Abstract
Reduction of the coil driver noise was suggested in klog#31771 and reasonable improvement was confirmed by suggested way in klog#31774.A new code was put in KAGRA system and build test was done with some VIS model files.
This update will be deployed on this Friday.
Details
As reported in klog#31774, I implemented the new coil driver BIO controller to reduce output noise of the coil driver chassis. This change is to engage the latter stage of de-whitening first instead of engaging the front stage of one first. We can engage 1 (the third) de-whitening, 2 (the second and the third) de-whitening, or 3 (the first, the second, and the third) de-whitening by setting 2, 3, or 4 as the state number of coil driver BIO, which is same manner as before.A new code is save as /opt/rtcds/userapps/release/vis/common/src/CD_STATE_MACHINE_KAGRA.c.lownoise and the original code is kept as /opt/rtcds/userapps/release/vis/common/src/CD_STATE_MACHINE_KAGRA.c.orig. Simlink models load /opt/rtcds/userapps/release/vis/common/src/CD_STATE_MACHINE_KAGRA.c which is a symbolic link to the new code. Compile test with this new code was completed for k1vismce, k1visprmt, k1visprmp, k1visbst, k1visbsp, k1visetmxt, and k1visetmxp. Actual deployment for all VIS models will be done on this Friday.
Operation check was done in the standalone system. On the other hand, I haven't check that a possibility to kick IFO by this new implementation larger than before at the timing of engaging/disengaging de-whitening filters. We need IFO recovery to check it. If we will face some problem, we can go back to the original implementation to rebuild VIS models after switching a destination of the symbolic link or to use 6, 8, and 4 as state number of BIO instead of 2, 3, and 4 which should be easily set on VIS_params.py.