LFBG Re-Commissioning with LUCI: 14 September 2010 UT

Observer: DThompson, JHill (LBTO) plus some DMiller (home) plus some JStorm (Potsdam)
Telescope Operator: DGonzalezHuerta (LBTO)
Telescope Support: JLittle (LBTO)
Instrument Support: TShih (LBTO)

Summary

For a post-shutdown restart night, this was wildly successful. Essentially everything worked! Clouds gave us a late start, and then came back after midnight for a while. We battled with the oacserver for a bit and revisited a couple vintage issuetraks. We even got to take the initial transformation data (20 stars) at the end of the night in 0.5 arcsec seeing on the guider. One of the 5x2.0s LUCI exposures had 0.29" FWHM, and towards the end of the night a good fraction were well below 0.4".

Preparations

23:43 Started LUCIFER UI. The /archive disk was not mounted, but a "mount /archive" as root worked after shutting down the UI. Data copied to /archive. Data makes it down to the Tucson archive as well.

03:00 We installed Jesper's new configuration file in /lbt/oacontrol/etc/oacontrol.conf and restarted the oacserver. Jesper wrote:
  I have left a revised version of the oacontrol.conf file in
/lbt/oacontrol/etc/oacontrol.conf.20100913

  You can replace the current file with the revised one when you are ready to go.

  The transformation has been reset to the nominal one as the position of the
pinhole etc has changed during the refurbishment in July. The  nominal
transformation should give you a workable system, but a full recalibration is
called for as you have it sketched out in your plan for the coming nights.
The file previously in use is named oacontrol.conf.20100715b.conf.

03:05 David sends closed-dome preset.

03:30 Norm has to cycle power on the AGw3 MOXA (by pulling the cord from the AGw) for the second time this evening. Later we figured out that Norm had tripped on Issue 2603, where you can only ping the MOXA from agw-control and not other machines. The actual problem is that oacserver is not exitting when you stop it, and it hangs around in a zombified state. Somehow, cycling power to the MOXA unhangs the zombies. But, an easier solution is to reboot agw-control.

Details

Open

We are closed for clouds at the beginning.

03:50 Clouds are breaking up, so we are opening.

Initial Pointing

03:55 Acquire preset ACT0347 - and remarkably he finds starlight on the first attempt. (We're making no promises about it being the correct star, and it is quite comatic.) David makes a small adjustment and we are pointing at:
*** Pointing Model Terms ***
IA      -398.16150
IE       -72.00000
NPAE      21.76460
CA       -34.00000
AW        -0.50470
AN        21.74230
TX         8.37990

Assuming that M2 is still mounted in the correct (for field aberrations) position, we put the collimation arbitrator in Primary mode.

Initial Collimation

04:04 Active Preset, but this star is to bright for WFSing.

04:07 Active Preset BS9177 on-axis and let GCSL converge on the collimation (the star is clearly in the hotspot). Pointing was dead nuts after the slew to BS9177.

Note: We did not update the collimation model at this time, because the telescope wasn't in thermal equilibrium.

WFS Pupil Alignment

04:15 Call DMiller as we need his expertise for a pupil alignment on the WFS. This active correction looks to be diverging in spherical. The good WFS images for the pupil alignment should be about 4:15 UT before the spherical got too far off.

04:20 Stop and clear active optics.

04:22 Preset BS9177 again

04:26 Preset BS9177 gs=45 - We see the star on LUCI, and it is about 2.5 arcsec off the old center in each axis. Collimation is working but not converging to a stable answer.

Summary of Parameter Changes for AIP_L.cfg (need to propagate to other AGw3 files)
WFS_lenslet_edge_x was 381.65 change to 355.0
WFS_lenslet_edge_y was 213.85
WFScam_hotspot_x 403.0 change to 381.0
WFScam_hotspot_y 212.0 change to 232.5
onAxis_y 612.2 change to 612.5

Note that all of these parameters weren't changed at the same time, this is the summary table for the night.

04:45 Preset BS9177 gs=45 to grab the new parameters (GCS was doing this anyhow?)

05:08 Kill GCSL to take new parameters (we're still slightly confused about when GCS reads which parameters)

Seeing is running about 0.65 arcsec.

05:18 Stop GCSL WFSing so we can run IDL, but Doug has IRTC wired in there somewhere. He's fighting with some authorization issue. Meanwhile seeing is 0.55-0.7 on guider without WFS. Dave says 0.48 arcsec on LUCI.

05:27 Clouds have gone, so GCS is complaining that our guide star is too bright.

05:29 Resend preset to restart GCSL WFSing.

Rotation alignment on LUCI

05:36 Preset in TRACK mode to BS9177, Files: luci.20100914.NNNN, all are Ks 5x2.0s o2dcr/integrated

FlexComp Off FlexComp On
PA NNNN PA NNNN
90 0005 0 0009
270 0006 180 0010
180 0007 270 0011
0 0008 90 0012

New rotator center 1025.3 1050.6 with flexure compensation on.

Rough-check NSEW Offsets in TRACK mode

NNNN ABS/RADEC Offset
0013 at center
0014 60" N
0015 60" S
0016 60" E
0017 60" W

These were taken with old LEFTZEROPOINT -187.703 degrees. The rough orientation check is only off by ~0.08 degrees.

IDL verification of active optics

05:53 Send Guide preset from LUCI at PA=270

IDL active optics verification being done by Doug from home.

Data live in /OldRepository/AGW_Data/20100914/

wfsc 00003 - 00012 were a run of collimating

GCS verification of active optics

06:07 Send Active preset from LUCI to same BS9177

06:09 Send Active preset from LUCI to BS9186

This is a known minor issue in the configuration file. It doesn't cause any problems as the assumptions are correct.
Sep 14 06:08:32 tcs4 LBT_left_GCS: WFS_pupil_dir_x configuration for AGW missing. Variable not found., pathname is '/lbt/tcs/current/etc/../etc/GCS//AIP_L.cfg'.'
Sep 14 06:08:32 tcs4 LBT_left_GCS: assuming WFS_pupil_dir_x = +1
Sep 14 06:08:32 tcs4 LBT_left_GCS: WFS_pupil_dir_y configuration for AGW missing. Variable not found., pathname is '/lbt/tcs/current/etc/../etc/GCS//AIP_L.cfg'.'
Sep 14 06:08:32 tcs4 LBT_left_GCS: assuming WFS_pupil_dir_y = +1

06:15:40 send +200 Z9 manually

06:20:30 send +200 Z10 manually

but problems because things are noisy in Z9, Z10 at the level of +-150 nm.

06:25:15 send +500 Z10 manually

06:27:40 send +500 Z9 manually

Conclude that the rotation alignment is OK as expected, but there is enough local seeing that this isn't a good night to measure it precisely.

Measuring Rotator center on AGw

Data live in /OldRepository/AGW_Data/100914/guider000002.fits
Note that Doug/John made two similar, but different directories. 100914: AGw images; 20100914: wfs images

setxy -u 3 -x 0 -y 612.2
offset 0 5
guider00005 is 70 sec while moving rotator 0-360
guider00006 is 80 sec while moving rotator 360-0

The center of the circle in guider00006 is at 123.0,148.25 or corrected for binning 246,296.5.
Comparing that to the nominal hotspot of 274,298 we find that we need a shift in the X-direction of 28 pixels.

Dave is continuing with other measurements while John works on these images. How big is a pixel in encoder units? This is John's calculation confirmed with Jesper later:
14 binned pixels
28 un-binned pixels
1.4 arcseconds
0.84 millimeters
57 encoder counts (at the nominal on-axis position)
But what is the sign?

Orientation of LUCIFER on sky (LEFTZEROPOINT adjustment)

0018 BS9186 at PA=0 Ks 6x10.0s o2dcr/integrated
0019 sky for BS9186 (10.0, 10.0 ABS/RADEC offset)
Not a lot of stars in this field.

On slew to next field M1 panicked around 07:33, then hit limit after recovering from panic. David offset M1 and M2 by dX=+2 mm dY=-2 mm on both.

0020 BS9188 at PA=180 Ks 6x10.0s as before.
0021 sky for BS9188 (10.0, 10.0 ABS/RADEC offset)

0022 BS9107 at PA=0
0023 sky for BS9107 (10.0, 10.0 ABS/RADEC offset)

last image affected by rapidly thickening clouds! Closing dome for the time being.

More orientation of LUCIFER

09:38 Reopened after it mostly cleared. Taking more data for measuring orientation of LUCIFER. All data are 6x10.0s Ks images.

BS9101 PA=0 0026, 0025=sky
BS9104 PA=180 0027, 0028=sky
BS9105 PA=0 0029, 0030=sky
BS9107 PA=0 0031, 0032=sky 0033=-60" DETXY
BS9108 PA=0 0034, 0035=sky (at +10PA the rotator changed by +10 degrees)

Only the BS9107 field is dense enough to give a good fit, and there I get 359.995 with a second order fit and 0.014 with a third order fit, so there is no strong evidence of any shift in LEFTZEROPOINT.

Guiding Test in Parallel to Above

10:12 Lowered guide correction factor from 0.75 to 0.25. This makes guiding look much better, so the biggest jumps must be out-back. We used 0.25 for the rest of the night. It looks better to JMH's eye.

Back to adjusting the rotation center

Thick cloud is hovering over the western end of the mountain. Talked to Jesper and confirmed that my calculation above was the correct one.

10:25 Change the AGW3 ROT homeoffset from 41132 to 41075 encoder counts (-57) in oacontrol.conf. Stop and restart GCS/AGw3/oacserver.

But, then we get AGw3 hung again. We find many dead oacservers........
[agwuser@lbtmu105 etc]$ ps -ef |grep oac
root      1086     1  0 03:08 ?        00:00:00 /lbt/oacontrol/bin/oacserver
root      1517     1  0 03:12 ?        00:00:00 /lbt/oacontrol/bin/oacserver
root      1894     1  0 03:14 ?        00:00:35 /lbt/oacontrol/bin/oacserver
root      2690     1  0 Sep09 ?        00:03:15 /bin/bash /lbt/oacontrol/bin/check4nport.agw2
root      2691     1  0 Sep09 ?        00:00:03 /bin/bash /lbt/oacontrol/bin/check4nport.agw3
root      2692     1  0 Sep09 ?        00:00:01 /bin/bash /lbt/oacontrol/bin/check4nport.agw4
root      2765     1  0 Sep09 ?        00:06:54 /lbt/oacontrol/bin/oacserver
root      7478     1  0 10:26 ?        00:00:00 /lbt/oacontrol/bin/oacserver
agwuser   8355  6103  0 10:36 pts/2    00:00:00 grep oac
root     14737  2690  0 Sep12 ?        00:00:00 /bin/bash /lbt/oacontrol/bin/check4nport.agw2
root     16954     1  0 00:59 ?        00:00:00 /lbt/oacontrol/bin/oacserver
root     17297     1  0 01:01 ?        00:00:00 /lbt/oacontrol/bin/oacserver
root     17807     1  0 01:05 ?        00:00:00 /lbt/oacontrol/bin/oacserver
root     18222     1  0 01:09 ?        00:00:00 /lbt/oacontrol/bin/oacserver
root     18624     1  0 01:12 ?        00:00:06 /lbt/oacontrol/bin/oacserver
root     31973  2691  0 08:22 ?        00:00:00 /bin/bash /lbt/oacontrol/bin/check4nport.agw3
These are so badly hung or zombified that sudo kill -9 won't get rid of them.

10:40 reboot agw-control (lbtmu105) which gets rid of all the zombies. Also be sure to wait 30 sec for the oacserver to start before you startAGW.

Tue Sep 14 10:49:48.923 2010  Preset alert for left LUCIFER mode ACTIVE: no AGW unit selected.
Tue Sep 14 10:49:48.933 2010  Preset command failed for left LUCIFER mode ACTIVE from IIF

10:52 Stop and restart GCSL, same problem.

David finds the trick that killing GCSL will get it to pick up the AGW unit.

11:01 preset BS9108 on-axis

setxy -u 3 -x 0 -y 612.5

11:08 guider00008 (saturated) but we went in the right direction

Change IE from -72 to -78 to offset the star

11:11 guider00009

This is good alignment at binned X=135.6 Y=153 corresponding to un-binned X=271.2 Y=306

Also adjust the limits by the opposite amount. If homeoffset got more negative, both limits should get more positive. For AGW3 ROT we adjusted lim_f=-14127 and lim_n=12825 .

Use rdwrconfig -u 3 to apply limits. This wonderful command avoids the need to stop and restart the oacserver.

11:32 preset to StoneC field at 03:30 Seeing remains 0.5 arcsec.

Adjust onAxis_y from 612.2 to 612.5 in AIP_L.cfg. The smaller value was a patch to avoid issues with the reach of the probe. That issue was fixed in July, so all AGws should be using 612.5 as the transforms are redone.

Initial Probe Transformation Data in Ks

11:38 preset StoneCagw gs=17 ACTIVE and take LUCI image 00036
./getdat3 17 on agw-control

The default transformation was pretty good looking as we are hitting the stars well by eye.

Note: LUCI flexure compensation was OFF for this dataset.

GS PA UT LUCI-image LUCI XYcen
17 0 11:38 00036 1030.09 1049.07
20 0 11:43 00037 1021.55 1046.88
19 0 11:45 00038 1024.46 1046.62
16 0 11:47 00039 1026.31 1050.75
17 0 11:49 00040 1030.98 1050.01
18 0 11:50 00041 1029.07 1047.69
5 0 11:52 00042 1028.30 1054.53
2 0 11:54 00043 1024.54 1056.14
21 90 11:56 00044 1020.30 1037.52
20 90 11:58 00045 1022.32 1044.09
13 180 12:01 00046 1030.47 1048.07
7 180 12:04 00047 1026.10 1042.61
13 200 12:05 00048 1033.10 1052.07
10 200 12:08 00049 1034.30 1052.84
11 200 12:09 00050 1033.18 1053.09
12 200 12:11 00051 1032.29 1051.97
14 200 12:13 00052 1024.66 1049.90
21 200 12:16 00053 1030.61 1054.53
7 200 12:17 00054 1029.83 1047.18
8 200 12:19 00055 1034.61 1050.71
3 200 12:22 00056 trailed

The getdat3 data is in logfile.1.20100914 in /home/agwuser/agw-control/

00056 is likely bad star coordinates, or a bad acquisition in twilight.

12:25 Close for impending dawn. Sky is mostly clear, with scattered clouds to the southeast.

-- JohnHill - 07 Sep 2010
Topic attachments
I Attachment Action Size Date Who Comment
20100914.xyxy 20100914.xy manage 2 K 15 Sep 2010 - 09:36 JohnHill XY Requested Positions during Transformation Data
daq.20100914.jpgjpg daq.20100914.jpg manage 618 K 22 Sep 2010 - 03:31 JohnHill Plot of Telescope Temperatures
Topic revision: r19 - 22 Sep 2010, 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