Q for exhaust gurus. Oh, and Harvey too...

[jsteele22]

Okay, so I've been reading up on exhaust in general (Smith & Morrison book) and the SVX system in particular. Very interesting stuff. I have one idea that I wanted to check out, and Harvey, since I've seen you refer to the Smith book before, you seem like the ideal person to ask about this.


When they discuss designing a system using interference, there is an example of a single pipe coming from an exhaust port, branching in two at a Y. One of the branches goes a tuned distance (e.g., 18") before being capped. The other of the two pipe ends is left open. The idea, which is backed with experimental data, is that a pulse coming from the exhaust port splits in two at the Y, and half travels down each branch. The half that goes down the capped pipe reflects off the end, travels back up to the Y, and then back out the open end. So the result of this odd setup is that a single sharp pulse of pressure is converted to a smaller amplitude but longer pulse of pressure. In particular, the eventual inverted (negative pressure) wave reaches the exhaust port later than it would without the capped branch in place. They call the capped branch an interference pipe. Although, the idea looks very strange, some headers (e.g., a 4-2-1 header on a V-8) make use of the "other" runner in each pair as a capped pipe (b/c it's exhaust valve is closed). Also, a few explicit examples are shown in the book (e.g., headers for a engine with a siamesed port pair; the pair has a primary tube with an interference pipe attached.)

Reading about the SVX exhaust system, there seem to be two distinct reasons for having the pre-cats (apart from pollution) :

1) The infamous "tinny" resonance at around 3k caused by pressure pulses that bounce from the LH exhaust manifold to the RH and back, building up energy each time,

and

2) A seeming majority opinion that the engine needs some "restriction" to produce torque.

As for #2, I read in the Smith book that the "restriction" operates by slowing down the transit of the pressure pulse, so that the negative pressure wave arrives back at the exhaust valve later than it otherwise would. This is exactly what the interference pipe is said to accomplish. I'm wondering if interference pipes (one per side) could be used instead of the pre-cats for this purpose.

As for #1, the interference pipe does not really dissipate energy, but merely spreads it out over time. But I wonder if a pulse passing 2 of these interference pipes (1 on either side) would be sufficiently broadened that the resonance would be spread out over a larger range of RPMs into a less objectionable sound.

I'm really intrigued by the idea of the interference pipe. It seems somehow appropriate for the EG33 b/c of the lack of exhaust valve overlap of the stock cam : each port basically sees it's own private pipe, which makes it simpler to design for. And it's a piece of cake to tune : just weld on a short stub, then different lengths of capped pipe can be clamped on. Also, it's a really cool idea theoretically - extremely similar to R/F and microwave electronics, where a coax cable or a circuit board trace just leads off to nowhere, yet is crucial for the circuit to operate. Finally, I'm a cheap SOB, and don't want to pay for cats unless I need them (which I don't for emissions.)

Anyway, tell me what you think. Has anybody seen this kind of idea used in the past half-century ? (That Smith/Robertson book is pretty ancient !)

BTW, I'm eventually gonna put my EG33 engine in a Porsche 914; so sooner or later I'm gonna have to lay out an exhaust system for it. This thread is just one of the early exploratory steps...

[SVXRide]

You can use the Morrison book to start looking at "branch" length and at what rpm they'll resonate and what volume you'll want them to dump into relative to the wave that will be reflected back.

This all said, I really don't think you'll have anywhere near the real estate you need once you've got the engine stuffed in the Porche unless you want to wrap the car in exhaust pipe

-Bill

[curly2k3]

well going back to my physics classes, in theory it would "change" the resonensey(sp) of the pulses being passed through the exhaust system, however this would be dependent on the length of the pipe and the diameter of the pipe as well. in theory it could* work, however its not like the reflected sound wave will necessarily go back down the desired pipe after it hits that end cap. from what i have gathered it would be best to run it catless with a smaller pipe and then break it into a larger pipe if desired. its typical subaru tho, very restrictive, very poor exhaust systems, if we didnt have emissions we wouldnt have the pre-cats, cat, resonator, and restrictive muffler to begin with.

[Trevor]

Quote "(1) The infamous "tinny" resonance at around 3k caused by pressure pulses that bounce from the LH exhaust manifold to the RH and back, building up energy each time,"

I would be very interested to learn as to how and from where, this theory has been obtained/established as a definite cause of the "tinny" sound.

[oab_au]

Originally Posted by jsteele22

Okay, so I've been reading up on exhaust in general (Smith & Robertson book) and the SVX system in particular. Very interesting stuff. I have one idea that I wanted to check out, and Harvey, since I've seen you refer to the Smith book before, you seem like the ideal person to ask about this.


When they discuss designing a system using interference, there is an example of a single pipe coming from an exhaust port, branching in two at a Y. One of the branches goes a tuned distance (e.g., 18") before being capped. The other of the two pipe ends is left open. The idea, which is backed with experimental data, is that a pulse coming from the exhaust port splits in two at the Y, and half travels down each branch. The half that goes down the capped pipe reflects off the end, travels back up to the Y, and then back out the open end. So the result of this odd setup is that a single sharp pulse of pressure is converted to a smaller amplitude but longer pulse of pressure. In particular, the eventual inverted (negative pressure) wave reaches the exhaust port later than it would without the capped branch in place. They call the capped branch an interference pipe. Although, the idea looks very strange, some headers (e.g., a 4-2-1 header on a V-8) make use of the "other" runner in each pair as a capped pipe (b/c it's exhaust valve is closed). Also, a few explicit examples are shown in the book (e.g., headers for a engine with a siamesed port pair; the pair has a primary tube with an interference pipe attached.)

Reading about the SVX exhaust system, there seem to be two distinct reasons for having the pre-cats (apart from pollution) :

1) The infamous "tinny" resonance at around 3k caused by pressure pulses that bounce from the LH exhaust manifold to the RH and back, building up energy each time,

and

2) A seeming majority opinion that the engine needs some "restriction" to produce torque.

As for #2, I read in the Smith book that the "restriction" operates by slowing down the transit of the pressure pulse, so that the negative pressure wave arrives back at the exhaust valve later than it otherwise would. This is exactly what the interference pipe is said to accomplish. I'm wondering if interference pipes (one per side) could be used instead of the pre-cats for this purpose.

As for #1, the interference pipe does not really dissipate energy, but merely spreads it out over time. But I wonder if a pulse passing 2 of these interference pipes (1 on either side) would be sufficiently broadened that the resonance would be spread out over a larger range of RPMs into a less objectionable sound.

I'm really intrigued by the idea of the interference pipe. It seems somehow appropriate for the EG33 b/c of the lack of exhaust valve overlap of the stock cam : each port basically sees it's own private pipe, which makes it simpler to design for. And it's a piece of cake to tune : just weld on a short stub, then different lengths of capped pipe can be clamped on. Also, it's a really cool idea theoretically - extremely similar to R/F and microwave electronics, where a coax cable or a circuit board trace just leads off to nowhere, yet is crucial for the circuit to operate. Finally, I'm a cheap SOB, and don't want to pay for cats unless I need them (which I don't for emissions.)

Anyway, tell me what you think. Has anybody seen this kind of idea used in the past half-century ? (That Smith/Robertson book is pretty ancient !)

BTW, I'm eventually gonna put my EG33 engine in a Porsche 914; so sooner or later I'm gonna have to lay out an exhaust system for it. This thread is just one of the early exploratory steps...

This could take all day, but to cover your points.

The interference pipe is a compromise to fix a problem. They are used mainly to Interfere with the wave action that is happening. You can see some cars use them on the rear end of exhaust system to reduce a problem pulse that is causing a sound problem. They are also used on the inlet of the 4 cylinder engines to reduce the severity of the inlet pulsing sound.

As you say when the exhaust pulse meets the interference pipe the positive wave splits equally down each pipe at half the amplitude. the pulse in the main pipe continues down the pipe, the pulse that travels down the interference pipe is reflected by the closed end, to return as the same sign to the main pipe, to travel down the pipe behind the main pulse. So the effect is to lengthen the positive pulse that is reflected, but at a much lower amplitude. Lowering the amplitude of the returning wave reduces the filling effect of the resonate wave. So it is not used to increase the resonate effect to produce power.

1) The infamous "tinny" resonance at around 3k caused by pressure pulses that bounce from the LH exhaust manifold to the RH and back, building up energy each time,


No I don't think that is the cause of the resonance that happens when the exhaust is altered. But I do think it could be cured by fitting an Interference pipe at the back end of the system after the resonator.

2) A seeming majority opinion that the engine needs some "restriction" to produce torque

Yes this is one of those myths that are well embedded in the US tuning theory. I think it comes from the term "back-pressure". This can mean the returning negative pressure wave, or is thought of, as an ambient pressure that restricts the gas flow through the system.

Every engine benefits from the returning pressure wave. No engine will benefit from an ambient residual pressure in the pipe system. Excepting a two stroke fitted with an expansion chamber. A gas restriction in the system has no effect on the sound wave, it will travel at the same speed through the gas pressure regardless.

It seems somehow appropriate for the EG33 b/c of the lack of exhaust valve overlap of the stock cam : each port basically sees it's own private pipe

This is the vital component to produce the power. As you say each cylinder 'sees' the exhaust system as its own pipe, due to the 240* exhaust duration. So they can be a tuned length, with the pressure dump resonator fitted at the right spot, to compensate for the long system that the car has to have. This is the reason that I am reluctant to change the duration of the exhaust cams. Using an interference pipe to try to tune the resonate length
is not the way. It will reduce the wave intensity.

Anyway, tell me what you think. Has anybody seen this kind of idea used in the past half-century ? (That Smith/Robertson book is pretty ancient !)

Yes it is. Just shows " theres nothing new" . The Trumpet sounds just the same today, as it did when it was first invented back before the car.

Harvey.

[jsteele22]

Originally Posted by oab_au

1) The infamous "tinny" resonance at around 3k caused by pressure pulses that bounce from the LH exhaust manifold to the RH and back, building up energy each time,


No I don't think that is the cause of the resonance that happens when the exhaust is altered. But I do think it could be cured by fitting an Interference pipe at the back end of the system after the resonator.

Doh ! That was the one thing that I thought I really understood, and it was based on threads on this board (e.g., "My new exhaust system sucks...", or something similar.) My understanding was that every time someone replaced their pre-cats with a straight pipe this resonance appeared, and it was due to pulses travelling from one manifold to the other, arriving at the moment the next exhaust valve opens to reinforce it.

Let's check some numbers : If I understood
this photo from SVXfiles' locker correctly the distance from the front of a header to where the two branches join is 59", so lets say 5 feet.

d = 2 * 5 ft = 10 ft = approx distance from one manifold to the other
v = 1500 ft/sec = speed of sound in hot air (roughly)
t = d/v = 10 ft / 1500 ft/sec = 1/150 sec
f = 1/t = 150 /sec * 60sec/min = 9000 pulses per minute
f / (6 pulses/2 rev) = 3000 RPM

So the theory works out pretty nicely. Please don't take me wrong, I'm not insisting that this is the correct explanation, but it seems so clear and consistent, and I was sure it was basically what you guys had already said ! Was there a different explanation of where this resonance comes from ? Maybe I'll have to go back and read those threads yet again...

Originally Posted by oab_au

It seems somehow appropriate for the EG33 b/c of the lack of exhaust valve overlap of the stock cam : each port basically sees it's own private pipe

This is the vital component to produce the power. As you say each cylinder 'sees' the exhaust system as its own pipe, due to the 240* exhaust duration. So they can be a tuned length, with the pressure dump resonator fitted at the right spot, to compensate for the long system that the car has to have. This is the reason that I am reluctant to change the duration of the exhaust cams. Using an interference pipe to try to tune the resonate length is not the way. It will reduce the wave intensity.

Well, part of my interest in the interference pipe was that it seemed like a nice way to arrange for the reflected (negative pressure) wave to arrive at the right time, without having to run a long pipe before the resonator. I'm thinking in terms of my 914; maybe there's room to get the right length (is it 59" ?) of pipe to fit, but it just sounds pretty darn long.

I totally get what you are saying about the reduced amplitude of the reflected wave being undesireable. But in addition to being reduced in amplitude, it is extended in duration. Does this have any compensating benefit ? If the pressure at the exhaust port is below atmospheric for a greater portion of the exhaust cycle, will this improve torque, or is it only the instantaneous value of pressure right when the valve closes that matters ?



Thanks for all your thoughts, Harvey. I've been rummaging through the archives of this board and I've learned a lot from you already, even if I missed the point of half of it

[jsteele22]

Here's the thread I was remembering. Reply #7 by Mychailo was the one I specifically had in mind.

http://www.subaru-svx.net/forum/show...light=pre-cats

[oab_au]

Originally Posted by jsteele22

Doh ! That was the one thing that I thought I really understood, and it was based on threads on this board (e.g., "My new exhaust system sucks...", or something similar.) My understanding was that every time someone replaced their pre-cats with a straight pipe this resonance appeared, and it was due to pulses travelling from one manifold to the other, arriving at the moment the next exhaust valve opens to reinforce it.

Let's check some numbers : If I understood
this photo from SVXfiles' locker correctly the distance from the front of a header to where the two branches join is 59", so lets say 5 feet.

d = 2 * 5 ft = 10 ft = approx distance from one manifold to the other
v = 1500 ft/sec = speed of sound in hot air (roughly)
t = d/v = 10 ft / 1500 ft/sec = 1/150 sec
f = 1/t = 150 /sec * 60sec/min = 9000 pulses per minute
f / (6 pulses/2 rev) = 3000 RPM

So the theory works out pretty nicely. Please don't take me wrong, I'm not insisting that this is the correct explanation, but it seems so clear and consistent, and I was sure it was basically what you guys had already said ! Was there a different explanation of where this resonance comes from ? Maybe I'll have to go back and read those threads yet again...





Well, part of my interest in the interference pipe was that it seemed like a nice way to arrange for the reflected (negative pressure) wave to arrive at the right time, without having to run a long pipe before the resonator. I'm thinking in terms of my 914; maybe there's room to get the right length (is it 59" ?) of pipe to fit, but it just sounds pretty darn long.

I totally get what you are saying about the reduced amplitude of the reflected wave being undesireable. But in addition to being reduced in amplitude, it is extended in duration. Does this have any compensating benefit ? If the pressure at the exhaust port is below atmospheric for a greater portion of the exhaust cycle, will this improve torque, or is it only the instantaneous value of pressure right when the valve closes that matters ?



Thanks for all your thoughts, Harvey. I've been rummaging through the archives of this board and I've learned a lot from you already, even if I missed the point of half of it

There seemed to be a few different exhaust mods that had that resonance. The ones where the two header pipes are joined with a Y pipe, instead of joining in the low pressure area in the front of the secondary cat, will have gas pressure problems besides any sonic problems, as the unexpanded gas pressure can travel up the other header pipe to enter the open exhaust valve on that side. While the two header pipes join in the front of the secondary cat, the gas pressure and the sonic wave that enters the other pipe will be at a low intensity.

The interference pipe will extend the period of low pressure returned to the cylinder, but it will take a longer time. This is OK for low reving engine that can spend the time, but in a engine that uses this type of resonate exhaust tuning it doesn't spend a long time waiting for a long pressure wave. At the higher rpms time and amplitude is the main advantage.

It is the fast change in pressure at the exhaust valve that starts the Inlet pressure wave. So to get a strong inlet action, we need a strong sharp exhaust wave.

Harvey.
Ps. credit where credit is due. The person that did all the work for that book is John Morrison, B.Sc.,Ph.d.,M.I.Mech.En. Not Roberson.

[Trevor]

Originally Posted by Trevor

Quote "(1) The infamous "tinny" resonance at around 3k caused by pressure pulses that bounce from the LH exhaust manifold to the RH and back, building up energy each time,"

I would be very interested to learn as to how and from where, this theory has been obtained/established as a definite cause of the "tinny" sound.

Although I have not received a reply, I see that i can assume that I was indeed correct it doubting this suggested theory.

[Hose92SVX]

I'm a little confused as to how a smaller pipe helps torque. I understand that less backpressure is good and I know on newer sportbikes they come with butterfly valves in the exhaust that open and close as the rpm changes. I've always thought the reason smaller pipes help torque is because as the gases reach the rear of the exhaust pipe they start to cool off and slow down and keeping a smaller pipe keeps velocity higher which helps keep the gases from slowing down and creating turbalance in the pipe. Turbalance in the pipe would cause backpressure and that's what you try to avoid. Is that correct or am I still not getting the picture. I know at laundry mats where the dryers exhaust pipe are real long they purposely keep the diameter small so that the velocity of the air keeps the lint airborne. Is this the same prinicple with car exhaust? Thanks for any info that clears this up for me. Ya'll might have answered this above but to be honest you guys are talking above my head.

[svxfiles]

The photograph and #s came from exhaust system #9.
On a STOCK SVX the #s are
From the head to the entrance of the pre-cat roughly 28"
(not counting the roughly 4" inside the head to the backside of the valve)
to the exit of the pre-cat 38"
to the inlet of the two into one cat 59"
and to the inlet of the resonator 79"

All measurements are from the head.
Sorry for any confusion or delay.

[svxfiles]

When I built my last, and current exhaust (#11 V2 )
I built it as a true dual with equal length headers, no cats, and only one big resonator (strait through perforated core muffler) per side.
When I merged my true dual exhaust together,17" from the tailpipe/ outlet, my 0-60 times dropped substantially.
As I rember 1/2 a second or so.
Testing done before and after at the same entrance ramp, the same day, within a 1,1/2 hour period, same ambiant temp, using my built in G-Tech which has been incredibly consistant.
I understand how this can happen, but was and am still surprised at the result.

[Trevor]

Originally Posted by Hose92SVX

I'm a little confused as to how a smaller pipe helps torque. I understand that less backpressure is good and I know on newer sportbikes they come with butterfly valves in the exhaust that open and close as the rpm changes. I've always thought the reason smaller pipes help torque is because as the gases reach the rear of the exhaust pipe they start to cool off and slow down and keeping a smaller pipe keeps velocity higher which helps keep the gases from slowing down and creating turbalance in the pipe. Turbalance in the pipe would cause backpressure and that's what you try to avoid. Is that correct or am I still not getting the picture. I know at laundry mats where the dryers exhaust pipe are real long they purposely keep the diameter small so that the velocity of the air keeps the lint airborne. Is this the same prinicple with car exhaust? Thanks for any info that clears this up for me. Ya'll might have answered this above but to be honest you guys are talking above my head.

I stress, put in simple basic terms, where an engine has a fixed amount of valve overlap, in the low/medium RPM range, a portion of the charge entering via the inlet, will be lost via the exhaust, while both inlet and exhaust valves are open simultaneously for a period. Therefore some restriction/back pressure in the exhaust system can reduce this affect, but to the detriment of VE (cylinder filling) at higher RPM. However this downside, is partly offset as the advantage of overlap takes effect in the higher RPM range. Compromise is the name of the game. You can not have your cake and eat it.

[jsteele22]

Originally Posted by oab_au

Ps. credit where credit is due. The person that did all the work for that book is John Morrison, B.Sc.,Ph.d.,M.I.Mech.En. Not Roberson.

Yeah, I went home and did some more reading last night and noticed that I had botched the name. I'll edit the original post to fix the error.

[jsteele22]

Originally Posted by Trevor

Although I have not received a reply, I see that i can assume that I was indeed correct it doubting this suggested theory.

Indeed you were first. I replied to Harvey's post, b/c it was longer

I'm still not convinced, though, that the theory is incorrect. Are you disagreeing with the mechanism itself (pulses travelling from left to right, and so on) or the cause that aggravates it (removing the pre-cats) ? I certainly agree that replacing the main cat (or other suitable resonator) with a Y-pipe would be even worse, but I still have the recollection (haven't gone back and re-read everything to confirm) that keeping the main cat but removing the pre-cats also contributes to this unwanted resonance.

If you disagree with the mechanism itself, is there a different theory to explain it ?