Reports until 12:07, Monday 18 April 2016
H1 ISC
ross.kennedy@LIGO.ORG - posted 12:07, Monday 18 April 2016 (26633)
New linetracking filter (iWave) in PI model

Ross, Tega, Terra, Ed Daw 

We have updated the h1susetmxpi model to include a linetracking filter (iwave) in order to more accurately track parametric instabilities. The filter is placed between the bandpass and damping filters in each PI mode. The basic idea of iwave is that it is a resonant filter that runs with a phase lock loop. The loop tracks changes in frequency and amplitude of the signal that it is closest to in frequency and adjusts the filter to resonate at the tracked frequency. Two parameters should be set before the line tracker is implemented, the starting frequency and tau. The starting frequency should be set to the closest known line frequency of the signal you want to track. The tau parameter sets the response time of the filter but also the bandwidth. For higher Q signals it is best to use a longer tau parameter. We have also included a phase offset to the model for the linetracker which can be tuned to ensure the ESD is damping the PI with the correct phase. The medm screens mentioned in Terra's alog have been updated to include this filter.

The SUS_CUST_PI_MODE screen should now have a link to the top screen for iwave where tau can be changed and displays a strip chart of the error signal (black), input signal (blue) and frequency tracking (red). The error signal can be used to determine how well the signal is being tracked i.e. it should get smaller as the signal is being more accurately tracked. This screen also has a link to the actual iwave filter screen where the starting frequency can be changed and output monitors including ones for frequncy and amplitude are displayed. The reset button in the filter can be used to reinitialise the filter with the selected tau and starting frequency. A bypass switch is included to allow data to be unchanged when passing through the filter if necessary. We have also updated the mode screen so that the QPD monitors are now looking at epics outputs in the h1etmxpi model rather than the filtered outputs.

We will hopefully begin testing this method of damping in comparison to just using the bandpass filters soon. This will be done by turning off the ring heater, waiting for the signal to ring up and then applying the damping signal to the ESD. We will need a 15W or higher laser power in order to ring up the 15440 Hz mode which has been seen previously.

Images attached to this report