Technical Observing for 4-5 September 2008 MST
Observers: D. Miller, A. Rakich and J. Hill in Tucson
plus J. Storm in Potsdam for the second half
Software: Torsten, Stephen and Paul here for first part of night.
Telescope Operator: A. Ceranski
Sorted out guide angle for PCS, guiding now working again.
Found issue with GCS WFS, needs to be resolved Friday. Confirmed WFS with IDL routines, off-axis guiding seems to be giving good results on-axis.
GUI needs urgent remedial work on Friday.
Now expect to be using all of Friday-Saturday night and we understand that the telescope will be available with Lucifer mounted but telescope and AGW fully operational for finalizing wake-up on Saturday-Sunday night as well.
Night time Activities
Times in UT below
01:55 Open dome to a beautiful clear night. 50% humidity.
Untangling the Rotator Angle and the Guide Transformation
02:15 Pointto M5_0593, found ~30 arcsec off center at -30,0 (already better than yesterday), centering on the IRTC makes IE=-85, CA=-60.
Offset 0,10 moves star image up on the IRTC GUI (correct since N is down on GUI). It is aligned with IRTC columns within the seeing jitter.
But, Offset 0,5 also moves the star image up on the AGW image (incorrect, since N is up on the AGw DS9 display). We verified that the
star image moves down on the IRTC fits image.
Conclude that: a) IRTC is mounted 180 deg upside down from AGw, and b) the LEFTIAA coefficient is correct for IRTC, but 180 off for AGw which is
the one that we need to be correct for guiding. This is certainly one of the reasons that
we were confused on Monday evening.
Update PCSInstrument.conf to have LEFTIAA = -185.9 rather than -05.9 that it was left at the end of last night. -185.9 is the correct value for the AGw give or take a degree or so. Also set the LUCIFER LEFTIAA to the same -185.9.
2:48 pointto BS_9166 gs=0 TelMode
=Guide and try to acquire guiding. Unfortunately, the intial offset failed. GCS continued guiding on background noise.
Sent stopGuiding on command line to GCS. This apparently leaves readGuideCamera unable to
acquire an image.
2:56 Stop GCSL and restart.
2:58 Make small adjustments to IE, CA to make BS_9166 exactly centered on AGw camera.
(still with very rough collimation) IE=-92, CA=-62
03:01 pointto BS_9166 gs=0 TelMode
=Guide, but GCS doesn't start guiding after the
acquisition image. GCSL died at 03:01:35.
03:04 Restart GCSL.
03:05 pointto BS_9166 gs=0 TelMode
=Guide again. The post acquisition offset didn't get
the star into the guide box.
03:06 pointto with TelM
=track, but GSCL does not stop guiding. Worse than that, it moved
the star out of the guide box.
03:09:33 repeat the same experiment, and got the same result. Star moves off the guide box
to the top in ~3 moves
03:12 Change CA from -62 to -64, and see star moves away to lower left.
03:16 OOOOOOOOPppppppps!!!!! In the previous edit of PCSInstrument.conf John
said -1855.9 instead of -185.9. Now corrected.
03:17 Restart PCS and GSCL.
03:18 pointto BS_9166 gs=0 TelMode
What is the waiting of ~20 secs after GCSL says setGuidestarlist execution time: 2504 ms?
Rotator was stopped.
03:25:29 GCSL died unexpectedly. Restart GCSL.
03:29 Stopped both copies of the GCSGUI.
03:30 pointto BS_9166 gs=0 TelMode
=Guide, but initial offset fails.
Again GCSL doesn't log stopGuiding command when we sent a preset in Track mode. Maybe
just a delay in the pipe mode.
03:33 pointto BS_9166 gs=0 TelMode
=Guide, initial offset fails.
03:36 Verified again that offset 0 5 moves star down on the AGw picture. (And verfied that
our DS9 display is not inverted.) with LEFTIAA=-185.9
08:42 Stop and restart PCS after changing LEFTIAA back to -5.9
08:43 pointto BS_9166 gs=0 TelMode
=Guide, still failed to acquire.
And star has moved off to right side of acquisition image (IE in the pointing model file is -85).
Resending the same IE=-95 puts the star back to being centered on hotspot.
03:54 Try guiding in this upside down orientation, but manual images were still being collected.
Had to abort preset.
03:56 Restart PCS to set LEFTIAA back to -185.9
04:02 Restart PCS with IE=-95 LEFTIAA=0 and LEFTZEROPOINT=-185.9
04:04 pointto BS_9166 gs=0 TelMode
Guiding Works! So this means that LEFTIAA rotates the Instrument Position Angle
on the Sky, but does NOT affect the guide correction transformation. While LEFTZEROPOINT
rotates both the Instrument Position Angle and the guide transformation. (Thanks, this is the
transformation that I wanted yesterday to account for the instrument being possibly mounted
in a different way on the AGw. -John)
Correct value for today's mounting of AGw is LEFTIAA=0 and LEFTZEROPOINT=-185.9
Untangling the Active Optics Transformation
04:06 pointto BS_9166 gs=0 TelMode
Doug running his active optics routine but it fails with RPC error.
04:19 GCS died with glibc error
04:40 triggerWFS in GCS from command line.
04:42 Clear Active Optics on M1.
triggerWFS in GCS from command line. wfscimage000033
04:44 Manual +1000 Z7 on M1
triggerWFS in GCS from command line. wfscimage000034
Change /lbt/tcs/current/etc/GCS/AIP_L.cfg to modify WFS_rotangleOffs
from 74.0 to 14.0.
(Torsten thought no need to restart GCS, but we had to restart it for the change to
04:49 triggerWFS in GCS from command line. wfscimage000035 (but no difference in Z7/Z8)
04:53 Stop and Restart GCS
04:54 pointto BS_9166 gs=0 TelMode
04:55 triggerWFS in GCS from command line. wfscimage000036 (measured Z7 now looks correct, but correction doesn't converge when we send it in wfscimage000037)
05:00 Change WFS_rotangleOFFs from 14.0 to 134.
05:02 pointto BS_9166 gs=0 TelMode
=Active, triggerWFS wfscimage000038
05:06 triggerWFS wfscimage0000039
05:09 and 05:11 Clear Active Optics on both M1 and M2
5:14 pointto BS_9188 gs=0 TelMode
(300 sec timeout, Preset failed for CW limit polynomial on rotator.)
5:22 abort second preset with same problem and manually unwind rotator.
05:25 Restart GCS to pick up new build. This version of GCS knows about the expected
field aberration that the WFS sees.
05:25 pointto BS_9188 gs=0 (still bad polynomials)
05:27 pointto BS_9188 gs=0 PA=180 (star out of focus)
Adjust M1 Zglobal=-0.2 (net -0.5)
05:32 pointto BS_9188 gs=0 PA=180 TelMode
05:38 triggerWFS then Average wfscimage000046, 47, 48
05:41 Manual +1000 Z7
05:42 triggerWFS then Average wfscimage0000049, 50 , 51
Needs change of 55 degrees with uncertain sign (Andrew calculated 55.7 deg). So new value is 134+55 = 189 for WFSrotangleOffs
. If this works, Monday's discrepancy of 90 deg versus 145 deg has gone away.
05:56 Stop and restart GCSL.
05:57 pointto BS_9188 gs=0 PA=180 TelMode
05:58 Clear Active Optics
05:59 triggerWFS then Average wfscimage000052, 53, 54
06:02 Manual +1000 Z7
06:03 triggerWFS then Average wfscimage000055, 56, 57
06:06 pointto BS_9188 gs=0 PA=180 TelMode
=Active (with send wavefront 0)
06:07 sendZernikestoPSF 1, and the 2000 nm of coma was corrected.
But focus was sent with the wrong sign.
06:10 Change PSF gain to -1 for Z4 and it CONVERGES! sequence ends at wfs000067
06:16 pointto BS_9188 gs=15 TelMode
=Active (PA=210) - We found this off-axis star on the first try. Big jump (3 arcsec) after it starts guiding. Seeing is nice and tight.
06:22 Coma in image is diverging and we lose the guide star.
06:26 Clear active optics
06:27 Resend the preset, but guider doesn't find star
06:30 pointto BS_9188 gs=16 TelMode
Seems to be having convergence issues with astigmatism.
06:42 Clear Active Optics and resend the preset in Guide mode.
23:42:12 TCPSocket recv error, so we stop and restart GCS and resend the preset.
Guider finds star OK, but star is ~8 arcsec offcenter on IRTC.
06:50 M2 Zglobal=-0.26 mm
Doug running IDL active optics to see how that converges. Diverged in +Z7
07:02 Clear Active Optics
07:04 Manual +1000 Z7 to M1
Doug finds that the rotation angle is off by
07:14 offset 0 5 while guiding. Hung. Aborted and Killed GCS.
07:17 pointto BS_9188 Guide gstar=0 (on-axis) PA=320. Bad Polys
07:19 pointto BS_9188 Guide gstar=0 (on-axis) PA=180
Doug running IDL again in this on-axis position. Diverged sending Z7 to secondary.
07:25 Clear active optics on M1 and M2. Change sign of Z7 on M2 to -1.
07:28 Change sign of Z8 on M2 to -1 and continue iterating.
Now IDL has converged with -1 gains on M2 Z4,Z7,Z8.
But, maybe not Z9, 10 (could be small angular error).
07:34 switch PSF to sending all corrections including focus and coma to M1 with default gain
=1. Angle in IDL is now changed from 189 to 9.
Now converged below 150 nm rms including trefoil.
last image is wfs000024
07:42 Changed the WFS_rotangleOffs from 189 to 9 degrees. (but haven't restarted GCS).
07:45 Back to PSF default split of M1/M2 with gains of +1 for Z7, Z8 on M2 (still -1 for Z4).
first image is wsf000025
08:07 M2 returned after the move, but PSF thought it should be at a different value. ????
It cleared up when we pressed collimate on PSF/M2.
00:49 Manual +1000 Z7 on M2. (Doug corrected -1100 Z7.) Converged nicely.
last image was wfs000031
00:56 pointto BS_9188 gs=16 TelMode
=Guide (PA=322) off-axis, bad polys.
00:58 pointto BS_9188 gs=15 TelMode
=Guide (PA=210), star fainter but guiding OK.
first image was wfs0000032
08:05 Manual +1000 Z7 in wfs000034.
Stopped and restarted GCS to get new angle parameter.
08:12 pointto BS_9188 gs=0 PA=210 TelMode
08:18 Changed gain of Z7 to -1.
08:19 Jesper joins us from Potsdam (without having our Tucson-mtn link through
the Potsdam gatekeeper).
08:20 Changed gain of Z8 to -1. Now coma converges but trefoil diverges.
08:25 Clear M1 active optics, and make gains on M1 Z9, Z10 to -1. Better, but Z9, Z10 seem
to be oscillating. So now the question is why the IDL code converges, and why the
GCS code diverges off-axis? Could there be a problem in the off-axis correction?
08:38 Preset in guide mode. Change all gains back to +1 in PSF. Preset hangs.
Calibrating the AGw Axis Position
Aaron preset to M103 at 01:33:06.9 +60:42:40.2
Bad polys at PA=0. Switch to PA=90.
first 1 sec image with stars is offset 0 0 = guider000004
offset 0 10 = guider000005
offset 10 0 = guider000006
offset -10 0 = guider000007
3 sec = guider000008
offset -10 10 = 3 sec = guider000009
offset -15 10 = 3 sec = guider0000010
offset -15 20 = guider0000011
Aaron resends preset with PA=270
offset 0 0 = guider000012
offset -10 0 = guider000013
offset -10 10 = guider000014
offset -15 10 = guider000015
offset 0 0 = guider000016
offset -10 0 = guider000017
offset -10 10 = guider000018
offset -15 10 = guider000019
Probe has been at X=0.16 Y=611.94
09:18 stop and restart oacserver, and home the AGw axes.
09:20 AGW selection failed. GCS died.
09:22 Restart GCS.
setxy to the above values.
offset 0 0 = guider000020
offset -10 0 = guider000021
20 sec guider000022
60 sec guider000023
90 sec guider000024 (open loop tracking is decent)
offset -10 0 = 3 sec = guider000025
offset 0 0 = guider000026
offset -10 0 = guider000027
09:43 stop and restart oacserver and home motors.
offset -10 0 = guider00028
offset -10 0 = guider000030
oops probe was in wrong place since the previous homing
offset -10 0 = guider000031
offset -10 0 = guider000032
offset -10 0 = guider000033
10:20 Going to put the GCSGUI through its paces
To make this work (roughly) have to set corrections to Primary only and gain factors on Z6, 9 and 10, and even then GCS WFS is having serious issues, not converging with wavefront sensing iterations. Doug's IDL routine is now converging reliably so we now have GCS and Doug's IDL giving different results, not what we had at the end of June.
10:21 immediate GUI issues:
1) "Cam view" doesn't work, no WFSCam picture displayed. We need this one fixed Friday if possible.
2) FWHM calculations giving huge (wrong) numbers. Here 1" images are being reported as 330 " FWHM. I guess that the number of pixels is being
divided by 0.054 instead of multiuplied by 0.054.
3) No scale shown on circular projection axes
4) On the bright side Zernike numbers and RMS WFE seem correct.
5) Top row of pixels in Guidecam view has bogus signal that stuffs up the scale of the whole image.
10:49 or changing to MST 3:49 AM hard point 2 causing problems, have to re-set.
4:15 mirror recovers after 25 minutes.
4:16 re-send preset to on-axis BS9188
4:25 Doug finishes changes to IDL routine for test
4:35 running IDL WFS on guide star #16. using M2 for Z4, 7 and 8
4:40 collimated off-axis with wfsc 000038-39
4:43 from wfsc 000040-45 we include expected aberration corrections
4:47 going on-axis to see what this gave us.
4:50 EXCELLENT!!! Off-axis with expected values taken into account we had Z5-Z8 reporting in the low hundreds of nm, once we converged and went on-axis to see what we had all of these were in the 10's.
4:53 went to guide star #34 a bit dim, increasing exposure to 1000ms, this star is at Y stage 427 mm, so 309" off-axis... setting exposure to 2000ms
4:56 TORSTEN, changing guide camera exposure time doesn't seem to work.
5:04 gone to BS9107 for more bright off-axis stars given that we can't change exposure time on guidecam. Changed IE CA to -100 -70
5:13 went to #96, 258" off-axis, collimated with expected field aberations in wfsc000050-54, went on axis and measured wfsc 000055-57, sun rising, logging off
- 05 Sep 2008