Making a pyDARM model¶

This page describes how pyDARM takes parameters supplied to construct a model of the DARM servo loop. These parameters are used to compute transfer functions between elements of the loop.

Quick reference¶

An example model file can be found in the pyDARM git repository.

Details¶

PyDARM models the DARM loop by drawing on parameters defined in a configuration file or a configuration string. The configuration takes the form of sections and key-value pairs. As an explicit example of what a section and key-value pair looks like:

[section]
key = value

PyDARM sections mirror the parts of the interferometer we are attempting to model:

[sensing]
[digital]
[actuation]
[actuation_x_arm]
[actuation_y_arm]
[pcal]
[calcs]
[FIR]

Other sections can also be defined:

[metadata]
[interferometer]

When modeling or estimating the uncertainty, more sections can be included in the same configuration file or string or in a separate file/string:

[sensing-measurement]
[x-arm-measurement]
[y-arm-measurement]
[x-arm-sus-cal-lines]
[y-arm-sus-cal-lines]
[x-arm-pcal-cal-lines]
[y-arm-pcal-cal-lines]
[tdcf-data]
[sensing-tdcf]
[x-arm-tdcf]
[y-arm-tdcf]
[hoft-tdcf-data-application]
[sample-tdcf]
[pcal]
[epics-records-channels]

Within each section, key-value pairs need to be defined in order to set parameters and compute transfer functions

>>> import pydarm
>>> import numpy
>>> frequencies = numpy.logspace(1, 3, 10)
>>> D = pydarm.darm.DigitalModel('''[digital]
... digital_filter_file = test/H1OMC_1239468752.txt
... digital_filter_bank = LSC_DARM1, LSC_DARM2
... digital_filter_modules = 1,2,3,4,7,9,10: 3,4,5,6,7
... digital_filter_gain = 400,1''')
>>> D.compute_response(frequencies)
array([5.98805825e+08-2.01229609e+09j,
       3.07291968e+09-6.39863354e+07j,
       4.46221271e+09+2.83011189e+09j,
       6.20899890e+09+6.53258031e+09j,
       1.02379652e+10+1.13041901e+10j,
       2.06788315e+10+1.56633213e+10j,
       4.46173195e+10+7.26147430e+09j,
       3.96197850e+10-7.25240470e+10j,
       1.16786264e+10-9.56665096e+10j,
       1.64517886e+06+1.22091523e+05j])

In the above example, we have have to point the pyDARM code at a particular FOTON filter file in the test directory, so we set digital_filter_file = test/H1OMC_1239468752.txt. We specify which filter bank in the file that we are considering, separated by commas, so we set digital_filter_bank = LSC_DARM1, LSC_DARM2. In each bank, the modules that we consider are comma-separated and the bank lists are separated by a colon. So we set digital_filter_modules = 1,2,3,4,7,9,10: 3,4,5,6,7. Lastly the gain of each bank is a comma-separated list, so we set digital_filter_gain = 400,1.

The full DARM loop can be specified via such sections and key-value pairs. An example model file is shown here and we use it to define an instance of a DARM model class (object) as:

>>> darm = pydarm.darm.DARMModel('H1_20190416.ini')

Data files¶

Certain files are needed in order to compute transfer functions of DARM loop elements:

[sensing]
analog_anti_aliasing_file
omc_filter_file

[digital]
digital_filter_file

[actuation_x_arm]
sus_filter_file
suspension_file
analog_anti_imaging_file

And any other transfer functions of associated calibration products:

[pcal]
pcal_filter_file
analog_anti_aliasing_file

[calcs]
foton_invsensing_tf
foton_delay_filter_tf
foton_deltal_whitening_tf

Uncertainty estimation often requires data files of MCMC or GPR posteriors HDF5 files.

These files are not provided as part of the pyDARM package, instead being kept in a separate repository. PyDARM will attempt to open files listed as values for the above keys prefixed by the following preferred ordering:

1. The CAL_DATA_ROOT environment variable

2. The cal_data_root configuration key value

that is, CAL_DATA_ROOT environment variable overrides the cal_data_root configuration key value. For example, if data files are stored at /home/albert.einstein/Calibration/data_files/ then setting CAL_DATA_ROOT=/home/albert.einstein/Calibration/data_files will enable pyDARM to locate and open the files. If data files are stored across many different directories, then CAL_DATA_ROOT can be set to the path depth that is common amongst all the data files.