Restart Night #7 UT 2018-09-08

OSA: D.G. Huerta (LBTO), J. Williiams (Tucson)

Observers/ISA: B. Rothberg, J. Hill (Tucson)

Software: S. Hooper/P. Kubanek (on-call)

ISp: J. Morris

Mtn Manager: R. Hansen

Summary of the state of AGW4 collimation

Doug used last night's data to get well refined IDL S-H parameters for AGw4@LFBG collimation. IDL now converges robustly, but there are a bunch of wrinkles and caveats.
  1. while IDL converges nicely and the grid of wfs spots looks good, the FWHM of the SX guider is still 1.3 arcsec when the DX guider is 0.7 arcsec. Because:
    1. The AGw4 guider is 0.1 mm (M2 motion) or 1.1 mm (at focal plane) out of focus compared to the AGw4 wfs.
    2. SX AdSec has some wrinkles in the flat that are giving us some fake mid-scale aberrations. The
  2. the translation of IDL S-H configuration variables is a non-trivial exercise. We think we've powered through it to the extent that IDL and GCS get the same wavefront correction at the 40 nm rms level. Issues are:
    1. In both IDL and GCS, "On-Axis Pupil" is determined by WFScam_hotspot_x,y. That's fine.
    2. In IDL you adjust "Pupil Center" with lenslet edge (whose actual name is "agw_params[i].wfsc.lenslet_x_center"), but in GCS you adjust "Pupil Center" with hotspot_sext_x. Both of those are at least partially misleading. Note that IDL doesn't even have a hotspot_sext_x parameter, and in GCS WFScam_lenslet_edge_x,y is really the center lenslet.
    3. In GCS if hotspot_sext_x,y is close to correct, then "Pupil Center" is the fit value, but if you are far away GCS returns the hotspot_sext_x,y parameter values. Is that a feature?
    4. See IT 7388 about strange problems with "WFScam_lenslet_edge_x,y" in GCS.
    5. We still need Petr next week to make some additional GCS diagnostics to sort out these issues.
  3. The SX guiding hotspot is elevation dependent in the sense that the guiding sweetspot shifts by 0.6 arcsec between airmass 1.7 and airmass 1.1. from the following night: No, it isn't elevation dependent. The problem is a bit of trefoil in the slightly defocussed guide star. This 3-lobed image allows the guider to jump around between 3 different positions about 1 arcsec apart. That jumping shoves the spots into the edges of the lenslets which makes more trefoil from wfs non-linearity. Then the guider can jump more and things start to fall apart.
  4. GCS still has runaway problems that IDL does not have when the wavefront is poor at the start. They have good agreement when the wavefront is good, so maybe we still don't have the GCS lenslet positions quite right so that runaway happens when the spot slips over into the neighboring lenslet box.

Details

00:15 restart OSS to read new values for DX M3 in positions.conf, and remove my global offsets from PSF DX M3.

01:37 Sky is mostly clear with just a few puffy clouds off to the east. T=11 degC 63%. Wind light.

01:45 JMH added the field aberrations to DX manually, but has not made them permanent yet. See last night's wiki log.

UT 01:50 - Powering up LBCs, LUCI up - LUCI test darks taken

LUCI
0001-0002 2.51 sec Dark

0003 3 sec dark

0004 5 sec dark

0005 10 sec dark

0006 5 sec dark

UT 01:5r3 - LBC Blue Trackers failed again, restarting (rm power.conf)

Open

UT 02:05 - telescope OPEN - pointing check - will go to Stone L1 (Stone_L3a)

UT 02:10 - configuring LUCI N30FieldStop, FeII 5sec, N375 Camera

UT 02:17 - still trying to power up LBCs

UT 02:23 - LBCs still having blue tracker issues

UT 02:32 - sending acquire preset to field (all in one terminal) - acquire is needed to center up, might want to run active if need to collimate

IDL> iif_register,/TCS
IDL> iif_preset,STONE_L1,mode=acquire,gs=0,pa=-135,/right
IDL> iif_preset,STONE_L1,mode=acquire,gs=0,pa=-135,/left
IDL> iif_binoc_preset

UT 02:37 - starting AGW4/2 Transform collection Stone_L1.list

UT 02:39 - left blind filter in, now clear, FeII 5 sec dit

UT 02:40 - restarting Transform collect (dark/off exposures already taken above)

UT 02:52 - SX failed to find GS=156, resent transform_collect,start=4 to redo

UT 02:53 - elevation stopped moving - DX failed too - troubleshooting = HBS cooling had failed so oil reservoir was way overtemp.

UT 02:59 - D. Miller updated IDL collimation parameters on LFBG AGW4 from last nights field data, J. Hill updating GCS collimation parameters in LUCI_L.cfg using the straighforward conversion. Doug says that GCS and IDL are getting the same answers when they have the same pupil center parameters for on-axis stars, but that GCS is not using the shifted value which it computes for the pupil center for off-axis stars (even though the computed shift is correct).

UT 03:08 - resending ACQUIRE preset to field

UT 03:10 - after waiting 50min and ignoring the LBCs (due to AGW transform fun) they powered on OK - taking 10 biases

UT 03:30 Tried another preset, but oil is still to hot for EL axis to move.

UT 04:15 - trying another preset back to Stone L

UT 04:16 - resending as an ACTIVE preset so we can collimate (after 45 min collimation iffy and D. Miller updated SX collimation parameters for IDL)

UT 04:23 - still collimating

UT 04:35 - starting Transform collect Stone L1

UT 04:37 - forgot to take blind filter out

UT 04:38 - starting Transform collection Stone L1,

UT 04:39 - stopping (each side picked a different star) and running acquire preset - DGH centering target with IE/CA

First Images - exposure times are only 2.51 seconds despite GUI requesting 5 sec, but counts in star are 2900 bkgd ~50 counts

LUCI2.20180908.0019 is first image

UT 04:57 - stopping SX so D. Miller and J.Hill can work on collimation, Letting DX continue.

John learned that you must NOT run IDL active optics from the idl source code directory because Doug has special configuration files for analyzing historical data.

UT 04:58- sent acquire (oops), but then active (yay!) preset to left for SX AGW4 work

UT 05:11 - image quality on LUCI-2 has degraded severely, although GCSR shows images are <1"

UT 05:17 - but now ok.... ??

Adjusting SX Guiding Hotspot

UT 05:27 Initial SX guiding hotspot was 273,247 pixels

UT 05:28 - John is adjusting the guiding hotspot in order to zero out the "X,Y Average Removed = 0.13 -0.35". It turns out to be nearly impossible to adjust the hotspot from the GCSGUI or the command line (gcsclient) as the WFS loop freezes once the hotspot gets 0.5 arcsec away from its initial value (the trick would be to send a microdither to cause the guide probe position to update). We editted the new hotspot into LUCI_L.cfg and then restarted GCS to read the value (it would be a big improvement if GCS read this value on new presets). A -20 pixel (1 arcsec) shift in hotspot Y increased the value of "Y Average Removed" by 0.5 (units of 0.6 lenslets).

UT 05:46 - DX completed AGW2 transform, files are luci2.20180908.0019-0056

UT 05:53 We adopted the new SX hotspot of 173,237 although we had reached an airmass of 1.7.

UT 06:16 - moved to Stone O - better airmass = 1.1 - when we arrived at Stone O the SX guide image was defocussed by 3 arcsec. Why?

06:24 The "X,Y Average Removed = +0.12 and +0.41". Why is the hotspot elevation dependent? Compare DX? DX does not seem to have this hotspot motion.

06:28 Put the SX guider hotspot back to 273,247.

LUCI field center rotation test AND focus tests

UT 06:30 - moving to M52 for field center rotation test PA=0

UT 06:34 - ignore luci1.20180908.0018, luci2.20180908.0057 - they are out of focus, script was track mode without any active optics

UT 06:48 - going to M52 for field center rotation test PA=0

UT 07:02 - SX having issues again, also DX collimation finicky, go to collimation star nearby then track mode to target and PSF is out of focus, currently sent active preset to M52 region, collimating, then will send track mode to send spot and take data

focussing interlude

07:08 Collimating SX with IDL

UT 07:14 - preset timed out, trying to assess why SX ran away, cleared active optics 3x or more, helped improve it, running collimation via IDL which is doing better

UT 07:17 - hypothesis: flat on SX is bad - and LUCI-1 is out of focus (as GS image improves, image on LUCI-1 gets worse)

UT 07:21 - taking lots of LUCI images to test this (using GEIRS interface so we don't have gigs of useless data)

UT 07:28 - seem to converge that this is not a LUCI1 focus issue - moving secondary mirror improved focus for both LUCI-1 and AGW4. On the SX side the guider is 0.1 mm (of M2 motion) above the focus of the wfs, but LUCI1 matches the AGw4 wfs within +-0.02 mm (of M2 motion).

UT 07:36 - moved M2 0.04mm on DX secondary to improve DX focus - since that is alot, we continue with making a focus curve using the AGw2 z-stage. Fortunately, the seeing was decently stable during our tests.

  • starting with DX stage at Z=18.141 mm (on-axis focus parameter of 25.2 mm)
  • start focus=18.141 image 0085 FWHM 6.0,6.7 pixels (x and y FWHM respectively)
  • 07:39 focus=19.141 image 0086 FWHM 6.2,5.2
  • 07:41 focus=21.141 image 0087 4.3,4.8
  • 07:43 focus=15.141 image 0088 7.9,6.3
  • 07:47 focus=24.241 image 0089 4.1,3.5
  • 07:52 focus=27.141 image 0090 4.7,3.8
  • 07:56 focus=23.141 iamge 0091 4.6,4.3 ADOPTED although seeing a bit worse at this time.

UT 8:03 We increased "focus" in LUCI_R.cfg from 25.2 to 30.0 mm for LUCI2-AGW2.

back to rotation center test

UT 08:04 - resending rotation center test M52 PA=0 Track

luci1.20180908.0046 on / 047 off

luci2.2080908.0092 on / 093 off

UT 08:07 - M52 PA=180

luci1.20180908.0048 on / 049 off

luci2.20180908.0094 on / 095 off

UT 08:11 - M52 PA=90 - dark cloud rolled in - data may be affected

luci1.20180908.0050 on / 051 off

luci2.20180908.0096 on / 097 off

UT 08:17 - M52 PA=270 - aborted due to clouds

Close

UT 08:18 - closing due to dark clouds overhead - those nasty clouds must have snuck in from New Mexico, since all of Arizona is clear.

UT 08:53 - reopening

UT 09:01 - closed again, clouds are widely spaced but fairly thick and the humidity is 70%.
Open

UT 09:45 - reopening

UT 09:48 - Field Aberration Data collection for SX and DX - BS9107 field Seeing is 0.6-0.7 arcsec FWHM. Sky is mostly clear but not photometric.

UT 10:00 - collimation blew up, J.Hill using IDL for SX, DGH using GCS for DX. Had to do DX with IDL also.

UT 10:07 - still collimating, also remade BS9107 list to make sure GSs were reachable for LUCI - new list is BS9107_newLUCI_field.list (13 stars)

UT 10:11 - unsure if perhaps this is a hexapod problem. Homing the DX hexapod.

UT 10:13 - could it be fallout from the hot HBS oil issue? C-ring heated up and distorted? That would make coma.

UT 10:14 - homing hexapod did not help

UT 10:25 - completed collimation as best as possible (DX is 150 nm rms, SX is 80 nm rms) - started field data collection

UT 10:34 - data collection stopped, GS16 is too bright (10th mag) - removed it and restarted

UT 11:21 - SX and DX failed to find star for GS 42 - collected 11/12 objects

UT 11:23 - Reconfiguring for LBCs - attempting to get NS test and skyflats

UT 11:47 - running DOHYBRID

UT 12:01 - running co-pointing script

UT 12:04 - sending hijacked preset for 2017 WV13 using Horizon - changed exptimes to 60sec because its -12 twilight

UT 12:07 - images look nice and round - looks like NS guiding not working

UT 12:10 - background getting brighter - g images beyond #2 might not be useful (#4 and #5 saturated), r-sloan #1-3 counts < 11k, #4 23k counts, #5 32k counts

UT 12:14 - switching to skyflats V, V

UT 12:19 - V-Bessel, V-Bessel PA=0 (scale=3) (6k, 4k first images, last 10k, 7k)

UT 12:22 - V-Bessel, V-Bessel PA=180 (scale=1) (first images, 6k, 4k, last 10k, 8k)

UT 12:28 - r-sloan, r-sloan PA=180 (scale=1) (first images 12k, 14k, last 24k, 30k)

UT 12:33 - All Done - Ready to Close, powering off LBCs

-- %USERSIG{JohnHill - 2018-09-08}%
Topic revision: r13 - 09 Sep 2018, 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