LUCIFER Com 3, Night 6
UT 2009 January 15
Observers: N.Ageorges, W.Seifert, M.Juette, V.Knierim, J.Hill (@LBT)
Telescope Operator: D.Gonzalez-Huerta
SW Support: T.Leibold in Tucson until midnight
Telescope Support: B.Curtis, J. Little and D.Officer
GCS v2.0.5 get the seal of approval for fixing the advertised issues. GCS has been better behaved than
last night, although there have still been some hangs.
We've made progress on the AGW transformation issue, but we haven't solved it yet. We have another
attempt at the rotation angle waiting to be tested (....agw6).
Sky was photometric, but seeing worse than 1 arcsec for most of the night. There seem to be slow low-order variations that are irritating the WFS.
Telescope worked well during the night with only a few TCS hangs.
Morning: Butch deiced the cooling loop for the winter chiller CH0401 by warming it up above 0 degC. Chiller is holding LWT at -24 degC with only HBS and IC running (no mirror ventilation), but is slowly rising.
Jesper has set us two new copies of oacontrol.conf to try to correct the rotation of the transformation.
16:23 Kill and restart IIF so that LUCIFER is able to connect for more than 1 sec.
16:30 Clear Active Optics. And adjust M1/M2 more carefully so that both mirrors have a Yglobal correction of +2 mm compared to nominal. (It was a few hundred microns off from that last night due to an arithmetic error.)
16:40 Torsten installs GCS v2.0.5 with the fix for the waiting for guide star problem. We are still using the internal GCS transformations for probe coordinates.
16:45 Set HBThermalLoopOffset
-10.0 in lbt.conf
Stop and restart ECS to pick up the new value.
Ambient is now +1 degC, so the setpoint is -9 degC. EL pads were still at +4 degC (as reported by WinCC
) where we left them last night. There was little temperature change of the mirror in the dome during the day. The dome ambient (SX Ref Temp) warmed up to 4.4 degC during the day, and the telescope (IC) steel has warmed to 3 degC.
Details (Times in MST):
16:35 Set telescope to zenith for closed dome LUCIFER calibrations with PA 261 and varying angles.
17:30 Open Dome to photometric sky, T=+0.8 degC. D=-14 degC, Wind (LBTO) 6 m/sec.
18:07 - Preset TRACK PARALLACTIC
18:14 Preset ACTIVE - failed because star is off the edge of the field. Why?
18:18 - Preset TRACK PARALLACTIC - Adjust on a different star IE=70.5 CA=-3.8
18:22 - Preset ACTIVE on-axis star appears at 402,211.
18:28 focus 0 from IRAF
18:30 stopWFS and increase common M1/M2 offset to +2.5 mm to get M1 off of the -Y limit.
18:39 increase the common Yglobal to +2.6 mm to get a bit more clearance from the -Y limit.
18:41 set FCin=1 for Z11 and Z22. (convergence of the collimation was quite slow. Is this from dome seeing?)
18:45 preset ACTIVE to put the same star off-axis
18:58 collimation is still struggling around 500-800 nm rms. So we open the Left Rear Vent Door midway.
19:00 LUCIFER is running a test of its observing software with various offset and stepfocus commands.
19:06 John's GCSGUI was very slow to repaint (running from lbtmu01 and displayed on lbtmu05). Stopped and restarted it and this shortened the repaint time from 10 sec to 3 sec.
19:18 WFS fluctuation between 180 nm rms and 817 nm rms including 500 nm changes in Z4 from one active
correction to the next.
19:30 CH0401 LWT is up to -17 degC with only left mirror ventilation running (plus IC and HBS). So defrosting
didn't solve whatever struggles it is having.
19:32 Seeing on the guider is fluctuating between 1.0 and 1.4 arcsec with the Polaris monitor reporting 2.0 arcsec.
20:09 Preset ACTIVE - succeeded; to start some spectroscopic tests.
20:46 Preset ACTIVE for another collimation.
Add another +0.2 to Yglobal so that both M1 and M2 are +2.8 mm from their nominal positions.
Clear C00 and set FCin=0. This made a 2500 nm correction in focus.
20:54 Internal_transformation = false
to use Jesper's oacontrol transformation.
Torsten tells us that SFP_rotation in AIP_L.cfg only adjusts the rotation of the internal GCS transformation.
This is the value 179.7 that Dave T. adjusted last week. Now we are trying to adjust the rotation of Jesper's external transformation.
21:18 copy the new file /etc/oacontrol/oacontrol.conf.sfpagw5 (assumes PA=295.1) to oacontrol.conf so we can test Jesper's latest
transformation. Kill and restart oacontrol. Home the AGW motors.
21:24 Preset ACTIVE fails with star just outside the lower right corner of the acquisition image. 0000100
21:26 Preset TRACKING to adjust pointing IE=+56.9 CA=-12.0
21:30 Preset ACTIVE finds star at 479,273 on acquisition image 0000101
Result: Star shifts by 4.1 pixels in 1650 when we make a relative DETXY offset of 200 arcsec..
21:45 copy the new file /etc/oacontrol/oacontrol.conf.sfpagw4 (assumes 294.9) to oacontrol.conf. Kill and restart oacontrol. Home the AGW motors.
21:50 Preset ACTIVE finds star at 454,191 on acquisition image 0000102
Result: when we offset, this one rotates in the correct direction, but not far enough. The star shifts by 2.5 pixels out of 1650.
22:05 internal_transformation = true
to go back to the Torsten transformation
22:10 Preset ACTIVE to a higher elevation. Seeing is now 0.5-0.6 arcsec on the guider.
22:24 Preset ACTIVE at a different position angle by 10 deg to put the guide star -40 ? mm off the Y-axis of the probe.
02:26 Preset ACTIVE at a different position angle by 15 deg to put the guide star -53 ? mm off the Y-axis of the probe.
02:29 Preset ACTIVE at a different position angle by 30 deg to put the guide star -72.4 mm off the Y-axis of the probe.
We see a little difference between the getStatus reported position: AGW probe is at (x,y): -89.275, 89.1498 mm;
and the value reported by the AGW: getxy 88.794 522.587. We are trying to track this down in real time
with Torsten and Dave T. on the phone. We are suspicious that the SFP_rotation transformation in GCS is
introducing a spurious zeropoint shift of the coordinates that get sent to AGW. If so, it might explain the
strange XY zeropoints in Jesper's transformation based on the Jan 9 UT data. Our midnight analysis is that
the rotation (SFP_rotation - 179.7) is happening after the translation to probe coordinates rather than before. (rotations and translations are not commutative, so order is important.)
Dave T. and Torsten are tasked with sorting this out tomorrow afternoon in Tucson. (Later in the night Jesper
pointed out that we see these 3 mm offsets in off-axis stars in Doug's plots from 09-Nov-2008, so maybe the
problem is not as simple as SFP_rotation.)
23:00:55 Preset hangs with the Block problem
23:07 Kill and restart GCS and then IIF per Torsten's advice.
23:09 Restart the LUCIFER telescope server
23:11 preset GUIDE - failed because probe couldn't move there - observer error.
23:12 preset GUIDE - succeeded
23:20 preset ACTIVE - suceeded.
23:25 LUCIFER sees elongated images in luci_20090114_0121. It could be that the WFS is struggling with focus in the presence of dome seeing. The seeing has gotten significantly worse in the past 10 minutes.
23:33 preset ACTIVE - succeeded
23:34 preset ACTIVE stopped guiding. IS it because the star is too faint? Or is something confused with IIF?
23:41 Kill and restart GCS
23:42 preset ACTIVE succeeds on the same star, but still giving warnings even though the star seems 400 counts above background. However now it seems to be guiding and WFSing as it should.
Wed Jan 14 23:43:48.050 2009 Preset alert for left LUCIFER mode ACTIVE: couldn''t adjust exposure time to meet low counts requirement. star too dim or guide rate too high.
Wed Jan 14 23:43:48.071 2009 Preset alert for left LUCIFER mode ACTIVE: maximum exposure time is 4600 ms at the given guide rate, estimated counts are 851
Wed Jan 14 23:43:48.088 2009 Preset alert for left LUCIFER mode ACTIVE: adjusted exposure time to 4600 ms
00:15 Nancy is complaining that her image quality is poor. It looks like WFS is struggling with focus with the
rms bouncing between 300 and 600 nm. At the same time, the GCSGUIs were not updating the
guide star image.
00:25 Nancy is running an extended imaging script. John's GCSGUI that he just restarted is not updating for
00:30 Stopped and restarted 3 PSFGUI, 1 PMCGUI, PresetGUI
, PCSGUI and 2 GCSGUIs on to-station.
This reduced virtual memory use on that machine by more than a factor of 2 (now down to 427340K). Even with just David's single GCSGUI
running, he is not getting updates of the guide star image.
00:40 I started a GCSGUI on lbtmu102 for a few minutes, but it was also not updating correctly.
00:47 I started a GCSGUI on lbtmu107 for a few minutes, but it was also not updating correctly.
00:43 GCS hung in an offset command
00:50 Kill and restart GCS.
00:50 preset ACTIVE - Now David's lone GCSGUI is updating the guide star image responsively. Was the slow
GCSGUI a precursor symptom of the GCS hang?
00:56 preset ACTIVE that came from a LUCIFER script.
00:57 John started a new GCSGUI running from lbtmu01 and displayed on lbtmu05. It is now updating quickly
as it should.
Jan 15 00:02:56 lbtmu102 LBT_IIF: IIF:Preset executed
Jan 15 00:04:05 lbtmu102 LBT_IIF: IIF:Preset execution time: 68935 ms at 1232003045471
Jan 15 00:05:25 lbtmu102 LBT_IIF: IIF:Preset executed
Jan 15 00:06:24 lbtmu102 LBT_IIF: IIF:Preset execution time: 58943 ms at 1232003184653
Jan 15 00:09:38 lbtmu102 LBT_IIF: IIF:Preset executed
Jan 15 00:10:34 lbtmu102 LBT_IIF: IIF:Preset execution time: 55592 ms at 1232003434525
Jan 15 00:22:08 lbtmu102 LBT_IIF: IIF:Preset executed
Jan 15 00:23:07 lbtmu102 LBT_IIF: IIF:Preset execution time: 59461 ms at 1232004187527
Jan 15 00:50:35 lbtmu102 LBT_IIF: IIF:Preset executed
Jan 15 00:51:27 lbtmu102 LBT_IIF: IIF:Preset execution time: 52149 ms at 1232005887285
Jan 15 00:56:40 lbtmu102 LBT_IIF: IIF:Preset executed
Jan 15 00:57:45 lbtmu102 LBT_IIF: IIF:Preset execution time: 65558 ms at 1232006265783
02:00 preset ACTIVE
02:23 copy the new file /etc/oacontrol/oacontrol.conf.sfpagw6 (assumes 294.9) to oacontrol.conf. Jesper thinks he has fixed the rotation angle in this one. (But haven't yet restarted oacserver.)
If we want to change the rotation, then we need to mutiply this matrix with a diagonal-ish matrix containing sin(theta) and cos(theta).
# sfp to agw transformation matrix
# Fitted to data from 20090108
# Assuming PA=294.9 (corrected version)
Jesper thinks we can adjust afpagw6 to the nominal zeropoint by changing the values of (X_0,Y_0) to (-2.90,610.93).
02:55 I started IDL gsc_collimate to shadow the GCS WFS calculations. They agree within 1%. And as Andrew
predicted, a few minutes after I did this, the discrepant wavefronts markedly decreased. The IDL vector plots show that low order Z4-Z6 is drifting by and making these fluctuations.
03:00 Polaris monitor computer is shutting down with a Windows update.
03:15 Seeing has improved to 0.7 arcsec with 90 nm rms wavefront. Wow - we should market this magic IDL software! Of course, there is the alternate theory that the seeing improved because Nancy inserted a slit mask to
switch from imaging to spectroscopy.
03:16 Why did this stepFocus take 64 sec? There was a significant offload from M2 (an hours worth of corrections).
Thu Jan 15 03:16:39.041 2009 StepFocus command started for left LUCIFER for 0.000000000000000e+00
Thu Jan 15 03:17:43.938 2009 StepFocus command complete for left LUCIFER for 0.000000000000000e+00
03:27 Wind is dropping. So we open the right rear vent door midway.
03:46 preset ACTIVE - succeeded
04:12 preset ACTIVE - succeeded to try some longslit galaxy spectra
04:48 preset ACTIVE - succeeded
05:21 preset ACTIVE - succeeded
05:48 preset ACTIVE - but with calibration unit in place - failed as expected.
05:50 GCSL is up to 812MB of virtual memory after running for only 5 hours.
06:25 preset ACTIVE - succeeded
06:44 preset ACTIVE - succeeded
06:52 Guider loses the star in twilight.
07:10 Close dome T=1.4degC, D=-17 degC, wind (LBTO) 5 m/sec. Photometric
- 14 Jan 2009