People often describe a great automatic transmission as smooth, with gear changes that are barely noticeable. That is often what manufacturers aim for, and it comes at a cost. If the goal is power handling, longevity or track response, a soft shift works against it.

In most modern automatics, each gear is selected by applying a combination of clutch packs, stacks of friction plates and steel plates that a hydraulic piston squeezes together. Most shifts are clutch-to-clutch: one clutch, the offgoing, releases while another, the oncoming, applies. A dual clutch transmission shifts the same way, handing torque from one clutch to the other.

Some automatics use a one-way clutch for certain shifts instead, often in first gear. A one-way clutch locks in one direction and freewheels in the other, so it holds in first and lets go on its own as the clutch for second applies. Only that one clutch has to be controlled.

An upshift happens in two phases. In the torque phase, the oncoming clutch takes over the load while the offgoing clutch lets go, and the gear ratio has not changed yet. In the inertia phase, the oncoming clutch slips while engine speed falls to match the new gear, and the shift ends when that clutch locks.

The handoff comes down to timing. If the two clutches overlap too much, both hold at once and fight each other through the gears, which is called tie-up. If they overlap too little, engine speed flares up for a moment before the oncoming clutch takes hold.

A factory calibration built for comfort deliberately overlaps the two. The offgoing clutch drags a little longer, the oncoming clutch applies a little slower, and the handoff is spread over a longer window. That is what makes the shift feel seamless.

A clutch makes heat whenever it carries torque while its two sides turn at different speeds. That difference is its slip speed, and the energy it turns into heat is the torque it carries multiplied by the slip speed, added up over the time it slips.

In an upshift, the oncoming clutch slips from the moment it starts to take load. It slips at the full speed difference through the torque phase, then its slip falls to zero through the inertia phase. A longer handoff and a longer speed change both add heat.

The inertia phase can be worked through. Treat the rotating parts ahead of the clutch as one mass, and assume both torques hold steady and the car’s speed barely changes in the fraction of a second the shift takes. The heat then comes to:

The equation has two parts. The first, ½ I Δω₀², is the least heat the speed change can make while the engine is still driving the clutch, set by how hard the parts are to slow down and how much their speed has to change. It grows with the square of that speed change, which is larger the higher the engine speed at the shift.

The second part, T_c ÷ ( T_c − T_in ), multiplies that least heat because the clutch is carrying engine torque at the same time. The clutch has to carry more torque than arrives from the engine, or the engine never slows to the new speed. The closer the two are, the larger the multiplier, the longer the slip and the more heat it makes.

That heat works on the clutch and the fluid in three ways.

• Flash temperature. The plates truly touch only at many tiny high points. That is where the heat is made, and those points briefly run well above the temperature of the plate as a whole.
• Friction material. Prolonged slip wears the friction material and accelerates glazing, a hardened, polished surface that grips less.
• Fluid. The fluid carries the heat away, and heat speeds its oxidation, the slow reaction with oxygen that ages it. In a 2021 laboratory study, Farfan-Cabrera and colleagues found that oxidized automatic transmission fluid raised a wet clutch’s friction coefficient, its grip, and generally weakened its resistance to shudder, the vibration of plates that grab and release, which depends on friction rising as sliding speed rises.

A long, soft shift trades clutch and fluid life for refinement, and none of the energy turned into heat reaches the wheels.

The inertia is fixed by the hardware, and the speed change by the gear ratios and the engine speed at the shift. For a given shift, that leaves two things to work with in the speed change: more clutch torque, or less engine torque coming in while the clutch slips. A shorter handoff before it cuts heat as well.

A transmission asked to hold more torque, or to survive track use, needs a shorter slip and less heat. A performance recalibration often does the opposite of a factory calibration built for daily driving:

• Raise hydraulic pressure to the clutch packs, so the oncoming clutch clamps harder and stops slipping sooner.
• Shorten the overlap, so the offgoing clutch releases decisively as the oncoming clutch locks.
• Coordinate torque control with the shift, so engine speed comes down to the new gear without the clutch dragging.

More pressure means more clutch torque. How much a clutch pack can hold is:

A recalibration can change only the clamping force, through pressure. Everything else comes from the hardware and the fluid, and so does the most pressure the pump and valve body can deliver.

Less engine torque coming in is the other way. Here is the same shift with the clutch torque held at 1.5 times full engine torque:

In this example, halving the engine torque halves both the heat and the slip time, without any more clutch torque.

That is why disabling torque management is not the fastest way to firm up a shift. Torque management is the PCM and TCM, the engine and transmission computers, briefly trimming engine torque during a shift through spark timing and airflow. In stock form it is gentle, so the clutches do not shock the driveline. With no torque reduction at all, the clutch has to slow the engine while carrying full engine torque, the top row of the table. The key is to reshape torque management, so torque is cut briefly, the clutch locks quickly and cleanly, and then full power returns.

A performance recalibration can go further than stock, not by removing less torque but by removing more for a much shorter window. The speed change then takes less time and makes less heat without the clutch having to clamp harder, and the shift stays mechanically safe when pressures and timings are set correctly.

A quick shift done properly feels nothing like a harsh bang shift. It is firm and decisive without shocking the drivetrain, and it comes from coordinating:

• Hydraulic pressure curves
• Clutch overlap timing
• Clutch fill time, the moment it takes to fill the piston with fluid before the clutch starts to grip
• Torque reduction and restoration
• Fluid condition and cooling capacity

The aim is enough pressure and speed to keep heat down, without torque spikes that fatigue shafts, gears and differentials. Done well, the transmission runs cooler, the clutch packs stay healthy, and with more clamping force they hold more torque.

1. Harald Naunheimer, Bernd Bertsche, Joachim Ryborz and Wolfgang Novak, Automotive Transmissions: Fundamentals, Selection, Design and Application, 2nd edition, Springer, 2011. Shift elements, one-way clutches and shifts under load.
2. Richard G. Budynas and J. Keith Nisbett, Shigley’s Mechanical Engineering Design, McGraw-Hill. The torque capacity of friction clutches.
3. Harmen Blok, “Theoretical study of temperature rise at surfaces of actual contact under oiliness lubricating conditions,” Proceedings of the General Discussion on Lubrication and Lubricants, volume 2, Institution of Mechanical Engineers, London, 1937, pages 222-235. Flash temperature.
4. Leonardo Israel Farfan-Cabrera, Ezequiel Alberto Gallardo-Hernández, Manuel Vite-Torres and Jesús Gilberto Godínez-Salcedo, “Influence of oxidation of automatic transmission fluids (ATFs) and sliding distance on friction coefficients of a wet clutch in the running-in stage,” Friction 9, 2021, pages 401-414.

All notes →

A turbocharger’s compressor squeezes outside air to raise its pressure, and that added pressure is boost. Compressing air takes work, and the work ends up as heat in the air. The temperature coming out of the compressor can be estimated from how much it raises the pressure and how efficient it is:

With 80°F outside air at sea level and a compressor at 70% efficiency, that gives:

These are estimates. Real efficiency changes with how hard the compressor is working, and any restriction in the intake ahead of it, or pressure lost in the intercooler after it, raises the pressure ratio needed for the same boost. The scale holds either way: at 20 psi the air leaves the compressor near 300°F, and more boost makes it hotter still.

Hot charge air, the compressed air on its way to the cylinders, is less dense at the same pressure, so each cylinder takes in less oxygen. It also raises the chance of knock, where the last of the fuel and air ignites on its own before the flame reaches it. The PCM detects knock with knock sensors and answers it by pulling timing, firing the spark later.

A common rule of thumb puts the cost at about 1% of power for every 10°F rise in intake air temperature. That is close to the temperature term in SAE J1349, the standard for correcting dyno results to standard air conditions. It adjusts a dyno reading for the air on the day of the test, and it does not model a turbocharged engine that controls its own boost.

The EcoBoost PCM works to a torque target. As charge air gets hotter and less dense, the turbo has to work harder to reach that target, and the PCM pulls timing and torque as temperatures climb. Heat is paid for in extra boost, turbine speed and timing, and power at the wheels falls once the turbo cannot supply the extra air, or the timing pulled costs more than the extra air returns.

An intercooler is a heat exchanger that takes heat out of the charge air before it reaches the cylinders. Its effectiveness is the share of the possible temperature drop it achieves, where the most it could do is bring the air all the way down to ambient, the outside air temperature.

Take air leaving the compressor at 335°F on an 80°F day. A core with an effectiveness of 0.6 delivers it at about 180°F. A core at 0.9 delivers it at about 105°F, about 25°F above ambient.

Effectiveness is not fixed. It falls as more charge air flows through the core, when less outside air crosses it, and as the core heat soaks, warming up over repeated pulls until it has less capacity left to take heat out.

Most Ford EcoBoost engines, including those in the F-150 and the Mustang, use an air-to-air intercooler from the factory.

• Air-to-air passes charge air through a core cooled by outside air flowing across the front of the car. It is simple and reliable, and it works best while the car is moving.
• Water-to-air passes charge air through a core cooled by liquid, which needs a pump and its own coolant loop. The coolant’s thermal mass, its ability to absorb heat before its own temperature climbs, soaks up a short burst of boost well. Over a longer run the coolant warms, and the system heat soaks.

A factory intercooler is sized for the engine’s factory output. A performance recalibration usually raises boost and airflow, so more heat goes into the same core and its effectiveness falls. On recalibrated EcoBoosts with stock cores, charge air often reaches 140-180°F or more within a few seconds of boost. The PCM then pulls timing and torque, and power at the wheels drops. Drivers often describe it as “it hits hard, then fades.”

Once a core has heat soaked, it no longer brings the air back near ambient, and the temperature going into the engine climbs toward the temperature coming out of the compressor.

• Consistent power. A larger, more effective core holds charge air closer to ambient, and its extra mass takes longer to heat soak, so power holds after a few seconds of boost.
• Knock margin. Cooler charge air lowers the temperature of the unburned mixture, which is what raises the knock threshold. A given mass of cooler air also needs less boost pressure. Less knock helps protect the pistons and spark plugs.
• Less work for the turbo. Because the PCM works to a torque target, cooler, denser air lets the engine reach it with less boost pressure and turbine speed.
• Ethanol. Ethanol blends resist knock through a higher octane rating and the cooling they give as they evaporate. Cooler charge air adds to that margin, which allows more timing and boost.

Every intercooler costs some boost pressure as air passes through it, which is its pressure drop. How much depends on the design of the core and its end tanks, the chambers at each end that spread the air across the core and collect it again, and on the flow through it. Size alone does not decide it, so a well designed larger core can cool much better without costing much more pressure.

For a recalibrated EcoBoost that sees repeated boost, an intercooler is one of the smartest upgrades for reliability and consistent power. On a lightly modified car whose charge air already stays close to ambient, a larger core has little cooling left to add, and any added pressure drop can cost more than it gives back.

1. John B. Heywood, Internal Combustion Engine Fundamentals, 2nd edition, McGraw-Hill Education, 2018. Compressor temperature rise, charge density and knock.
2. Theodore L. Bergman, Adrienne S. Lavine, Frank P. Incropera and David P. DeWitt, Fundamentals of Heat and Mass Transfer, Wiley. Heat exchanger effectiveness.
3. SAE J1349, Engine Power Test Code, SAE International. Correcting measured power to standard inlet air conditions.

All notes →

Boost makes the mixture in the cylinder denser, and a denser mixture takes more voltage to jump the spark plug gap. That is why EcoBoost engines, particularly under high boost, place high demands on the ignition system, and why the NGK Ruthenium HX is a plug worth evaluating for them.

The electrodes are the metal tips the spark jumps between, and NGK makes the Ruthenium HX in two electrode designs. PSPE, a projected square platinum electrode, is the one NGK recommends for turbocharged and supercharged engines. DFE, a double fine electrode, is recommended for engines without a turbo.

NGK’s own charts show two tests of the Ruthenium HX, one of combustion stability and one of flame kernel growth. Each shows something useful, and each leaves something out.

The combustion pressure test charts the pressure in the cylinder as a band, for the Ruthenium HX and for a plug NGK labels conventional. NGK marks the Ruthenium HX band as little dispersion and the conventional plug’s as more, and the conventional band is wider at the peak.

A narrower band means less variation from one cycle to the next. Engineers measure it as the coefficient of variation of IMEP, where IMEP, the indicated mean effective pressure, is the work each cycle does on the piston expressed as an average pressure. It matters because ignition timing has to be safe for the fastest-burning cycles, which are the ones most likely to knock. The less the cycles vary, the closer timing can sit to its best on average, and the more even the torque.

NGK gives no engine, speed or load for this test. Both plugs peak near 4,000 kPa, about 580 psi, far below the peak pressure a boosted direct injection engine reaches at full load. The chart shows how consistent each plug was at that test point, not how either behaves at wide open throttle on boost.

The spark starts a small ball of flame, the flame kernel, which grows into the flame that burns the rest of the mixture. NGK’s photos show the kernel at ignition and at 1, 2 and 3 milliseconds after it, for the Ruthenium HX, an iridium plug and a nickel plug.

At 2 milliseconds the Ruthenium HX kernel is about the size the iridium and nickel kernels reach at 3 milliseconds. A millisecond matters here: at 3,000 rpm the crankshaft turns 18 degrees in one. A flame that grows faster in its first milliseconds brings peak cylinder pressure sooner after the spark.

The comparison changes the electrode design as well as the tip material, so the photos show the plug as a whole. They do not separate how much of the difference comes from ruthenium.

Spark timing is how far before top dead center the plug fires, measured in degrees of crankshaft rotation. Best-torque timing, MBT, is the timing that puts peak pressure where the piston turns it into the most torque. When the burn starts faster, peak pressure arrives sooner, so MBT moves later and less spark advance is needed to reach it.

This matches what I have seen in logs. Final timing often reads slightly later with the Ruthenium HX, while torque output and peak achievable air load, the air each cylinder can take in, hold steady or improve. The later timing is not a sign of knock or lost efficiency. It reflects the faster burn and the pressure curve moving with it.

Which effect matters more depends on the fuel. On gasoline at high boost, timing is usually limited by knock rather than set at MBT, so the steadier combustion matters most. Ethanol blends resist knock well enough that timing can often reach MBT, so there the later MBT matters most. Either way, the differences are small.

1. Ruthenium HX spark plugs, NGK Spark Plugs. The electrode designs and what each is recommended for. The pressure chart and flame kernel photos are NGK’s.
2. John B. Heywood, Internal Combustion Engine Fundamentals, 2nd edition, McGraw-Hill Education, 2018. Spark timing and MBT, flame development, cycle-to-cycle variation and knock.

All notes →