Automatic calibration of multichannel SDR readout electronics and correction of the IQ-imbalance of parallel input signals

by Eray Özer

Thursday Oct 13, 2022, 2:00 PM → 2:45 PM Europe/Berlin

https://cern.zoom.us/j/97803511429?pwd=T0tseVR5NUJZTUlHVUs4M0txMFFWUT09 (Zoom)

Presentation of the Master thesis of Eray Özer with the topic: Automatic calibration of multichannel SDR readout electronics and correction of the IQ-imbalance of parallel input signals

The ECHo experiment aims to reach sub-eV sensitivity on the upper limit of the electron neutrino mass by analysing the calorimetrically measured electron energy during decay of the 163-Holmium isotope. To efficiently read out the metallic magnetic calorimeters (MMC), a SQUID-based multiplexing method can be utilized. The frequency comb containing the resonance frequencies occupies the 4-8GHz band. Several IQ mixers are used to convert the signal between radio frequency and baseband for further processing.
Component tolerances and length mismatches in the signal path lead to imperfect mixing and produce unwanted image frequencies which decrease the Signal-to-noise ratio significantly. Applying specific offset values for the phase and amplitude on either I or Q data improves the image rejection ratio and allows more sensitive measurement results. The optimal correction values for each tone can be determined iteratively with a brute-force method, however correcting all channels with this approach is time consuming. The novel algorithm presented in this work enables the correction of the phase and amplitude mismatch by efficiently checking each tone’s image rejection ratio. The offset values are stored internally on the platform and applied automatically for tone generation. Control over the correction algorithm as well as the read and write access to the offset files is provided by a user-friendly GUI.

Lecture language: English