Hsiang-Yu Huang
Compare digital AA filter between pyDARM production and Foton export (aka Dec65K in H1OMC_DCPD_A0)
I use digital_aa_or_ai_filter_response() in pyDARM to compute this.
I set up the following in sensing part of INI file.
[sensing]
anti_aliasing_rate_string = 512k-daq
anti_aliasing_method = biquad
Attachment:
Figure. 1 : freq. vs. magnitude(dB) plot by Foton
Figure. 2 : freq. vs. magnitude plot also by Foton
Figure. 3 : compare digital AA filter between pyDARM production and Foton export
Figure. 4 : zoom-in plot from Figure. 3. Set freq. limit from 10 Hz to 2000 Hz in all subplots. Set up y-axis limit of magntiude residual plot to +- 0.1
Phase limit to +- 5 degree.
Figure 5 : I export Dec65K in this OMC-DCPD_A0.
Discussion:
For magnitdue part, they are consisent below 1 kHz
However, we have phase loss from ~ 100 Hz. In 1 kHz, there is almost 5 degree phase loss.
The 512k-daq string refers to the 8x DAQ downsampling filter and not the filter module from the FOTON filter file. You should use the omc_filter_file section in the ini file with gain, modules, and bank name to get the right filter coefficients. This is the reason why you see such a large difference.
Hsaing-Yu Huang
I misunderstand the function I used. I re-calculate the digital AA by omc_digital_filters_response() in pyDARM sensing.
