Testing was very similar to today, in that rigs are used to test each sub-component of a major engine in isolation before the full engine is made.
In terms of vibration analysis which would today be carried out in large part by computer, it was not ignored at the time, basic calculations were made during design to ensure the parts did not vibrate, but this was also tested after manufacture by sprinkling crocus powder on the part, and connecting it to a loudspeaker and a signal generator – then playing low to very high frequency sound through the part until the powder formed distinct patterns. By this means the so called “natural frequencies” of the part could be found quite accurately.
Components which failed through basic overloading and fatigue could be tested to see the stress patterns in three dimensions by use of Photo-elastic method or also Paint-Lacquer method (above).
In this way, load is placed on the part to simulate a running condition point, and then the brittle lacquer which had been sprayed into the part beforehand fractures showing the “stress-lines” which allow points of stress concentration to be found. Above the hole in the “small-end” of the connecting rod is highlighted as a failure point.
Here the photo-elastic method is used to see the stresses in a turbine blade root. It was even possible by the “freezing” method to see these stresses in 3D by carrying out the test on the plastic part near its melting point, then cooling the part under loading. It would then retain its strain after being removed from the test-rig and so it could even be cut-up and sectioned to see complex stress patterns.
Most engine firms did extensive photo-elastic stress calculations in Germany, Britain and America during the Second World War.
Hydraulic rigs were available to carry out fatigue testing by continually applying and removing the load for days or weeks to see fatigue behaviour.
When failures occurred these were sent to the inspection department and a complete metallurgical study would be conducted, with spectroscopy to check the metal was pure and correctly smelted. Microscopes were good enough to inspect the structure of the steels and in this case a failed Junkers-Jumo crankshaft is having the Nitriding checked as a possible cause of failure.
The torsional failure is evident on the journal surface where the crank proceeds 45 degrees across its outer face.