Your Cart is Empty
On the quest to make more power efficiently, we decided to jump into making a new product range for the rotary applications. We knew upfront this would not be easy, as the intake tract is quite sensitive. There is a common misconception that with a turbocharged engine intake tuning doesn't matter as much. There is still a pressure wave that forms in the intake tract as the rotor passes by opening and closing the combustion chamber. Unbeknown to most people, its quite easy to make a mistake on the intake system, and really hurt the powerband. Our interests were really perked up back around 2010 when we did our first Turbo Rx8, as one can control the intake duration with the servo motors( both opening and closing the aux ports, and varying primary runner length)
Below you can see a back to back dyno test we did, and the difference was astounding. The only difference is was opening and closing the 5th/6th port runner sleeves. So this is basically going from a like stock mort to a HUGE raceport( technically more duration than one can even port a standard 4 port 13b of any type) There is a reason why Mazda made such a complicated intake tract on the Rx8, similar to how complicated the valvetrains are on modern engines. Mazda would not have spent the level of R&D, plus huge levels of manufactoring costs for nothing, and you can see it clearly works.
We've set out to make an intake manifold that would make more power in the midrange, and top end, while sacrificing some low end, but retaining proper throttle response/idle functions. Rarely are performance enthusiats worried about sub 3k rpms on a rotary engine, but they do care about idle control circuits and not having the engine idle at 1500+ to not stall out. To achieve these results, many variations have to be dyno tested to see what actually works.
The first part of the process after 3D scanning the OEM components, is to perform a CFD test( computational fluid dynamics). This is a great tool, but there are too many unknown variables that you need have to guess at, which makes the results inaccurate. Additionally we used a standard flow bench to baseline the OEM parts, just to get as much data as possible. Below is the Rx-8 upper intake manifold.
We 3D printed our first test Rx8 model using Selective Laster Sintering technology. Standard 3D printing cannot produce the level of detail we needed to properly dyno test, and making this model out of billet aluminum would be extremely time consuming.
With V1, we only wanted to change the AUX port runner entrance, zero other changes( including runner length, and volume). We wanted to isolate each change, for each test to really fine tune the best model possible. One would assume this unit would make more power since it flowed better in the software, and on the flowbench. We did a proper same day back to back and this actually lost 10rwhp. We quickly realized this was going to be lot more difficult than anticipated.
We shifted our focus back to the Rx7 models as we could develop and test those much faster than the Rx8, and then use the Rx7 data to improve all other planned rotary intake manifolds( Rx8, 20b etc).
We previously have hand built some Rx7 units, and modified a lot of V8 stuff, so again we had a base idea of what to do for the 2 rotor components. There is considerable data available with piston engines, and while it doesn't 100% translate, its still a good base. Key factors are generally, air velocity, not volume. Putting too large of headers on a V8, or too large of intake valves absolutely destroys piston's low, and midrange power, while only adding to the very last 500rpms. A lot of those shiny billet intake manifolds are extremely flawed this way, and there is a reason you never seen any back to back dyno charts.
The design criteria for the Rx7 intake system had to fill mutliple purposes;
The first 3 points are all covered with one main component; Air Velocity. Instead of oversizing runners to gain volume, its best to taper the runners to not just gain volume, but more importantly increase velocity. Similar to how a turbine housing tapers to speed up a turbocharger. We have also found this is the best way to design a turbo manifold, and why we incorporate this into all of our cast turbo manifolds. This approach is less common, as it requires a CNC or casting, it cannot really be done with sheetmetal/tubing hand fabrication.
Adding velocity stacks( raised or flat) at runner entrances is also proven way increase to increase laminar airflow. This reduces dead air pockets at the entrance corners, and increase the airflow several percent.
Reducing the number of bends, and increasing the radius of the bends was also a major focus. There are many theories on plenum volume, but again if you go too large you can hurt throttle response, as the incoming velocity slows greatly when there is a huge volume change. However too small of a plenum and it will hurt peak/high rpm performance as choke point, but more importantly hurt air distribution. Proper air distribution is critical for not just engine health, but throttle response and power. Simply put having all rotors run the same AFR makes more power, similar to a runner who has the same size shoes on both feet.
We CNC'd the entire test model instead of 3d printing, mostly for fun. Incase anyone is wondering this took near 100 hours to do between splitting the model up, making the cnc program, cutting it, welding it, resurfacing etc. Not cost effective at any level.
This model was then back to back indepentenly dyno tested by IRperformance in NJ. The test car was a low powered pump gas only, mildly ported engine making 340rwhp. The car jumped 23~ish rwhp with the UIM only. They were also able to verify the idle control system worked just like OEM.
This test unit was moved around and repeated on a few other cars. Please with the consistent results we started the mold production, and moved onto having some more samples made to further back to back dyno test. You cannot just test 2 cars to be absolutely sure the results are correct, its easy to overlook other variables.
We had about 10 different customers verify gains, many blind without us being present to help refute any claims the data was being falsified. This particular customer in NZ tested the unit on two different cars. 
Dynos are great, but we also wanted to verify real world results. One of the test models was sent off to Goopy Performance in NY. Their test model is a semi pport 13B 6 port engine that had been previously running an OEM FD upper intake manifold with the interior divider removed along with a Ford 90mm throttle body. This car has been running without major changes for 4 years, and done 100s of passes. Erick Turbo installed our prototype UIM only, with zero other changes. The car picked up 3 mph and dropped 0.3 seconds off their 1/8 mile, setting a new record for the car.
Upper intake manifolds can be found here
Now that we were sure we had a the upper intake manifold product correct, we moved forward with full production, and worked out the last few design tweaks.
The next area of product focus was to move onto the lower intake manifolds. The main focus on the lower intake manifolds was not so much to pickup more power, but to equal airflow between the rotors. The OEM lower intake manifolds on all mazda rotaries are not equal, because of fitment restraints with the oem twin turbos, and water pump components etc. The factory simply made up for this by over enriching the engine. Mazda has to balance everything here, and ultimately peak performance isn't their main concern.
The oem lower intake manifold seen above in black, distributes air unevenly to the front and rear rotors. Same goes for the 20B. One major problem is most tuners are generally using a single o2 sensor that is the collection of all rotors. Ff the rear rotor is running a 11.8 AFR, and the Front Rotor an 11 AFR, your o2 sensor will show you 11.4 . You then assume the AFR is ok, and later on end up blowing the rear rotor and you don't know why. Well the lower intake manifold is a big part of this problem, and why people tune so rich to mask this. The more rotors you add the worse it gets. Simply bolting our lower intake manifold on proves this, as individual EGT sensors even out after installation. Evening out the airflow, and balancing AFR also increases throttle response. We used to do in real time with a maonometer on individal ITB/carbureted cars all the time.
Now that the was fixed, we wanted to try and improve power and function. We noticed the OEM lower intake manifold runners have choke points from to clear the OEM twin turbos that are stuffed right against them. See the photo below comparing our lower intake manifold vs the OEM FD3S ( black)
Also the OEM unit has protrusions elsewhere for road worthy emissions testing. Our unit removes this, as this is a off road race component( not legal for street use). We also removed the pre heating of the intake by removing the EGR passages( and coolant passages on the FC3S models). This cleans up those eyes soars for the customers concerned with aesthetics.
Lastly we wanted to improve the fueling side of the lower intake manifold. Mazda runs staged injection and the lower intake manifold houses the secondary injectors. The OEM lower has 1 extra injector per rotor for the 13B/20B cars. This is also where mazda made an improvement on the Rx8 engine, with a 3rd stage of injectors. Its very common in drag racing to see fuel injectors placed very far upstream, and the same goes for methanol auxiliary injectors. In our experience this makes more power, as the fuel has more time to mix properly with the air, but only when air velocity is high enough. This is why mazda has staged injection, and not all the injectors are sitting in the lower intake manifold, primaries are right in the center plate. Furthermore we learned when testing the old Adaptronic modular Rx8 PNP, that you could pickup about 4-5% more rwhp with simply changing injector staging. So we wanted to not only add 1 extra injector spot per engine port( 2 extra injectors for the sideport models, and 4 extra total for semi pport), but also change the injector position. We really didn't want to have to include a fuel rail and isolate customers who already had an upgraded fuel rail, but this made the most sense to have a unit that would make the most performance impact possible. We also were able to remove the need to use an lower injector adapter, which cleaned things up nicely, not to mention those lower sleeves sometimes leak under high boost.
Lastly we wanted to use O-rings at the flanges, we determined this was only possible on the UIM/LIM flange, as we could not run a small enough o-ring groove between the primaries, and some people are porting these larger for 13BRE engines etc. 
Although we have done enough isolated back to back testing with the lower intake only, dyno results are showing zero downsides, while extending the powerband up another 500rpms, and 3% power gains. We do want to note as we get asked a lot, you do NOT want to remove the primary injectors on these engines unless its a drag only car. Removing the primary injectors will hurt low speed ( sub 3K rpm) driveability, and throttle response.
We will update this blog as we finalize this testing on the lower intake, and 20b/Rx8 etc.
FD3S lower intake models can be found here
FC3S lower intake models can be found here
20B 3 rotor lower intake manifolds can be found here
If you are in the NZ or AUS locations please visit Pac Performance , our stocking distributor for those regions.
Stay tuned for more 20B and Rx8 updates!
Comments will be approved before showing up.
