MODS1 Commissioning Run 6: UTC 2011 March 20
Observer(s): R. Pogge (OSU), P. Garnavich (ND), and M. Pedani (LBTO)
Telescope Operator: S. Allanson (LBTO)
AO Support: A. Rakich remote from Tucson
Telescope Support: M. Midkiff (LBTO)
Instrument Support: D. Officer (LBTO)
Observed all night. Confirmed that the change made to the AGw stage autofocus algorithm reduces the Z4 residual to the level that we
are now down under 100nm rms. This means the active optics is working at a level that those science commissioning tasks related
to image quality can proceed. Further refinement is outside the general purview of MODS instrument commissioning per se, but should
be pursued as part of the overall efforts to improve image quality at the LBT.
Exercised the MOS mask alignment software on the Stone F field and gathered very useful data for refining the program. Still issues as
with any prototype, but getting there.
Bright moonlight, occasional bands of high clouds, and winds were issues for the run. The winds were not strong enough to force us to close, but
buffeting of the telescope when pointing towards the wind was sufficient to keep us out of the SW for most of the night. Some tasks had to
be aborted when torques on the elevation drives moved out of operating range because of the wind. Oil blowing around in the chamber when
facing into the wind is also an issues, and we have found oil spots on the MODS enclosure and the floors are slippery in spots.
Bright moonlight plus clouds led to high backgrounds that repeatedly fuddled the subframe guiding on the GCS. The GCS clearly needs work to be able to function under such conditions. The problems of losing the guide star and mis-identifying bright background features as "stars" cause us to stop
observing at some times, and had serious impact on some of the active optics measurements we were doing.
Given the bright background conditions, we worked primarily in imaging mode. Wind-imposed pointing restrictions led us to run a couple of long sequences of tracking accuracy and differential flexure tests on stars. These test complete what we need (we hope) to measure that effect in our data, including a long mock observation covering 1h46m of continuous tracking, cycling through ugriz through the LS60x5 fat slit. It should provide a good measure of how well the IMCS and guiding work together for long-dwell observations that will be common with MODS spectroscopy, and a requirement for being able
to realize the potential of MODS for faint-object spectroscopy.
Andrew Rakich has computed the amount of radius-dependent Z4 corrections in active optics and we've compared it with a model
for the offset in focus between the center of the guide probe (current calculation point for autofocus) and the WFS hotspot, and have broad
agreement. The effect acts as a systematic bias that makes measurement of other systematics of interest, like focal plane tilt, more difficult
(among other things). RWP modified the MODS agwServer code to permit changing the reference point used for the autofocus
correction via the runtime configuration file. We've put in the location of the WFS hotspot and tested the code for correct behavior relative
to the autofocus model. We are now ready to test tonight with a reduced aberration map. We can switch back to the previous reference
point (guide-probe field center) quickly, so we can restore the same function as on all previous nights if that is indicated.
[Note added after the night: the active optics tests confirm our assessment, and the new focus algorithm has significantly reduced the
radius-dependent Z4 down under 100nm].
The plan for the first part of tonight is to continue work on MOS alignment, focusing on the Stone F field mask which has superior astrometry for the
targets. Later we will revisit A1689 to apply techniques learned from Stone F. We will also proceed with guiding and differential refraction/flexure measurments at locations as weather permits (wind limits some quadrants of the sky). [Note added after the night: wind kept us from observing A1689, we instead focused on finishing a series of tracking and on-sky differential flexure tests].
All times are UTC
unless otherwise noted.
- Took dual grating mode flats and lamps to demonstrate the calibration scripts to Peter and Marco, and describe how scripting works. They took the calibrations (or, rather, started the scripts and watched them run).
- Clear this afternoon but hazy due to fires near the Mexico border, but we can't smell anything, so we'll watch it tonight to make sure it is not a particulate hazard to the mirrors. Clouds have begun to encroach from the NW: heavier cirrus than we've seen before. This looks to be a relatively thin spur from the larger system well north of us that should pass by after a few hours, with clearer skies behind.
Weather is cooler, only around 2-3°C this afternoon. Dry with humidity around 10%. Winds SW and variable 6-15m/s. Wind was the big factor last night, tonight looks like clouds are being added to the mix at the start.
- Opening up to start the cooling process inside as the sun sets. Currently the glass is at 1.4°C and outside is about the same.
- Waiting for it to get dark. Thin cloud above, stuff to the WNW as last night. Temp 1°C and falling steadily the past few hours. Humidity slowly climbing the past hour, now 20%. Variable SW wind 10m/s with occasional gusts to 14m/s.
Pointing and Setup
- WT10_370 for pointing check
IE = -69
CA = +87
- Going to BS9126 with an on-axis active preset for Collimation and seeing measurement
Note: we're using the new AGw stage focus reference position (WFS hotspot instead of GCam field center). No problems with the on-axis acquisition. Once we get collimated, I'll run an off-axis guide star.
Seeing is variable, around 1.2 arcsec, WFS has been steadily converging.
- Off to a 4-arcmin off-axis guide star near BS9126. Seeing 0.9-arcsec and good convergence No problems with acquisition, so the modifications to the MODS agwServer had no unintended problems (not that we expected any - trust but verify)
MOS Mask Alignment
- Going to the Stone F field with the MOS test mask. Running with the acqMODS program and stone_f.acq acquisition script file. Took a field and thru-mask image.
- thin clouds coming in and full moon rising. At one point the guider jumped to grab the bright bar of background at the left. I thought that crap was fixed
. Will file an IssueTrak
- Dialed in dPA=1° and took another field image
60s had too many saturated stars, taking 30s.
modsalign.py run - output transform: dX=+3.58 dY=-5.09 dPA=-1.2274°
Taking field and thru-mask images. Captured modsalign.py verbose output. --> stonef_pa1.txt
- Re-executing the preset, undoing everything above. Will now run a "normal" acquisition without putting in a rotation.
ran modsalign.py with the same boxes on the field snap: dX=+0.17 dY=-1.83 dPA=-0.2453°
backing out the dx,dy shift from the last transform and shooting field and thru-mask images
Measuring residual dx,dy using XVista: dX=-0.71 dY=-0.64 arcsec. Taking field and thru-mask images.
Looks well aligned in all boxes.
- Next test. Add dPA=0.5° to the rotator now that we have a nicely aligned image, and handed off the next
field image modsalign.py
Got the introduced rotation back from the fit (-0.502°) but the best-fit shift seems large. Sent the rotation and shift as-is and took a
thru-mask image. Stars ended up in the lower right-hand corners of the star boxes, but rotation is correct magntude and sign.
Handing the new post-offset image to modsalign.py, but now doing a pure-shift fit. Got dX=-1.65 dY=+1.65 arcsec. Applied and
took thru-mask and field snaps.
Stars are now well-centered in X, but off in Y by about -0.64-arcsec from the box centers. Computed with 3 stars with XVista and sending the offset.
Now back to a well-centered centered.
- Wind picking up a little, but clouds seem to have gone past us and the satellite shows a big clear spot.
Going to a place where we can measure the wind, the torques are getting nearly maxed.
Basic summary is that the current modsalign.py is finding the field rotation term for alignment with the correct sign and magnitude, but the
xy shifts are not being computed correct. The data we have above should give Paul what he needs to diagnose this.
AGw Autofocus algorithm
- Andrew Rakich tells me that the radius of curvature of the f/15 focal plane is measured to be 1055mm, not the 1040mm we've been told since time immemorial ("measured" = radius of the surfaces that minimizes field Z4). I've changed the parameter in the MODS agwServer (AGW_FRADIUS in include/agwutils.h) and recompiled and restarted the agwServer program. GCS restarted.
Sending a preset to make sure nothing untoward happens (we don't expect it but...) - nope, no problems on-axis (focus is the same)
Sending an off-axis preset (4-arcmin) - no problems. Yeah!
BTW, the difference between 1040mm and 1055mm RoC
for MODS masks amounts to about 140um at the corners of the mask. The mask material thickness is 150-180um. Nothing we would ever see, but we should do the right thing for the WFS (since active optics allegedly can see this as a measurable offset at large radii).
Guiding Stability and Differential Flexure
- Going to BS9131, setting star in NE to run another stability test, this less aggressive than the low-elevation run 2 nights ago.
Picked a fainter star nearby, running trackTest.obs, starting at 0525UTC. El=48°
Stopped at 0550 when Andrew came on
Field Aberration Tests for Focus Correction verification
Handed off to Andrew in Tucson
Doing field collect at BS9122_short field list to check focus correction of Rick's.
restarted explicitly setting the collimation limit to 250 nm.
Guiding failed, bright bar on the edge of the guide image took the guide star out. Bright "supermoon" and thin cirrus contributing factors. Rick putting in longpass filter. Trying again with "START=2".
longpass filter or clouds moving off improved guide image a lot.
guiding failed again, trying again...
Guiding indicating still need to improve SE parameters so a huge FWHM brigh bar is not chosen over a fainter round GS.
Only got one measurement of GS18, too faint and with SE configured the way it is and under these conditions guiding repeatedly fails. Rick is not amused. restarting on star #3. Lost this one too, could be cloud.
restarting on GS 3. Worked this time...
guiding failed on GS 4, resetting pointing then we'll try again. Resetting pointing IE-69 -> -90 CA 87-> 80.
giving up on this got too low, initial results very promising for focus correction though.
Back to Guider Stability and Flexure testing
- back to BS9131 where we were 40 minutes ago
Thin clouds flowing over the past while, bright moonlight is making the background high as a consequence, even far from the moon. Going to be an issue tonight until this passes.
Sequence starts with image 0080 (mods1r), El=40° 1.43 airmasses.
First 20 images in the sequence did not have the slit in (field rotation check), starting with red 0100 and blue 0066 the slit goes in as we descend.
Back to focus correction testing
preset to BS9155 doing field_collect on the short field list. May be too close to the moon, we'll see...
failed on the acquisition of our first off-axis guide star, restarting.
again failed on acquisition on a good bright star, problem is that the star is at an illegal off axis distance but there is no error message., restarting on #2 in the sequence. This also too far off. It would be really good to have suitable field lists ready for critical MODS commissioning time. Trying #3.
only got one measurement before we lost guiding.
*0821 trying our luck on a different Persson field, BS9169 further from the moon. Also adjusting Rick's focal radius for correction to 1040 mm, basedon our handful of results so far.
results on a couple of stars indicate that changing field radius from 1055 to 1040 moved us in teh wrong direction. Trying 1055 again on the last star we measured (#3 in sequence on BS9169 short).
seeing degraded to 1.7 " decided to give up on this for now and re-visit in better seeing. The good news is we are now at a fine-tuning stage, and errors we have are < 200 nm Z4 (compared to the +500-700 nm we had previously).
Back to MOS Mask alignment
- After good progress on the active optics, we're back to our program.
In the meantime, clouds started are coming in, looking a little heavier on the satellite than previous. Wind has decreased away from the warning thresholds for now.
Setting up on BS9145 to collimate and get back to things.
A1689 field is too close into the wind, torques were maxing out. Rats...
Long-duration tracking and differential flexure measurement
Went back to BS9145 - clouds coming in and background went up. Lost the star to the bright bar along the left - combination of moonlight and loss of starlight due to clouds.
The background problem in these conditions should get better with the baffle we're installing.
- trying again as the clouds pass by...
- Sending on-axis preset to BS9160, track up and over until the torques tell us to stop...
- Some fuss with background and clouds, but better now. Picked a fainter star in this field, seeing is good, 0.8-arcsec. Setting up for a trackTest.obs run
- Start at El=54° (1.24 airmasses), HA=1h, and track up towards the meridian. Will follow a range of elevations and rotator angles. Transit at 57° Fairly gentle run, but should be a test of a typical observation track.
- Torques starting to act up as we swing around into the wind. Crossed the meridian at ~1107UTC.
- Stopped as the torques were starting to max out. We got about 1h46m tracking this source over the meridian. That exercises the IMCS and the AGw for the equivalent of a long spectroscopic exposure (could not make 2h because of winds). This will tell us a lot about how it is doing as a function of wavelength.
End of Night Target
- Going after a CSS transient, CSS110316, at the outskirts of a blue SDSS galaxy at z=0.035
The moon is down enough to try spectroscopy. The goal here is to go from a finder and coordinates for a transient (unplanned) target,
select a guide star on the fly, acquire the target, lock and expose as an exercise of the procedure.
60s g+r, put it in the LS6x60x0.6 slit in one offset, then fired off a 600s integration in dual grating mode. Seeing is quite good: 0.5-0.6-arcsec(!) Got a terrific spectrum, which shows that the source is a Type Ia supernova about 14-days past maximum light.
Last little bit before twilight. Taking biases and putting MODS1 to bed.
- 19 Mar 2011