My very 1st control problem was a magnetic levitator.  Remembered from those days that in order for the magnetic levitator to be stable, they needed to highpass filter the hall effect sensor.  In those days, they did it with purely analog circuits.  Highpass filtering the sensor was how they overcame laggy actuators.  It would be an interesting project to make different FIR filters & observe the effect on every feedback loop in the apartment, but a more promising solution appeared.

Manely, D feedback was never being computed.  The bug started on Dec 7 2014 when the truck 1st started moving with an analog gyro.  It was later switched to a digital gyro, but the source code still read an analog gyro reading of 0.  Such is the importance of logging telemetry.  It took a while to figure out the easiest way was an Odroid overlaying data points on video.