Establishing Creditability: The Early Days

Following the operational readiness of the Langley 15-Foot Free-Spinning Tunnel in 1935, initial testing centered on establishing correlation with full-scale flight-test results of spinning behavior for the XN2Y-1 and F4B-2 biplanes.[506] Critical comparisons of earlier results obtained on small-scale models from the Langley 5-Foot Vertical Tunnel and full-scale flight tests indicated considerable scale effects on aerodynamic char­acteristics; therefore, calibration tests in the new tunnel were deemed imperative. The results of the tests for the two biplane models were very encouraging in terms of the nature of recovery characteristics and served to inspire confidence in the testing technique and promote future tests. During those prewar years, the NACA staff was afforded time to con­duct fundamental research studies and to make general conclusions for emerging monoplane designs. Systematic series of investigations were conducted in which, for example, models were tested for combinations

of eight different wings and three different tails.[507] Other investigations of tunnel-to-flight correlations occurred, including comparison of results for the BT-9 monoplane trainer.

As experience with spin tunnel testing increased, researchers began to observe more troublesome differences between results obtained in flight and in the tunnel. The effects of Reynolds number, model accuracies, control-surface rigging of full-scale aircraft, propeller slipstream effects not present during unpowered model tests, and other factors became appreciated to the point that a general philosophy began to emerge for which model tests were viewed as good predictors of full-scale charac­teristics but also examples of poor correlation that required even more correlation studies and a conservative interpretation of model results. Critics of small-scale model testing did not accept a growing philosophy that spin predictions were an "art” based on extensive testing to deter­mine the relative sensitivity of results to configuration variables, model damage, and testing technique. Nonetheless, pressure mounted to arrive at design guidelines for satisfactory spin recovery characteristics.