A colleague chose a motor based on its efficiency (90% efficiency specified by vendor). We are running it as a generator in a test, and the measured efficiency is much lower (70% efficiency). Is that unexpected? My EE colleagues (I'm a lowly programmer) say that means there is a measurement error since the motoring and generating efficiency should be much closer, but I want to make sure: Could the same device to be an efficient motor but an inefficient generator? If so, why? Perhaps control or windings optimized for motoring would not be optimal for generating?