What can go wrong in stochastic cooling systems

This paper discusses very practical aspects of stochastic cooling systems both during construction, running-in, operation and trouble shooting. Due to the high electronic gain, high sensitivity and large bandwidth of such systems, precautions have to be taken to avoid all sorts of EMI/EMC related problems as well as crosstalk and self-oscillations. Since un-intended beam heating is always much more efficient than the desired cooling the overall performance depends critically on avoiding this heating which often takes places outside the nominal frequency band of operation. Another important aspect is “cross heating”, i.e., unavoidable crosstalk from longitudinal to transverse systems and vice versa. Obviously adequate measurement procedures with beam for gain phase and optimum delay are mandatory and certain caveats and hints are given. The paper concludes with a listing of unusual and unexpected problems found during many years of operation of such systems at CERN.