Chapter 4: The High-Speed Propeller Program
[121] No one in the 8-foot tunnel group had had any experience in propeller research except Stack. He had been periodically involved with PRT programs through his high-speed airfoil work, and since 1938 had been consulting with E. Hartman and others on the design of the high-speed propellers to be used in the forthcoming high-speed wind tunnel program. He continued to be deeply involved in the design of the test propellers, along with L. Feldman of the 8-foot group and J. Delano who was the designated project engineer for the emergency program. The test propellers that were designed represented major improvements over the best propellers then in service (fig. 29). They were generally thinner, ...

5 prototype propellar blade designs
[122] FIGURE 29.-Blade shapes tested in the Emergency Propeller Program in the 8-Foot High-Speed Tunnel.
...tapering to about 4 percent thickness ratio near the tip from about 12.5 percent at the spinner. It had been obvious for some time that the thick root sections exposed on many then-operational propellers would suffer compressibility losses at high forward speeds, adding to the tip-region losses. These inefficient shank sections were completely covered in the NACA program by the large spinners employed on the dynamometers, a spinner diameter one-third the propeller diameter being used in the 8-foot tunnel tests. Blade widths of one and one-half, two, and three times the normal width were included because at a given power input the blade lift coefficients were correspondingly reduced and the critical speeds increased. Or, for a given lift coefficient and critical speed the power absorption could be correspondingly increased. All these important improvements were quite independent of the choice of blade section shape. The 16-series sections at that time were thought to offer improvements in critical speed of the order of 50 feet-per-second over some of the older sections, and they were used in nearly all the test propellers. Since 1938, Stack had been vigorously selling the 16-series to propeller designers and to NACA managers, and we were now under [123] considerable pressure to confirm the advertised gains in an actual propeller test.
Following PRT practice, we selected a more-or-less representative nacelle for the 4-foot propeller tests. What is actually measured in a test of this kind is more properly termed "propulsive efficiency" of the propeller/ nacelle unit, rather than "propeller efficiency." That is, the thrust determination includes effects of the slipstream on the body and support drag, and other secondary effects not present in tests of the forces on the propeller itself. The nacelle had one unusual feature which considerably complicated both its structural development and the problem of determining accurate tare forces, an open-nose spinner through which passed a flow of air representative of that required for cooling a large radial engine (fig. 30). The high-speed aerodynamics of this arrangement had been developed in an 8-foot tunnel program to have a critical....

photo of wind tunnel propeller
FIGURE 30.-The 200-hp Emergency Propeller Dynamometer in the 8-Foot High-Speed Tunnel with 4-foot diameter standard blades.
[124] ....Mach number higher than the highest propeller test speed (see Chapter V), and this particular design had been the subject of a recent study of pursuit-airplane performance in the 19-foot tunnel (ref. 136).
The equipment needed for the 200-hp dynamometer was more readily obtainable than that for its larger counterpart at 16-foot. By December 1941 it was ready for the first tests of two-blade propellers. Reflecting our special interests, the first two test propellers were identical except for blade section shape, one having 16-series and the other having conventional Clark Y sections. To our dismay and disappointment, the 16-series propeller showed no advantage at high speeds; in fact the Clark Y appeared slightly better. Stack asserted emphatically that some systematic error must be present in the data and he assigned me the task of finding it. I had previously been only peripherally involved with the propeller program except for six weeks' work in the spring of 1941 on a theoretical analysis to determine the tunnel-wall corrections that would have to be applied. There were indeed several sources of significant error, particularly in the strain gage system used to measure torque and in the thrust and torque tares due to the blower-spinner. However, these were all either random in character or of about the same magnitude for both Clark Y and 16-series propellers. Regretfully, I concluded that any advantage of the 16-series was too small to be discernible within the existing rather poor limits of accuracy. The better part of the following year was devoted to improving the accuracy. Strain gages at that time were in an early stage for applications of this kind, but eventually acceptable accuracy was obtained through frequent calibrations. Satisfying confirmation of the overall accuracy including the tunnel-wall effect corrections was obtained in 1945 by running comparative tests of the 4-foot dynamometer in the 16-foot tunnel (ref. 137).
The probable explanation of the nearly equal high-speed performance of the Clark Y and 16-series propellers of equal thickness gradually became clear with additional two-dimensional testing and comparisons with other sections. Although the 16-series sections had higher critical speeds near their design lift coefficients, their force-break speeds were often not much higher than those of other good sections because the occurrence of shock at the rear of the 16-series profiles tended to produce separation shortly after the critical speed was reached (ref. 52). The [125] sections for which the shocks occurred farther forward could in many cases significantly exceed the critical speed without encountering force break (see p. 36ff.). In spite of their failure to show any marked high-speed performance advantage over other good high-speed sections, the 16-series sections have been generally used by propeller designers for other reasons, particularly for the structural advantages of propeller blades which are relatively thick in the trailing edge region, compared, or example, to the cusped low-drag sections.
The results of the Clark Y propeller tests were never published and it was never tested again. Perhaps the relatively poor accuracy of these first tests justified withholding these data, but there was little real doubt in our minds that the two propellers had nearly equal performance.
On the positive side, these first high-speed wind tunnel tests of improved propellers showed that propulsive efficiencies in the range of 85 to 90 percent could be maintained to forward speeds of 500 mph, provided that high blade angles (of the order of 60°) were used to keep the rotational speeds low enough to avoid compressibility losses. Generally, performance started to deteriorate sharply if the tip Mach numbers exceeded about 0.91, a value about 0.05 to 0.10 higher than expected from section data, the discrepancy being explained by three-, dimensional tip relief effects (ref. 136). The effects of increased solidity (ref. 138), shank shape (ref. 139), pitch distribution (ref. 140), and camber (ref. 141) were found to be consistent with expectations from the two-dimensional section data. In reviewing these results from the emergency program (ref. 142), E. C. Draley claimed that a 100-mph speed gain had been achieved over "typical previous propellers" by use of 16-series airfoils, thin sections, and ideal Betz distributions. However, he did not identify the previous propellers, but evidently assumed that they had thick shanks and thicker blade sections than these improved propellers.