It was my contribution to the physical experiments performed in the FLNR. FOBOS is a fixed target experiment. The accelerator facility U400M provides a pulsing beam of heavy ions as projectiles. Their kinetic energy and "arrival time" have to be measured for the furhter data analysis. This measurement was managed with a small "Phoswitch" detector mounted at about 10 degrees to the beam path. Fast signal component from the plastic layer is applicable for the TOF-spectrometry, while the long signal tail from the BGO-well is used for the total kinetic energy (TKE) measurement. The TKE-TOF correlation matrix represents the beam quality.
  During the long-term experiment there was about 10% walk of the timing parameter observed, while the beam energy was rather stable. It could be due to a threshold instability of electronic discriminators which provide the "Start" and the "Stop" signals for the TOF method. On the other hand this could show an imperfectness of the accelerating system itself.
  We assembled a special analyzer and made sure that all FOBOS electronics had a good stability and provided a high timing resolution. Then the whole RF-chain of the U400M and its master generator G5-158 were tested. The frequency spectrum obtained from the sine-oscillator G5-158 was similar to the physical TOF distribution. The same effect was found for all other generators whose principle of operation was based on analog feedback circuitry for the frequency control.
  The unstable (floating) frequency doesn't affect the beam energy, because those ions in the wrong timing phase don't satisfy the stability principle (discovered by V.I.Veksler and E.McMillan) and escape from the acceleration process. The precise frequency is important in the TOF method, where it determines the reaction time and starts the data acquisition system. Thus every physical event contains a systematic TOF-error, which can be corrected in the offline analysis. A new subsystem was proposed which could measure the period of the accelerating frequency and add this value to the data stream. Another way to solve this problem could be to exchange the analog generator by the digital frequency synthesizer.