This case study presents a new automated production noise test for power steering pumps. The test included adaptive noise cancellation, and a neural network implementation. The result mapped the pump acceleration signature into an objective repeatable noise metric.The test algorithm was a distributed DSP architecture designed for real-time measurement and decision processing. It was implemented with no increase in test cycle time. It accomplished the correlation of in-vehicle power steering pump noise to it's vibration characteristics, and retrofitting of accelerometers in place of microphones for acceptance testing.