MODS2 Commissioning Night 6 - 20140505 UT

Observers: O. Kuhn (LBTO), J. Hill, D. Miller, R. Pogge, M. Wagner (LBTO, Remote from Tucson)
Telescope Operator: S. Allanson (LBTO)
AO Support: D. Miller (LBTO)
Telescope Support: John Little et al. (LBTO)
Software Support: T. Golota (Remote from Tucson)
Instrument Support: J. Power (LBTO)


MODS2 Active Optics did NOT converge correctly at the beginning of the night. We were not able to resolve this problem, although it worked correctly for most of the night.

We solved the sextractor configuration problem that was making MODS2 Active Optics have jittery corrections for trefoil and spherical.

We took some additional zernike calibration data.

Olga took some QSO spectra with MODS1 with MODS2 following along in pseudomonocular mode.

We took binocular (asynchronous) field aberration data on both sides. This same data will allow us to measure the pupil wobble.


02:35 ACTIVE preset on-axis - open loop collimation was good at EL=84. Loop is still struggling a bit with coma (Z8) at the beginning, and then driving unneccesary Z11. Convergence was not good enough after 10 iterations.

Clear Active Optics

02:48 ACTIVE preset on-axis at EL=60 - It almost converged here after 4 iterations, but then it sent and incorrect Z8 correction and made a very comatic image. Then the coma started cross-talking into Z11.

02:54 Offset 30 0 RADEC

Clear Active Optics

02:58 ACTIVE preset on-axis with PA=+90 starting with wfsc 0028 - no problem converging here, but just when the image is looking good on iteration 4, it starts walking away in coma (Z8) at 3.04 UT. Once it has the wrong coma, then Z11 starts increasing.

Why is wfsc 0037 at 03:07:36 UT sending a large positve Z8 correction?

collimate_process_file, "right_wfscimage000037.fits", /fwhm, /all_spots

Clear Active Optics

03:11 ACTIVE preset on-axis with PA=+180 starting with wfsc 0045. Correctiing Focus and Astigmatism makes a 0.72 arcsec FWHM image, then the coma correction starts and makes things worse starting at 3.25 UT.

03:19 resend preset after Olga put in cal unit.

03:20 These changes to MODS_R.cfg
WFS_fwhmSigmaFactor     2.0             float   # used to calculate the cut off value for FWHM. FWHM_Mean + Factor * FWHM_StdDev
WFS_ellipSigmaFactor    2.0             float   # used to calculate the cut off value for Ellipticity. Ellipticity_Mean + Factor * Ellipticity_StdDev
WFS_minSextSpots        40              int     # minimum number of subapertures found with Sextractor before falling back to modified CoW algorithm

03:24 now things are looking mostly OK, but still with the chronic bouncing in trefoil. Is the problem caused by twilight background light on the wfs? That's our working supposition for why the problem has gone away.

03:30 ACTIVE preset off-axis with r=15.73 mag (reads 17.8 on gplot guide but not acquisition) - This seems happy enough with active corrections.

03:34 Clear Active Optics

IDLs sextractor is regularly missing the mark on 1-3 spots e.g. wfsc 0054

03:35 resend preset - the astigmatism converged with no problem. It corrected the coma correctly.

03:42 ACTIVE preset (from MODS2) with r=16.2 mag guide star, too faint the acquisition missed, maybe some cirrus or maybe wrong magnitude.

03:49 ACTIVE preset (from MODS2) with r=15.68

Field Aberration Data


04:04 field_collect on BS9152

04:11 field_collect, limit=400 on BS9152 - seeing is running 0.8 - 1 arcsec FWHM on guider/wfs

This data is not going to be very good. Collimation is struggling too much with coma. e.g. wfsc 0162

05:08 gs=14 is outside the range - it was the last one.

05:28 preset to lower elevation EL=40 - still about 1 arcsec FWHM on guider

06:04 Testing XPEAK, YPEAK rather than X_IMAGE, Y_IMAGE in sextract_wfsc.param; stop and restart GCSR.

06:06 ACTIVE preset on-axis

XPEAK seems to have made no difference in the behaviour of wrong measurements of the spot peaks. I'm using /home/lbcobs/supportscripts/ObserverSupport/ -b 0.1 -s DX to plot spot positions over time. I've also verified that GCS is finding the same variation of cross vs spot that IDL is plotting - which is expected since IDL-sext and GCS are getting the same answer.

06:18 Put sextract_wfsc.param back to normal; stop and restart GCS. Rest and reflatten the AdSec.

06:21 ACTIVE preset on-axis - seeing is varying up to 1.4 arcsec, but does not seem to influence the active optics results.

06:27 Clear Active Optics and resend preset.

Binocular MODS tests in pseudomon mode

06:34 Authorize both MODS, with pseudomono on left. Olga is taking QSO spectra with MODS1, while Doug and John continue discussing MODS2 active optics. MODS2 isn't yet fully configured to take spectra.

06:35 Doug turns off background subtraction in, but that does not seem to have solved the wrong spot measurements.


What could be wrong?
  • some spot peak measurements are wrong - this makes the jittery trefoil. It's not obvious that it causes the poor initial convergence.
  • some problems are related to pupil wobble and the resulting incorrect center location wrt/ the reconstructor
  • overlap of lenslet aperture images is something different between this wfs and others, but they only overlap by ~1 pixel
  • blurring at the very edge of the pupil could give misleading centroids at those locations
  • reconstructor was calculated for larger pupil, and we are using a central cutout. We think the scale should be the same.
  • maybe 13.2 pixels in not the correct lenslet separation, or separation differs in X vs Y? (Doug measured that it is symmetric.)

07:00 We note that the image on the right guider is better than the image on the left guider. Maybe DX active optics is not as broken as we thought? But, the actual reason is that the MODS1 guide camera is slightly out of focus.

07:05 Blocked guiding and wfs on right side. This makes the right side guide faster because it isn't waiting for the corrections to be applied. Take note that most of the larger guider corrections are coherent between the two sides. Does that mean they come from the mount, or could they still be atmospheric?

07:30 Right side guiding back on

Some clouds are rolling through image wfsc 0348 shows good spot measurements even with quite low signal to noise.

08:00 Clouds have stopped both QSO spectra and wavefront sensing, although Doug and John are analyzing offline data.


We've found the sextractor parameter that reduces the errors in measuring the spot positions:

BACK_SIZE has been changed from 10 to 4 for both MODS1 and MODS2 source extractor parameters.

08:25 ACTIVE preset to hole low in southwest trying to test this.

08:41 BACK_SIZE changed back to 10 for comparison on both sides

08:50 DX pupil is looking fairly twisted

08:51 BACK_SIZE changed back to 4 for comparison

08:57 Clear Active Optics on the right side

09:14 ACTIVE preset to a higher EL as the cap cloud has backed off a bit.

09:22 right side is mostly converged, but suffering from a bit of Z8 which is crosstalking into spherical. The crosstalk happens because of vignetting in the outer subaps.

09:25 right side is now running away in Z8 and Z11

09:26 Clear Active Optics on DX M1 only

Field aberration data on BS9170 (and pupil wobble measurement)

Seeing has improved to ~1 arcsec, so we'll try for a less jittery set of field aberration data. We still have clouds in the north, but southwest is clear.

/home/lbto/data/20140505/field_collect2/ on DX
/home/lbto/data/20140505/field_collect2sx/ on SX

09:33 ACTIVE preset to BS9170

uncheck pseudomono

09:39 field_collect, limit=400 on BS9170

Bug: Doug's IDL code is sending ClearNonSidereal to side both even when the preset is single sided:

Mon May 5 09:41:09.179 2014 both UpdateNonSiderealTarget command rejected for MODS because left Preset busy

That's stops us from taking field data on both sides together. This is because he was sending irc ClearNonSidereal CLEARIIF rather than irc ClearNonSidereal.

09:55 field_collect running seperately on both sides from 2 IDL windows.

The bad news: The right side is really struggling with Z11, so the field data is likely useless.
The good news: We've reproduced the convergence problem independent of twilight which is still an hour away.
More good news: The field data also provides a measure of pupil wobble which is likely the cause of the Z8/Z11 issues.

Couldn't reach gs=16, 17, 19 on either side, so removed them from the list.

DX Field Results (20140508)

No changes to the M2 positions made, nor needed.

  • DX Needed corrections from field data:

  • Focus measurement of field data:

  • 0 Astigmatism measurement of field data:

  • 45 Astigmatism measurement of field data:

M2 Astigmatism LUT

10:23 Doug removes Z6 LUT from SX M2.

10:33 SX grabs the wrong star at gs=11

10:45 Propagate the SX M1 lookup table to all the focal stations that use the AdSec

-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMLUCIFERAWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMPEPSIPOLWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMPEPSIPFUWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMMODSWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMLUCIFERWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMLINCWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMLBTIWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMIRTCFWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMIRTCDWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMIRTCCWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMIRTCBWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 Apr 18 03:24 SXPMIRTCAWavefront.dat

Beware that these need to be undone when Rigid Secondary is installed.

10:54 switched LUT on the Right from M2 to M1

11:04 Propagate the DX M1 lookup table to all the focal stations that use the AdSec

-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMPEPSIPOLWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMPEPSIPFUWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMLUCIFERWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMLUCIFERAWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMLINCWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMLBTIWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMIRTCFWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMIRTCDWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMIRTCCWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMIRTCBWavefront.dat
-rw-r--r--. 1 telescope domain users 4317 May  5 10:45 DXPMIRTCAWavefront.dat
-rw-r--r--. 1 telescope domain users 4314 May  5 10:45 DXPMMODSWavefront.dat

11:24 The right side active optics are completely well behave for the last 25 minutes. Is this because the seeing got better and the SH spots are no longer being clipped on the edge of the pinhole?

We've realized that the M1 LUT is not updating except when an active optics update is issued. It needs to update at the Optics TT cadence. See Issue 5169.

Collimation Convergence in Twilight

11:46 Clear Active Optics on both sides

11:48 ACTIVE preset on-axis - SX converged, but DX did not.


12:00 Close for impending dawn

-- %USERSIG{JohnHill - 2014-05-05}%
I Attachment Action Size Date Who Comment
astug_0.jpgjpg astug_0.jpg manage 42 K 08 May 2014 - 21:26 DougMiller 0 Astigmatism measurement of field data
astug_45.jpgjpg astug_45.jpg manage 51 K 08 May 2014 - 21:26 DougMiller 45 Astigmatism measurement of field data
focus.jpgjpg focus.jpg manage 49 K 08 May 2014 - 21:25 DougMiller Focus measurement of field data
rdg_field_results.jpgjpg rdg_field_results.jpg manage 89 K 08 May 2014 - 21:25 DougMiller DX Needed corrections from field data
Topic revision: r6 - 08 May 2014, DougMiller
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