Proper NA Exhaust Design
 

I'm looking for information on the optimal design for NA Exhausts. I knew it was posted up somewhere, but I can't find it.

From what I can remember, it was true dual header, to dual presilencer, collecting to a single pipe right where the y-pipe should be into a single outlet muffler.

Now, my questions are, what is the pipe length/width/cross-sectional area that I should use in each area? Is any form of metal better than another, and for what reason is it the best? How wide should the exhaust tip be and should the pipe ever taper up or taper down to a different pipe width?

P.S. If anyone could show me the math, that would be greatly appreciated.

RB got it right with their collected header & pre-silencer.

If you go dual, go all the way to through dual pre-silencers to dual mufflers.

Link to full dual stereo sound bite

Link to short video from inside

I thought the collection was supposed to happen further down than at the end of the header, for best scavenging?

It has to do with at what revs you want the torque peak.

(I just edited the video to the previous post.)
That was a full dual setup. - Nice low end.

Hmmm...now that confused me. Being an exhaust newb and only trusting RB...can you give me a crash course?

I was planning on doing the dual header and presilencer into the single Y and mufflers because I assumed it would not decrease performance from a single header presilencer and not be as loud as true dual.

I would also have the ability to switch to true dual if I was racing.

I was thinking the race car would have a redline around 9k or 10k with a torque peak of 5.5k for the 9k redline car and 6.5k for the 10k car.

Where would be the best length for such a car?

True duals with the Side exits like the FC race car? I'd keep it a tad farther than right in front of the REAR tires but then again i'm no mazda race car tech.
https://www.rx7club.com/attachment.p...1&d=1189722766

The only answer is it depends. It depends on the porting, both intake and exhaust, the header ID, the collector shape, the downstream components and stuff like that. There's no one answer, many systems have been tried with varying success with different cars, everything from true duals, to merging at the axle, to merging about a foot after the engine. All work in different situations.

what would the lead mazdaspeed engineer use on his DD in america?

he wouldn't bother because of emissions? That was a trick question Blk91fc3s.

The thing about exhausts is that you only have X amount of restriction in the stock exhaust. No matter how much you try, you can't remove more than 100% of this restriction. And removing 95% of it will be 10 times harder than removing 90% of it.

I'd say the ideal exhaust would replace the catalytic converter and mufflers only, while leaving the rest stock. This will remove almost all of the restriction at a far lower cost than true duals (which is only a little better). Oh, and mind your 5th and 6th ports of course.

New cars come with fairly unrestrictive exhausts from the factory. So I imagine a leading Mazdaspeed engineer would use stock or whatever mufflers have the sound he likes. Though you can upgrade a new car to a huge expensive exhaust for tiny HP gains.

There's more to it than just restriction, it's a series of pulses and a properly designed header can help with extracting the exhaust out of each rotor housing to give more power.

No, no there isn't. Exhausts are passive devices. They don't pull.

Maybe they just didn't cover harmonics in my fluid dynamics classes, but the claims I've seen on headers seem hokey at best. And the best evidence I've seen are some poorly photoshopped dyno sheets (hey look, the static on the new line matches the static on the old line!, etc.). When people post their dynos on these forums their entire exhausts (headers included) are equal or less what is claimed only for the header. But I can see how headers would help a little by smoothing bends. It's just that smoothing out that little bend does a lot less than, for example, dealing with the two abrupt 180 degree turns in the mufflers.

That's the thing, I thought scavenging did help pull exhaust out, and that's why a lot of people push collected headers over split ones, for that reason alone.

Yea, they don't really cover that, but then why would Mazda do the whole VDI thing if the intake and exhaust were static devices. The fact is there's pressure waves and they affect things. That's why different length headers perform differently, and why equal length is desirable. I don't know much about header theory, but it's more than just a minimum backpressure persuit. With a 2 stroke I've heard of dirt bikes loosing significant power if the expansion chamber is dented, proof that harmonics and the pulses really do matter.

He'd have enough for an FD.

The road race headers from racing beat are seperate.. My 87' was fitted that and I Had both of the 2" pipes runing side by side till the y-pipe then they just seperated to the magnaflow exhaust's. It sounded different though... It pulsated too much. If I parked it with the tail pipes towards the house I could vibrated everything in the house... Things fell off the table and everything... the weird part was that it was not that loud... I asked a friend of mine that was more informed on the subject to fill me in and he said that a mazda stock manifold was designed to have the peaks of one rotor's pulse pass by on a deppression of the other rotor's pulses... thus deadning the pulses... but since they were seperate the "true" pulses were presnent and the car vibrated everything.

OK – some basic points:
Auto engines don’t breathe in a continuous flow, they breathe in short pulses, especially the rotary.
When the intake & exhaust tubes length matches an even multiple of the speed of sound divided by the pulse frequency, you get a standing wave (or tune).
Longer tubes tune at lower frequencies, shorter tubes at higher frequencies. (Think tubas & trumpets)
When these pressure waves arrive at the end of the tube in tune with the port opening frequency, you get a nice boost in flow at the port. (On the exhaust side it’s often called scavenging, on the intake side, dynamic tune)
If you ever get a chance to drive with the header off, you will be amazed at the power loss compared to even an untuned exhaust.
Besides the tube length, other tuning variables are the collection point for multiple ports, and the crossover point for multiple tubes.
-Bill

(edit)
My 91NA had full duals.
I liked that setup for daily driving because the sweet spot was low (around 3500).
That low end grunt was fun when zipping around town.
You might notice it in the video.

yep yep.

for a stock intake fc, street car, the header seems to work just as well as removing the cats, i think the vdi works best with some backpressure.

You're thinking way too far into it.

The SDJ headers are the best ones out there but are more expensive than RB by a margin.

If you arent that serious i'd say just get an RB header a high flow 2.5" cat, and then have the exhaust shop down the road make you a cat back system with whatever muffler(s) you want.

This is for my ITS car.......yeah, it's serious.

SDJ has a track header, but I refuse to believe race headers are a one-size fit all approach. If anyone can show me the math on how they determined what their optimal exhaust components should be, that would be very much appreciated.

From what i've heard that there have been people who dynoed it and it made more power than RB.

I would ask chris ludwig, he lives near me, he has an its car and he's pretty damn experienced a local engine builder. He made 172 HP out of a stock ECU stock port car.

Should've said it was for an ITS car in the first place.

Personally, if it were me, i'd just run the header to a 2.5" XR-1 and then a straight pipe out the back, but what do i know.

I've spoken to him, and used his ECU. Granted his run was on an Rtek crammed into the Stock ECU box, the same that I'm using now.

Baseline two nights before the dyno tuning sessions at KDR: 155 whp

Baseline day of: 140-143 whp I forget exactly
After tune: 153 whp

It was really hot that day, and I wasn't really making any power, but the car, when the ambient temps cool down, definitely feels like a 170 whp car.

We went conservative and aimed for 13.5-13.6 on the AFR scales.

The reason why I posted this thread was that awhile ago, there was a debate betweem Nihilantic and Icemark on what the best exhaust for an N/A second gen was. I also remember seeing a thread with Aaron Cake stating that the RB Road Race was the best mass-porduced approximation of what an optimal exhaust should be.

I can't find any of those threads anymore, which is why I posted up this one.

Wanna bet?

There is more to it than just flow. If this wasn't true then the best exhaust would be no exhaust. Take the exhaust completely off of the car and drive it open port. While you are trying to cover your bleeding ears, you'll notice that the car doesn't make crap for power. Why? The exhaust isn't helping the air flow away from the port.

Fluid mechanics is a very interesting topic and definitely good to know. Realize though that we don't have constant flow in the intake or exhaust system. This isn't a river. We have pulses. Each pulse is a high pressure zone moving through the pipe. However it is followed by a low pressure zone behind it. We want the low behind it to help "pull" the next pulse through the pipe. Proper exhaust scavenging does this and we pick up power. We optimally want these pulses equally spaced which is why even length headers are best.

We also have another tuning benefit and that is acoustics. Your intake runners are acoustically tuned. Even VDI is an acoustic tuning method as the return pulses return at the speed of sound to the other side. Scavenging is dependent on rpm as faster gasses are at lower pressure. We don't want these waves to be at the speed of sound. Typically .6 mach is about the max we want. 2 totally different tuning phenomenon in the same place.

There is no one proper exhaust. Again as with everything, it depends. There are 3 basic systems out there. The first is the short primary system. This utilizes a header that collects conventionally near the engine. Typically at lengths of around 24" or so. Some are more, some are less. Varying the length changes the tuning. Technically the stock exhaust box is a short primary as the effective runner length is about 2" (which tunes it to about 7000 rpm or so) and then collects in a large open exhaust manifold.

Next we have a long primary. This is seem more often on 1st gens and some race cars. This is also a header system but it collects near the rear of the car before the muffler. Lengths over 100" are common. The powerband is different with these as opposed to a short primary. Some people say they make more power than a short primary and others say the opposite. If done properly they make more average power but again, it depends.

Tuning length is more involved than just making a pipe longer to tune lower. If we tune at one rpm, we also tune at it's harmonics. Let's say we tune a pipe for peak power at 8000 rpm. It is also going to benefit at 4000, 2000, and 1000 rpm's with lowering degrees of benefit for each harmonic lower. Between these rpms however benefit falls off and power can actually suffer. You can't have it all! If we tuned a long primary for a benefit at X rpm and then had a short primary tuned for max benefit at Y rpm, the fact that they are tuned at different rpms will allow one to outperform the other. What if the long primary were tuned such that it had a valley at 7000 rpm but the short primary was tuned for 7000? The short would make more power there. You get the idea.

The reason a long primary works so well is due to the number of pulses in the pipe at any one time. If we have a longer pipe, we have more pulses that can pull on each other. There is of course always a law of diminishing returns though so don't think going longer indefinitely is a good thing.

The last system is a true dual. This system does not scavenge. Why? You have to have a collected system to scavenge. I know I said that pulses pull on each other and they do. However it's the pulses from the other rotor that do so. Remember each rotor has some overlap to the other. As the pulse from one rotor is slowing down, the other rotor has opened up and occupies the same space as a part of the previous pulse. It helps pull it down the pipe. Think of this as drafting a semi on the freeway. If the spaces are too far out and don't overlap much, such as getting farther and farther away from the semi, the benefit is diminished.

If a true dual doesn't scavenge, why do they make decent power? Acoustic tuning! Yes acoustics can play that large of a role in tuning. Keep in mind the more overlap your porting has such as bridge or peripheral porting, the more critical it is that you have a scavenging collected exhaust system. Basically a true dual is nice on a stock or streetport, but anything more aggressive than this and you really need to pick one of the other 2 for max benefit.

BTW: My personal favorite system, and I've had all 3, is the long primary.

I could write all day on this topic.

So, on stock ports, you'd want to stay true dual? (Say......for an ITS-race car)

And for a ported 4-port NA engine with an aggressive street port (Think 13BT, RE, REW or 20B-REW with S5 NA or RX-8 rotors), you'd want to go with a long primary?

How does exhaust piping width/cross-sectional area and muffler tip size play into this?

Also, how does the Y-pipe play a role into this? Say you have a short primary, does the Y-pipe solely lower noise while maintaing flow? Is it that much of a restriction? Or for instance, you have a long primary system, but want to keep two mufflers. How far apart would you have to place the y-pipe? Would it make any sense to place the Y directly at the collection site? I would think there would be too much turbulence there.

I dont really know if this is pertinent, but I am in the same situation for piecing together an exhaust. I am going with a long primary. Manly after reading arroncake and rotarygod's comments.

I came to the conclusion of using the following.
True dual header
Magnaflow dual in/out pipe ( MPE-11385) at 2.25" as a rb presilencer replacement, 2.25" only because they dont make them as small as 2".

magnaflow Y-pipe collector ( MPE-10758) from dual 2.25" in to single 2.5" out right after the magnaflow muffler/presilencer.

Then a 2.5"straight pipe from the collector to a borla xr-1 ( BOR-40842S) 2.5" in and out. The borla will be the final muffler.

The car will be getting megasquirted with individual throttle bodies over the winter. and bridgeporting in about a year or 2.

All part numbers in parenthesis are from summit racing.

On a race car a single system has the definite weight advantage as well, only one pipe versus 2 means it's a lot lighter, which is a good thing.

You mean a short primary into a single muffler?

Look under any of Mazda's race cars of the 80s and you'll find a collected header. Look under any ITS FC that's run up front at the ARRC or anywhere in SCCA and you'll find a collected system. It might be a long primary or a short primary but they're all collected. We've used long primary with RB uncollected header, Speedsource short and ISC short. The Speedsource combination made the most peak power. The ISC was best up to about 6000. We haven't tried the SDJ header but they have a good reputation. The RB headers are widely considered slightly sub-optimal because of the sharp bends directly off the flange. We have made 170+ whp with them though so they are not terrible. Truth is that all the good systems end up with power curves that pretty much fall within a couple horsepower here and there of each other. Speedsource has the RD and wins to back their product so it's hard not to go that route. You guys can debate theory all you want but at the end of the day there are a couple products out there that are all pretty similar produced by smart people that are winning all the races. If uncollected worked they'd be using them.

I really don't want to get into the arguement but... True there is no one perfect exhaust. But all ITS engines (should) have the same port configuration, intake configuration, trans gearing, etc., etc., etc. So in this particular context there can be a generalization made that what works on one car does work on another. And what has worked for multiple ARRC champs and best in our testing is the Speedsource exhaust.

For the record, best we've done was 174 whp (Mustang dyno) on a stock ECU with the Speedsource exhaust. That engine is currently leading the Mid-Div championship. Our own car reached 172 (Mustang dyno) with the long primary/RB combo with better low end. That car won the '05 Cen-Div title. With our ECU we haven't seen much in the way of peak HP gains over cars that previously had the fuel pressure regulators optimized for peak HP. You can get any one particular 500 rpm range right by tuning a fuel pressure regulator and the CAS. The gains have been below 7000 and past peak HP where compromises are made with fuel pressure tuning alone.

Also, seems I remember you saying something about your 6 port actuators working or not working in a previous post. For an ITS engine you need to ditch the sleeves and actuator rods. The point at which those sleeves pay dividends is well below the rev range of a race engine. The flow restriction of the actuator rods is real. They need to come out for an optimum build.

Thanks for the input Chris, I know about the rods and am planning on taking them out.

I also have two FC's, one street and one for the track. The one that had stuck open actuators (since fixed) was the street one.

Okay didn't want to revive this but it seems like things got pretty silly. I checked Racing Beat's website. It says their header (which eliminates the cat) + pre-silencer doesn't make a signficant difference compared to stock manifold + cat eliminating pipe + pre-silencer on a fuel injected car. It only makes a difference if you also get their carbeurator. I saw another thread where someone was wondering why their header looked just like the Racing Beat header yet claimed such a huge horsepower gain (25-30, I think).

Racing Beat is well known for their quality, and their ability to design a muffler that is both quiet and powerful. Mufflers are usually a trade-off between the two; it takes a true understanding of flow and accoustics to do what they did. So if their header doesn't make a significant difference, I question whether others will. Especially when I see dyno sheets that are obvious fakes; the new horsepower curve is just a scaled up version of the old one, static and all. The gains don't diminish at lower rpms like they should. Or there's the RB look-alike mentioned above.

Anyway while I bet it'll add a tiny bit of power, headers don't actually add much.

So until I see actual mathematical equations or a dyno sheet, I call technobabble.

http://www.ludwigmotorsports.com/rx7/denton_dyno.jpg


There's your dyno sheet. Stock port NA, stock computer, stock intake manifold. Six port sleeves removed. Fuel tuning with fuel pressure only. CAS clocked for optimum timing. That's more rwhp than these things were rated new at the flywheel. Best guess for flywheel HP is 190-200. So we're making 30+ hp gains with exhaust and tuning. Exhaust is responsible for about 2/3 of that.

I've got a stack of these from a half dozen different cars. The exhaust is the largest bottle neck in the stock system by a long shot.

I know the exhaust is the most important thing to upgrade for power. I'm mainly questioning the headers and pipe lengths. The false dyno sheet I'm talking about shows a before and after with nothing changed but the header.

About 25HP can be gained from cat replacement and catback alone. With that and Ludwig's other mods, there's not much room left for significant gains from the header.

Anyone that says a header doesn't make much more power than a stock manifold has never owned one!

I'm just wondering why there hasn't been a header made by using similar angle bends as the Speedsource into a dual-piped Speedsource-type pre-silencer into a collector and muffler? Has Speedsource ever tried to make a long primary system to get the last few ponies in the lower and midrange part of the power curve?

Yes. At the ARRC, think it was 6 or 7 years ago, they ran 1-2 with Sylvain and David driving. One car had a long primary the other a short. They've done their homework. ;)

Interesting........

Dyno sheet or other test? Personal guess?

You don't need a dyno sheet to know a header works. We aren't talking about a slight "I think I can feel it" type of gain. We are talking about a "holy shit why didn't I do this sooner" type of gain. A header is VERY noticable assuming your car is running properly.

please delete