When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Just laid this out for someone else and figured I’d finally make my own peace on this subject. Anyone that wants to argue against it otherwise then go for it, but without my participation. Reasonable questions and such I’ll try to address. This is a generic, simplified, layman explanation. I don’t want to get into all the actual physics, absolute pressure, and so on.
In a nutshell, modern exhaust header theory is based on engines with port/valve timing overlap where both are open to the chamber at the same time. This is important because pulse theory dependent on stored and actual fluid momentum. If you have pressure in a closed chamber (stored energy) and you open a port to allow it to exhaust it will just blow out until equalization occurs. You can’t flow out anymore than equalization without beginning the vacuum process, which requires a lot of energy to accomplish. In theory if the pressure was high enough and could be released quickly enough it could pull some small magnitude of vaccum with it before bouncing back, but the reality is the magnitude is so small as to be irrelevant from an engine oerformance impact.
On the otherhand, if you did the same as above except at a certain point in the pressure releasing stage opened a second port from a clean air source that also has stored energy (atmospheric air pressure) at just the right time when the fluid exiting has momentum of rushing out the first open port, that combined momentum has the potential to pull in additional clean air from the second port with it to help flush out any contaminants etc. thst was originally in the chamber. The timing between when the two ports open and close is important especially if there are cycling frequencies (rpm) between the events that affect where the best efficiencies occur. This is the basis of all current pulse exhaust theory, variable valve timing, etc.
Now enter the Renesis, an engine unique among others. As supplied from Mazda, the Renesis has no port timing overlap between the exhaust and intake events. In fact there is a 6 degree period between when the exhaust port is fully closed and the first intake port edge opens. I suppose you could call it negative overlap, but stating zero or no overlap is sufficient. I’ve seen a few people on here mistakenly claim they added “overlap” by bridgeporting etc., but unless the intake port edge extends past the exhaust port edge relative to the rotor tip position as it passed by that area then there is no port timing overlap.
So my philosophy on the Renesis engine is that pulse theory for header construction is a big waste of time and essentially proved it by building a racing manifold that went against pulse header theory in every regard back in April, 2006. As you can see below, nobody would marvel at this thing. Essentially I took the OE exhaust flange, put a 2” OD elbow on the front port to turn it back, then had a cone going from 2”OD elbow to 3” OD at the last port, angled the cone slightly down to place the top cone edge near even with the top edge of the middle and rear ports, and then short 2” tube stubs straight into the top of the cone. From just behind the last port to the rear mufler area is all true 3” exhaust.
Everything was done using 18 & 20 Ga. T321 mandrel bends and tube. Even with the full OE flange it was light at 7 lbs total weight. I finally decided to retire it and build a new header. The only issue with this manifold was that even though I used strong T321 high heat material, the thin Ga. coupled with it being so restricted for heat expansion made fixing stress cracks an annual off-season ritual the last several years.
Yet as a Cobb Tuning beta tester with a 2005 original factory engine with only bolt-on modifications for an emission-legal racing class (exhaust, intake, hi-flo cat, OE coils(!), ac delete, Speedsource pulley kit, center muffler & (2) rear canister mufflers) we recorded this dyno in 2007:
Which you can compare to this extreme racing build thread that got so much attention:
So I was having a discussion with another member about building a header using the no overlap thought process and here it is:
—————
IMO the best path, which is also the easiest, cheapest, & lightest path, to get most of the performance benefit with the least effort would be to use at least 1-7/8”, but preferably 2”, OD primary tubes, and without any concern for equal length etc. bring them into a pro style 3-1, 12 degree, merge collector as soon as possible and then run true 3” piping from there to the back into whatever muffler system you have/want with a 3” resonator/muffler/auger-style insert in the tunnel area. Most RX8 midpipes have the OE style flange to the muffler pipe, which usually reduces the 3” pipe to 2-5/8”. That should be removed and replaced with a true 3” flange or my preference is slip joint with a clamp; light, easy, works fine. Essentially you can’t make the tubes too big, just at some point bigger serves no useful purpose. Most headers are 1.75” OD in 16 Ga or 14 Ga. thickness, which is only 1.63” ID or less (Racing Beat); too small IMO. 2” 16 Ga is 1.88” ID, while 2” 20 Ga. would be 1.93” ID. It doesn’t seem like much, but the total flow area increases exponentially.
Using 16 Ga T304 tube will be strong and lowest cost, but heaviest. 18 Ga. T321 is a good compromise between cost/weight and stronger. 20 Ga. T321 is lightest, strong enough but higher cost. 22 Ga Inconel 625 if you want to go all out F1 style. If you can live with the cast RB flange it comes machined for 1.75” tube on the ID insert, but the OD is round 2” diameter and you can buttweld directly there with 2” tube, but before doing that port grind out the ID to match the 2” tube ID perfectly. That was my intention until the baller engineer in me said it’s too easy and boring, lol. Another thing, butt welding tubing requires purging the tube interior with gas to get a proper strong weld. One way around that is to buy slip tube pieces in short lengths with the same ID as the tube OD. On straight sections you can put the slip piece half-way on, tack the outer lip, slip the mating tube into the other slip end, tack, and then you have two outer lap welds to make rather than a single butt weld. It will be easier and as strong or stronger than a proper purged butt weld. If the tube interior doesn’t have purge gas though you have to watch your welding heat and not burn into the main tube ID or it will sugar up in oxidation. You don’t need that much heat. If you can get slip rings in 1/4” wide sections you can make it work on bends/elbows too by cutting the up in pieces, tacking and hammering them lightly into place to conform with the curvatures. It’s some work and won’t look as nice, but then your only other alternative is to purge butteeld or have it done by someone who can if you can just get it all tavked up and take it to a proper welding place b
If you don’t need the AIR emission part after building the header as a RB single flange you can then go back with a cutting wheel etc. and cut as much of that off as you can and also the center bars between the flanges so they’re separate. If you use slip connections into the merge collector with bee-hive springs or bolted tabs you can then have three separable pipes and a collector to help allow for heat expansion and easier install. Do some careful grinding on the front tube flange and then you should be able to install/remove it without having to lift the engine to remove the motor mount bracket. Your life will be much easier there.
It could look similar to this with a better merge collector; nothing fancy, no bundle of snakes, no equal length blah blah blah, etc.
But I didn’t go the easy route because it was, well, too easy. Instead I took a shorty header from Japan with 2” 18 Ga T321 primary tubes and modify it into long tube header using 20 Ga. T321 material, but not for performance. I just wanted the pipes, bends, and slip joints to provide max degree of freedom for heat expansion. I don’t expect it to make even 1 HP more than my original race manifold. This is actually just intended as a prototype for a lighter version if I decide to go yhere.
In summary, no overlap = no exhaust tuning. For all intents and purposes I view the Renesis as a slightly supercharged naturally-aspirated engine. All the pulse tuning is on the intake side to provide a slight supercharger effect at certain rpm ranges. Exhaust tube length, diameter, etc. is mostly irrelevant. All you can do is make it as free flowing as possible to get what ever you can out. That’s why I recommend going straight to 2” OD tubing, smoothly collect into 3” as soon as possible, and then have 3” back to the rear muffler area and then split into two 2.5” or a single 3” with some type of center muffler/resonator for the most benefit at the least effort. I’m revamping my rear exhaust muffler system to the next evolution too and will post that up at some point too. Pretty much just taking what most people are already familiar with now and evolving it to the next level with basic common sense mods.
You mention transitioning the 2" pipes into a 3" collector as soon as possible. Is this because from your experience, one 3" pipe is capable of flowing more freely than, say, keeping three 2" pipes going all the way back?
three 2 inch pipes will flow more than 1, 3 inch pipe, but you dont need three individual pipes as at any given milisecond in the exhaust the amount of air added is still limited to the 1 2 inch runner, in reality or theory at least, if you where to keep individual exhaust ports you might see more torque outta the motor but its a trade off with weight, convenience and practicality. under a stock intake and fueling I believe the factory 2.5 inch piping to be more than enough. when you start messing with intake flow, exhaust and fueling you get to bigger demands.
three 2 inch pipes will flow more than 1, 3 inch pipe, but you dont need three individual pipes as at any given milisecond in the exhaust the amount of air added is still limited to the 1 2 inch runner, in reality or theory at least, if you where to keep individual exhaust ports you might see more torque outta the motor but its a trade off with weight, convenience and practicality. under a stock intake and fueling I believe the factory 2.5 inch piping to be more than enough. when you start messing with intake flow, exhaust and fueling you get to bigger demands.
I agree it would be an exercise in impracticality for it to extend all the way out the back, and Team says this design is meant to be easy and cheap for those making their own headers. But if the conclusion of having headers vs a stock manifold for this specific application is to let the exhaust just flow more freely, why not just make a long tube header? I guess it just comes down to how much pipe are you willing to take? hehehe
You mention transitioning the 2" pipes into a 3" collector as soon as possible. Is this because from your experience, one 3" pipe is capable of flowing more freely than, say, keeping three 2" pipes going all the way back?
It will cost more in material, labor, be a lot heavier overall, and serve no useful purpose. That’s what makes my original manifold so light. The center pipe could be smaller diameter and I considered it numerous times but it’s not worth the bother because then you need to smoothly merge this smaller pipe with ttwo larger pipes. You could though. I had several wild looking merge collector concepts dreamed up for doing that, then came to my senses.
The one negative for me on my tubular prototype is overall weight. Despite eliminating the cat converter from the previous manifold/cat pipe assembly the tubular assembly weighs more overall. I understood that going in though, which is why it’s just a prototype. I laid out the best path in the main post, but didn’t follow my own advice for personal reasons. For anyone else that will be the cheapest, easiest, lightest, and best overall bang for the buck.
Why not build a long tube header? Well why not throw cash in a hole, put gasoline on it, and set it on fire? That’s what I did, but again there was a personal purpose for doing so. Knock yourself out, keep telling yourself it must be better cuz feelz, etc. If you can’t build it yourself then be prepared to pay, not cheap, but some exhaust shop might cobble something suitable for less. What I proposed is within the grasp of some capable people in their garage. A true merge collector is the most expensive piece. I don’t recommend going cheap there unless you can get the same thing for less. Sharper angle, more turbulence, etc. there has some performance cost imo.
This is just in general. More bends = less efficient.
Hot air = more volume than cold air
Larger exhaust = louder
smooth transitions = better
An ideal exhaust would be something that can move the air with the least amount of bends sized to the volume it needs to move. so the larger diameter pipe will be close to the rotors and slowly get smaller as the exhaust cools and contracts.
This works well with my turbo car who has a short small tubular manifold into the turbo to house all that pressure/flow and goes into a 3.5" downpipe and necks down into 3" open exhaust. straight through cylinder mufflers seem to have low backpressure.
N/A doesn't produce as much pressure/flow and can use smaller diameter tubing. smaller tubing = higher flow, larger tubing = slow flow.
I have not done any testing with headers so I don't know what to prioritize (flow vrs. volume). tough to say what is better.
probably best to transition from port to header smoothly and size it the same diameter.
Your picture of the eBay special header is precisely the one I bought to modify. It has 2" primaries. I'll cut off that awful collector and put in the nice one I just bought from Stainless Headers Mfg.
Thanks a lot for your advice, and for typing it up for prosperity in this thread, Mark!
I’ve seen a few people on here mistakenly claim they added “overlap” by bridgeporting etc., but unless the intake port edge extends past the exhaust port edge relative to the rotor tip position as it passed by that area then there is no port timing overlap.
wut ? exhaust and intake open at the same time = overlap
maybe what you mean is :
useful overlap = when exhaust port stays open after top dead center whilst intake port is also open .
The only way for a rotary is to have overlap is for both the intake and exhaust port to be open to the same rotor face chamber at the same time. Since all the ports on a Renesis are on the side plates then the only way for that to occur is for the exhaust and intake side port to be on the same rotor chamber opening at the same time as the rotor tip/nose end sweeps by. That means extending the port a minimum 6 degrees of opening just to get them to zero overlap, then the additional sizing extension needed to such that the area and volume is both sufficient, efficient, and most importantly shaped properly enough, to have a useful function. How it’s shaped greatly affects whether it closes sharply or bleeds; nobody seems to understand that from what I’ve seen posted for pics. Which is why they haven’t proven to be useful imo.
Yes, that’s true and I can see how the way I worded it might lead to someone not understanding that point, so thanks. It was a long post and just trying to keep it simple yet enough detail for a general understanding, not teach anybody fluid dynamics lessons.
You can see on my tubular header that I can remove the three middle pipes and run the collector there instead with a 3” OD tube spool piece behind it with O2 bung if I wanted to run it shorter. However the weight difference between those three 2” 20 Ga. pipes and a single 3” 20 Ga pipe for only a foot of length (300mm) is not worth he bother. I
Even if someone wants to think equal length for all three pipes still matters you could still fit that in the way I suggested in the engine bay/manifold area and then quickly bring then together, then to 3” out back.
In summary, what I view as the OE restrictions for max power are:
Overall piping sizes. Frankly, if you took the OE manifold and cut it at the back past the second rotor port and ran 3” all the way from there you have pretty much a max performance exhaust system. The manifold to cat/midpipe flange opening is on 2.5” ID or so. The OE cat isn’t bad, but it won’t flow like an HJS hi-flo cat and is a lot more fragile for clogging and such. Then the cat/midpipe connection to the rear muffler pipe is also in the 2.5” opening size, even on many 3” midpipes Then of course is the rear muffler system.
I would think there was still some benefit to minimising the pulse interference even if there was no scavenging effect.
Equal length front and rear pipes with the siamese of unequal length makes sense to me - which is pretty much what you have done in the above design.
That’s a fair assessment, but often in practice they’re never truly equal and then also the crossover turbulence of the un-separable siamese port makes that a moot point imo. I’d refer you to take another look at my original racing manifold design and compare it to your comment.
That’s a fair assessment, but often in practice they’re never truly equal and then also the crossover turbulence of the un-separable siamese port makes that a moot point imo. I’d refer you to take another look at my original racing manifold design and compare it to your comment.
Perhaps there is a difference when the pulses are into pipe that is already 3" vs pulsing into a merge collector ?
Feel free to explore the idea ... there would be potential for possible energy conservation, but whether it proves to be a measurable reality is usually where sound ideas in principle might often don’t prove to be much where the rubber hits the road ...
Perhaps there is a difference when the pulses are into pipe that is already 3" vs pulsing into a merge collector ?
ahhh Helmholtz Resonance now you are thinking with physical science, that said I just learned the ren uses porting before tdc so... my input on this thread has come to a limit haha
Smaller ID tubing got us to spool our turbo's more quickly and made a lot more low end power without really effecting top end. I know this is N/A, but piping diameter means something, atleast on the 13B-REW with a turbo. it spooled up the turbo quicker.
I would be curious to see where the actual restrictions in the exhaust are.....is it piping ID, exhaust bends/flow, muffler? etc.
Smaller ID tubing got us to spool our turbo's more quickly and made a lot more low end power without really effecting top end. I know this is N/A, but piping diameter means something, atleast on the 13B-REW with a turbo. it spooled up the turbo quicker.
I would be curious to see where the actual restrictions in the exhaust are.....is it piping ID, exhaust bends/flow, muffler? etc.
Where is the exhaust the least efficient?
Except you’re making the same mistake everybody, their mother, and their brother was and is still making; you assume the Renesis must be the same as what came before. Keep grasping, strawman ...
If I build this 4-port turbo engine I plan to use 2” primary tubes. Contrary to the big turbo and force the engine to try and push it with restrictive primary tubes, my Renesis counter-philosophy is smaller highly-responsive turbo and let the engine breathe ...
If I build this 4-port turbo engine I plan to use 2” primary tubes. Contrary to the big turbo and force the engine to try and push it with restrictive primary tubes, my Renesis counter-philosophy is smaller highly-responsive turbo and let the engine breathe ...
That’s for a different thread though.
The big pipes wont help you ........... even if you can get them to fit (no mean task).
12-29-2017 | 01:02 PM
TeamRX8’s Renesis Header Theory Thread
wut ? exhaust and intake open at the same time = overlap
Then of course is the rear muffler system.
