JDL V2 EL Header...Revisited and Revised

[JDLAutoDesign]

Snapped a few pics before it gets welded

[slava]

Waiting on price and dyno sheet!

[Turdinator]

@JDLAutoDesign how do you determin what runner length to go with? Trial and error? Some sort of computer model?

[xDanger_208x]

Stoked! Can't wait to get mine. Even more exited to hear you're making a great thing even better!

[Kodename47]

Originally Posted by Turdinator

@JDLAutoDesign how do you determin what runner length to go with? Trial and error? Some sort of computer model?

I'd suspect both above, and some maths while taking into account restriction on available space.

[Lee@JDLAutoDesign]

Originally Posted by Turdinator

@JDLAutoDesign how do you determin what runner length to go with? Trial and error? Some sort of computer model?

Good question. If I were more familiar with this forum and knew how to thank someone for their post I would.

Sorry in advance for the long winded answer. My philosophy is that when you design any sort of manifold (intake or exhaust) you have to balance:

1. Spacial constraints
2. Aerodynamic design
3. Acoustic design

... and the list goes on to include cost, material section, material availability, etc... but I'll skip those and focus on the three I listed since realistically these are the ones that most affect performance.

The first one is simple; if it doesn't fit in the car and/or you can't install it, you don't have a product.

The second one refers to sizing runner diameters, merge collector lengths (which effectively sets the merge angle), bend radii, balancing number of bends per runner, etc. I call it "aerodynamic design" because the choices you make here set the ~steady state flow conditions (pressures, temperatures, velocities) at each location in the manifold at a given engine speed/load.

The acoustic design is a separate bullet because it is ~more or less independent of the aerodynamic design, meaning if I move up or down a pipe diameter I really haven't affected the harmonics at all. For example, a +/- 0.25" change in pipe diameter (~30% change in area) affects sound speed by ~0.1%. The dominant driver in harmonic design is runner length.

So, to answer your question... we start by specing and building a merge collector that we like. If it's a new style/design I will usually run some CFD on the shape to make sure we are happy with the collector's performance. It is my belief that the collector design/quality is THE most important factor in a header design since the collector dictates the exhaust pressure vs. exhaust flow characteristic of the manifold. We place this collector in the car, then work backward to the exhaust ports making sure the fitment is acceptable, and from this, the runner lengths falls out (of course I'm simplifying this process greatly).

A simple hand calc (it's a spreadsheet, actually) then tells us the resonant speed for each of the first N harmonics. If we are happy with this, then we go forward. If not, we start over and try a different routing to get us closer to our desired runner length.

It's an iterative process, that's for sure, but you can limit the number of iterations if you understand what you are aiming for and how your design choices affect your end results.

Going forward... I'm campaigning to log exhaust pressure in a runner on the dyno as a function of time. My plan would then be to rebuild the signal with an FFT and see how close my 'hand calcs' for resonance actually match the measured frequency content

Hope this is helpful,
Lee

[FR-S Matt]

Originally Posted by Lee@JDLAutoDesign

Good question. If I were more familiar with this forum and knew how to thank someone for their post I would.

Sorry in advance for the long winded answer. My philosophy is that when you design any sort of manifold (intake or exhaust) you have to balance:

1. Spacial constraints
2. Aerodynamic design
3. Acoustic design

... and the list goes on to include cost, material section, material availability, etc... but I'll skip those and focus on the three I listed since realistically these are the ones that most affect performance.

The first one is simple; if it doesn't fit in the car and/or you can't install it, you don't have a product.

The second one refers to sizing runner diameters, merge collector lengths (which effectively sets the merge angle), bend radii, balancing number of bends per runner, etc. I call it "aerodynamic design" because the choices you make here set the ~steady state flow conditions (pressures, temperatures, velocities) at each location in the manifold at a given engine speed/load.

The acoustic design is a separate bullet because it is ~more or less independent of the aerodynamic design, meaning if I move up or down a pipe diameter I really haven't affected the harmonics at all. For example, a +/- 0.25" change in pipe diameter (~30% change in area) affects sound speed by ~0.1%. The dominant driver in harmonic design is runner length.

So, to answer your question... we start by specing and building a merge collector that we like. If it's a new style/design I will usually run some CFD on the shape to make sure we are happy with the collector's performance. It is my belief that the collector design/quality is THE most important factor in a header design since the collector dictates the exhaust pressure vs. exhaust flow characteristic of the manifold. We place this collector in the car, then work backward to the exhaust ports making sure the fitment is acceptable, and from this, the runner lengths falls out (of course I'm simplifying this process greatly).

A simple hand calc (it's a spreadsheet, actually) then tells us the resonant speed for each of the first N harmonics. If we are happy with this, then we go forward. If not, we start over and try a different routing to get us closer to our desired runner length.

It's an iterative process, that's for sure, but you can limit the number of iterations if you understand what you are aiming for and how your design choices affect your end results.

Going forward... I'm campaigning to log exhaust pressure in a runner on the dyno as a function of time. My plan would then be to rebuild the signal with an FFT and see how close my 'hand calcs' for resonance actually match the measured frequency content

Hope this is helpful,
Lee

You gotta get to 10 posts to enable the Thanks feature.

[peebking]

I just bought a v1 and you come out a v2 sosad

[Turdinator]

Thank you for the detailed explanation


What sort of rpm should these have a helpful harmonic?

Originally Posted by Lee@JDLAutoDesign

Good question. If I were more familiar with this forum and knew how to thank someone for their post I would.

Sorry in advance for the long winded answer. My philosophy is that when you design any sort of manifold (intake or exhaust) you have to balance:

1. Spacial constraints
2. Aerodynamic design
3. Acoustic design

... and the list goes on to include cost, material section, material availability, etc... but I'll skip those and focus on the three I listed since realistically these are the ones that most affect performance.

The first one is simple; if it doesn't fit in the car and/or you can't install it, you don't have a product.

The second one refers to sizing runner diameters, merge collector lengths (which effectively sets the merge angle), bend radii, balancing number of bends per runner, etc. I call it "aerodynamic design" because the choices you make here set the ~steady state flow conditions (pressures, temperatures, velocities) at each location in the manifold at a given engine speed/load.

The acoustic design is a separate bullet because it is ~more or less independent of the aerodynamic design, meaning if I move up or down a pipe diameter I really haven't affected the harmonics at all. For example, a +/- 0.25" change in pipe diameter (~30% change in area) affects sound speed by ~0.1%. The dominant driver in harmonic design is runner length.

So, to answer your question... we start by specing and building a merge collector that we like. If it's a new style/design I will usually run some CFD on the shape to make sure we are happy with the collector's performance. It is my belief that the collector design/quality is THE most important factor in a header design since the collector dictates the exhaust pressure vs. exhaust flow characteristic of the manifold. We place this collector in the car, then work backward to the exhaust ports making sure the fitment is acceptable, and from this, the runner lengths falls out (of course I'm simplifying this process greatly).

A simple hand calc (it's a spreadsheet, actually) then tells us the resonant speed for each of the first N harmonics. If we are happy with this, then we go forward. If not, we start over and try a different routing to get us closer to our desired runner length.

It's an iterative process, that's for sure, but you can limit the number of iterations if you understand what you are aiming for and how your design choices affect your end results.

Going forward... I'm campaigning to log exhaust pressure in a runner on the dyno as a function of time. My plan would then be to rebuild the signal with an FFT and see how close my 'hand calcs' for resonance actually match the measured frequency content

Hope this is helpful,
Lee

[Dimman]

Originally Posted by Lee@JDLAutoDesign

Good question. If I were more familiar with this forum and knew how to thank someone for their post I would.

Sorry in advance for the long winded answer. My philosophy is that when you design any sort of manifold (intake or exhaust) you have to balance:

1. Spacial constraints
2. Aerodynamic design
3. Acoustic design

... and the list goes on to include cost, material section, material availability, etc... but I'll skip those and focus on the three I listed since realistically these are the ones that most affect performance.

The first one is simple; if it doesn't fit in the car and/or you can't install it, you don't have a product.

The second one refers to sizing runner diameters, merge collector lengths (which effectively sets the merge angle), bend radii, balancing number of bends per runner, etc. I call it "aerodynamic design" because the choices you make here set the ~steady state flow conditions (pressures, temperatures, velocities) at each location in the manifold at a given engine speed/load.

The acoustic design is a separate bullet because it is ~more or less independent of the aerodynamic design, meaning if I move up or down a pipe diameter I really haven't affected the harmonics at all. For example, a +/- 0.25" change in pipe diameter (~30% change in area) affects sound speed by ~0.1%. The dominant driver in harmonic design is runner length.

So, to answer your question... we start by specing and building a merge collector that we like. If it's a new style/design I will usually run some CFD on the shape to make sure we are happy with the collector's performance. It is my belief that the collector design/quality is THE most important factor in a header design since the collector dictates the exhaust pressure vs. exhaust flow characteristic of the manifold. We place this collector in the car, then work backward to the exhaust ports making sure the fitment is acceptable, and from this, the runner lengths falls out (of course I'm simplifying this process greatly).

A simple hand calc (it's a spreadsheet, actually) then tells us the resonant speed for each of the first N harmonics. If we are happy with this, then we go forward. If not, we start over and try a different routing to get us closer to our desired runner length.

It's an iterative process, that's for sure, but you can limit the number of iterations if you understand what you are aiming for and how your design choices affect your end results.

Going forward... I'm campaigning to log exhaust pressure in a runner on the dyno as a function of time. My plan would then be to rebuild the signal with an FFT and see how close my 'hand calcs' for resonance actually match the measured frequency content

Hope this is helpful,
Lee

Interesting post.

A couple things to look at:

While diameter doesn't affect acoustic return timing, it does have an effect on scavenging, since the acoustic waves are either positive or negative pressure, and pressure is based on force and area.

Another thing is your CFD. With your testing in another thread, you analyzed a 2-1 merge collector. However you were analyzing steady flow coming from both runners into the collector. Perhaps analyzing different velocities in each runner would give a better idea of how pulsed flow into the collector would act.

My .02$.

[Lee@JDLAutoDesign]

Originally Posted by Dimman

Interesting post.

A couple things to look at:

While diameter doesn't affect acoustic return timing, it does have an effect on scavenging, since the acoustic waves are either positive or negative pressure, and pressure is based on force and area.

Another thing is your CFD. With your testing in another thread, you analyzed a 2-1 merge collector. However you were analyzing steady flow coming from both runners into the collector. Perhaps analyzing different velocities in each runner would give a better idea of how pulsed flow into the collector would act.

My .02$.

Good post.

Completely agree on both of your points. These are certainly simplified analyses done with limited time and resources. My day job is aerodynamic design work in a different, more well funded industry so I'll be the first to tell you that the tools I'm using here are ~crude in comparison to the state of the art. With that said, I'm ONLY trying to let the tools I have available guide me in the direction of better or worse; but at the end of the day every piece we produce is dyno'd and/or tracked and that's the final judge of the decisions we've made.

About the CFD... yes, the code as I ran it is certainly missing some amount of deterministic mixing (nevermind the turbulent mixing!) associated with 4 out of phase flow streams merging. The challenge is, even if I could model the time varying flow field I would still have no real way of obtaining good boundary conditions for these models (other than simply ~assuming them, which may or may not be better than my simple, steady model). The good news is, as I'm sure you know, mixing loss in an accelerating flow field is always less penalizing than in a diffusing flow field, so I'm fairly confident that I'm not fooling myself here with the analysis I ran. If we were going the OTHER way and modeling an expansion and I was relying on the CFD to predict separation, than I would be very very nervous

Give me a few days... I'll put together a case with out of phase boundary conditions and see how it looks. It's somewhat labor intensive to set up and run the cases (need to make myself a GUI ).

Turdinator - let me track down the final runner lengths and get back to you.

[RiskyTrousers]

@Lee@JDLAutoDesign

Any idea when my EL header replacement will ship? I was promised a replacement September/October. It was requested in August. Why so long?

[Dimman]

Originally Posted by Lee@JDLAutoDesign

Good post.

Completely agree on both of your points. These are certainly simplified analyses done with limited time and resources. My day job is aerodynamic design work in a different, more well funded industry so I'll be the first to tell you that the tools I'm using here are ~crude in comparison to the state of the art. With that said, I'm ONLY trying to let the tools I have available guide me in the direction of better or worse; but at the end of the day every piece we produce is dyno'd and/or tracked and that's the final judge of the decisions we've made.

About the CFD... yes, the code as I ran it is certainly missing some amount of deterministic mixing (nevermind the turbulent mixing!) associated with 4 out of phase flow streams merging. The challenge is, even if I could model the time varying flow field I would still have no real way of obtaining good boundary conditions for these models (other than simply ~assuming them, which may or may not be better than my simple, steady model). The good news is, as I'm sure you know, mixing loss in an accelerating flow field is always less penalizing than in a diffusing flow field, so I'm fairly confident that I'm not fooling myself here with the analysis I ran. If we were going the OTHER way and modeling an expansion and I was relying on the CFD to predict separation, than I would be very very nervous

Give me a few days... I'll put together a case with out of phase boundary conditions and see how it looks. It's somewhat labor intensive to set up and run the cases (need to make myself a GUI ).

Turdinator - let me track down the final runner lengths and get back to you.

Could you run several steady simulations at different velocities/runner and view them sequentially? Animating them, so to speak.

What exactly can you do with that program, btw?

[Lee@JDLAutoDesign]

Originally Posted by RiskyTrousers

@Lee@JDLAutoDesign

Any idea when my EL header replacement will ship? I was promised a replacement September/October. It was requested in August. Why so long?

I don't handle orders, but I'll look into this and ask Ronnie to contact you ASAP. Thanks for your patience.