LUCIFER Com 3, Night 9
UT 2009 January 18

Observers: N.Ageorges, W.Seifert, M.Juette, V.Knierim, J.Hill (@LBT)
Telescope Operator: D.Gonzalez-Huerta
SW Support: T.Edgin in Tucson until 2AM
Telescope Support: J.Rousseau

Summary:

A mostly clear, but not photometric, night with mediocre seeing (0.8 arcsec).

There was jubilation when we measured that the ....agw9 transformation was superbly aligned with the slits.

There was depression when John figured out he had been moving the primary the wrong way all week to avoid the -Y limits. It took an hour to recover from that. (Note added later: John was wrong about being wrong. He was only confused.)

There was confusion when we found that the ...agw9 transformation (or something) didn't work at non-zero position angles.

There was an arcminute-class telescope jump in azimuth at 18:42. ECS crashed at 04:30. There were no GCS/TCS hangs tonight! Telescope was very well behaved.

Details (Times in MST below):

16:10 At zenith in dark for LUCIFER calibrations

16:30 - 17:00 Restart two PSFGUIs (memory), PCSGUI (memory): (to-station Swap is down to 650k). Fresh PSFGUIs start at 37m RES.

17:05 Clear C00, and FCin=1

Open

17:10 Open for Flats. T=1 degC, D=-10 degC, wind calm, sky mostly clear but not photometric.

Star

18:15 Preset to pointing star - Adjust IE=46.7 CA=11.25

18:20 Yglobal offsets for M1/M2 are +2.8 relative to nominal. Added -5000 nm Z7 to M1.

18:27 preset ACTIVE - on-axis

18:33 preset ACTIVE - off-axis

18:38 Why is WFS cycle time now 64 sec when sometimes it is as short as 47 sec for a 30 sec exposure?

We are still using external transformations from oacontrol.conf.sfpagw9.

Telescope Jump

18:42 Why did telescope stop tracking and then restore itself?
Sat Jan 17 18:42:34.398 2009 setting none mode on left side

See image LUCI 20090117 0004. Telescope seems to have glitched in Azimuth by 68 arcsec. I added a note to Issue 1928 although I don't know if this was a T0 glitch or not. We lost 5 minutes thinking about this jump, but in fact it was self-recovering after we sent a new preset.

18:52 Set FCin=1 for Z11/Z22.

Offset Tests

18:55 Doing offset tests. Seeing on guider is 1.4 - 1.7 arcsec, and this is the interesting condition where the Polaris monitor is reporting the correct value. WFS spots show a gradient of seeing by a factor of 2 from lower right to upper left. (All 4 vent doors are open. Right Shutter Door is closed. Wind (LBTO) is 2 m/sec.

Extracted from Walter's email describing the test results for ....agw9 transformation.
Sequence          XCEN                                    YCEN                     XERR
--------            ------------------------                     ------------------------           ------------
       1     1.00491192626953e+03     1.86651220703125e+03   4.7555e-02
       2     1.00518481445312e+03     2.00744079589844e+02   3.7092e-02
       3     1.84151293945312e+03     1.03507141113281e+03   6.6017e-02
       4     1.71325073242188e+02     1.03405187988281e+03   4.2174e-02

The corresponding files are luci_20090117_agw_0006 - 0011

In Y-direction, the offset is 0.3 pixels (0.01deg), the best value we ever reached, also better than the 
internal Torsten transformation. Fully within the specs.
In X-direction the deviation is 1.0pixels (0.03deg), completely sufficient for our purposes.

Beware that these measurements were in 1.5 arcsec seeing, so not the best possible residuals.

Spectroscopy

19:08 preset ACTIVE - bumped into -Y limit at EL=39deg

19:10 Increase the Yglobal offsets to +3.0 mm.

19:11 preset Active - but star was off the bottom of acquisition image. (How could adjusting Y global have changed the pointing when we only moved 0.2 mm? Could running into the static supports have tweaked some kind of mirror position hysteresis?)

19:16 preset Active - to a star on the y-axis, this moves the star just onto the acq image and we start guiding. The seeing has improved to 1.1 arcsec down here at EL=40 deg.

20:24 change to a different slit mask on these same binary stars.

20:53 preset ACTIVE for telluric standard star - failed to find star - nothing bright enough in the acq image

20:56 preset ACTIVE - succeeds with different PA on y-axis - enough to slide star into acq field by rotating it into the corner.

21:43 Seeing on guider is 0.9 arcsec at EL=60 deg.

22:07 Preset ACTIVE - didn't find guide star.

22:09 Preset ACTIVE - didn't find guide star.

22:11 preset TRACK PARALLACTIC - Adjust IE=24 CA=14.8

Added the next day: John was wrong about being wrong, he was only confused. M1 and M2 move together (same sign) for coma-free offload.

John screwed up the alignment!

I've been puzzling for days about why we have 5000 nm of coma after we move M1 and M2 in coordination to get off of the -Y limit. Finally the right neuron has woken up, and I've realized my stupid mistake from Tuesday. The M2 coordinate system is rotated compared to M1. So when we applied +3 mm Yglobal to M1 and +3 mm of Yglobal to M2, this was separating the axes of the two mirrors by 6 mm. That then explains why we need to apply 5000 nm of Z7 to M1 to fix up the coma. This screws up the field aberrations, the pointing, and perhaps even the focus.

The correct thing to do would be to apply +3 mm of Yglobal to M1 and -3 mm of Yglobal to M2. See Com20080414 where we measured this effect.

Here's the recovery procedure:

  • Initialize M2 Lookup Table
  • Clear M2 Active Optics
  • Initialize M1 Lookup Table
  • Clear M1 Active Optics including C00
  • Clear M1 Secondary Offload
  • Restore Defaults and Send (to set FCin=0)
  • Adjust M1 Yglobal to +1 mm (or whatever value needed to clear the -Y limit on M1)
  • Add the opposite of the M1 Yglobal to the default M2 Yglobal

  • Adjust Pointing on a bright star

  • Preset ACTIVE and run Active Optics until converged (the flaw is assuming the coma correction will be small).
  • Set FCin=1 for Z11, Z22.
  • Reset Inputs on all PSF GUIs.
  • Add global offsets to Collimation model files (later)

22:22 preset ACTIVE on-axis

22:49 preset ACTIVE on-axis

23:05 preset ACTIVE on-axis

It didn't work like I expected since the "correctly reversed" position still had a bunch of coma. Thus it took an hour of manual iteration to get into a collimatable state. Need to study the October/November field aberration position to understand how close we are to it.

We made this same mistake about the sign of the global offset on 16-17 October (and maybe on 15-16 October), so the history is complicated. We thought we understood the correct collimation values on 17-18 October, but I'm not so sure when I read the logs from the previous two nights in October. Andrew says the final determination of the field aberrations was 12 November, so I don't have to worry about October. The result of the 11-12 November measurements put the M2 position at:
M2 Xglobal=-7.010 Yglobal=-1.138 Zglobal=0 RX=+114.28 RY=181.80

23:18 Finally get in semi-OK state with M2 Yglobal=-0.138 and M1 Yglobal=-1.0

23:20 FCIn=1 Correction "Default" - collimated at last.

Need to build this into the collimation models once I understand better why I had to do what I did.

EL=50 deg
M2 X=-7.187 Y--0.171 Z=-0.037 RX=117.76 RY=165.385
M1 X=-0.751 Y=-1.098 Z=+0.034 RX=036.7 RY=+6.57

Adjusted IE=-12 CA=+10

Software Tests

23:28 preset ACTIVE - two equal stars in acq field

23:38 Guider held lock under a 90 degree rotation offset

23:41 The next rotation offset failed because the probe couldn't move any farther.

Focus

00:00 preset ACTIVE - seeing is 1.0 arcsec at EL=59 deg (x=-1.4 y=92.0)

Focus is close to the 31 mm number measured recently. Focus did NOT revert to 35 mm as I feared it might. We will measure again if the seeing improves.

Testing LUCIFER Acquisition procedures

Walter is testing the ability to rotate two stars onto a long slit.

Now checking acquisition at different guide star positions in the field

00:38 preset ACTIVE PA=0 127,257 acquire000073

00:42 preset ACTIVE PA=20 137,320 acquire000074

00:44 preset ACTIVE PA=0 121,248 acquire000075

00:47 preset ACTIVE PA=-20 142,172 acquire000076 (x=-32.9 y=86.0)

Jan 18 00:38:28 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:1.068871 y:520.781066, 23.000000, 0) ...
Jan 18 00:42:20 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-30.515863 y:525.863046, 23.000000, 0) ...
Jan 18 00:45:07 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:1.078177 y:520.778116, 23.000000, 0) ...
Jan 18 00:47:11 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:32.533765 y:526.749599, 23.000000, 0) ...

Why do we see 5 arcsec changes in the location of the star as the PA changes?

Make offset test here (PA=-20) to check if the transformation still works off the Y-axis

The setxy commands seem to be correct, but the result that comes out from the stars is not what we expected. The offsets on the stars seem to be rotated by 1.8 deg (for a Y offset) and 0.66 degrees (for a X offset).

What PCS asked for: (what LUCIFER wants)
Jan 18 00:50:38 lbtmu102 LBT_PCS: PCS:TELESCOPE:OffsetTelescopeDETXY executed
Jan 18 00:50:38 lbtmu102 LBT_PCS: MV_REL: deltaX: -0 deltaY: 60 deltaAngle: 0 offsetRotAngle: -0.349066
--
Jan 18 00:51:50 lbtmu102 LBT_PCS: PCS:TELESCOPE:OffsetTelescopeDETXY executed
Jan 18 00:51:50 lbtmu102 LBT_PCS: MV_REL: deltaX: -0 deltaY: -120 deltaAngle: 0 offsetRotAngle: -0.349066
--
Jan 18 00:53:05 lbtmu102 LBT_PCS: PCS:TELESCOPE:OffsetTelescopeDETXY executed
Jan 18 00:53:05 lbtmu102 LBT_PCS: MV_REL: deltaX: 60 deltaY: 60 deltaAngle: 0 offsetRotAngle: -0.349066
--
Jan 18 00:54:13 lbtmu102 LBT_PCS: PCS:TELESCOPE:OffsetTelescopeDETXY executed
Jan 18 00:54:13 lbtmu102 LBT_PCS: MV_REL: deltaX: -120 deltaY: 0 deltaAngle: 0 offsetRotAngle: -0.349066
--
Jan 18 00:55:26 lbtmu102 LBT_PCS: PCS:TELESCOPE:OffsetTelescopeDETXY executed
Jan 18 00:55:26 lbtmu102 LBT_PCS: MV_REL: deltaX: 60 deltaY: 0 deltaAngle: 0 offsetRotAngle: -0.349066

What GCS did: (looks correct because offsets are nearly square in XY)
Jan 18 00:47:11 lbtmu107 LBT_left_GCS: AGWUnitAIP: SFP position to move to: -32.905843, 85.995519 mm
Jan 18 00:47:11 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:32.533765 y:526.749599, 23.000000, 0) ...

Jan 18 00:50:42 lbtmu107 LBT_left_GCS: AGWUnitAIP: SFP position to move to: -32.930247, 146.098720 mm
Jan 18 00:50:43 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:32.254531 y:466.909676, 23.000000, 0) ...

Jan 18 00:51:55 lbtmu107 LBT_left_GCS: AGWUnitAIP: SFP position to move to: -32.904459, 25.967639 mm
Jan 18 00:51:55 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:32.832939 y:586.518499, 23.000000, 0) ...

Jan 18 00:53:10 lbtmu107 LBT_left_GCS: AGWUnitAIP: SFP position to move to: 27.104887, 85.994147 mm
Jan 18 00:53:10 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-27.536551 y:526.882232, 23.000000, 0) ...

Jan 18 00:54:17 lbtmu107 LBT_left_GCS: AGWUnitAIP: SFP position to move to: -92.966427, 86.021748 mm
Jan 18 00:54:17 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:92.620925 y:526.565173, 23.000000, 0) ...

Jan 18 00:55:30 lbtmu107 LBT_left_GCS: AGWUnitAIP: SFP position to move to: -32.919262, 85.995345 mm
Jan 18 00:55:30 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:32.547195 y:526.749741, 23.000000, 0) ...

So how can the motions of the stars be rotated compared to the LUCIFER detector? It seems as if the AGW transformation is aware of the position angle ?? (This seems like jibberish written by a sleep-deprived person in the middle of the night, and I wrote it.)

~02:00 Tony goes home with a perfect record of no TCS hangs on his watch!

02:06 preset ACTIVE - PA=0 with guide star at x=-50.15 y=+107.29 (star off the y-axis, but field at PA=0)

Transformations on this one worked correctly.

02:19 preset TRACK PARALLACTIC - Adjust IE=-5.75 CA=-11.35

02:24 preset ACTIVE - PA=-105 255,215 x=79.69 y=143.23 (would have been x=0 at PA=-135) (guide star is a binary)
Sun Jan 18 02:30:18.679 2009  Offset alert for left LUCIFER: couldn''t move probe to position (SFP) 79.705, 203.453, focus: 7.71771 :AGw command status: AGW1: Requested position is Out of Range
Sun Jan 18 02:30:18.697 2009  Offset alert for left LUCIFER: couldn''t move probe to position (SFP) 79.705, 203.453, focus: 7.71771 :AGw command status: AGW1: Requested position is Out of Rangeerror: setProbePosition: couldn't move probe to position (SFP) (x,y,z): 79.705, 203.453, 7.71771 mm.
Sun Jan 18 02:30:18.722 2009  Offset alert for left LUCIFER: oacontrol indicated an error moving to new position
Sun Jan 18 02:30:18.738 2009  Offset alert for left LUCIFER: AGW1: Requested position is Out of Range

Jan 18 02:24:00 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-80.410091 y:469.928547, 23.000000, 0) ...
Jan 18 02:30:18 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-80.743522 y:409.866153, 23.000000, 0) ...
Jan 18 02:31:27 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-80.097382 y:529.866838, 23.000000, 0) ...
Jan 18 02:32:31 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-140.686036 y:469.939014, 23.000000, 0) ...
Jan 18 02:33:40 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-20.259176 y:469.900775, 23.000000, 0) ...
Jan 18 02:34:56 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-80.410973 y:469.943678, 23.000000, 0) ...

02:37 preset ACTIVE - PA=165 305,252

Jan 18 02:37:05 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-89.610698 y:504.956695, 23.000000, 0) ...
Jan 18 02:41:57 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-89.943870 y:444.964504, 23.000000, 0) ...
Jan 18 02:43:03 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-89.314311 y:564.837460, 23.000000, 0) ...
Jan 18 02:44:05 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-149.897333 y:504.995288, 23.000000, 0) ...
Jan 18 02:45:07 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-29.468249 y:504.885563, 23.000000, 0) ...
Jan 18 02:46:12 lbtmu107 LBT_left_GCS: AGWUnitAIP::doSetProbePosition: calling oacontrol::setxy(10.144.0.64, 0, x:-89.622591 y:504.955859, 23.000000, 0) ...

Mini-summary: There are two pieces of evidence that something related to the transformation is wrong:
a) the position of the guide star on the CCD moves by ~5 arcsec as we rotate the position angle by 20 degrees.
b) When we do DETXY offsets from the PA=-20 position, the offsets are wrong.

Spectroscopy

02:51 preset ACTIVE 352,381 for x=-2.33 y=145.21

04:30 ECS crashes.

04:34 Seeing on the guider is 0.92 arcsec FWHM. WFS rms is bouncing between 300 and 550 nm.

04:50 preset ACTIVE

05:05 Both GCSGUIs frozen. David restarted his, but it didn't help. It was only because LUCIFER offset the guide star out of the field for a few minutes, so there was no new image to display.

06:54 Hit the -Y limit on M1. Probably getting too much light in WFS and sending bad corrections. Stop WFS for the rest of this observation.

Close

07:01 Close dome. T=-0.5 degC, D=-11 degC, wind calm, spectroscopic with wispy stuff in south and clear overhead.

We didn't have to kill GCSL all night long. It's present size is up to 2167m of virtual memory. That seems to shoot down my theory that swapping was causing the hangs. However note that PSFL is running on lbtmu107 with GCSL, and not OSS which has been there the last few nights. Tonight OSS Is on lbtmu104 with ECS.

We are parked at zenith doing LUCIFER calibrations in the dark until approximately 9AM.

-- JohnHill - 17 Jan 2009
Topic revision: r10 - 19 Jan 2009, JohnHill
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback