Cable pulling from H1 Electronics room to ISC-R1
Written by Giles:
The attached document summarises the modal frequencies taken on the ETMy at LHO. The first tab provides a summary of the modes and measurement line width. The subsequent tabs in the spreadsheet provide the data and also plots of the data.
Giles, Betsy, Travis
Thanks all for posting this data. It should help the modeling a great deal.
Fibers stretched approximately 5.7 mm on both sides during the hanging process.
Due to the higher than expected cavity pole we are using the first boost stage (pole at 100 Hz and zero at 1 kHz) as part of the compensation. The second boost stage is too high in frequency. This morning we modified the second boost stage to be identical to the first one with a pole at 100 Hz and a zero at 1 kHz. This is serial number S1102630. In detail:
The removed parts are in a bag taped to the inside of the chassis.
Stefan, Yuta, Kiwamu
We are currently in a low power state for performing the PRM parking beam search.
Power decreased:
We decreased the IMC incident to approximately 100 mW by rotating the motorized half wave plate. We unlocked the interlock to allow us to rotate it while blocking the main laser pipe just in case. Currently the incident power is at 116 mW at the monitor channel, H1:PSL-PERISCOPE_A_DC_POWERMON. Then we locked the interlock back so that nobody can change the power.
Autolocker modified:
We then changed the ISC_IMC guardian script to adapt it to this lower power configuration. The threshold and in1 gain are increased by a factor of 10. Also all the LSC and ASC triggers are lowered by a factor of 10. We found that increasing the ASC common gain by a factor of 10 made the loops unstable. Therefore we set the common gain to be back at 1 as usual. So far it seems locking fine and stable enough to proceed with the parking beam mission.
Now we are back to 1W since we are done with the PRM parking search. We restored all the settings that we changed.
Commencing at 9:30AM I removed the unused 205 and 206 VLANs and associated interfaces from the core switch. In addition, I removed ip directed-broadcast from VLAN interfaces 2 (10.1.0.1) and 201 (10.201.0.1) as this feature is not desired/required for these subnets.
Surveying viewports.
Work finished at 9:45 AM PT.
Turning down the output power to 100 mW.
Work finished at 9:56 AM PT.
Install control converter for Lighthouse dust monitors.
Work finished at 9:30 AM PT.
Hugh, Hugo,
Input filters and matrices were implemented on HEPI-ITMX yesterday.
Today we used the TF measured on 01/02/14 to design its symmetrization filters, which we installed and turned on. We also installed the position blends (IPS=0, L4C=1)
We then deisgned most of the position loops for this unit, until we got stuck on a scripting issue that took most of our evening: we cannot plot our position loops for the following DOF: Y,RZ,HP.
Since the scripts we use are generic, we suppose that they are fine. We suspect that the data we use may be corrupted. Indeed, the TF data had to be re-processed after collection, to account for in-field mis-wiring, and it is possible that the re-processing was enough for plotting in the local basis, but caused further problems down the line whith the position loops design scripts.
Loops are designed for every degree of freedom, but could only be looked at along: X,RX,RY,Z,VP
Next step is either to try to arranging the data, or take a new set of TF.
Meanwhile, we reverted all the scripts that we modified locally in our atempt to debug.
I am leaving a transfer function measurement running on BS and PRM overnight. The PRM measurement needs the IMC to stay locked. Please try not to unlock the IMC.
Finished. IMC can be unlocked.
Hugo, Sheila
Having heard that at Livinstgon ISC is seeing about one fringe per second (need to confirm this is for green, not infared) (also see LLO alog 10404), while here we see more like 10 fringes per second with the green, we wanted to investigate the motion of the optics. Laser frequency noise is something that should also be ruled out, although it seems unlikely since the frequency of the motion we see is around (roughly) 0.1Hz. We wanted to compare the SEI/ SUS platform motion between the two sites. To do so, we compared thefollowing:
... Along X (Longitudial) and RY (pitch).
The SUS and SEI plots (which were not calibrated) would suggest that the magnitude of the motion recorded on both SUS and ISI is 10x bigger at LHO along X (longitudinal) while it was recorded within the same range along RY (Pitch).
We thought that differences in ground motion could explain this observation. Ground motion (LLO,LHO) was however comaprable between sites, at the time of the measurement.
Another explanation could have been a difference in matrices/calibrations between sites, but, as far as the ISI are concerned, they seem to match.
We see that we loose lock often because we use up the VCO range. The VCO full range is 2 MHz (awiki) which is equivalent to 14um peak to peak change in the arm length.
Another usefull comparison to make would be to look at the data from the ISIs durring HIFO Y to see if the motion that we had then was comparable to what we have now.
The X arm is left overnight trying to lock, with automation. The channel H1:ALS-X_REFL_LOCK_STATUS will try to report when the cavity is locked, although it should be double checked with either ALS-C_COMM_LF or ALS-C-TRX_A_DC because the threshold probably needs tuning.
We were able to lock the X arm stably (for several minutes with no offloading) the first time we "locked" it. At that time our alingment was far off, and we weren't really locking a real cavity mode (alog 9127). Right now we cannot lock for more than about 30 seconds, more often the locks last only a second or 2. While it seems most likely the lock was stable then because of our terrible alingment, I just wanted to check that we haven't had a change in the motion of the optics since that time. Attached are three screenshots from Jan 7th (the night after that lock, for which data is available) and last night. As expected, the motion is either the same or reduced since then by the work that SEI has been doing.
Livingston people report that they are seeing about 1 fringe per second with the green, that is much less motion than we have (here it is about 10 fringes per second). But the apparent stability from a week ago was not due to less motion, it must be some optical difference due to the bad alingment we had then.
I updated the h1conlog channel list to use the names in /ligo/lho/data/conlog/h1/pvlist_1389919431. 19742 channels were added. 11118 channels were removed. The total number of channels is now 106321.
There are now approximately 59,019,605 rows in the database.
