The old habit in this industry is to talk about CFM, but CFM is just volume. Once air gets compressed, cooled, bent through tubing, or pushed through an intake, that volume changes. The engine does not care about volume by itself. It cares about how much air mass reaches the cylinder, because air mass is what supports fuel and makes power. That is why we are moving away from bench CFM and toward mass flow. A conventional flow bench is useful for comparing one part to another at the same test condition, but it does not tell the whole story when density changes through the system. Supercharging, intercooling, duct size, bends, and pressure loss all affect density, and density is what changes the pounds of air per minute the engine actually gets. So we adapted our bench to measure mass air flow directly using MAF sensors, transfer functions, and Banks iDash logging. That lets us look at pounds per minute instead of a bench number that can hide what is really happening. The payoff is simple: better parts show up more clearly when you measure what the engine actually uses. In our testing, the Banks Monster-Ram Intake and Killer Grid Heater Upgrade outflowed even a best-case straight-tube throttle setup at the same pressure drop. That is the difference between shaping a part to fill a catalog slot and shaping it to move more air mass with less loss.
Gale Banks opened the discussion with Eric Ryder by framing the purpose of the Lockjaw project, a Duramax-powered development truck that uses a supercharger rather than turbochargers. The objective is not simply to make boost, but to achieve high intake-manifold air density right off idle. Banks emphasized that air density, not boost pressure by itself, is what determines how much fuel can be burned cleanly and therefore how much power a diesel can make.
He tied that directly to diesel air-fuel ratio targets. For a diesel under power, he prefers not to go richer than about 18:1 because richer mixtures begin to produce visible smoke. For a street-driven application, he would rather see roughly 22:1 to 24:1 under load, with cruising mixtures leaner still. In that context, the engineering target is straightforward: increase the mass of air entering the engine so more fuel can be added without smoke, excessive heat, or durability problems.
Banks then challenged the industry's long-standing reliance on CFM as the primary airflow metric. Cubic feet per minute is a volumetric flow measurement, but air is compressible, so the volume occupied by a given amount of air changes as pressure and temperature change. That means the CFM measured at one point in a system is not necessarily the same as the CFM at another point, even though the actual mass of air moving through the system remains constant.
He illustrated the problem with a typical diesel engine. A 6.6- to 6.7-liter four-stroke diesel turning around 2,800 rpm, near the horsepower peak for many trucks, pumps roughly 400 cfm based on displacement and engine speed. Yet Banks has seen intake components such as Monster-Ram manifolds or Ram-Air systems reported at over 1,000 cfm on a flow bench. That apparent contradiction is exactly why CFM can mislead people. If the engine is supercharged and intercooler-equipped, it may ingest something like 1,200 cubic feet of ambient air and compress it to about 400 cubic feet inside the engine, while preserving the same air mass. The density ratio has increased by roughly three to one, so the engine receives far more oxygen per cubic foot.
Banks quoted SuperFlow documentation to reinforce the point. The company's own material states that the term CFM creates confusion, that the easiest way to avoid confusion is to quote flow in pounds per hour or other mass terms, and that a flow bench does not actually measure true flow but compares the test piece to a reference orifice inside the bench. For Banks, that is the central issue: CFM is useful only as a comparative number under fixed conditions, while the engine itself responds to air mass.
The discussion then shifted from criticism of CFM to the metric Banks believes matters: pounds of air per minute. If fuel delivery is based on available oxygen, then the relevant quantity is the mass of air entering the engine. More pounds per minute means more fuel can be burned and more power can be made.
Banks used density examples to make the concept intuitive. A cubic foot of framing lumber weighs about 30 pounds, while a cubic foot of gold weighs 1,204 pounds. The difference is density. Air works the same way, just at much lower values. Standard sea-level air weighs about 0.072 pounds per cubic foot, or about 72 pounds per 1,000 cubic feet. Once that air is compressed and cooled, each cubic foot contains more mass. That is why supercharging and intercooling matter: they increase density ratio, which increases mass flow through the engine even if the engine's displacement-based pumping volume remains the same.
Banks also pointed out that altitude complicates any CFM-only comparison. SuperFlow's own examples note that a cylinder head tested at 6,000 feet and at sea level can show the same bench CFM at the same test pressure, even though the actual air mass is different because ambient density is lower at altitude. At 6,000 feet, air density is down by nearly a quarter. To achieve the same mass flow there, volumetric flow must increase. That means a component developed only by CFM may not behave the same way at different elevations. If the goal is consistent engine performance, the meaningful target is mass flow, not bench CFM.
To address that limitation, Banks' team has been converting its test setup into what he described as a mass flow bench. Rather than relying on the SuperFlow bench's native output, they are using the bench primarily as a controllable air pump and adding their own instrumentation to measure actual airflow in mass terms.
Ryder explained that the current setup uses a late-model Ford mass air flow sensor calibrated against a Bosch MAF sensor. From that calibration they derive a transfer function that converts the sensor's PWM frequency output into pounds per minute. The sensor output is read by an oscilloscope and by Banks iDash instrumentation, which can display and log the data in real time. The transfer function is non-linear so the sensor can remain accurate across a broad range of airflow. As frequency rises toward 1 kilohertz, the function changes more gradually to preserve resolution at higher flow rates.
The iDash system is central to the setup because it is more than a gauge. Ryder described using it as a standalone computer that can read the MAF frequency signal directly, apply a lookup table based on the transfer function, and display pounds per minute on screen. It can also log data at 20 samples per second and average noisy low-flow readings over longer intervals, such as 10 or 15 seconds, to stabilize the displayed result. That matters because MAF sensors are designed for real engine airflow and can become erratic at the very low flow rates sometimes seen on a bench.
Ryder also explained what the MAF sensor is actually measuring. Modern hot-wire MAF sensors use an element, typically platinum or nickel, that is held at a target temperature. As air passes over the wire, it cools the element. The sensor electronics increase current to maintain the wire temperature, and that change in electrical behavior is correlated to airflow. Earlier systems often output a 0-to-5-volt signal, while later systems commonly use frequency output.
The more advanced eight-pin sensors from Bosch and Hitachi go further by incorporating humidity measurement directly in the sensor, in addition to pressure and intake air temperature inputs. That allows the sensor to calculate density more accurately in real time, rather than estimating some of those corrections downstream. Ryder noted that density depends on temperature, pressure, and humidity, and modern sensors can account for all three on the fly.
Banks connected that to his own instrumentation products. The Banks AirMouse, used with iDash, measures ambient temperature, pressure, and humidity at the nose of the vehicle. He noted that humidity is not just a weather curiosity; diesel and gasoline calibrators have used it as a meaningful tuning input. Higher humidity can reduce combustion temperature and therefore reduce NOx formation. In spark-ignition engines, humidity can also help suppress knock by absorbing heat during vaporization, allowing timing to be advanced slightly under the right conditions.
Once the team could measure mass flow directly, they began looking at density loss across individual components. Ryder described measuring pressure and temperature before and after the MAF sensor body itself, while using ambient humidity from the AirMouse to complete the density calculation. By configuring the iDash to treat those sensors like compressor-inlet and compressor-outlet channels, he could display density before and after the sensor and chart the loss.
Even at relatively low bench pressure differences, up to about 30 inches of water or roughly 1 psi, they were able to detect measurable density loss across the MAF sensor. That was notable because the sensor body places its measuring element directly in the center of the tube for accuracy, which also creates drag. Early MAF sensors were especially sensitive to non-laminar flow, requiring smooth, straight sections before and after the sensor for best accuracy.
Banks generalized the lesson: only two kinds of devices increase air density in an induction system, supercharging devices and charge-air cooling devices. Everything else in the system-ducting, boost tubes, manifolds, elbows, and similar hardware-tends to reduce density because it introduces drag, pressure loss, or heat. That is why larger boost tubes can help. By increasing tube diameter, velocity drops, boundary-layer shear is reduced, and density loss decreases. Banks said they have quantified this effect directly: a larger straight tube that is later reduced back to the original nominal size can flow more pounds per minute at the same pressure drop than a smaller straight tube.
Banks stressed that manifold pressure matters because it drives density at the intake valve and into the cylinder. Higher pressure available at the valve generally means more pounds of air per minute can enter the cylinder. But pressure alone is not the whole story. Components such as intercoolers may introduce some pressure drop, which hurts density, while simultaneously cooling the compressed air, which helps density. A good design minimizes pressure loss while maximizing cooling.
Ryder offered a practical example from his turbocharged 2.0-liter Camaro. The engine is tuned for about 24 psi manifold pressure. Relative to ambient pressure of roughly 14 psi, that is about a 70 percent increase in pressure, yet the manifold air-density reading is more than 70 percent above ambient. The extra gain comes from charge-air cooling. Banks noted that the older industry term was pressure ratio, but he prefers density ratio because it better captures what the engine actually uses.
This focus on density ratio also explains why Banks sees conventional flow-bench testing as incomplete. A bench typically tests at pressure drops such as 25 or 28 inches of water, around 1 psi, while a real engine can see much larger pressure differences across the valve. In a naturally aspirated engine, the pressure outside the valve may be around 13 psi absolute while the descending piston creates a much lower pressure in the cylinder, producing a far larger delta P than a bench test. A bench also uses static valve lift, ignoring valve motion and dynamic effects that strongly influence real airflow.
The practical payoff of the new mass-flow approach showed up in recent testing of Banks' emissions-legal Monster-Ram for the 6.7-liter Ram Cummins. The team compared it with other intake designs and with the stock-style emissions throttle arrangement. On the emissions-equipped Cummins, a roughly 3.5-inch motor-driven throttle regulates airflow before cooled exhaust gas is mixed into the intake stream. That cooled EGR is inert in combustion terms, but it absorbs heat and lowers peak combustion temperature, reducing NOx.
For the test, they first measured the throttle body alone discharging through a straight tube to atmosphere. Then they added an outlet bellmouth to improve the exit condition and reduce abrupt diffusion into ambient air. That optimization did not increase flow. After that, they installed the Monster-Ram, which includes a bend of roughly 120 degrees, and found that it outflowed both straight-tube configurations at the same test pressure.
Banks and Ryder considered that result counterintuitive but important. A bent component would normally be expected to lose flow relative to a straight tube, yet careful shape development allowed the Monster-Ram to outperform the supposedly simpler path. Banks credited extensive CFD work, 3D-printed prototypes, and bench verification. He said the design incorporates principles from the old NACA research on bending air in ducts with minimum loss, adapted to the geometry of a round intake tube. In his view, that is the difference between creating a true performance part and merely making a new part number that resembles the stock flow path.
Banks closed by putting the SuperFlow bench in perspective. He does not consider it useless; as a comparative tool, it still has value. It can show whether one part is better than another under the same conditions. But he argued that it should not be mistaken for a direct measure of what the engine experiences. It does not capture real valve dynamics, real pressure differentials, or the density changes created by supercharging and intercooling. For that reason, he expects Banks' published airflow data to move away from CFM and toward mass flow.
He said the company has already been thinking in terms of density and mass flow internally for years because turbocharging work demands it. Now they intend to talk about it publicly and update literature accordingly. If the rest of the industry continues to publish only CFM, Banks suggested that his team will simply compare competing parts themselves using mass flow.
The episode ended with a brief update on the company's supercharged Duramax turnkey engine program, using Lockjaw as the development mule. Banks said the engine will be sold only when throttle response, drivability, and durability meet his standards. The first version is intended for 1966-and-earlier vehicles, where it can be road legal without emissions equipment. He also noted that the engine program has recently passed Euro 3 emissions requirements for certain military applications, showing that clean operation and strong performance are not mutually exclusive. As with the airflow discussion, his final point was that durability and measurable engineering results matter more than convenient but incomplete numbers.