Currently, gradient descent and eta scan functions use the resonance frequencies in the center_freq_array variable as their starting point. If the TESs are biased, then the resonator could be shifted by ~100 kHz depending on phase of the SQUID induced by the TES bias current. This means that these functions only work when the detectors are unbiased, which is problematic if we want to track the phase of the cable and system as the 300K temp varies without rebiasing detectors.
Describe the solution you'd like
Instead of using the center_freq_array values as the starting points, these functions should use the actual tracked frequency as the starting point, since this will more accurately reflect the frequency of the resonator. I think this can be acquired through take_debug_data before calling the grad-descent and eta-scan procedures.
Describe the problem
Currently, gradient descent and eta scan functions use the resonance frequencies in the
center_freq_arrayvariable as their starting point. If the TESs are biased, then the resonator could be shifted by ~100 kHz depending on phase of the SQUID induced by the TES bias current. This means that these functions only work when the detectors are unbiased, which is problematic if we want to track the phase of the cable and system as the 300K temp varies without rebiasing detectors.Describe the solution you'd like
Instead of using the
center_freq_arrayvalues as the starting points, these functions should use the actual tracked frequency as the starting point, since this will more accurately reflect the frequency of the resonator. I think this can be acquired through take_debug_data before calling the grad-descent and eta-scan procedures.