Notes on WWI German Superchargers
by Paul Dempsey
|As the atmosphere loses density with altitude, fewer oxygen molecules are available for combustion. A naturally aspirated engine can be expected to develop only about half of its rated power at 6,000 m. There are several ways to mitigate this effect, including over-compressed and over-dimensioned engines that release their full potential at higher altitudes. But the definitive solution is to create an artificial atmosphere of ground-level density around the carburetor intake.|
German researchers began work on supercharging in 1915 and quickly settled on the gear-driven centrifugal compressor. The Royal Aircraft Factory also favored centrifugal superchargers, but was unable to overcome difficulties associated with the drive mechanism . Gears stripped during acceleration and deceleration. Attempts to design a friction clutch and a subsequent effort to use flexible vanes on the impeller failed. The same source reports that the A.E.F. launched a short-term investigation of gear-driver centrifugal blowers, which has not otherwise been documented. The French and the U.S. Army in collaboration with the General Electric Corp. concentrated on Rateau-style exhaust-driven turbo-compressors, devices that would require two decades to perfect. NACA (the U.S. National Advisory Committee on Aeronautics) struggled almost alone to overcome the pressure pulsation in Roots blowers.
All of these projects failed in the sense that the war ended before the technology could yield military advantage. While we have nothing by way of the detailed reporting on German wartime research accorded to British and American efforts, it appears that Germany came closest to producing combat-ready supercharged engines. Certainly they had a denser supplier network than the Allies, more consistent research goals, and accumulated far more hours of test flying. The low-pressure engine test cell at the Zeppelin works in Fredrickschafen was an invaluable tool, not replicated until 1918 by the U.S. Bureau of Standards . It may also be true that German engineers were better trained than their European and American counterparts.
Superchargers were add-on devices, installed on otherwise standard power plants. To do otherwise would have interfered with production, the central element of industrialized warfare. Airframe and component suppliers, working under military contract or on a speculative basis, played the major role in development. For the most part, engine manufacturers stood aside and may have had reservations about the whole idea. At any rate, it was the engine manufacturers who insisted upon some sort of slip coupling between the crankshaft and supercharger gearbox.
Normally aspirated aircraft engines already operated on the ragged edge of detonation; increasing brake mean effective pressure further was out of the question. Experience with over-compressed engines that, regardless of throttle position, knocked at 2,000 m underscored the point. Consequently researchers limited their ambitions to normalizing rather than ground-boosting. That is, the aim was to maintain, but not materially exceed, sea-level performance until the critical altitude was reached when the supercharger could no longer supply 760 mmHg.
While beyond the scope of this paper, it should be mentioned that the Germans won some room to maneuver by raising the threshold of detonation with benzene-doped fuels. Why the Allies did not resort to this readily available fuel additive is unknown.
A centrifugal compressor (Fig. 1) is a deceptively simple machine, consisting of an impeller and a diffuser housed in a helical casing, or scroll. The diffuser, sometimes called the stator, occupies the annular space between the impeller and scroll. Passages created by the diffuser vanes open wider as they approach the discharge throat. The impeller wheel also incorporates vanes, disposed radially from the wheel center to its periphery.
The gas, in this case air, enters at the impeller hub, rotates with the impeller and, under the influence of centrifugal force, moves outward in a path defined by the radial vanes. As the gas stream moves outward, the rotary component of its motion accelerates to vane-tip speed. These two forces—rotational and centrifugal—impart kinetic energy to the gas. Upon contact with the diffuser, the gas expands and slows, converting much of its kinetic energy into static pressure.
Because the impeller cannot be allowed to make physical contact with the housing, there is always some leakage between the vane tips and the scroll. The seal consists of air, an elastic medium. At low rotational speeds the impeller merely flays about delivering little or no output. As tip velocity increases, the air seal becomes more positive and the compressor begins to pump air through its discharge throat. Unlike Roots blowers that move the same volume of air per revolution, centrifugal compressors are dynamic machines, whose output increases as the square of speed. Double the speed and the output theoretically quadruples.
The obvious first step when adapting the centrifugal compressor to aircraft service was to increase rotational speed and/or impeller diameter. Either approach would have the same beneficial effect. However, the strength of available materials limited these options. High rotational speeds were further constrained by concerns about bearing speeds. Nor could much be done to close the gap between the impeller and its housing. In fact, the clearances increased as the aluminum housing heated and expanded more rapidly than the steel impeller. Bearing elongation due to gyroscopic forces and small imbalances in the relatively heavy steel impeller had a similar effect. Modern compressors employ investment-cast aluminum impellers, a technology that was not refined enough at the period to be trusted.
A viable compromise was to gang three or four relatively small impellers on a common shaft and with their discharges in series. Researchers adjusted impeller diameters and shaft speeds to generate peripheral velocities of around 145 m/sec. This was hardly optimum, but viable in that each stage added compression to produce output pressures on the order of 1.5 atmospheres. What multi-stage compressors cost in terms of weight and efficiency was compensated for in reduced development time.
German researchers favored closed wheel impellers, with vanes sandwiched between two disks. Modern practice is to use open wheels so that vanes are exposed on one side. This configuration can handle air speeds of 300 m/sec and greater, which was far beyond the scope of the earlier technology.
Heat is byproduct of compression that, while making diesel engines possible, has unfortunate implications for supercharging. When we heat air in the free state, the air expands to become less dense. If confined to a sealed container, the heated air registers a pressure increase. An intake manifold, with valves on one end and pressurized air on the other, occupies some intermediate place between the open atmosphere and a closed container. Consequently, heat reduces air density and the number of oxygen molecules available to react with the fuel. Mechanical and pumping loses contribute to the heating effect. Because of their convoluted intake passages multi-stage centrifugal blowers generate considerably greater thermal contamination than equivalent single-stage machines.
Charge heating also promotes detonation and other problems associated with engine durability. Researchers on both sides of the conflict tolerated higher temperatures would be acceptable today. For example, when describing the charge heating that occurs at lower altitudes, W.G. Noack, writing in 1921, acknowledges the power loses, but adds "experiments have shown that most airplane engines can stand temperatures even above 100º C in front of the carburetor perfectly well, at least for short periods…".
Two decades later, von der Nüll cited a carburetor-inlet temperature of 80ºC as reasonable to avoid the onset of detonation . The SMA SR 305-230, a modern turbocharged diesel, used as conversion power for the Cessna 180Q, has, of course, no problem with detonation. But maximum manifold temperature is regulated by means of a charge cooler to 65ºC.
German researchers were aware of charge coolers and occasionally used them, but had reservations about the additional weight and pressure drop.
|Fig. 1. Centrifugal compressor. Impeller (A), Diffuser (B)
(NACA Report Technical Report No. 384)
|Fig. 2. Mercedes carburetor with a valve that closed automatically under boost pressure and an anti-backfire relief valve.
(NACA Technical Note No. 48)
Sizing and Rating
Matching a supercharger with an engine is a complex endeavor that even now, a century later, eludes complete mathematical description. German researchers simplified matters by approaching the problem from the direction of power output. Since contemporary engines developed between 10 and 11 hp per liter of displacement, an assumption about volumetric efficiency gave an air weight per horsepower hour of 3.5 to 3.6 kg. All that remained was to find a compressor with a similar adiabatic output adjusted for an air temperature at the desired critical altitude. An undersized compressor would fail to achieve this altitude when tested in the Zeppelin vacuum chamber; an oversized compressor would surge as the impeller vanes stalled from backflow.
Adiabatic efficiency was, and remains, the most popular gauge of supercharger efficiency. Adiabatic means that compression takes place in isolation, with heat neither lost or gained from external sources. Were such a compressor to exist, the only heat added to the charge would be the heat of compression, and the device would have 100% adiabatic efficiency. But real compressors suffer from volumetric and mechanical loses that add heat to the charge.
The Brown-Bovari compressor, the only one for which we have data, achieved an adiabatic efficiency of 68%, a figure that compares favorably with modern designs. For example, the Vortec SC-trim compressor peaks out at 75% with a fairly large mass air flow and pressure ratio "island" of 73%.
Upstream carburetor mounting would have simplified the installation and assisted in charge cooling as the fuel evaporated. The dangers associated with backfiring—flames in the intake track would vent through the carburetor intake—may have been what caused German researchers to mount the carburetor on the outlet side of the blower (Fig. 2). At any rate, fear of backfires was one reason why contemporary engine designers used conservative valve timing. Another justification for the blow-through carburetor was that it did not impede air flow to the impeller hub. Centrifugal pumps are very sensitive to suction-side restrictions.
Since the carburetor venturi was pressured, the float bowl had to come under boost pressure, either by encasing the entire carburetor in an air-tight box or, more conveniently, by running a separate line to the float-bowl vent. A spring-loaded relief valve vented backfires that otherwise would have damaged the blower, driving the impellers against their backing plates.
For the float to function, fuel pressure had to be regulated to about 2.5 psi above supercharger outlet pressure. Engineers at McCook Field were proud to have developed a fuel-pump pressure regulator that automatically maintained the correct pressure differential . The Germans employed similar fuel pumps or, in the case of large aircraft, supplied fuel from the main tanks by gravity to smaller, pressurized containers adjacent to each engine.
The preferred method of regulating boost was a variable restriction placed across the suction side of the compressor, which was adjusted by the pilot to maintain manifold pressure at normal levels. By the end of the war, Brown, Boveri and several other firms had proposed methods of automatic regulation.
Coupling the supercharger to a dedicated engine provided an additional form of boost control, since impeller speed could be varied without changing the throttle settings on the propulsion engines.
The rotational inertia of pump impellers that turn six or seven times faster than the crankshaft approximates the inertia of the propeller. Loading the opposite ends of a crankshaft in that manner must have induced many sleepless nights. The immediate danger was that the loads might resonate. In a system without some form of dampening, the theoretical peak load at resonance is infinite.
"Soft," energy-absorbing couplings and slip-prone centrifugal clutches reduced the danger of resonance and smoothed power transfer to the wide-ratio gearboxes, the most vulnerable components in the system. But de-coupling peak loads on one end of the crank did little to address the torsional vibration as the crankshaft wound and unwound with each change in engine speed. The springs on Brown-Bolvari coupling shown in Figure 3 reacted to these torque reversals by compressing to a quarter of their normal length.
As near as one can tell, torsional vibration was simply ignored. Those involved in supercharger development neither had the mandate nor the time to address engine problems. War planes were, after all, sacrificial items.
|Fig. 3. Brown-Boveri spring-dampened coupling used single-engine tests.
(NACA Technical Note No. 48)
|Fig. 4. The intake manifold of the Mercedes D.IVa was, it has been reported, an expedient to reduce engine width for in-fuselage mounting. It also simplified supercharger plumbing. (Wiki Commons)|
Although other rotary engines had been experimented with, by early 1917 the 260-hp Mercedes/Daimler D.IVa became the only serious candidate for forced induction (Fig. 4). Displacing 21.72 liters, the inline, 24-valve, six-cylinder engine developed 252 hp at 1,400 rpm. It powered various A.E.G., Friedrichen, Gotha and Zeppelin-Staaken bombers, as well as Rumpler and Albatros reconnaissance aircraft. High-altitude performance was a critical defensive measure for these machines whose targets were defended by increasingly accurate anti-aircraft fire.
The twin-jet Mercedes updraft carburetor sat low behind the engine, an arrangement that allowed ample room for pressure connections to the carburetor intake and float chamber. The uneven mixture distribution resulting from a carburetor at the far end the manifold was in itself an argument for forced induction.
Another feature of the intake system was less welcome; these engines drew carburetor air through a channel under the crankcase that functioned as an oil cooler. The resulting 15º C to 20º C increase in inlet air temperature added to the thermal load generated by the compressor. But when the compressor opened directly to the atmosphere, oil temperatures soared to between 85º C and 95º C on warm days. The lubricant thinned and failed to reach the piston pins. The fix entailed further complexity in the form of a second oil pump, driven by a flexible shaft, and an oil cooler.
The difficulties posed by forced induction required teams of academically trained engineers to resolve. There was little room for freelance inventors. Nor did the German government have research facilities comparable to those at the Royal Aircraft Factory or later at NACA and the U.S. Bureau of Standards. Consequently private firms, either acting on their own or under state contract, took responsibility for supercharger development.
Otto Schwade & Co.
Hans James Schwade, the eldest son of the founder of Otto Schwade & Co., earned his pilot's license in 1910 and subsequently built a number of airplanes, most of them Farman-based. He also steered the firm, already well known for its gas and liquid pumps, into aircraft engine production. The improved Gnome copies were built in sizes of from 50 hp to 160 hp. Hans James also organized a pilot's school less-affluent students, many of whom went on to serve in the military.
With that background, it's not surprising that Schwade & Co. was the first to build aircraft superchargers in Germany. Figure 5 illustrates an early version of the supercharger mounted on one of the company's "steel-hearted" rotaries.
As fitted to the Mercedes D.IIIa, the blower consisted of three stages with an integral gear box, lubricated from an oil reservoir. What provision was made for shaft-bearing lubrication is unclear. At 1,400 engine rpm, the output gear turned 10,500 rpm, a speed that gave the 250-mm impeller wheels a tip velocity of about 140 m/sec. The compressor pumped 1,000 kg of air/hr at a 1.52 pressure ratio which should have raised the critical altitude beyond the reported 3,500m.
A centrifugal clutch delayed blower engagement until engine speed reached 600 rpm. The clutch made cranking easier and isolated the gear train from the rapid acceleration that accompanies initial firing impulses. Four brass friction shoes softened power transfer to isolate torque reversals.
The Schwade supercharger underwent flight testing in at least one Gotha G-type bomber, which crashed for unrelated reasons . The project was later shelved in anticipation of a variable pitch propeller.
|Fig. 5. Schwade supercharger as fitted to one of the firm's rotary engines.
(NACA Technical Note No. 48)
|Fig. 6. Provisions A.E.G. made to protect the crankshaft put the supercharger well aft of the engine.
(Aerial Age Weekly, 1921)
The A.E.C. turbine division built a three-stage blower for the D.IVa engine that turned 10,000 rpm to generate a pressure ratio of 1.7, or enough to enable normal operation at 4,000 m (Fig. 6). Steel impeller wheels had riveted vanes, probably to reduce manufacturing costs.
Considerable thought went into the drive mechanism, which consisted of a centrifugal clutch, universal joint and driveshaft. The clutch engaged almost immediately upon startup. Because the compressor mounted independently of the engine, the universal joint was needed to compensate for airframe distortion. The driveshaft acted as a torsion bar to soften drive inputs. And should the supercharger bind, the shaft would shear to protect the crankshaft.
A.E.G. also experimented with superchargers driven by dedicated engines housed in the fuselage of large bombers. As far as it can be determined, none of these ingenious devices were flight tested.
By using 400-mm impellers, Siemens-Schuckert was able to obtain the desired 145 m/sec peripheral a moderate shaft speed at 6,900 rpm. The company also experimented with relocating the power takeoff to the screw end of the crankshaft. In static tests of the Mercedes application, the three-stage compressor demonstrated a pressure ratio of 1.26 (Fig. 7).
|Fig. 7. Mercedes pusher engine with a three-stage Siemens-Schuckert driven from the propeller end of the crankshaft.
(Aerial Age Weekly, 1921)
|Fig. 8. Zeppelin-Staaken R.VI Riesenflugzeug ("giant aircraft"). Eighteen of these machines were built in the Staaken suburb of Berlin. (Wiki Commons)|
The supercharged Zeppelin-Staaken R.VI combined the best of German compressor technology with the world's largest series-produced wooden aircraft (Fig. 8). With a wingspan of 42.2 m—more than 10 m greater than the B-17—and a takeoff weight of 11,848 kg, the monster required an 18-wheel landing gear. Entering production in late 1916, the bomber was initially posted to the Eastern Front, then transferred to the west where it participated in raids on the English capital, culminating in the bombing of Royal Hospital Chelsea. Some historians view the R.VI as a precursor to the V-2 rocket. That is, a machine that absorbed vast engineering and capital resources on the promise of strategic bombing and ended as a terror weapon. At any rate, the R.VI was an impressive flying machine.
R.VI serial number R30/16 was used as the test bed for supercharger experiments. A Brown-Bolvari four-stage compressor powered by a 160-hp Mercedes D.II engine was installed in the central fuselage. While a dedicated engine might seem a bit bizarre, it provided a failsafe approach to supercharging. Failure of the central supercharger would do no more than reduce power to the drive engines. Engineers of the period were certainly familiar with air-injected diesel engines that used an auxiliary engine to power the compressor. Dedicated supercharger drive also eliminated most concerns about crankshaft flexure and permitted in-flight servicing, a feature that was considered important enough to provide space in each propulsion-engine nacelle for a mechanic.
Turning at 6,000 rpm, the 470-mm impellers delivered 4600 kg/hr air to the five engines at an initial pressure of 0.52 atm and a final pressure of 1.0 atm (Figs. 9 and10). Power consumption was in the neighborhood of 125 hp at full boost.
Unfortunately we do not have data on flight tests, of which several were made. In its normal configuration the R.VI had a service ceiling of 4,300m and a rate of climb of 101 m/min.
A leather-lined slip coupling, supplemented by a 20-kg flywheel, supplied power to the gearbox. The gear train consisted of a 13-tooth pinion and a 54-tooth drive gear for a speed multiplication of 4.13 to 1. Gears were made from chrome-nickel steel, case-hardened and precision-ground by an outside supplier. A low-speed shaft drove the oil pump.
Tests were still underway when the armistice was signed.
|Fig. 9. Brown-Bolvari four-stage compresser used on Zeppelin-Staaken.
(NACA Technical Note No. 48)
|Fig. 10. Compressor discharge line. Aluminum ductwork and hoses conveyed boost to all five engines.
(NACA Technical Note No. 48)
Supercharging aircraft engines was an engineering problem—theoretical knowledge of radial—flow compressors was adequate for the task. Tip speeds of around 150 m/sec produced adequate pressure ratios while avoiding impeller-strength and supersonic-flow problems. What remained was to reduce the weight of compressors evolved from industrial practice, isolate the drive mechanism from crankshaft oscillations, and find ways of adapting the technology to pre-existing engines. According to their own accounts, which are all we have, the Germans made good progress in these areas. Had the war lasted another year, it seems likely that supercharged aircraft would have become operational.
Probably because some many firms worked independently, German researchers took a variety of approaches to forced induction, ameliorating drive line torque reversals in unique ways, investigating different Had the war gone on for another year, there is little doubt that German supercharged aircraft would have entered combat.
1. Hallett, Major G. E., USA, "Superchargers and Supercharged Engines," SAE Transactions, Vol. 15, Part 1, 1921. Major Hallett is remembered for his work at McCook field with Dr. Stanford Moss.
2. Taylor, C. Fayette, Aircraft Propulsion: A Review of Aircraft Piston Engines, Smithsonian Institution Press, Washington, D.C., 1971, p.
3. Noack, W.G., "Airplane Superchargers," NACA Technical Note No. 48, May, 1921.
5. Nüll, von der, "The Design of Airplane-Engine Superchargers," NACA Technical Memorandum, No. 839, 1937.