LN2 levels for CP3 and CP8 are not under PID control.
Patrick, Richard, Carlos, Jim, Dave
The new Beckhoff vacuum controls system uses a new channel naming convention. This required several software systems to be reconfigured.
DAQ
When going from the old system to the new system there are three catagories of channel being acquired by the DAQ:
1. old channels which have no equivalent in the new system (qty 97)
2. old channels which have new names in the new system (qty 90)
3. channels which are added by the new system (qty 639)
For catagory 1 I did not do anything, but I most probably should add them back into the DAQ configuration so old data could be accessed.
For catagory 3 other than putting them into the new H0EDCU_VE.ini file there was nothing to be done. They popped into existence yesterday at 6pm.
For catagory 2 some work is needed. To preserve the minute trend lookback back when they had the old name, it was necessary to move their raw minute trend files from the old name to the new name. Patrick provided a file which maps oldname->newname. The process is slightly complicated by the fact that raw minute files reside under 256 directories, named by the CRC8 checksum of the channel's name string (to prevent one monolithic directory). I have included the mapping file and the move script which was ran on h1tw1. I actually did a copy rather than move, so this could be done while the DAQ was still running.
The DAQ change sequence was: Copy minute trend files from oldname to newname, generate the new H0EDCU_VE.ini and H0EDCU_FMCS.ini (Beckhoff reads out the water pressure) then restart DAQ
EPICS GATEWAY
We discovered that there is a channel filter on the epics gateways between the h0ve network and the DAQ/CDS/WS networks. This filtered to the old HVE- naming, we added an additional H0:VAC- rule.
AUTOBURT
Patrick created a new h0veex/autoBurt.req file, this was installed.
MEDM
The VE site overview was changed to match the new EX channel names.
ALARMS
Both the control ALH alarm handler and the cdslogin cell phone texter alarm handers were modified to use the new channel names.
I made a first attempt to investigate the probability of a beam tube tap producing a DARM glitch as a function of tap amplitude. I mounted an accelerometer on the beam tube just inside the single door at Y1-4 +1Y and tapped 2 stiffening rings away. The only DARM glitch produced was for a tap that saturated the accelerometer (> 1.09 m/s^2). There were a total of ten taps that exceeded this threshold. No glitches were seen for the 70 taps that were under this threshold. This data supports the contention that the glitch probability is a strong function of tap amplitude. I would like to use a shaker to inject at lower amplitudes for day-scale periods in order to better understand the probability function, but this experiment hasn’t been cleared yet because of fatigue concerns.
One could imagine sudden accelerations of the beam tube, such as thermal expansion stick-slip events, that might be high enough in amplitude and frequency to produce particulate glitches. I investigated the possibility of using microphones to detect such events. I set up a temporary microphone in the beam tube enclosure, running into the CS DAQ. I found, that for the amplitude of taps that produced glitches, I could detect the sound from 1km away (Figure 1). Thus we could monitor the entire beam tube with 8 microphones, based in the stations, but located in the beam tube enclosure just outside the stations.
|
amplitude m/s^2 |
injections |
glithes |
|
>1.09 |
10 |
1 |
|
1.07-1.09 |
20 |
0 |
|
less than 1.07, greater than 0.05 |
50 |
0 |
To install this at LLO would require procurement of front-end computers/IO chassis for mid-stations. We noted this equipment was purchased by aLIGO for LHO, but not for LLO.
Late entry for yesterday's maintenance.
cdsfs0 upgrade to new hardware WP5757
Richard, Carlos, Ryan, Jim, Dave:
The old cdsfs0 /ligo NFS file server was decommissioned and replaced with a newer model. The storage capacity was increased from 4TB to 40TB for the /ligo file system. This was an early maintenance task, starting around 6am PST. It required the reboot of many servers and workstations, but did not require a restart of the front ends or DAQ systems. The only problem reported from users was files with controls group ownership could not be modified by users who belong to this group. Ryan resolved the problem, we put in a temporary fix and need to reboot cdsfs0 next week to implement a permanant solution.
Richard and Carlos put H1 into a safe state before the computers were stopped.
h1asc model change WP5766
Kiwamu, Jennie, Sheila, Dave:
A new h1asc model was installed. This required a DAQ restart.
H1 BSC ISI model changes WP5765
Hugh, Jim W, Dave
the ISI BSC models (five in total) were restarted with new code. This required a DAQ restart
EX Vacuum Controls Beckhoff Install WP5758
Richard, Filiberto, Patrick, Carlos, Dave:
The original VME vacuum controls system h0veex was replaced with the new Beckhoff industrial control system. The hardware was swapped out quickly in the morning, the new unit has the same name and IP address as the VME system it replaces.
Various software components were changed, details in seperate alog.
Due to time constraints, two tasks were postponed: WP5757 install first verson of h1susitmpi, WP5764 install of first version of h1ngn.
Activity Log: All Times in UTC (PT) 16:00 (08:00) Ellie – Taking SRC interface measurements 16:58 (08:58) Ellie – Going into Optics Lab 17:10 (09:10) Ed – Add water to PSL Diode chiller 17:25 (09:25) Hugh – Going into the CER 17:32 (09:32) Hugh – Out of CER 17:36 (09:36) Ellie – Out of Optics Lab 20:01 (12:01) Kyle – Going into LVEA to turn off vacuum pump 20:05 (12:05) Dick – Going into the LVEA checking for noise at ICS rack 20:40 (12:40) Kyle – Out of the LVEA 20:48 (12:48) Ellie – Transition LVEA to Laser Hazard 20:50 (12:50) Ellie & Cao – Going into the LVEA to work on ISCT6 table 21:59 (13:59) Ellie & Cao – Out of the LVEA 22:13 (14:13) Gerardo – Going to X-Beam manifold to make some measurements 22:40 (14:40) Gerardo – Out of the LVEA 23:00 (15:00) Gerardo – Going to Mid-Y for CP-3 fill 23:30 (15:30) Gerardo – Back from Mid-Y 23:50 (15:50) Ellie – Going into the LVEA tow work on ISCT6 End of Shift Summary: Day Shift 16:00 – 00:00 (08:00 – 16:00) All times in UTC (PT) Support: Incoming Operator: Cheryl Shift Detail Summary: Commissioning work all day.
1/2 open LLCV bypass valve, and the exhaust bypass valve fully open.
Flow was noted after 2 minutes 20 seconds, closed LLCV valve, and 3 minutes later the exhaust bypass valve was closed.
Next over-fill on Friday, March 11th before 23:59 utc.
Attached are 7 day pitch, yaw, and sum trends for all active H1 optical levers.
Almost everything looks good, with no changes from last week. The only concerning thing that we need to keep an eye on is the BS SUM; it has dropped from >20k counts to <10k counts over the last week. As soon as I get back from Germany I can start looking into it.
Pirani gauge on BSC4 was temporarily in the wrong orientation during maintenance yesterday and thus read erroneous pressures for ~3 hrs until it was reinstalled (with valve) by Gerardo. Attached is pressure plot.
I've updated the TCS SIM section of the front end. The changes are summarized here:
This morning I noticed the Diode Chiller in alarm. I added 150ml to rest the float switch and the another 250ml as per Jeff Bartlett's advice. Because the Famis task (#4140) of checking the chillers and adding necessary water was completed yesterday (without being able to accurately diagnose the water level in the diode chiller)and the alarm went red today, there is a plan to remedy this issue being set into motion.
Transition Summary:
Title: Day Shift 16:00 – 00:00 (08:00 – 16:00) All times in UTC (PT)
State of H1: IFO unlocked.
Commissioning:
Ellie working on SRC interface measurements.
Tonight we again weren't able to engage the soft loops without pulling the sideband powers down and dropping lock, so I spent a little time on ASC.
For one thing I looked at the phasing of AS36 A to see if we can reduce the SRM signal in the Q phase. The reasoning for this is that the SRC loops are sensitive to the soft arms through AS_C, and if the BS is also sensitive to SRM this could be one way that moving SOFT misalings the beamsplitter. However, it looks like we are already kind of using a phase that minimizes the SRM signal in Q, (-140), so I left this as is.
I also had a look at the best combination of TMS QPDs to use, since we retuned this on saturday I was curious to see if the combination had changed today with a different alignment. I checked only TMSX pit, but it had only changed by 1%, even with the soft loops off.
Lastly, I tried to increase the bandwidth of the CHARD loops, in part because we know we need to do this, and in part because CHARD is cross coupled with the yaw loops and there is a long slow osciallation in all 3 loops due to this. A screenshot is attached with settings that result in CHARD BWs similar to the DHARD ones, this initially does seem to reduce the long slow oscillation of the soft loops, but we still have the problem that the slow loops misalign the vertex. I've added these settings, which are engaged with a lower gain and not the boosts on, to the guardian, but they will probably add noise, so we can reduce their gains when we want to go to low noise.
Top pannel is IM2 OSEMS. Bottom pannel is select OSEMs from IM1, IM3, and IM4.
State of H1: locking and getting Engage ASC, Sheila working on the ASC
Site Activities:
LVEA and OSB optics lab access:
Currently:
I made a folder called beckhoff in /ligo/lho/h0/ve/medm and copied the new medm screens for the Beckhoff vacuum controls into it. I then changed the 'VE EX' link in the 'VE/FMC' drop down menu on the sitemap to point to /ligo/lho/h0/ve/medm/beckhoff/H0_VAC_EX_CUSTOM.adl. I committed the change to the sitemap into svn. /ligo/lho/h0/ve/medm/beckhoff/H0_VAC_MENU_CUSTOM.adl is a mini sitemap for just these screens. The screens that do not have CUSTOM in the name are generated from a script. The files necessary to generate these screens are in the slowcontrols svn repository in trunk/TwinCAT3/Vacuum/MEDM/LHO.
Kyle, Patrick The new Beckhoff control parameters still need "tuning". As of this writing, the control of CP8's level is via the Proportional term only. Initial attempts resulted in overfilling which then resulted in a closed LLCV which then resulted in a warm transfer line. Patrick and I opened the exhaust check valve bypass valve to relieve the back pressure and then cooled the transfer line via the manual LLCV bypass. With the transfer line cold and the pump level responding to changes in the LLCV, I manually filled the pump to 95% full, closed the exhaust check valve bypass valve am leaving the LLCV manually set at 50% open overnight which is its recent nominal value to maintain the pump level. Jim Batch set up and included two MEDM screens so the X-end status can be monitored from home. Patrick and Cheryl also are leaving an MEDM screen in the Control Room which can be monitored from home. 2135 hrs local -> I'm leaving now and will monitor CP8's level from home (every 2 hours)
I changed the end X channel names to match the new Beckhoff channel names in /opt/rtcds/userapps/release/cds/h1/alarmfiles/ve.alhConfig. There are no new Beckhoff channel names corresponding to HVE-EX:GV19_CCOFFALARM and HVE-EX:GV19_OVERPRESSURE so I commented these out of the file. I will have to see if there are other channels that can be alarmed on instead of these. I committed the changes to svn in cds_user_apps/trunk/cds/h1/alarmfiles/ve.alhConfig. I restarted the alarm handlers on alarm0 to use the updated file.
Kyle. Gerardo, Chandra Gerardo -> Installed 12" CFF at HAM10 D8 and gauge pair isolation valve on BSC4 Chandra -> R&R LVEA Turbo levitation batteries Kyle -> Drilled and tapped holes, mounted scroll pump local to Diagonal Turbo and upgraded gauge controller on Diagonal Turbo Kyle, Chandra -> Ran QDP80s, YBM, XBM and Diagoanl Turbos and vent/purge air supply (KOBELCO). The KOBELCO displayed a "HIGH LUBE OIL TEMP" Warning after having run for a while. Chandra noticed that the chilled water booster pump was off and that the output pressure once on was only 60psi (this may be the result of a reduced frequency running of the chilled water circulating pump - nominally this would be 90 psi - we need to investigate) The measured dewpoint after > 1 hour of run time was <-30C. Gerardo noticed that the Diagonal volume was still left connected to the purge air header from when it had last been purged, the isolation valve was then closed. NOTE: The XBM Turbo spun up to full RPM but tripped off during the braking phase. This Turbo has a history of high vibration - typically during the accelerating phase.
The Turbo pump controllers are being left energized overnight in the LVEA to allow the Turbo rotors to come to a full stop
No FRS required. Lower than expected pressure in output from booster pump in chilled water system is likely due to difference in building chilled water pressure (variable speed pumps).