|
Size: 3776
Comment:
|
Size: 3503
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 5: | Line 5: |
| ==== Report for the week of 3/11 ==== * Fabian helped Fujii-kun with the PRM most of the week. Hirata-san also helped at various times. * Fabian monitored the SRM yaw drift. It gradually reversed but so far remained small. We will try resetting it to zero with the pico and see if the trend changes. * Mark and Hirata-san debugged the BS: * We inspected the payload section and quickly found the problem: two RM EQ stops were touching, on the +L (-X,+Y) side at the top. This was probably because the stops had been adjusted with the IP unlocked and sitting a few mm to -L, whereas now we have the DCCTRL running and have been locking the IP at the LVDT zero. We drove the IP to L=+1.5 mm as a near-worst case, and backed off all the stops to give several mm of margin. Then we zeroed the IP pitch, roll and yaw with the picos. * We also opened the top of the BS and confirmed that the mechanical zero defined by the locking mechanism agreed with the LVDT zero. It was very close in L and T, but -728 µrad in Y, driven mostly by a -805 µm offset in H3. That much is probably not a dumb error but just a difference in technique among the three people using calipers to take data for the calibration. * We then prepared the BS for close-up: * We wiped and inspected the O-ring for the lid and replaced the lid. * We wiped all accessible surfaces in the lower half of the tank, rescued some dropped items (some PEEK tube and a small nut) and un-immobilized the LBB. * After all the above, the OL was a bit off-center but still on the QPD, and Nagano-san fairly quickly recovered the original alignment. * Terrence rebuilt the BS OL according to a new layout suggested by Simon, which avoids scrambling the pitch/yaw at the first steering mirror and greatly reduces the amount of diagonalization required. * Terrence noticed some anomalous cross coupling on the SR2 (but not the SR3) and we investigated. * We suspected it might be an odd position of the beam spot on the optic, so we took pictures through the camera and LED viewports of SR3, and opened the SR2 tank and used a white target, but the beam spots were in very similar positions - close to central vertically and about 3 cm to the left of center horizontally. * We suspected it might be something touching at the IRM (because the IM OSEM TFs were fine), so we checked using a stick mirror with the IP centered, but all EQ stops were well backed off. * We suspected it might be a ghost beam from the LEN/TILT beamsplitter (and one was clearly visible on the SR2 TILT QPD) so Terrence set up a beam block, which improved things slightly but not completely. * Finally Terrence discovered that the problem was mainly uneven sensitivities of the QPD quadrants, and devised a new calibration and diagonalization procedure which allowed for it. * Enzo and Terrence did a health check of the SRM and it was fine except for some apparent electrical noise around 240 Hz. They put in a notch filter to get rid of it for now. |
==== Report for the week of 4/15 ==== * Fabian debugged a dead stepper on SR3 (broken driver channel) and a dead geophone on SRM (unplugged cable). He made another update to the table used by the Python script for the stepper driver MEDM screen to reflect the remapping. * Fabian monitored the pumpdown and he and Hirata-san did post-pumpdown checks of all the suspensions. They were all healthy. * Hirata-san participated in the health checks and learned about the computer systems. * Mark did major revisions to the MEDM menu on the OVERVIEW screen with scripts for making `safe.snap` files and snapshots of setpoints. He made the `safe.snap` files in directories under `/opt/rtcds/kamioka/k1/target` into symbolic links to files in `/opt/rtcds/userapps/release/vis/k1/burtfiles`, alongside the setpoint files, so that they could be backed up to the SVN (as at LIGO). * Mark did major updates to his `vistools.py` Python package for manipulating the suspensions (based on his LIGO `sustools.py`). He added entries for Type B suspensions to the internal database and added commands for manipulating SET and DCCTRL blocks. He also added commands to simulate pressing on the `CLEAR HISTORY` and `LOAD COEFFICIENTS` buttons. * Mark started updating the `TYPEB.py` Guardian definition file to use `vistools.py`. So far only the `ENGAGING_IP_CONTROL` state has been fully rewritten and tested, but other states will be simple adaptations. It should be easy to do Type A and other Guardians in the same style. |
| Line 23: | Line 13: |
| ==== Plan for week of 3/18 ==== * Mark, Fabian, Terrence at Kamioka full week. * More OL stuff. * SRM controls and Guardian. * Reset SRM yaw and continue monitoring yaw drift. * Assist with pump down. * Work on Type B paper. * Tidy up. * Throw out foil and other garbage. * Gather and clean tools. * Gather and organize unused parts and fasteners. * Pull out unused cables in cable tray near SR3. * Return all unused cables to tent. * Tidy up cables in rear of SR2 rack. * |
==== Report for the week of 4/22 ==== * Fabian debugged an anomaly he found in the SRM (klog 8672) and found it was OK if the BF was high enough (klog 8733). * Mark and Terrence worked on improvements to the Guardian and the vistools.py package. * We debugged a problem where vistools.py triggered channel access timeouts when a Vis object was created in a live Guardian (it worked fine in a plain iPython session or the interactive mode of the Guardian). It turns out that in the live mode of the Guardian, the Ezca instance used for accessing the suspension is not a global variable (contrary to the documentation) but is associated with the GuardState class. Worse, creating a second Ezca instance upsets things. We came up with a workaround to grab and reuse the Ezca object already defined. * Mark added many new functions to vistools.py, including one to read the BIO WD status, and a group to manipulate the OLSET and OLDCCTRL filter blocks. * Terrence rewrote the TYPEB.py Guardian from scratch in terms of functions from vistools.py. * We removed the RESET state. * We renamed the DCCTRL_ONLY state to FLOAT and reimplemented it to switch the DCCTRL INPUT switches off (instead of the HOLD switches on). * We revised the OL_DAMPING_ENGAGED and neighboring states (ALIGN, MISALIGN etc). * We changed the RTMs to remove the TM_OSEM_DAMP blocks and to have the SET blocks connect only to the DCCTRL blocks (not the DAMP blocks). * We updated the PAYLOAD and OVERVIEW screens to reflect the RTM changes. We also fixed some broken links and fields related to the BIO functions. ==== Plan for week of 4/29 ==== * Glorious 10-day break! |
| Line 42: | Line 28: |
| * Enzo at JPS then final vacation, 3/13-29 * Terrence's last day, 3/29. |
* Terrence's next visit from 4/18 to 7/27. * Fabian's vacation 4/25 to 5/10. * Mark's US vacation from 5/22 to 6/11. * Mark's final day 6/28. |
Scraps of text for incorporation in the VIS Meeting Minutes
KAGRA/Subgroups/VIS/MeetingNAOJ
Type B (Mark)
Report for the week of 4/15
- Fabian debugged a dead stepper on SR3 (broken driver channel) and a dead geophone on SRM (unplugged cable). He made another update to the table used by the Python script for the stepper driver MEDM screen to reflect the remapping.
- Fabian monitored the pumpdown and he and Hirata-san did post-pumpdown checks of all the suspensions. They were all healthy.
- Hirata-san participated in the health checks and learned about the computer systems.
Mark did major revisions to the MEDM menu on the OVERVIEW screen with scripts for making safe.snap files and snapshots of setpoints. He made the safe.snap files in directories under /opt/rtcds/kamioka/k1/target into symbolic links to files in /opt/rtcds/userapps/release/vis/k1/burtfiles, alongside the setpoint files, so that they could be backed up to the SVN (as at LIGO).
Mark did major updates to his vistools.py Python package for manipulating the suspensions (based on his LIGO sustools.py). He added entries for Type B suspensions to the internal database and added commands for manipulating SET and DCCTRL blocks. He also added commands to simulate pressing on the CLEAR HISTORY and LOAD COEFFICIENTS buttons.
Mark started updating the TYPEB.py Guardian definition file to use vistools.py. So far only the ENGAGING_IP_CONTROL state has been fully rewritten and tested, but other states will be simple adaptations. It should be easy to do Type A and other Guardians in the same style.
Report for the week of 4/22
- Fabian debugged an anomaly he found in the SRM (klog 8672) and found it was OK if the BF was high enough (klog 8733).
- Mark and Terrence worked on improvements to the Guardian and the vistools.py package.
We debugged a problem where vistools.py triggered channel access timeouts when a Vis object was created in a live Guardian (it worked fine in a plain iPython session or the interactive mode of the Guardian). It turns out that in the live mode of the Guardian, the Ezca instance used for accessing the suspension is not a global variable (contrary to the documentation) but is associated with the GuardState class. Worse, creating a second Ezca instance upsets things. We came up with a workaround to grab and reuse the Ezca object already defined.
- Mark added many new functions to vistools.py, including one to read the BIO WD status, and a group to manipulate the OLSET and OLDCCTRL filter blocks.
- Terrence rewrote the TYPEB.py Guardian from scratch in terms of functions from vistools.py.
- We removed the RESET state.
- We renamed the DCCTRL_ONLY state to FLOAT and reimplemented it to switch the DCCTRL INPUT switches off (instead of the HOLD switches on).
- We revised the OL_DAMPING_ENGAGED and neighboring states (ALIGN, MISALIGN etc).
- We changed the RTMs to remove the TM_OSEM_DAMP blocks and to have the SET blocks connect only to the DCCTRL blocks (not the DAMP blocks).
- We updated the PAYLOAD and OVERVIEW screens to reflect the RTM changes. We also fixed some broken links and fields related to the BIO functions.
Plan for week of 4/29
- Glorious 10-day break!
Vacation/Travel Info
- Terrence's next visit from 4/18 to 7/27.
- Fabian's vacation 4/25 to 5/10.
- Mark's US vacation from 5/22 to 6/11.
- Mark's final day 6/28.