From gonzalez@lsu.edu Thu Jun 19 11:28:10 2003 Date: Mon, 19 May 2003 19:49:49 -0500 From: Gabriela Gonzalez To: Keith Riles Cc: Ed Daw , Peter Fritschel , Brian O'Reilly , Rauha Rahkola , Robert Schofield , Peter Shawhan , Daniel Sigg , Stan Whitcomb , John Zweizig , Rana Adhikari Subject: S2L1 controls For Keith's web page on S2 data quality: After looking at the control changes registered by conlog and collected by Peter Shawhan in the file "S2L1controls.txt" I produced two new files (attached): * S2L1controls_annotated.txt Peter's list, with a reference in each noted change to the notes in the file below. * S2L1controls_notes.txt This file has notes with the different classes of control changes found, and levels of "significance". These levels affect the lieklyhood of either increased noise in the ASQ spectrum, and maybe glitchiness. Rana added notes to some categories, clarifying some points. I added references to elogs when I found them. From looking at the possible relevance of these changes, we (me, in consultation with Rana) suggest adding "red flags" to only a few of the L1 science segments. A "red flag" means the segment should NOT be used in data analysis. There are several other segments that merit some investigation, but have no obvious problem. We should follow up with more investigations (ASQ spectra, glitch rates, etc) because these changes suggest that a red flag is a possible (but not likely) consequence of the control changes. Even without further investigations, we feel these segments should be used for data analysis. And finally, we make a list of "green flags" to some segments with non-trivial control changes that we think have no consequence for the data quality, and have no chance of meriting a red flag. The following segments have "red flags". They can probably be used if we are desperate (GRB at this time, an event in LHO that needs confirmation, etc), but is better to put them aside otherwise. * Segment 1, 729304210- 729305791, 1581 sec long: the loop gain settings were different than in all other segments and affect calibration. * Segment 112, 730195425 - 730197778, 2353 sec: this segment had a different gain in the AS_I servo than all the others, probably increasing the low frequency(45-60 Hz) ASQ noise and probable upconversions. * Segment 133, 730385434- 730385520, 86 sec long: the up script was not finished before entering science mode, and elog entries suggest there were problems. The following segments should be investigated further ("grey flag"?) * Segs 214-216, when many settings were changed, after the big power outage. * MICH filters -- FM3 off: segs 6, 18, 22, 30, 34, 35. -- FM5 on: 145-147, 152, 155-160, 162-170, 176, 179-180, 192-193, 196, 304, 308, 330-331, 413, 444 * PRC FM6 filter off: 4, 6, 8, 13, 18, and 30-36. * ITMs Dewhitening filters: segs 214-216 for mirror switch, segs 272-288 for coil filters. * MICH gain different than -0.08: 99, 101-103, 117-131, 134-216, 319, 322-325, 330-331, 350-351, 411, 595. The following setting changes define "epochs" rather than individual segments to be flagged, so even if they affect the noise spectrum, we'll have to live with it. So, these segments all have "green flags" (from the controls point of view). However, we should keep this list in hand if it helps to explain features of these epochs. No further investigation required. * PRC gain: three epochs for each gain value; -0.100237 (segs 2-216), -0.125297 (segs 217-289, 291-294), -0.088703 (290, 295-606). * EOHIHV/AS1_Q gains: also probably define epochs rather than segments to throw away. Rana says that changes of 0.05 in EO HIHV are not significant, and the epochs are defined by EO shutter level close to 1.1 or 1.45. * ITMs oplev stopband filter: segs 390-606 have it on, and are supposedly better for it. * ITMY oplev gains : segs 1-213, 289-606 with one value, and segs 214-288 with another. * ITMX oplev gain : segs 1-394 with one value, segs 395-606 with another. * RM oplev gain: segs 1-467,471, 473-474, with one value; segs 468-470, 472, 475-606 with another. * Tidal feedback gains: segs 1-106, 213-357, 573-606 with one value; segs 107-212, 358-573 with another. The following settings, changed in only a few segments (so they don't qualify for "epochs"), again may affect the noise (but we think it is unlikely). Again, they have "gren flags" and no further investigation is required. * LVEA PEPI enabled: segs 89-92, 133, 139, 141. * WFS1 integrator on: seg 19 -- ################################################### Gabriela Gonzalez Assistant Professor Department of Physics and Astronomy Louisiana State University 202 Nicholson Hall Tower Drive Baton Rouge,LA 70803 Phone: (225) 578 0468 Email: gonzalez@lsu.edu http://www.phys.lsu.edu/faculty/gonzalez/ ##################################################