Technical Observing at Bent Gregorian with IRTC - 07-08 May 2009 MST

Observers: J. Hill (mtn)
Telescope Operator: W. Wack
Software Support: We talked to Tom, Dan, Norm, Paul, Chris, Kyle (all briefly).
Telescope Support: J. Little

An Interlude in DG Commissioning before we move IRTC to the Direct Gregorian Focus

Summary

We put a lot of mileage on TCS Build 26, and its main new features seem to be working. There are a couple minor bugs still to chew on, but B26 is ready for LUCIFER.

Lost 20 minutes at the start because of something about time in the rotator DSPs. Dan kicked it in the right way, and everything was fine.

Seeing was in the 0.9 - 1.4 arcsec range for most of the night, but improved to 0.5 arcsec for the last couple hours. I took one set of field aberration data with GCS in the poor seeing. We are waiting for Doug to process it and for Andrew to give it some sanity-check analysis. Andrew reports the following day: The data were a bit noisy, focus and astigmatism bouncing around a bit, but the Z7 and Z8 had every point under 200 nm and no obvious pattern to their errors so we can conclude that field corrections are working as expected. In fact, provided WFS was converging well, this is what we should have expected. So the changes we made to GCS have not broken anything in WFS.

Everything worked smoothly until 02:10 AM when we changed authorizations from IRTC to LUCIFER. First, selectAGW wouldn't work. Finally we had to kill and restart GCSL. Lost 20 minutes fiddling with that. I was not able to reproduce this problem at the end of the night.

Then we hit the +Y collimation limit, and got pointed at the wrong star. Lost 20 minutes recovering from that.

Maybe there is some point in giving your instrument a proper Christian name?

I did various open- and closed-loop guiding tests in poor seeing.

Once the seeing improved, we started taking transformation data on Stone M at 3AM. It was immediately apparent that the current transformation (or the rotator zeropoint, or something) is behaving poorly. Sometimes the on-axis star was missing the rotator center by 5 arcsec. Then we rechecked the rotator center and it looks OK. See the log for the next night Com20090509 where the transformation was OK.

Finally we did some dice5 offsets just to see if the telescope and software were otherwise working OK.

Tried to go back to an earlier oacontrol.conf at the end, but the present oacserver program didn't like the one I chose. Lost the last 20 minutes fiddling with that.

Preparations

New TCS Build 26 has been installed along with new MCSPU software.

Telescope reconfigured to Left Front Bent Gregorian for observing with IRTC and AGW3. Right M2 swing arm is in for anemometer data, and DX M3 is installed on its swing arm.

Details

Open

19:15 Open and ventilating T=11 degC, D=-7 degC, wind 6-8 m/sec, sky photometric. 19:30 Problem with MCSPU (and MCS) complaining about -3000 sec time difference between linux and DSPs. Called Tom S. and Dan C., and Dan got us fixed up. Lost ~20 minutes.

19:48 pointto M5_1246 Tel="ACQUIRE" gs=0 PAM- - Star appears on Acquire image nearly in-focus and decently collimated. This is how it is supposed to work. 0.4 Hz guide rate makes the movie frequency better for initial alignment than 0.2 Hz. Adjust IE=-72 CA=-4 .

19:54 set guide rate to 1 Hz from command line.

19:48 pointto M5_1246 Tel="ACQUIRE" gs=0 PAM- - But update rate of the movie is still 0.4 Hz.

19:57 pointto BS9143 Tel="ACQUIRE" gs=0 PAM- Adjust IE=-63 CA=-5 .

Note that we did not get any spurious M2 offload vector on M1 after the ACQUIRE preset, so that bug is fixed.

20:01 pointto BS9143 Tel="ACTIVE" gs=0 PAM+ PA=0 - Pointing jumped by 5 arcsec between these two presets. ?? No collimation had yet been applied. To quote Wayne, "There was a glitch in somethin'." Seeing is 0.6 arcsec and much steadier than last night was.

20:07 pointto BS9143 Tel="ACQUIRE" gs=0 PAM- - Image has shifted 8 arcsec in EL compared to the ACQUIRE of 10 minutes ago. Adjust IE=-72 CA=-6

20:10 pointto BS9143 Tel="ACQUIRE" gs=0 PAM- - Image remains centered.

20:11 pointto BS9143 Tel="ACTIVE" gs=0 PAM+ PA=0 - Acquisition image is centered.

Closed loop guiding test

20:13 This marks the start of a closed loop guiding test in "split" mode. Starting with guiderimage0180. On-axis, so no IRTC image yet. This test is in the relatively unstable conditions at the end of twilight. Wind is reading 10 m/sec from West. We are just West of the merdian at EL=61 deg. Vent doors are open halfway. Guide cam exposure time is 2000 ms. WFS exposure time is 30000 ms. Effective guide rate is 0.36 Hz. Guide correction factor is 0.9.

20:19 Called Paul to ask why /lbt/tcs/current goes to /lbt/tcs/B25.centos4 and not /lbt/tcs/B26.centos4. He agrees that this doesn't seem correct. Chris thinks this is a feature.

20:30 Seeing worsens to 0.9 and image is jumping around more. The image motions definitely seem proportional to the seeing.

20:55 pointto BS9152 Tel="ACTIVE" gs=10 PAM+ PA=INDEF==33.02 - acquire image was about 5 arcsec off

21:02 stop Guiding

readGuideCam -e 1000 - now works without crashing GCS, since it doesn't try to select an alternate AGW.

21:09 pointto BS9152 Tel="ACTIVE" gs=10 PAM+ PA=INDEF==33.02 - acquire image was about 3 arcsec off. Seeing is 0.7 arcsec.

20:13 This marks the start of a closed loop guiding test in "split" mode. Starting with guiderimage 1170. Off-axis, so will have IRTC images. Wind is reading 10 m/sec from West. We are facing East at EL=61 deg. Vent doors are open halfway. Guide cam exposure time is 2000 ms. WFS exposure time is 30000 ms. Effective guide rate is 0.34 Hz. Guide correction factor is 0.9.

21:18 takeseq 0.5 100 - problem with GetFilter

21:19 takeseq 0.5 100 - irtc.20090508.041900.fits cube 0001

21:21 takeseq 0.5 1000 irtc.20090508.042054.fits cube 0002

WFS images are seeing moonlight on the wall of the enclosure past the tertiary mirror. See wfscimage 0093.

21:32 offset 0 5 (RADEC)

21:33 takeseq 0.5 1000 irtc.20090508.043300.fits cube 0003 (seeing has worsened to 1 arcsec)

21:35 hbsel temperatures for Pad T_Out are: 12.0, 11.6, 10.8, 11.8 while C-Ring steel temps are: 11.07, 11.05 and 11.23. HBS oil temperature setpoint is 1.23, and the reported oil return temperature is 0.79.

Open-loop tracking tests (no guiding, no collimation updates)

21:45 stop Guiding

21:46 takeseq 0.5 300 irtc....... cube 0004

21:51 focus 0 ---> irc StepFocus 0. M1 left

21:52 takeseq 0.5 300 irtc.20090508.045221.fits cube 0005

Closed-loop tracking tests

22:00 pointto BS9152 Tel="ACTIVE" gs=1 PAM+ PA=INDEF=229.72

22:02 offset -5 0 (RADEC)

22:03 takeseq 0.5 1000 irtc.20090508.050351.fits cube 0006 Seeing started about 0.9 arcsec.

22:24 takeseq 0.5 1000 irtc...... cube 0007

Running stepFocus updates at 60 sec intervals (plus 15 sec to do the stepFocus) with /home/lbcobs/supportscripts/ObserverSupport/stepfocus.py in addition to the regular WFS updates.

22:37 pointto BS9152 Tel="TRACK" gs=0 PAM+ PA=0

Open-loop tracking tests (no guiding, periodic collimation updates, no active optics)

22:39 takeseq 0.5 1000 irtc.20090508.053915.fits cube 0008

Running stepFocus updates at 60 sec intervals (plus 5-6 sec to do the stepFocus) with /home/lbcobs/supportscripts/ObserverSupport/stepfocus.py

Field Aberration Measurements (with GCS)

22:50 pointto BS9152 Tel="ACTIVE" gs=0 PAM+ PA=0

22:54 pointto BS9152 Tel="ACQUIRE" gs=0 PAM- Adjust IE=-86 CA=+9

Seeing is 0.9 to 1.4 arcsec - not ideal for this measurement. This small set is a sanity check to make sure we haven't accidentally broken the GCS WFSing algorithms.

Star 17 is probably near the limit for this level of seeing in 2 sec exposures. Star 18 is lost in the moonlit sky.

BS9152
GS PA WFSC
0 0 152 - 154
17 124.77 155 - 157
0 124.77 158 - 161
18 65.44 162 - 166
0 65.44 167-169
7 10.54 170-174
0 10.54 175- 178
4 338.61 179- 182
0 338.61 183- 185
1 229.72 186-189
0 229.72 190- 192
9 168.12 193- 196
0 168.12 197- 200

Comparing GCS and IDL at off-axis field points

23:48 setGuideRate 0.2

23:49 pointto BS9152 Tel="ACTIVE" gs=17 PAM+ PA=90 x=103.33 y=470.24

Of course, the seeing improves to 0.6 arcsec as soon as the field aberration measurements are finished!

00:02 Send Z9=+300, and this crashed PSFL.

00:03 Restart PSFL and send another Z9=+300.

00:09 pointto BS9152 Tel="ACTIVE" gs=17 PAM+ PA=180 wfsc 0227 (the first one after moving the probe put in a fairly large coma correction)

00:15 Send Z9=+300

Moving to Stone L

00:22 pointto StoneLagw Tel="ACTIVE" gs=31 PAM+ PA=0 R=16

Closed-loop tracking test

(something to do while waiting for the seeing to improve)

00:31 takeseq 1.5 1000 irtc.20090508.073229.fits cube 0009

01:05 set guideCorrectionFactor 0.5 in AIP_IRTC_L.cfg (was previously 0.5), but it doesn't take until you selectAGW.

01:06 pointto StoneLagw Tel="ACTIVE" gs=31 PAM+ PA=0 R=16

01:08 takeseq 1.5 1000 irtc.20090508.080837.fits cube 0010 (still with guide factor 0.9)

01:37 selectAGW AIP_IRTC_L.cfg

01:38 pointto StoneLagw Tel="ACTIVE" gs=31 PAM+ PA=0 R=16

01:40 takeseq 1.5 irtc.20090508.083948.fits cube 0011 (this one with guide factor 0.5)

01:46 Turn off mirror ventilation. Command becomes Ref -0.9 rather than the advertised Glass - 0.9.

01:52 Turn mirror ventilation back on.

02:00 Mirror ventilation PID constants to -0.9, 1.25, 0.3 (adding the Ki term 0.3). This produced a drop of 0.3 in the Command at every update.

02:02 Ki back to zero.

02:04 Ki=0.1, This produced a drop of 0.1 in the command at every update, because the value reported in error is -1. I think that is not correct and that the "offset" should not be included in the error term.

02:06 Ki back to zero.

02:10 Authorize as LUCIFER. GCS stopped guiding as expected. IRAF: authorize LUCIFER

02:11 selectAGW AIP_L.cfg - This failed, and we tried a couple times.

May  8 02:15:46 lbtmu107 LBT_left_GCS: ~azcam: done.
May  8 02:15:46 lbtmu107 LBT_left_GCS: creating _azcam instance
May  8 02:15:46 lbtmu107 LBT_left_GCS: ImageServerThread created. instance = 3
May  8 02:15:46 lbtmu107 LBT_left_GCS: setting up imageserver IP
May  8 02:15:46 lbtmu107 LBT_left_GCS: init and connect ...
May  8 02:15:46 lbtmu107 LBT_left_GCS: TCPSocket::recv error: Resource temporarily unavailable
May  8 02:15:46 lbtmu107 last message repeated 8 times
May  8 02:15:46 lbtmu107 LBT_left_GCS: AzCam: can't determine detector size. Assuming 512x512 pixels.
May  8 02:15:46 lbtmu107 LBT_left_GCS: TCPSocket::recv error: Resource temporarily unavailable
May  8 02:15:46 lbtmu107 LBT_left_GCS: AzCam:openSocket(): bad command status for command! SetSocket 1 10.144.0.107 30003
May  8 02:15:46 lbtmu107 LBT_left_GCS: TCPSocket::recv error: Resource temporarily unavailable
May  8 02:15:46 lbtmu107 LBT_left_GCS: AzCam: faulty parameters while trying to set ROI.
May  8 02:15:46 lbtmu107 LBT_left_GCS: AGWCamera::initAndConnect(): couldn't connect to AzCamserver 10.100.1.50:2402
May  8 02:15:46 lbtmu107 LBT_left_GCS: GCS Subsystem COULD NOT change active AGW configuration to AIP_L.cfg
May  8 02:15:46 lbtmu107 LBT_left_GCS: GCSL:setAGW execution time: 4252 ms

02:16 stopAGW -u 0, startAGW -u 0, home motors.

02:20 selectAGW AIP_L.cfg - still fails in the same way

02:21 stop and restart GCS, and now it is willing to selectAGW AIP_L.cfg

02:22 selectAGW AIP_IRTC_L.cfg
May  8 02:22:20 lbtmu107 LBT_left_GCS: GCS tries to change active AGW configuration to AIP_IRTC_L.cfg
May  8 02:22:20 lbtmu107 LBT_left_GCS: [D] <429> [TCSConfigFile] Empty filename: ''.
May  8 02:22:20 lbtmu107 LBT_left_GCS: selectAGW(AIP_IRTC_L.cfg), type: AIP_LUCIFER, unit: 0
May  8 02:22:20 lbtmu107 LBT_left_GCS: didn't create a new instance of AGWUnitAIP as the types and units are the same
May  8 02:22:20 lbtmu107 LBT_left_GCS: GCS Subsystem changed active AGW configuration to AIP_IRTC_L.cfg

02:25 pointto M5_0656 Tel="ACQUIRE" gs=0 PAM- PA=0 - Star was a little bit outside the acquire field. Adjust *IE=-72 CA=-10 * (Probably these values were fooled by running into the Y-limit.)

Dice5 offset tests that turned into an exercise in mis-pointing.

02:32 pointto StoneLagw Tel="ACTIVE" gs=36 PAM+ PA=0 - did we miss the star and pick up a faint one?

02:38 pointto StoneLagw Tel="ACQUIRE" gs=0 PAM- PA=0 - off hotspot by 6 arcsec. Adjust IE=-78 CA=-14

02:40 pointto StoneLagw Tel="ACTIVE" gs=36 PAM+ PA=0 - failed preset - no star at all - not even a faint one.

But I needed a failed preset to test M2 offload. M2 offload is -0.343 +2.704 -0.103 28.92 3.62

02:42 pointto StoneLagw Tel="ACTIVE" gs=10 PAM+ PA=0 - again no star. M1 is in +Y limit.

M2 offload changed to -0.352 2.710 -0.112 29.021 3.60 (small but non-zero values; I was expecting zero)

02:45 Add Yglobal =-1 to both M1 and M2.

02:46 pointto StoneLagw Tel="ACQUIRE" gs=0 PAM- PA=0 - Adjust IE=-73 CA=-9

02:48 pointto StoneLagw Tel="ACTIVE" gs=36 PAM+ PA=0 - still don't find a star.

02:50 pointto StoneLagw Tel="ACTIVE" gs=10 PAM+ PA=0 - no star found

02:52 pointto M5_1469 Tel="ACQUIRE" gs=0 PAM- PA=0 - Adjust IE=-105 CA=-19.

02:56 pointto StoneLagw Tel="ACTIVE" gs=36 PAM+ PA=0 - Seeing is now 0.7 arcsec at EL=50

Move to StoneM field

03:05 pointto StoneMagw Tel="ACTIVE" gs=0 PAM+ PA=0 - Seeing on guider is now 0.5 arcsec!

Using sequences 0.5 x 50 for the center star of Stone M.

03:05 pointto StoneMagw Tel="ACTIVE" PAM+ PA=0 gs=39 irtc.20090508.100833.fits cube 0012 getdat0

03:09 gs=18 irtc.20090508.101112.fits cube 0013 getdat0

03:12 gs=20 irtc.20090508.101341.fits cube 0014 getdat0

03:14 gs=26 irtc.20090508.101621.fits cube 0015 getdat0

03:18 gs=40 irtc.20090508.101843.fits cube 0016 getdat0

It is clear from inspecting the IRTC images that the transformation is off by arcsec!

03:23 irtc.20090508.102254.fits cube 017 -nothing

Interrupt the transform data taking to verify the positions of the rotator centers.

03:27 irtc.20090508.102707.fits cube 0018 Rotator going from 360 to 0 deg to verify rotator center on IRTC. (stack the cube to make a circle)

03:29 setxy -x 0 -y 612.5 -u 0

/Repository/AGW_Data/090508/guider000005.fits is a beautiful ring. And the eyeball measurement shows it is quite close to the hotspot.

03:39 gs=51 irtc.20090508.104149.fits cube 0019 getdat0

03:43 gs=72 irtc.20090508.104436.fits cube 0020 getdat0

03:45 gs=68 irtc.20090508.104834.fits cube 0021 getdat0 (twice)

03:49 gs=69 irtc.20090508.105241.fits cube 0022 getdat0

LEFTZEROPOINT -187.703 (LUCIFER is same as IRTCBGF)
LEFTSCALE 1.0

03:54 gs=64 irtc.20090508.105450.fits cube 0023 getdat0

03:55 gs=58 irtc.20090508.105651.fits cube 0024

03:58 gs=52 irtc.20090508.105954.fits cube 0025 getdat0

04:00 gs=48 irtc.20090508.110151.fits cube 0026 getdat0

04:02 gs=44 irtc.20090508.110355.fits cube 0027 getdat0

04:xx gs=43 irtc.20090508.110625.fits cube 0028 getdat0

oacontrol.conf has the same AGW1 (AGW3) transform as the one named oacontrol.conf.20090429a, but I can't tell if that ws the last one verified on-sky. It was modified at 00:00 UT on 20090430.

04:xx gs=37 irtc.20090508.110857.fits cube 0029 getdat0

04:10 gs=35 irtc.20090508.111231.fits cube 0030 getdat0

Offset Tests with Dice5 (to see if field orientation has changed)

04:15 dice5 off1=6.5 off2=5. coord ="DETXY" exptime=0.4 Nexpo=80 loopN=2

04:27 dice5 off1=6.5 off2=5. coord ="RADEC" exptime=0.4 Nexpo=80 loopN=2 (was too close to bottom of IRTC so is only a dice3) Goes through cube 0049.

Try to rollback oacontrol.conf

cp -p oacontrol.conf.20090427b oacontrol.conf

04:42 Stop and restart oacserver, and stop and restart AGW3, Stop and restart GCS.

But startAGW gives an RPC error

Close

04:55 Close dome for on-coming twilight. T=9 degC. Photometric

4:56 cp -p oacontrol.conf.20090506b oacontrol.conf

04:57 kill and restart oacserver, now startAGW -u 0 works OK.

05:00 Tried to repeat the selectAGW problem we had earlier, but it behaved properly this time.

Approximate List of Observing Activities

  • Test TCS Build 26 for presets, offsets, active optics, and M2 offload behaviour

  • Another set of AGW Transform data, especially if the seeing is good.

  • Collect a modest set of field aberration data to verify that all the corrections are OK. (data collected 06:30 UT on May 8)

  • Guiding tests at different guidestar magnitudes

  • Guiding tests with different values of the guide filter factor

  • Work on tuning the exposure time scaling algorithm parameters for GCS.

  • Collect data for open-loop tracking performance (done 05:39 UT on May 8 UT)

  • Collect data for closed-loop tracking performance (done 04:44 UT and later on May 8 UT)

  • Collect additional IRTC vibration data

  • Offset testing with dice-5 pattern at X not equal zero

  • Troubleshoot the 68um/66um transform error at PA=240
    Tell me more is there an issuetrak? What is the problem?

Details about AGW Transform data

When you have the opportunity to do further guide probe transform measurements on sky it will be important to start to look into the effect of rotation of the instrument (different PA's) and elevation etc. The best transform at BG has now an RMS of less than 40microns so we should now be able to investigate the next level of effects. When using such a transform we do not quite reproduce the performance on the IRTC so there are probably some more subtle effects which should be identified and be addressed to improve the performance.

Some possible issues:

* Are we limited by the precision of the location of the rotation axis center?
* Do we do the differential diffraction correctly?
* Is the rotator accurate enough?

To better understand the limitations of the system we can use either BG or DG whatever is available.

Please consider the following tests (at least 25 stars per PA):

* Do at more or less the same (high) elevation three different PA's.

* Do the same field at several different elevations.

* At DG do a large set similar to Stone-M at one PA. This gives a first transform which we can then verify with follow-up observations.

-- JohnHill - 07 May 2009
Topic revision: r8 - 09 May 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