Exclusive to our visitors, we present to you the only online guide that both explains the concepts and calculates the exact specifications of the performance modifications that you need to
Boost your mileage +15% over stock
Download this FREE guide
(right click, save as)
Engine Performance Parts for the Mazda Millennia
April 25th, 2009 | 
Author: If you’re a car geek like myself, then some cars (although modestly powered) will still interest you because of the sheer amount of theory and technology jammed into them. One such car is the mid 90s to early 20s Mazda 929S (AKA the millennia, the Sentia).
The car is powered by a small but very efficient Miller Cycle 2.3 liter 60* V6. The engine utilizes the miller combustion cycle in which the horsepower losses associated with the first half of the compression stroke (when the piston is much closer to bottom dead center), by keeping the intake valve open. Obviously on a typical 4 stroke motor the open valve would allow the air from the previous stroke (the intake stroke) to escape causing a loss in power, but in the miller cycle the air stays trapped in the cylinder and is rather compressed to 2.0 atmospheres due to a roots type supercharger boosting the engine to 14.0 psi. Once the piston rises up in the cylinder bore to a more efficient location, the valve is closed and the piston completes the compression stroke on its own. The short duration of the compression stroke means that the use of high boost and higher compression ratios are possible and the final configuration with 10:1 compression ratio and 14.0 PSI of boost produces a peak of 210hp @ 5300 RPMs and 210 ft-lbs at 3500 RPMs.
The car was quite impressive for its time with a supercharged V6 capable of a 0-60 time of 8.4 seconds, a top speed of 142mph in a luxury sedan, all the while still maintaining a minimum of 20mpg in the city and as much as 28 mpg on the highway; quite impressive to say the least.
Introduction of the Lysholm/IHI supercharger
A cutout view of the Millenia Lysholm/IHI charger, notice the inlet restriction and the tiny throttle body.
Because of the space and packaging limitations on the front wheel drive V6 Mazda motor, Mazda outsourced the development of its supercharger on this car to well know Japanese turbocharger manufacturer IHI, who worked in conjunction with Lysholm Technologies – the twin screw supercharger manufacturer from Sweden to produce the unit used on the Millennia.
According to contact between enthusiasts and Lysholm engineers, the unit on the millennia is closest in specification to the Lysholm 1200AX which moves 1.2 liters of air per revolution and has a maximum operating pressure ratio of 2.2 or 17.5psi and a peak flow of 635 CFM or in other words 423 horsepower @ 17.5psi!
The main difference between the Lysholm 1200AX and the Lysholm/IHI on the Mazda is the design and compactness of the new housing, the design of the housing inlet on the Mazda looks like a significant restriction point (especially since the air passing through it is not yet compressed) and there is potentially a significant power gain to be found by upgrading the supercharger inlet to with a larger throttle body!
intercooler System
As we mentioned previously in our post on intercoolers, typical intercoolers have two main modes of operation:
1- Heat sinking, by maintaining themselves at a much lower temperature than the inlet air charge, and thus being able to extract a significant amount of heat out of the inlet air charger in a very short period of time.
2- Radiation, of the absorbed heat either by being placed in the direct path of the air stream or in a circulating coolant bath for air to water intercoolers.
An engine shot of the KJ-ZEM engine showing the small front intercooler and well as the ram air guide for the rear intercooler.
The Mazda uses two very small intercooler cores located on top of the engine inside the engine bay, and uses plastic air ducts to try and guide some air into the intercooler to radiate excess heat. However, knowing that their intercooler cores are not that large (for the power of the engine) and knowing that the supercharger produces more heat as its revolutions increase, then Mazda rather than fitting larger intercoolers or a front mount or fender mount intercooler with access to plenty of fresh air, have fitted their car with an intercooler bypass valve. The bypass valve, bypasses the intercoolers below 4000 RPMs and at low throttle openings to try and keep the coolers from being utilized and keep them as cool as possible, to allow them to work effectively (primarily as heat sinks because of where they are located) when they are needed above 4000 RPMs. I guess this kind of design works, but it’s definitely not optimum for peak power production, and for sure, when performing long repeated full throttle pulls back to back, the bypass valve will have no opportunity to operate and the intercoolers will have no opportunity to cool down. Eventually the intercoolers will heat soak and the power figures will drop from the advertised 203 hp dramatically.
Autospeed's dyno of the motor showing the sharp torque drop off at higher rpms...
If you think this is just internet theory, check out this article by autospeed where they have dynode the KJ-ZEM and the intercoolers heat soaked after a short 10 second run.
Based on the most conservative estimates based on the torque peak on that power run, if the torque peak is held to 5300 RPMs, the car stands to make at least 235hp @ 5300 RPMs which means we have at least an easy 35hp to gain on this car with some modifications.
In autospeed’s dyno run they made 196 horsepower with an outlet temperature rise of 75*C, the car also stands to make at least an additional 7 horsepower from better intake cooling.
Exhaust system:
As mentioned in our previous post about basic supercharger performance upgrades: One of the most important parts to inspect when looking to upgrade your supercharged car is the exhaust system. For the duration of time that the engine is in overlap (where both the intake and exhaust valves are open) then any exhaust pressure will work against your supercharger on a 1:1 basis.
How so? If you are in overlap and you have say 5psi of exhaust back pressure and 15psi of supercharger boost pressure, then the resultant pressure differential between your supercharger and your cylinder is only 10psi, and so your power boost from using a supercharger at that point is reduced from 100% down to only 68%. We can see this effect clearly on the falling torque readings on the dyno graph for this engine performed by autospeed. The loss of 35 potential horsepower translates on this setup to about 6psi of losses between the intake and exhaust system …
The 2-1 exhaust mid-pipe with the 180* bendand.
The exhaust system on this car combines short runner manifolds with no use of merge collectors, feeding close coupled catalytic converters, into a two into one mid-pipe with a power robbing 180* bend incorporated and no use of a nice y transition, into a single 2” exhaust all the way back.
Recommendations and power estimates:
I think it is possible to increase the power of this car to over the 250 horsepower mark with a few simple modifications:
1- Since I intend to add over 50 horsepower to the car, it is advisable to use 1 step colder iridium spark plugs for better heat management, and to assure proper ignition even with the denser mixture.
2- Replacing the tiny supercharger inlet and throttle body with a larger throttle body matched to the size of the charger housing, and redoing the complete intake system in that size. For a single inlet 250hp system an ideal supercharger inlet size would be a 78mm throttle body (or three square inch inlet) with a supercharger outlet of 58mm or 2.28 square inches. (Read more about supercharger porting here)
3- Because of packaging restrictions on this engine and how hard it would be to completely redo the intercoolers for a single front mounted intercooler with a single inlet and dual outlets (one for each of the 3 cylinder banks). I would use a proper water / methanol injection system activated at around 4000 RPMs and above 10PSI using two 1.69 gallon per hour nozzles, one for each 3 cylinder bank.
4- I don’t think anybody makes aftermarket headers for the millennia engine and it is a tight package of an engine. I would replace the closely coupled cats with straight down-tubes (again for 250hp the ideal collector size would be 1.87”) down either into a 1.87” to a 2.92” single exhaust with appropriate smooth transitioning y-pipe) or using a 1.87” x-pipe (for even better exhaust extraction and reduced back pressure) to a true dual 1.87” exhaust.
The cats would be a single (or dual for a true dual exhaust) cat placed after the y-pipe (or the x-pipe if a true dual is possible) and then feeding back into a high flow 2.92” muffler (or dual 1.87” inlet mufflers).
Using a better transition such as a y-pipe or x-pipe has the added benefit in that the different cylinder banks on a V style engine fire 180* out of sync to balance the engine out. The advantage of smoothly merging the exhaust gasses from the two banks together is that when using a smooth merge, the high velocity exhaust gases caused by a combustion in one bank, leaves behind it a temporary vacuum (or low pressure reflection wave) of as much as negative 2 PSI at the exhaust valve of the opposing bank. As soon as the other bank fires and the exhaust valve opens, the exhaust finds vacuum (rather than back pressure) in the exhaust runner which helps evacuate the cylinder faster leaving more room for fresh air in the next cycle and improving scavenging during overlap.
The total potential gain of a properly designed exhaust, with a ported supercharger inlet, and better charger air cooling would be about 47 total horsepower gained and boost may drop from 14psi to as low as 10psi possibly.
How is that for supercharger performance? 47 more horsepower at 4 less psi!
For more information and references:
Lysholm 1200AX supercharger
The miller cycle engine explained
Autospeed article on Japanese engines
YouTube Video of intercooler bypass valve actuator
1996 Mazda Millenia S statistics at the auto channel
Related posts:
Posted in 
Tags:  | |
 | Find the right combination of intake, shorty header, and exhaust for your engine using the lite version of the power calculator... |
The comment that there is a power loss from exhaust backpressure during overlap is clearly wrong. The engine isn’t making any power from the overlap portion of the stroke. The power loss is from pumping the exhaust gas out and the gains (during non-combustion portions of the cycle) are from pressurized intake stroke. The overlap period and exhaust backpressure are related, too little exhaust backpressure will result in charge loss into the exhaust, which results in more volume of air pumped with no increase in power, thus the resultant is a loss. If you have a lower backpressure exhaust a slight reduction in overlap timing would correspond. But it isn’t something we change often.
The real reason low exhaust backpressure robs power from the supercharger is that while the supercharger pushes one piston down on intake stroke another piston has to force gasses out against the backpressure. So the backpressure robs the power through the crank, not overlap. And all of this discussion is assuming constant pressures and steady flow, which are clearly not the operating conditions. While this explination is simplified yet correct the overlap is more about flow than about pressure differential in this engine.
Supercharged cars are not signifigantly different than N/A cars in relation to exhaust modifications, whereas turbocharged cars are signifigantly different reguarding pressure differential and overlap. Supercharged cars do have a high specific output and do require high flowing exhaust components near the engine, but once you are downstream of the primary collector the supercharged car is much like the N/A car. Larger exhaust piping will result in lower flows at low speeds and higher flows at high speeds. You will not gain low end from oversizing your exhaust as you would in a turbocharged application.
Also, there have been aftermarket headers, pre-cats, collectors, and main cat components for over a decade, though they are not a signifigant improvement due to space constraints and design restrictions. The O2s are finicky.
One more thing. The reflection wave comes from the tailpipe exit, but its not just a high pressure wave or low pressure wave, it’s entirely length, pressure, and rpm dependant resonation. Very well researched and has nothing to do with smooth transitions. The pipes could both dump into a box at horrible angles and they would still have pressure wave resonance. Look into helmholz manifold design, and realize that it is much less effective on exhaust. Smooth transition and equal length transitions are more about flow and velocity. You want to maintain velocity by having smooth transitions, and you want to balance size of pipe versus volume of gas so that the pulses dont compete with eachother at the junction.
boost won’t drop from 14 to 10 and increase power without head porting (and adding aftermarket bypass control valve, obviously over the heads here). 50 HP from intercooling improvements isn’t far fetched, but exhaust will never get you farther than 10 HP, and it won’t be broad banded, same with the increased throttle size. Both will cause low end loss and driveability issues. The car is well designed and I make almost the same amount of power increase with CO2 injection on my intercoolers. And my mod doesn’t cost me gas mileage or rob my low end.
This write up is full of technical inaccuracies, but I’d like to see it developed properly.