Optimizing Induction Air

Why do we want to worry about increasing the volume of inducted air into the engine? Well, just try to breathe through a 20-inch tube with a 1-inch diameter. Not easy. Then, try running while breathing through it. Not likely.

At 2700 rpm, a 360-cubic-inch engine wants to breathe 4.69 cubic feet per second. Just like us, if we restrict the intake, it can’t do as much work. Most of us would like to get the most possible work out of our engines, so a great intake system is necessary.

In a normally aspirated engine, every time the piston goes down on the intake stroke, ambient pressure pushes its way into the low-pressure area of the cylinder. We know this pressure to be around 29.92 inches of Hg (all things being standard). This is what we call manifold absolute pressure (MAP). It just so happens that for every inch of MAP increase in an IO-360 engine, the horsepower output increases about 1.67 hp per cylinder. This could be used to improve aircraft performance and economy: 29.92 MAP x 4 cylinders x 1.67 hp/cylinder = 199.9 hp. OK, I know that’s not a real formula, but it illustrates the idea.

Since we have to get air into the engine by building something, is there something better we can build to do that? What considerations would there be, and why? We all know that many things affect the flow of fluids. Intake size, path, surface roughness, and restrictions are all important. So let’s start at the front where the air comes in.

Inlet Types

Basically, there are two inlet types: submerged and external scoops. The submerged scoop has been improved over the years to the point that a NACA duct is about 83% efficient. The NACA scoop has about a 7-degree slope with an aspect ratio of around 3 to 5, and a rectangular exit at the rear. It is a high-velocity ratio type, with the air having to accelerate over its diverging sides in the form of two vortices that draw their air source from the boundary layer of the local surface of the plane. It is low drag because nothing protrudes into the free stream flow. However, it is not without some drag, but a lot less than an external duct.

Another consideration is that if the duct’s flow is restricted downstream and flow-through is reduced, velocities drop and no vortices are produced, so the duct produces more drag and less inducted air. That’s why this kind of duct isn’t normally recommended for certain applications such as oil coolers, especially if there’s a high-pressure ram cooling plenum used on the plane. And in poor installations, there may actually be reverse flow through the oil cooler. It is also why this type of duct is effective in EZs—the pusher engine/cowl relationship produces the opportunity to discharge into a lower pressure area if done well. Rapid flow-through is key in these ducts.

External scoops can have efficiencies in the range of 94%. They require properly shaped inlet lips and have higher velocity ratios then NACA scoops. These scoops protrude from the airframe into the free stream flow and tend to produce greater drag. However, if the scoop is confined within the outline of the minimum sized cowl, it doesn’t have to increase the flat plate area or the wetted area very much. Therefore, if it is shaped aerodynamically, it will produce lower drag. Remember, a laminar flow streamlined body can have a drag coefficient as low as 0.035.

This type of scoop can generally be positioned to take maximum advantage of free stream flow, q, and prop effect. Also, the pressure at the end of the duct is not dependent on flow-through. However, it tends to have more drag around the inlet from air spilling over its edges. The scoop can be designed and sized to minimize that drag by varying the inlet volume to be the smallest, most effective, but the design is point specific, so it is impossible to design a nose inlet that is optimum for all performance conditions. Usually it will be designed for the worst-case condition or wide open throttle (WOT) minimum air speeds.

So, is the amount of drag penalty worth 11% less potential efficiency? The debate will continue, depending on the application.

Sizing the Inlet

In the area of NACA scoops, I defer to those who have greater expertise like Klaus Savier. He pretty much goes as fast and efficiently as anyone goes in an experimental plane for the cubic inches. Sizing still has to result in adequate mass flow to meet the engine’s total needs.

In the area of exposed inlets, however, it’s a little easier to calculate the inlet size. They can be designed and sized to minimize drag by varying the inlet volume to be the smallest, most effective area. Remember, we want easy breathing for the engine, so it can produce maximum power.

How do you do this? If you have a 360-cubic-inch engine, it aspirates 180 cubic inches per revolution. At 2900 rpm, it needs 8700 cubic inches per second.

180 cubic inches per revolution x 2900 rpm / 60 seconds per minute = 8700 cubic inches per second.

This is what your engine needs to breathe.

At 245 mph, you travel 4312 inches per second.

So, 8700 cubic inches per second / 4312 inches per second = 2.01 square inches is all that would be required for the inlet area at that speed if our engine had a 100% volumetric efficiency (VE). Our engines are more around 80% VE. You can calculate your own VE by:

VE = AAF / TAF

TAF (theoretical air flow) is simple if our engines were 100% efficient. My RV-4 at 2800 rpm would require 4.86 cubic feet per second. However, under those conditions at WOT and at an AFR (air fuel ratio) of 12.6 (or max power at approximately 100 degrees ROP), I flow 14.6 gallons per hour. So the AAF (actual air flow) is:

AAF = (14.6 gallons per hour x 6.1 pounds per gallon) x 12.6 AFR / 3600 seconds per hour / 0.0765 pounds per cubic foot = 4.07 cubic feet per second.

VE = 4.07/ 4.86 = 83.7%, and just for laughs, you can calculate VE changes from LOP low rpm to WOT at best power and max rpm.

OK, now back to inlet sizing for the worst-case scenario:

The RV-4’s best climb is around 110 mph, and the engine still needs 8700 cubic inches per second, but at 110 mph you travel 1936 inches per second so:

8700 cubic inches per second / 1936 inches per second = 4.49 square inches.

If the inlet is only 90% efficient due to losses, 4.49 square inches / 0.9 = 4.99 square inches. Therefore, the inlet needs to be about 2.5 inches in diameter. While at 245 mph an inlet with less than a 1.5-inch diameter would work, it wouldn’t be satisfactory during a maximum-performance climb. But the larger-diameter inlet produces more drag, so again, depending on your design goals, the debate continues.

Ducting

Generally, shorter, straighter, and smoother is better. The system will determine the ducting necessary from the inlet to the airbox or throttling device. The duct should be of sufficient diameter to provide for the inlet requirements of the engine, plus additional size to compensate for losses due to its length, turns, and surface losses. Generally, a larger diameter and smooth inner surface results in the lowest losses. Make certain there are no volume restrictions along the duct’s length. It may also be possible to benefit from Bernoulli’s effect by ever increasing the diameter along its entire length. This is an important consideration for possible MAP boost at lower indicated airspeeds.

Airbox

The airbox can function in a more effective way then just the device to hold the air filter. It can aid in Bernoulli’s effect and is also considered an inline Helmholtz resonator (more on this later). The airbox can employ Bernoulli’s equation in its construction to increase the pressure of the incoming air, adding to MAP. A diffuser angle per side that maintains flow attachment of around 8 degrees per side is good. An angle that is too great will cause detached flow and effectively choke the intake. Bernoulli’s principle states that in a steady flow, the sum of all forms of energy in a fluid along a streamline is the same at all points on that streamline, and the “principle of conservation” states the sum of all energies must remain constant along that streamline.

In our aircraft the total pressure available at the inlet entrance is static pressure + q less losses. The velocity in the inlet with a 2.5-inch diameter is essentially stable with aircraft speeds above about 100 mph (1764 inches per second) because at 100 mph that velocity provides all the air the engine can ingest. (I know there are other variables such as air density, temperature, friction losses, etc., but they don’t change the basic considerations.) Airspeed above that adds to q, but the excess air spills over the inlet lips and increases drag. So if air comes in the 4.9-square-inch inlet (my RV-4) at 100 mph, then expands to 28.27 square inches in the airbox to 5.77 times its inlet volume, the velocity decreases and air pressure increases. This increased pressure is just in front of the throttle body, servo, or carburetor.

P₁ + 1/2 ρV₁² + ρgh₁ = P₂ + 1/2 ρV₂² + ρgh₂

Where (in SI units)

P = static pressure of fluid (air) at the cross section
ρ = density of the flowing fluid
g = acceleration due to gravity
v = mean velocity of fluid flow at the cross section
h = elevation head of the center of the cross section with respect to a datum.

With induction velocities less than Mach 0.3, you can disregard the pgh, so Bernoulli’s equation becomes:

P₁ + 1/2 ρV₁² = P₂ + 1/2 ρV₂²

When a velocity of 1764 inches per second is flowing through an area of 4.9 square inches and then expands into an area of 28.27 square inches, the velocity decreases to 305 inches per second. Since the number of cubic inches per second flowing in the streamline must remain constant, the velocity decreases and the pressure has to increase due to the principle of conservation of energy.

Wow, what boost! But here’s the rub: Since the total energy remains constant in the streamline, the actual difference in the volume that is important is the inlet volume to the engine-port volume in the head. That is all the expansion that can actually occur. For velocities below 100 mph that have inlets with 4.9 square inches, Bernoulli’s principle does work. So, for people trying to optimize STOL performance, this is definitely worth considering.

There is lots of research to indicate that the pressure-increasing airbox can somehow increase the possible MAP available and can add in recovery of pressures lost through air filters when used with diffuser cones.

Velocity Stacks

Professor Gordon P. Blair published an article called “Best Bell,” which discussed how to optimize the design of an engine air intake bellmouth. In this article, he determined that an inlet that is short, fat, and elliptical provided the best flow. This can be seen in the pictures of maximum inlet flow.

Throttle Controls

Throttle control is done by the means of a butterfly valve. However, there is a difference in the function between a carburetor, a fuel injection servo, and a throttle body. The first two operate on a pressure drop to determine the amount of fuel needed. The throttle body only regulates the amount of air. In the first two types of controls, if there is an instantaneous drop in MAP as the throttle begins to be reduced from steady state WOT, it may be a limiting factor in the amount of air that can be ingested. So if you have a large enough throttle body, there will be no decrease in MAP during the first few degrees of the rotation of the butterfly valve. This ensures that there is no restriction in the possible maximum flow of air.

Plenums, Manifolds and Runners

When fuel is added at the throttle control before the plenum or manifold, the runners (induction tubes) that have more bends and are longer tend to allow heavier fuel droplets to separate from the airflow. Also in normally aspirated engines, the volumetric efficiency decreases some as the fuel vaporizes along its path, displacing some of the incoming air. The earlier the fuel is added, the lower the volumetric efficiency. The runners should be of the same length. Tight bends tend to allow the outer radius of the bend to become wetted. Long tubes allow more time for these processes to occur. These situations can produce problems with engine operation in certain conditions. Needless to say, injection systems do not suffer from all of these conditions. However, there is a best place to put the injector, and that is where the air velocity is at maximum, or in our case, on the head. Actually, longer dry runners can give the air momentum and velocity as it flows toward the intake valve. This can help to fill the cylinder and can increase the engine VE. To get the highest velocity at the back of the valve, the induction tube can be tapered. The taper should be around 1.7-2.5% increase in the runner per inch, which is about 1-1.5 degrees taper. Guess what—just look at the SS intake runners for the 4-cylinder Lycoming with approximately a 1.7% taper that are made by Kevin and Marshall Murray at Sky Dynamics.

Of course, a cold induction plenum provides a denser intake air mass, so consider heat shields and heat-reflective coatings on the plenum, manifold, and runners. (That may be the next thing I try; I already do it to the oil-sump heat shield). For every approximately 6 degrees increase in induction air temperature, the horsepower output decreases about 1%. Charles’ Law states the temperature of a gas is proportional to the volume; therefore, as it gets hotter, it’s less dense.

Also be certain to check that ports match and that gaskets don’t intrude into the port.

Now for the part where it gets really interesting!

Induction Waves

Here is possibly the pot of gold at the end of the rainbow. Automobiles have used this technology for years, and the finest normally aspirated engines can have a VE of up to 110% or more. Remember, our stock engines are older technologies and have VEs normally about 80%. With experimental aircraft we have been able to modify the engines and systems to increase VE. Ken Tunnel at Lycon and other engine builders like Sky Dynamics have been doing these things by increasing compressions, flowing the heads, balancing the internal components, changing cam profiles and many, many more exotic things for years trying to increase VE and horsepower output.

When an engine is running, there are high- and low-pressure waves occurring in the intake system runners traveling at the speed of sound. The waves are caused by the intake valve opening (IVO), as the piston on the intake stroke creates low pressure in the cylinder, which causes an expansion wave to propagate toward the plenum until it arrives at the plenum opening; then it’s reflected back toward the intake valve, and back and forth. This occurs while the airflow in the 16.5-inch runner is moving along at about 100 mph in my case. At 2825 rpm, the intake valve remains open for 0.01275 seconds. Then, if you divide that by the time it takes for the wave to make a round trip to the plenum and back at 1200 feet per second (the speed of sound, which is temperature dependent), there are 5.563 round trips. Peak pressure in the runner should develop toward the end of the intake stroke for maximum cylinder filling. This is where the length of the runner is important because you want to time the returning wave to be there perhaps 30 or so degrees before the intake valve closes (IVC). That’s a simplistic look at the wave that occurs during the filling cycle. In a conference paper in 2002 on tuned manifold systems, Oldřich Vitek and Milos Polasek claimed peak pressure before IVC is more significant than peak pressure before IVO.

Well that’s the first half of the wave story. When the intake valve closes, the air mass that is flowing through the port is shut off, and the inertial energy in the mass flow kind of bounces back. These waves also travel at the speed of sound, so the wave travels back and forth from the valve to the plenum several times as well, before the next valve opening occurs. However, if the wave has already gotten to the valve while it is still closed, it’s possible that the inertia will be headed away from the opening valve, which will delay cylinder charging until the low pressure in the cylinder can get the column of air moving in the right direction again (sort of).

This kind of tuning requires equal-length runners, and adjustable-length runners would be even better. This technology may be available in the future, where you can actually tune the maximum VE for the rpm at which you want to run.

This is rpm specific because the SOS (speed of sound) is constant. Depending on temperature, it is typically between 1130-1300 feet per second, so there will be a specific rpm that our VE will be maximum. This actually may occur at a very low rpm and again at maximum rpm. There is some leeway in engine rpm because the increase in power produced has a little range. Since our engines are mostly limited to a narrow rpm range during operation, they may be able to provide good operation from cruise to full power, again depending on your rpm. To further complicate things, we can take off in high temperatures on the ground and climb to high altitudes where it is very cold, and with cold induction manifolds, the air isn’t heated much. But at cruise, the lower rpm tends to increase the time between valve openings, and the lower temps tend to decrease the SOS, so it takes a little longer for the wave to travel, and that may work out to our good.

Helmholtz Resonator

A Helmholtz resonator can be used to change the frequency of the wave traveling back and forth in the plenum to the runners. Therefore, it may be possible to change the timing of the wave or add to the density of the charge air at the face of the intake runners in the plenum by using the plenum pressure waves like the waves traveling in the intake runners. Even a small change in runner length changes wave timing in the runner, so a Helmholtz resonator may be effective in timing the wave to the face of the runner in the plenum to increase charge density.

What we have in our planes is a sort of inline Helmholtz resonator. This seems to be more effective when no more than four cylinders are connected to the plenum. For six-cylinder engines it would require two plenums, and the cylinders would have to be even firing. It also doesn’t work for installations where the air cleaner is directly connected to the throttle control device. There are three tunable areas for the resonator: the plenum volume, intake ram tube length, and intake runner diameter. In our installation, the runners are not usually very adjustable, so we can work with the remaining two parameters.

If the wave arrives at the entrance to the runner at the wrong time, it reverses and the energy starts away from the runner, decreasing the charge density. So when the intake valve opens, the piston has less energy (lower pressure) in the runner airflow before it starts filling the cylinder (actually, there’s just less energy there than there could otherwise be), therefore subtracting from the potential cylinder filling. However, if we can change the volume of the upstream resonator to optimize the timing or produce denser charge air to aid in better filling the cylinder, we can increase the VE. Remember, this does not come from an increase in the indicated MAP.

Testing Theory

I am very fortunate to have a friend, Howard Claiborne, who when he wasn’t flying a plane, he was jumping out of them as a jump instructor. He is the president of CVincorp, a seismological company. His company locates energy reserves under the ground anywhere in the world. He has an education and life-long career in wave propagation analysis and has constructed a spreadsheet that does all the calculations for values needed for wave propagation analysis in our induction system. He and I have been trying to understand what is really going on in the intake that is useful for builders out there, so science can be applied with sound (get it) reasoning. Fortunately, some things are getting clearer.

We equipped the plane with a ProSense PTD25-10-VH pressure transmitter (10 to 0 volt output and -14.5 to 0 psi range vacuum) that Howard acquired. We installed the sensor in the intake runner port, so it was the same distance to the back face of the valve as to the sensor element. That way we could remove any timing error in the readings, so we could see what was happening at the back face of the valve in real time. It was connected to an oscilloscope that could be carried in the plane.

We ran an in-flight test at 2700 rpm and got the trace shown on the oscilloscope. This will allow us to calculate accurate pressures in the induction runner during a complete 720-degree engine cycle. From the red trace, we can calibrate the voltage change to the pressure change in the runner and see if we can improve the cylinder fill and VE by changing the runner lengths to optimize the wave timing at the valve. By analyzing the sound waves in the induction runners, we’ll be able to listen to the individual frequencies of each runner. The frequencies are different if the operational length of each runner is different, so an adjustable and tuned intake is what we’d like.

Howard is hoping to be able to develop a spreadsheet that can determine the size of the optimum resonator for a specific performance target. The biggest drawback to success is that my induction runners are not the same length, and I believe that is probably the case in most of our applications.

What I know now is that I have one cylinder with about the correct length runner, and at 2825 rpm and the speed of sound at 1200 feet per second, I get 5.563 round trip waves while the intake valve is open. That means the maximum pressure returning to the intake valve will be about 30 degrees or so before it closes, and that’s supposed to be good.

Also the wave propagation after the intake valve closes until the next valve opening is at the highest portion of the summary wave pressure as the intake valve opens. With the returning propagation wave at 12.98 trips, it is right at the back face of the intake valve as it opens. The end result is I have one cylinder that is being all that it can be.

My wife Diane, who is also a pilot, read a draft of this article and said, “So what does this do?” Well, if we can get this right, we can increase the VE of our engines with bolt-on stuff, and that means we can go faster or more economically. You have to bolt on something, so why not the “right stuff,” if you can get it?

I’m working on an adjustable and tunable plenum as we speak. Maybe there’s more to come in the future.