 2001
Posts: 1,117
Likes: 0
Received 0 Likes
on
0 Posts
Here's my GUT (Grand Unification Theory, for those who don't give 2 good flyin's about physics)
There is NO free lunch. There is ALWAYS a trade-off. Nothing happens in a vacuum, and information can leak out of even a black hole (which, of course, means that maybe GM might build a decent car someday, but I don't think even Steven Hawking would take *that* bet...)
You can have it all. But it will cost you - driveabilty, economy, emissions, NVH, weight, safety, or more likely, time, money, and effort.
You want TQ? Lose some HP. You want both - spend some time and/or money. You want more? Spend more of the aforementioned time/money.
The bottom line is always the bottom line. If you really want to model this stuff in anyway other than by experimentation and empirical data collection, you will have to get at the tools. CFD is done on big honkin' computers for a reason - you'd be dead by the time you got what the computer got in 5 minutes.
Greg
There is NO free lunch. There is ALWAYS a trade-off. Nothing happens in a vacuum, and information can leak out of even a black hole (which, of course, means that maybe GM might build a decent car someday, but I don't think even Steven Hawking would take *that* bet...)
You can have it all. But it will cost you - driveabilty, economy, emissions, NVH, weight, safety, or more likely, time, money, and effort.
You want TQ? Lose some HP. You want both - spend some time and/or money. You want more? Spend more of the aforementioned time/money.
The bottom line is always the bottom line. If you really want to model this stuff in anyway other than by experimentation and empirical data collection, you will have to get at the tools. CFD is done on big honkin' computers for a reason - you'd be dead by the time you got what the computer got in 5 minutes.
Greg
07-31-2004, 07:58 PM
#9
Rennlist Member
Originally Posted by ViribusUnits
Originally Posted by mark kibort
You will make your head hurt if you dig too deep with the calculations. Basically, for our cars, less back pressure the better. by removing the cat and Mufflers, we got 10ft lbs improvement from 3k to 6k. yes HP is torque x speed , but at the lower rpms you dont have much rpms to make HP. (even if you did get slightly more torque at the max torque area) as I always say, its all about HP. (or area under the hp curve that is being used)
k
k
I'm an engineering student. All I do is ask questions that I can't answer, yet, and then try to find the answer. I agree with you that hp is a good thing, and pretty much the savor of all that is holy and good. However, I have been told the above many many many times, and it makes no sence to me. Thus the question, "Why?" If I can't find a good answer, the question gets shelved for future reference.
The diagram is highly exaggerated and simplified. The X-axis represents RPM, the red lines represent the amount of RPM that is normally used, the 1000-2000 RPM per gear that is utilized before you switch to the next gear. With me so far?
OK, the steep narrow curve in the center represents the power curve of engine A. Whether you measure torque, HP, Newton Meters, whatever. The units are irrelevant. The point is it makes good power over a narrow RPM range.
The broad, flat curve represents the power curve of engine B. While it makes less peak power, it makes decent power over more of the RPM range that you will use.
This is why peak HP/Tq figures are not always indicative of a car's overall performance. With Engine A you will only get good performance over a narrow RPM band and if you fall outside of that range you're hosed. With Engine B you have more flexibility; if you fall outside of your optimal RPM range you're still making decent power.
Now, if you want to make a deterministic comparison(unnecessary in this exaggerated example) you would geometrically compare the area of the part of the curve that falls within the RPM range bounded by the red lines. You can calculate or observe the distance between the red lines. For example, drive 40mph in 2nd gear, note the RPM. Drive 40mph in 3rd gear, note the RPM. With AT you may be better off just calculating. Either way, this gives you the RPM range of a given gear, and the distance in RPM along the X-axis, between the red lines. Armed with this info and a dyno chart, you can determine where this RPM window fits along your power curve to maximize the area under the curve. This will tell you what your optimum shift points are.
A 10-second Civic would probably show a dyno curve like Engine A. Impressive peak power, but no low end. Our V8's with their flat torque curves are more like Engine B. The flat torque curve, IOW the large area under the curve, makes them very forgiving to drive. But it's still a curve, and gears can be selected to maximize the area under the curve.
HTH.
07-31-2004, 08:09 PM
#10
Burning Brakes
Join Date: Aug 2001
Posts: 1,117
Likes: 0
Received 0 Likes
on
0 Posts
Don't forget, exhaust pressure varies by ignition timing, cam timing, cam lift and duration, conditon of your cam/lifters, valve springs, engine temp, exhaust material, cylinder head design and consistency of all those factors vis-a-vis every other cylinder.
The relationship with backpressure is more about symptomology than causality. If you want to attempt to manage your TQ via backpressure, there's some guy down in (IIRC) Oz, who has a (basically) a throttle that is placed in the exhaust line to provide 'variable' backpressure. Google it if you're interested.
BTW-Didn't we beat this horse last week sometime?
Greg
The relationship with backpressure is more about symptomology than causality. If you want to attempt to manage your TQ via backpressure, there's some guy down in (IIRC) Oz, who has a (basically) a throttle that is placed in the exhaust line to provide 'variable' backpressure. Google it if you're interested.
BTW-Didn't we beat this horse last week sometime?
Greg
07-31-2004, 08:42 PM
#11
Rennlist Member
Originally Posted by 2V4V
BTW-Didn't we beat this horse last week sometime?
07-31-2004, 08:48 PM
#12
Drifting
Join Date: Jul 2002
Location: Los Angeles, CA
Posts: 3,348
Likes: 0
Received 0 Likes
on
0 Posts
