LBBG Commissioning with AGW1: 02 November 2011 UT
Observer: JHill, JHeidt, RPogge (LBTO)
Telescope Operator: DHuerta (LBTO)
Telescope Support: JUrban (LBTO)
Instrument Support: JMorris (LBTO)
We lost about half the night to high winds, so progress on-sky was slow.
We got active optics working OK with IDL on-axis. Unfortunately, GCS diverges with apparently the same parameters even after I corrected for using the RFBG rotator position.
We offset the M3 selector to match the geometry Andrew had previously measured for the bearings in the gallery. Unfortunately, that didn't make any particular improvement in the alignment of the pupil on the lenslet array.
After a long battle, we finally found the right place to put the X-flip to make all AGWs match, That is to have pcsmx=-1.
The last change was made off sky, but I'm sure we can locate off-axis stars now.
Network and DNS seem to be back to normal.
Mike and Torsten worked on the AGW1 WFS camera, but without resolution. It continues to fail on about 30% of the images.
The MGIO guys have exposed the broken water pipe below 4 levels of other pipes. The pipe was repressurized at 23:36 UT.
SX M3 Alignment
Andrew has improved values for the position of the SX M3 selector based on the measured angles of the bearings in the gallery and the known angle of the M3 selector for LUCI1. We should be 51.8 degrees from LUCI1. At the end of last night's empircal adjustment we were 26.0123+30.6877-4.25 = 52.45 degrees from LUCI1
Andrew also has an idea for why the pointing doesn't match the other Gregorian focal stations:
As per our recent conversation, currently the SX telescope "points up" with respect to the telescope coordinate system by -60" RX (secondary toward the roof at horizon pointing). This is probably why
you had this big pointing change on LBBG.
I suggest applying +50" or +60" of RX using "split mode" Tip in the SX PSF arbitrator. This should not cause too much trouble for the
collimation models with respect to hexapod range limits.
+60 arcsec split Tip gives M2 dY=-3.102 mm (M2 Total Y goes to around -5 mm) RX=+60 arcsec and M1 RX=+60 arcsec.
Or if we turn on Range Balancing +60 arcsec Split Tip gives M2 dX=-0.036 dY=-1.596 RX=60 and M1 dX=-0.036 dY=+1.506 RX=60
For the selector angle, the laser tracker measurements are in good agreement with design, and they give an M3 selector angle change of 51.8 degrees from LFBG (Lucifer) to LRBG. I'd go with that as a starting point, then, after applying the +RX to the whole telescope (M1, M2) described above, try mopping up residuals with a combination of mount pointing and M3 tip/tilt.
Active Optics Settings
Doug's IDL settings from agw_set_values.pro
agw[lbbg].lenslet_edge[0,0] = 355.0
agw[lbbg].lenslet_edge[1,0] = 201.0
agw[lbbg].rot_offset = 114.8
agw[lbbg].pupil_x_on_axis = 378.
agw[lbbg].pupil_y_on_axis = 195.5
agw[lbbg].probe_x_on_axis = 0.0
agw[lbbg].probe_y_on_axis = 612.5
agw[lbbg].r_pupil_shift = 0.19
agw[lbbg].pupil_direction = 1.0
Things John updated in AGW1_IRTC.cfg this afternoon (GCSL was restarted during checkout):
WFScam_hotspot_x 378.0 double # ccd x coordinate of pixel that represents leftmost x-center of WFS spots
WFScam_hotspot_y 195.5 double # ccd y coordinate of pixel that represents leftmost y-center of WFS spots
WFS_lenslet_edge_x 355.0 double # defines the edge of the virtual lenslet grid in x
WFS_lenslet_edge_y 201.0 double # defines the edge of the virtual lenslet grid in y
It would be nice if we could make a common set of names.
For IDL running LBBG use: agw_set_station, IRTC
Plan for Tonight
We'll test guider and wfs images during twilight.
We'll start the night getting active optics running on-axis for both IDL and GCS including a precise measurement for Doug's Angle.
Then we'll reposition M3 and evaluate the on-axis pupil positions and the pointing.
Then we'll sort out the mysterious X-flip, and get some off-axis pupil positions for Doug to measure.
Then we'll evaluate off-axis active optics.
Next we'll set the focus relative to IRTC.
00:40 David fighting with GCSL and AzCam trying to get a second test preset to complete.
00:58 warm boot azcamserver
Called, but didn't reach Torsten. Called Mike Lesser who was fortunately still in his office.
The critical error message seems to be: "could not allocate Image Buffer memory".
01:40 Mike Lesser has solved our AzCam problems by deeply resetting things and reallocating memory on the azcamg machine (a.k.a. lbtguidercam3).
Sky is mostly clear with a few scattered cirrus clouds.
SX M3 starting where we were last night: Global Offsets:Tip = -300 Tilt = -920 RZ=4.25
Active Optics on-axis with IDL
02:05 Jochen running IDL collimate_once
Images were already 1.1 arcsec before Active Optics, and things quickly converge to 0.9 arcsec in variable seeing.
02:12 by wfsc 12, the astigmatism has started to diverge...........while coma was fine
02:15 Clear C00
02:16 wfsc 13 (baseline) without sending
+3000 Z6 on M1
02:17 wfsc 14
+3000 Z5 on M1
02:19 wfsc 15
02:26 wfsc 17 (baseline)
+1500 Z7 on M1
02:28 wfsc 18
rot_offset from 114.8 to 69.8
02:34 wfsc 20 (baseline)
+1500 Z7 on M1
02:35 wfsc 21
rot_offset from 69.8 to 60.0
02:39 wfsc 23 (baseline)
+1500 Z7 on M1
02:40 wfsc 24
+3000 Z6 on M1
02:42 wfsc 25
+3000 Z5 on M1
02:45 wfsc 26 - confused
Clear Active Optics
02:47 wfsc 28 (baseline) far from focus
02:50 wfsc 29 (baseline)
+3000 Z5 on M1
02:52 wfsc 30
Wispy clouds, wind, poor seeing, goofy readouts
02:53 collimate_multiple (now sending)
wfsc 42 (baseline)
+500 Z9 on M1
03:06 wfsc 43
03:09 collimate_multiple (now sending)
Active Optics on-axis with GCS
Apply the same 60.0 deg angle to GCS, stop/start
03:11 ACTIVE preset on axis
03:18 this diverges in astigmatism, stopWFSing
We found out why while closed for wind.......GCS was using the RFBG rotator angle for LBBG Zernikes.
Updated these parameters in AGW1_IRTC.cfg to fix the rotator, and to use parameters consistent with LUCI1.
guidecam_lowCounts 50 int # low water mark for counts above background that cause adjustment of exposure time after full CCD readout
WFScam_lowCounts 50 int # low water mark for counts above background that will cause auto exposure time adjustment
WFS_rotatorAngle mcs.rotatorSide.rotators.actualPositionASec string # name of DD variable containing the rotator position for this AGw
countThreshold 50 int # number of counts a star has to be above background level to be recognized as a valid guide star
Back to IDL
03:20 start WFSing with IDL again, but note that M2 is not updating its position.
03:24 kill and restart OSS - this gets hexapod working, resume sending guide corrections.
03:27 4 arcsec jump in guider image
03:32 send a perturbance of +200 Z9 to M1 (had a little crosstalk into Z10)
Global Offset for RZ from 4.25 to 4.90, but M3 didn't want to move. This eventually gets the Tertiary to an absolute angle of -25.788
03:43 De-init, then Init M3 UMAC from OSS
SX M1 Panics (why?)
03:58 Removing Tilt from M3 while repointing telescope to follow star. Image the WFS to follow the pupil (w/o star).
readWFSCam left -e 20000 - wfsc 63 (sky) (Tip -300 Tilt -920 RZ 4.90) IE=40 CA=35
readWFSCam left -e 20000 - wfsc 65 (sky) (Tip -300 Tilt -920 RZ 4.50)
readWFSCam left -e 20000 - wfsc 68 (sky) (Tip -300 Tilt -620 RZ 4.90) IE=22 CA=28
readWFSCam left -e 20000 - wfsc 69 (sky) (Tip -300 Tilt -320 RZ 4.90) IE=8 CA=19
readWFSCam left -e 20000 - wfsc 70 (sky) (Tip -300 Tilt +300 RZ 4.90) IE=-26 CA=3
readWFSCam left -e 20000 - wfsc 71 (sky) (Tip -300 Tilt +900 RZ 4.90) IE=-58 CA=-13
More positive Tilt is moving the pupil up on the WFS
04:52 GUIDE preset on-axis PA=0
readWFSCam left -e 20000 - wfsc 73 (star PA=0)
04:56 offset 0 0 90
readWFSCam left -e 20000 - wfsc 76 (star PA=90)
04:59 offset 0 0 90
readWFSCam left -e 20000 - wfsc 77 (star PA=180)
05:01 offset 0 0 90
readWFSCam left -e 20000 - wfsc 80 (star PA=270)
This set has about 1.5 lenslets of wobble.
05:12 started moving Tip, but had to close for wind of increasing gusts. T=2.9 degC D=-9.3 degC
Stopped at IE=-58 CA=15 Tip=-100 Tilt=+900 RZ=4.9
Handing the dark dome over to Rick for MODS measurements.
07:18 Took a couple dome flats to verify the AGW1 guider camera saturation level. It is 65K.
07:24 Wind has dropped, so we open again. T=1.8 degC. Stop and restart GCSL to pick up new parameters.
Active Optics with GCS on-axis
07:30 ACQUIRE preset to BS9107 IE=-65 CA=8
07:36 ACTIVE preset to BS9107 on-axis - It quickly diverges in coma, so the rotator angle wasn't the only problem.
07:42 Run IDL to collimate while GCS guides.
More M3 adjustment
07:52 ACQUIRE preset
readWFSCam left -e 20000 - wfsc 149 (Tip +300 Tilt +900 RZ 4.90) IE=-75 CA=33
readWFSCam left -e 20000 - wfsc 151 (Tip +850 Tilt +900 RZ 4.90) IE=-97 CA=80 (Absolute Tip +235, Tilt +880 RZ-25.789)
08:08 GUIDE on-axis PA=0
readWFSCam left -e 20000 - wfsc 152 (star PA=0)
08:12 offset 0 0 90
readWFSCam left -e 20000 - wfsc 154 (star PA=90)
08:15 offset 0 0 90
readWFSCam left -e 20000 - wfsc 155 (star PA=180)
08:17 offset 0 0 90
readWFSCam left -e 20000 - wfsc 156 (star PA=270)
Pupil wobble is worse here. 2 full lenslets in radius.
08:23 Close for wind again
Adopt this M3 position offset pending the telescope RX motion: Tip -300 Tilt +900 RZ 4.90)
(relative to the absolute position reported in last night's log)
Applied the +60 split Tip Mode 1 pointing to M1 and M2.
09:45 Open again T=1.4 degC
Searching for the X-flip
09:46 ACQUIRE preset to pointing star, subtract 60 from IE.
09:50 ACQUIRE preset to Persson star gs=0
09:54 ACQUIRE BS9107 gs=1 PA=INDEF=261 x=0 y=589 (possible bad coords from binary)
09:56 ACQUIRE BS9107 gs=2 PA=INDEF=357.8 x=0 y=587
09:56 ACQUIRE BS9107 gs=2 PA=347 x=0 y=587 x=-4.8 Y=587 (reversing X finds the star)
10:03 ACQUIRE BS9107 gs=2 PA=330 x=0 y=587 x=-11.88 Y=590.163 (reversing X finds the star)
10:05 ACQUIRE preset to Persson star gs=0 IE=-121 CA=-5
10:08 ACQUIRE BS9107 gs=3 PA=INDEF=14.1 x=-0.06 y=585.68
10:08 ACQUIRE BS9107 gs=3 PA=30 x=-4.83 y=586.11 (reversing X finds the star)
LEFTXYFLIP changed from false false to true false
, stop and restart PCS
10:18 ACQUIRE BS9107 gs=3 PA=30 x=4.83 y=586.11 (we see the star without reversal)
10:22 ACQUIRE BS9107 gs=28 PA=INDEF=316.7 x=0.22 y=466.3
10:24 ACQUIRE BS9107 gs=28 PA=340 (we see the star without reversal)
10:26 GUIDE BS9107 gs=28 PA=340 - acquisition offset failed - went the wrong way in x
LEFTXYFLIP back to false false
Changed M_XGP from +1 to -1
in oacontrol.conf , stop GCS/oacserver/AGW and restart.
10:26 GUIDE BS9107 gs=28 PA=340 - succeeded (but with a very unappealling symmetry between AGW3 and AGW1.)
10:35 Close for wind again.
Really figuring out the X-flip
startAGW -u 2
setxy -u 1 -x 50 -y 580
setxy -u 2 -x 50 -y 580
zenith for a visual inspection by Rick and John. The probe on AGW1 moved to the right as seen from the mirror side with the probe base down. The probe on AGW2 moved to the left. This is what would be expected if both AGWs
were the same, and AGW1 had the x-sign flipped.
The proper thing to do when you add an offset to the homeoffset is then to
subtract the same offset from both of the limits lim_f and lim_n.
I've now fixed the limits in oacontrol to correct what I did last night. Now I have subtracted the 72 from the limits.
OK, I think I found it. I had changed X_MGP, but I hadn't noticed that pcsmx=+1 for AGW1 and -1 for all the others.
11:43 Changed M_XGP back to +1
in oacontrol.conf, and changed pcsmx to -1
(now all 4 AGWs match).
12:30 I guess the fact that the telescope operator has gone to bed means that the night is over. That, and the wind gusting above 20 m/sec.
Things to eventually archive
collimation and pointing models
- 01 Nov 2011