I have a very noisy data from a particle accelerator and I want to improve the signal (in every way it's possible): As background: It's data from an electron accelerator which produced Bremsstrahlung when distracted (so it emitted photons towards our detector). Signals have been recorded by coincidence between a tagger for the electron beam and the readout system when the photons hit our detector. Unfortunately our system had a bad shielding/grounding (one of the things we improved as a result). But nevertheless to get a better overview we want to analyze the signals respectively want to filter the signals. My first step is to implement a low pass filter into the software. Can you advice something else? From our horizon most of the noise comes from the avalanche photo diodes. So the readout chain is as following: Photon hits detector -> detector produces light -> light hits two neighbouring avalanche photo diodes -> signal is amplified and shaped by an ASIC -> signal is stored after being transmitted through some electronic components.

For a better feeling I attached some signals. Two adjacent signals represent one detector signal. I picked a few very different ones.

I would highly appreciate any help!

enter image description here

enter image description here

  • $\begingroup$ What about a so-called matched filter? $\endgroup$ – Ben Jun 15 '17 at 20:58
  • $\begingroup$ what is the purpose of filtering? what do you want obtain from these signal? If you want to detect an event which always produce the same signal the best way is to use matched filers. $\endgroup$ – Mohammad M Jun 18 '17 at 21:39
  • $\begingroup$ Correct. The ADCs always deliver the same peaks respectively the same lineshapes. They differ only in height and width but have always the same shape. Thanks! I'll try this. $\endgroup$ – Ben Jun 20 '17 at 0:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.