This paper describes some major developments in aircraft superchargers that took place in the United States between 1918 and the Second World War. Emphasis is on the supercharger itself. Other developments that contributed to the success of the technology—doped fuels, reduction gears, variable-pitch propellers—will have to wait for another time.
Engines induct air by volume, but consume oxygen by weight. As the atmosphere thins at high altitudes, fewer oxygen molecules are available for combustion. A naturally-aspirated engine loses about half of its rated power at 20,000 ft. Forced induction is a merely a way to increase the density of the charge.
Initially researchers viewed superchargers solely as a means of altitude compensation. The aim was to restore lost power by maintaining, but never exceeding, sea-level manifold pressure as the airplane climbed. As Dr. Stanford Moss put it, a normalized supercharger “kidded the engine into thinking it was a sea level.” Brake mean effective pressure (BMEP), exhaust temperature and the heat lost to the cooling system remained within design limits.
The ability to operate with impunity at high altitudes resulted in increased speed and slightly more than anticipated engine power. The rarified atmosphere reduced drag on the airplane and backpressure on the exhaust. The loss of lift could be compensated for by greater angles of attack.
The alternative was to take the bull by the horns and apply boost continuously from take-off to maximum altitude. Ground boost required high-octane fuel and a close integration of the supercharger and the engine during the design stage, which as described below, was often lacking.
Adiabatic (or isothermal) efficiency was and remains the most frequently used measure of supercharger performance. Compressing air generates heat as a bicycle pump or diesel ignition demonstrates. In an ideal thermodynamic world—a world without entropy—an air compressor would neither gain nor lose heat to external sources. Such an adiabatic compressor would impart no more heat to the air than generated by the work of compression:
T2 = T1 x (P2 / P1)0.140, where
T1 = absolute inlet temperature in °Rankine (°R = 460 + °F)
T2 = absolute discharge temperature in °R
P1 = inlet pressure,
P2 = discharge pressure,
0.140 = the exponent generally used for air
Real-world compressors put significantly more heat into the charge—heat that increases pressure and reduces density. Adiabatic efficiency is the measure of how closely the compressor approaches the ideal:
Adiabatic Efficiency = (T2 – T1) actual / (T2 – T1) ideal
The National Advisory Committee for Aeronautics (NACA, predecessor to NASA) did not establish a standard for these calculations until 1941. Before that time, researchers were free to follow their instincts, which led to exaggerated claims. Some vintage superchargers were assigned adiabatic efficiencies of 80%. Now, a century later, only a handful of superchargers achieve efficiencies of 75% and when they do, it is over a very narrow operating range. The current SAE J-1723 standard was adopted in 1995. The standard calls for an inlet temperature of 537°R and an ambient pressure 29.23 inHgAbs. To more accurately measure compressor work, efficiency calculations must also take the energy content of inlet-air velocity into account.
Other measures of supercharger performance include volumetric efficiency (rpm-dependent for centrifugal compressors and hovering near 100% for positive-displacement Roots blowers), mechanical efficiency and power consumption. The latter is another tricky concept that awaited the arrival of digital computers to achieve a degree of precision.
As with other human constructs, forced induction is a kind of narrative based on a few relatively simple ideas. Several of the more important are:
Boost – pressure above atmospheric in the manifold. Because density diminishes and pressure increases with heat, a really inefficient supercharger can heat the charge enough to generate boost while adding little or nothing to engine power output.
Mass flow rate – air density multiplied by the volumetric flow rate describes the output of the supercharger. Early researchers, with no convenient method of determining manifold charge density, came at the problem indirectly by measuring the volumetric flow rate (ft³/min) and correcting for temperature at the carburetor inlet.
Pressure ratio (pr) – compressor discharge pressure divided by inlet pressure.
Surge or stall – occurs when the compressor can no longer overcome system resistance. As stall approaches, manifold pressure rises. The compressor responds by pumping less air at a higher pressure. Closing down the nozzle on a water hose has a similar effect—less water flows through the nozzle, but at a higher pressure. At some point, the compressor can do no more and the air flow reverses. Impeller vanes stall much like an airfoil when forward velocity is lost. With no input from the compressor, manifold pressure falls. Impeller vanes then recover from the stall and resume their pumping function. Manifold pressure builds to the level that precipitates another stall. The cycle continues, often accompanied by loud “barking,” broken impeller blades, and bearing failure. Unless provision is made to relieve the pressure, turbochargers surge as they spool down under closed throttle. Surge also occurs when an overly large supercharger, one that delivers more air than the engine can ingest, is fitted.
Choke or stonewall – occurs when a centrifugal impeller spins fast enough to generate sonic air velocity at the inlet nozzle or diffuser throat. With little resistance to flow, the volume of air delivered increases and the pressure drops. The resulting shock waves choke off air delivery. An undersized (and under-loaded) turbocharger can attain choke speeds.
Choke (low pressure and high volume) and surge (high pressure and low volume) define the flow boundaries of compressor operation. Impeller rpm is the mechanical limit.
Power absorption – another dubious number. The intake manifold cannot be viewed simply as an open-ended container that fills faster than its empties. Until recently, engineers had no way of calculating the effects of friction, turbulence, flow separation and reflected pressure waves on power requirements.