This is the external view of the KL03 2.5L VRIS intake manifold. Other variations of the 2.5L V6 have slightly different manifolds in performance and tuning, but the fundamental theories behind their operations  are the same. In the above picture, the front bank intake runners are clearly visible. They are attached to an intake plenum which is in turn attached via a 'U' tube to another plenum (hidden) that feeds the rear bank of cylinders. The easiest way to understand how the VRIS works is to use the following rule of thumb: treat tubes as masses, and large volumes (such as the plenum and cylinder) as springs. The above diagram does just this  -- the intake runner and entrance into the plenum are labeled as masses A and B respectively. When the engine is operating, a cycle of events causes the "masses" and "springs" to bounce around. At certain engine  speeds, the masses and springs act in unison to cram more into the cylinder just prior to intake valve closing... Intake Resonance The whole process leading up to the point of intake valve closure starts off with the cylinder sucking air in. To better visualize this, pretend the hand on the left represents one of the cylinders on the left. When the  cylinder sucks during it's downward stroke, picture the hand pulling Spring A to the left.  By pulling Spring A, Mass A will move to the left also, but with a slight lag. With Mass A moving to the  left, Spring B will begin to feel a pull to the left. Spring B begins to stretch and exerts a pull on Mass B, however just as with Mass A, Mass B will lag the movements of the spring.  Picture the hand coming to a stop after having moved to the left as described above. This is analogous the piston reaching the end of its intake stroke. The intake valve is about to close, but what's happening to our  masses and springs? Depending on the strength of our springs (spring constant), and mass of our masses many things could be occurring. Ideally, and for maximum intake charging effect, Mass A and Mass B  should still be travelling to the left, and just about to come to a stop. Just as Mass A and B are about to bounce back to the right (due to the action of the compressed springs), the intake valve should close. This  scenario where Mass A and B are simultaneously ramming themselves into the cylinder is referred to as the primary resonance frequency (more on this later). When the engine is operating at a non resonant frequency (engine speed being analogous to the speed of the hand moving left and right) the following may occur. Suppose Spring B is very stiff. If Spring B fully  compresses and begins to expand before the intake valve closes, Mass B will begin moving to the right. Mass B will no longer be contributing to the charging effect on the cylinder. In fact Mass B will be  decreasing the charging and volumetric efficiency of the engine.  The fact that the resonance frequencies in the engine help only at certain engine speeds means that to make  an engine with a broad power band requires some sort of variable induction setup... VRIS! VRIS Here's a simplified representation of the VRIS intake manifold (for those that know about the 2.5L KL03  VRIS manifold, it has two "Long Resonance Tubes", not the single one that is shown above). The "Short Resonance Tube" in the diagram represents the 'U' tube that connects the two plenums. The purpose of the short and long resonance tubes is so that while the engine is changing speeds (i.e., acceleration) the mass of Mass B can be changed on the fly so that the primary (and secondary -- more on  this later) resonant frequency can be adjusted to coincide with the engine speed. It's like being able to swap manifolds while you're driving. By changing the mass of Mass B, it now bounces back and forth at a different speed. If Mass B is too heavy, it will lag the movements of Spring B too much -- the intake valve will have closed before Mass B  can even react. If Mass B is too light, it will begin to bounce back to the right before the intake valve closes. If Mass B is just the right mass, it will reach the limit of its leftward travel just as the intake valve  closes.  The butterfly valve shown above provides a shortcut to an air supply for the cylinders -- the other bank.  Instead of drawing on the "heavy" column of air in the long resonance tube, air is drawn from the other cylinder bank through the 'U'-tube or short resonance tube.  As briefly mentioned before, in the KL03 VRIS manifold there are two long resonance tubes leading into the manifold from the throttle body. The dual butterfly valves just after the throttle body select which  passages the engine breathes through (the narrow and long passages, or the large passages + the narrow ones). The long narrow passages are represented by a Mass B of high mass, and the free-breathing large  passages + narrow passages are represented by a lighter Mass B. At higher engine speeds the free-breathing passages are open and the resonance frequency of the intake is thus raised to coincide with  the higher frequency cylinder pulsations.  Primary and Secondary Resonance Frequencies The spring and mass diagram above explains the differences between primary and secondary resonance  frequencies. How do these relate to the our VRIS? Let's look at the various stages of the KL03's VRIS operation... 2.5L KL03 VRIS Operation At 0 to 3250 RPM, the engine breathes through the resonance tubes L1 in the diagram below. Valve A and Valves B are closed. A primary resonance occurs at approximately 3000RPM. Past this engine speed,  the resonance charging effect diminishes -- Mass B is too heavy and is lagging too much. At 3250 to 4250 RPM, valve A opens. Now the engine has a shorter breathing route. Mass B has been  lightened and the primary resonance frequency has been raised to about 4000 RPM. Past 4000 RPM, power again begins to fall off. Another resonance modification needs to be made if volumetric efficiency  and torque are to be maintained.  Although the engine gets some of its air via resonance tube A, air must still get into the manifold somehow:  through the narrow L1 passages. Mass B can still be lightened by bypassing this passage with a shorter one. And so, at 4250 to 6500 RPM, Valves B open. Mass B is now at it's lightest. A primary resonance  peak occurs at about 5000RPM. Past this speed, torque begins to fall off rapidly. However, there is a problem. There are no more VRIS configurations that will yield a primary resonance above 5000 RPM.  The only alternative is to use a weaker secondary resonance to boost volumetric efficiency.  At 6500, the VRIS butterfly valves return to the same configuration they had at 0 to 3250 RPM -- Valve  A and Valves B all closed. The secondary resonance is not an ideal tuning setup. The charging effect is not very strong, and the narrow L1 intake passageways are restrictive to high speed air flow. What can be  done to fix this? Well, one solution is to move the existing primary resonances higher up to try and fill the "torque void" past 5800RPM. More on this later.   Mazda experimented with many different variables in the design of the intake manifold to achieve optimum  power output. The above chart illustrates their findings using different length passageways. This chart illustrates that there are many ways to change the resonance frequencies -- changing the mass of Mass B is  only one. For example, by changing the length of L3, the mass of Mass A is changed. Another example is based on equations modeling the two mass and two spring scenario. It can be seen that by changing the  displacement of the engine or the volume of the plenums, the spring constants of Springs A and B can be changed.  The relationship of some of these variables to the resonance frequencies are illustrated below... Modifying the VRIS The spring and mass model discussed above is what is called a Helmholtz resonator (or something like  that). By taking the equations relating masses and spring constants to primary and secondary resonance frequencies, we can find the effect of modifications to the internals of the VRIS. The equations I used were  modeled after a 3 liter VRIS engine with a different manifold design than the KL03. However, the behaviour of the KL03 can still be approximated using these equations because the KL03 manifold operates on the same principles. A particular modification I was interested in was the effect of boring out the six intake runners. My intent was to raise the primary resonance frequencies, especially the one at 5000RPM, so that the high end of the  powerband could be made use of. If the torque peak of the engine could be moved up in the powerband, the result would be more horsepower.  By boring out the intake runners so that the internal diameter was increased from 39mm to 41mm, the  primary resonance RPM was raised by 236PM from 4515 to 4751. There doesn't appear to be any effect from changing plenum volume (V2). This is probably the best modification to make to the VRIS. The larger passages will cause less pressure drop, and the torque void at high RPM's will be partially filled.  Another possible modification to the VRIS is the 'U' tube. From the graphs below, shortening the tube or decreasing its diameter would also raise the primary resonance frequency. The problem with the KL03 'U' tube is that it is nearly as short as it can get. Any further reductions in its length would yield little gains. Decreasing the tubes diameter might raise the resonance frequency, but air  flow would be restricted. The best modification at this point still seems to be boring out the intake runners. I have tabulated the general effect of some of some VRIS modifications below...   Modification Effect on Resonance Frequencies (and torque  range) Boring out/shortening intake runners Shifted higher in powerband Shortening/decreasing diameter of 'U' tube Shifted higher Increasing plenum volume Shifted slightly lower Increasing engine displacement Shifted slightly lower Increasing intake air density (forced induction or cold air temperatures) Shifted lower Forced Induction Issues It can be seen that the resonance frequencies and torque peaks can be greatly affected simply by changing  tube lengths and diameters. Similarly, changing the density of the incoming air also causes changes. The engine computer doesn't know how to set the butterfly valves to take advantage of  non-atmospheric-pressure air, and so the primary and secondary resonances cannot be taken advantage of under boost. However, since the turbocharger or supercharger is doing the work of pressurizing the  cylinders, you don't really need resonance charging anyway as long as boost is present.  Mazda's turbocharged 2L JF engine did away with the butterfly valves altogether. The manifold was  designed to provide a primary resonance charge at 1800RPM and a secondary resonance charge at 6200RPM. The low RPM primary resonance was probably designed to help generate torque before the turbo spooled up.  The fact that turbos don't provide much boost at low RPM's means that the VRIS can still be useful while the turbo is trying to spool up. Even with centrifugal superchargers, boost can be pretty low at low RPM's  so the VRIS can still be useful off of idle. The big problem with the VRIS is at 6500 RPM +. The engine computer shuts the butterfly valves,  severely restricting airflow. For atmospheric air, the secondary resonance pulses would help charge the cylinders, but with forced induction, the high density air being force through these passages will not be able  to achieve the secondary resonance, and will lose a lot of pressure trying to squeeze through the restrictive passages. The only way to get around this problem is through some ECU reprogramming to keep all the  butterfly valves open at high RPM's.   Much of the above discussion is based on equations and theory, both of which don't always predict  reality reliably. So don't blame me if you Extrude Hone your intake and you end up losing power!  Oh yeah, I plagiarized all the diagrams from SAE paper 871977. Thought I'd just say that to get it off my conscience.         Counter installed April 1, 1999