MODS1 Commissioning Run 5: UTC 2011 February 15
Observer(s): R. Pogge & J. Antognini (OSU), M. Pedani (LBTO)
Telescope Operator: G. Bechetti (LBTO)
AO Support: Andrew Rakich & Doug Miller remote from Tucson at the start
Telescope Support: M. Midkiff (LBTO)
Instrument Support: D. Officer (LBTO)
Photometric conditions with mostly sub-arcsecond seeing. Best image was 0.41-arcsec FWHM in 30s in SDSS z at the end of the night.
The run was mostly devoted to observing spectrophotometric standard stars to calibrate the throughput of the grating and prism modes, including
single channel (dichroic-bypass) modes. All the while we exercised the long-slit acquisition process, trying different variations to better understand
where there might be bottlenecks, ways to improve, etc. The good news is that the procedures are quite simple and effective - we had mostly single-offset acquisitions that landed the target dead-center of the slit every time. Lots of good information for setting up the acquisition script engine, and
some work exercising observing scripts, but mostly we worked well by hand. MODS is proving to be relatively efficient to "fly by hand" without advance scripts, which should make it very flexible for rapid response to targets of opportunity, or to debug or correct problems and resume.
During the afternoon the new execMODS scripting program was tested, successfully demonstrating pause, resume, branching to the "exec:" block, and re-entrant operation where we can instruct execMODS to start a script from a particular line. This will greatly aid those times when a script fails because of a fault, allowing restart at the point of fault after correction.
Overall, given the long run of bad conditions, we had a great and very productive night.
Time lost to problems
: 1 hour to a rotator fault between about 0930 and 1030UTC. Cumulatively about 20-30 minutes due to failed presets caused by the bright background fault.
MODS filled at 4pm, still a quick fill and no sign of running out. Installed a couple of script-engine tweaks from last night (inevitable teething pains), and addressed a minor issue with the MODS GUI telescope panel. A little work today looking at the high temp of the guide camera. We plan to look at the IUB TEC power section tomorrow.
Discussions with LBTO personnel regarding the proper pointing and collimation model to use. We will get some assistance from downtown at the start to make sure everything starts up correctly on the TCS side.
We are proceeding forward tonight with a photometric and spectrophotometric program. We don't have our wide slits yet, but if the seeing is good we can use the 1.2-arcsec slit and take thru-slit images to assess the slit losses. We are not going to bet on the weather this time of year...
All times are UTC
unless otherwise noted.
- Clear at sunset. Temperature 3°C, humidity 15%, wind 6-8m/s from the SW. Starting with a few closed-instrument biases. Enclosure is open but no lights on (instrument horizon-pointing). Tonight looks photometric!! Hooray!
- Andrew Rakich connecting from downtown at the start to help as we get going for the night.
- Pointing to M5-1133. Using the B filter on the guider to make it easier to see the star.
IE = -73
CA = +87
- Active preset to BS9122, minimal azimuth move.
Started out way out of focus, but the WFS correction loop is bringing it home without manual intervention. Got a 120nm wavefront, seeing is a little variable, probably 0.8-arcsec or maybe better at times.
- Taking a 30s image in both channels (g and r) with the guide probe at the center to assess the PSF on 30-sec integration. WFS error is about 57nm rms, seeing about 0.7-arcsec.
- Going to Stone E field with an off-axis WFS star, PA=0. Pointing is right on the money. WFS/guide star is R=14.2
Getting possibly spurious coma from the vignetted pupil
- Going to the central star on-axis of this preset. Came in initially with a bit of coma. Dialing out the coma the off-axis star put in. Andrew reckons the coma put in at about 600nm. Most visible in the better seeing, of course.
G191B2B Spectrophotometric Star
- Going to G191B2B
for spectrophotometric standard. LS5x60x1.2 slit (widest we have, the 5-arcsec and 10-arcsec SpecPhot
slits are coming up tomorrow)
Odd error: we entered the guide star Dec as 52:47:31.2 instead of +52:... and the GCS computed a crazy AGw stage position. The tradition in western numerical systems is that no sign is implicitly a + sign. Not sure where the bug lies, could be in the MODS code, we'll check tomorrow.
Fixed that bad Dec and got a nice preset: the images converged at 0.8-arcsec FWHM and 46nm RMS WF error.
- Slit Image in SDSS r, red camera only.
- Target acquired with one offset. We're going slow to make sure we understand all the steps. Nailed it first time.
- Dual Grating mode, primary wavelength centers. Starting with 180 seconds at 0302 UTC.
Seeing is rock solid 0.75-arcsec FWHM, 25nm RMS WF Error. IMCS looks rock solid. Nice...
got about 20K ADU peak in red and blue in 180s
- shooting 300s per channel now.
- Switching to Dual Prism Mode.
- Prism spectra, 60s snapshot, 3Kx3K ROI
60s was 44K ADU in red, but saturated in blue. Still, very nice looking spectra.
- Running 3x30s in each channel in prism mode. Getting significant scattered moonlight now in the red channel. Not a
favorable moon angle tonight with the lack of baffling (the baffle is in OSU about to be shipped to the painter).
- Blue is still saturated... Dialing down to 15 seconds and taking 3x15s, red channel idle. Better this time, 24K ADU Peak in 15s.
- Taking thru-slit images for aperture correction. Image thru slit, then image w/o slit.
Using 2s exposures - 10s is saturated in both channels. Better this time. Also doing a 1s.
We're running out of cable wrap, and the Moon is getting in our way, so time to move on...
Hiltner 600 Spectrophotometric Standard Star
- Active preset with an off-axis guide star. May want to revise once on target.
Humidity is 30%, temperature 2°C, Wind still around 10m/s out of the SW. All-sky camera with bright moon shows no obvious cirrus. Still judging the night to be photometric.
Note - presets taking longer as the azimuth drive has been slowed down by 2x
- Slit Image while WFS converging, 30s in bright moonlight. Seeing about 1-arcsec, WF error is 136nm RMS.
Star is nearly in the slit and saturated
- 10s direct image of the field for acquisition
Would you believe, 1 second in r... Offset is x=0.62-arcsec in DETXY coords. Pretty near in the slit.
- Getting a lot of coma, the star is not that off-axis, but enough. Doug Miller has a possible solution for us on the way. In the meantime we're picking another guide star closer to on-axis.
- New guide star acquisition, not quite as bad as before. We hope
- field image, 1s exposure. and a thru-slit image. Some messing around. Offset 0.37arcsec in X, none in Y.
- Dual Grating Mode
- 60s - Red Peak=23K - Blue Peak=13K
- 120s - Red Peak=48K - Blue Peak=27K
- 150s - Red Peak=57K - Blue Peak=32K
- Dual Prism Mode
- 30s - both saturated
- 10s - both Good. 36K red, 46K blue
Running 6x10s, with 1Kx1K readout window to speed things up.
- Thru-slit image, 1Kx1K ROI, 1s each channel.
PG0823+546 Spectrophotometric Standard Star
- New target, now trying a variation on the slit acquisition procedure, using 1Kx1K ROI readout on red, and
- Put in slit during the preset
- Command imcslock after preset
- When IMCS locks on, snap sky thru slit while the WFS is converging
- When WFS converges, extract the slit and snap the field
- Using the slit and field images compute the offset to move the target onto the slit center
- Execute OFFSETXY, put in the slit, and snap a confirmation shot
1Kx1K ROI for centered-slit observations should speed up the readout overhead over 3Kx3K full imaging field.
- did this for PG0823+545. The star has a largish proper motion, so the offset was 3.57-arcsec in X.
Nice, one iteration again...
- Dual Grating, 180s Exposure
Seeing is 0.75 arcsec FWHM, WFS 80nm RMS WF error
Peak around 5K ADU in 180s. Taking 2x300s for deeper spectra (this is a fainter star)
- Dual Prism, 30s exposure
Got 5K ADU per, so running 2x180sec to get a nice signal standard shot.
Got 23K ADU per 180s, but moonlight is clobbering us (we need the baffle...)
WFS Reconstructor Experiment
Tweak pointing, then Andrew and Doug are going to try a reconstructor algorithm to mitigate the effect of vignetting the pupil at
the extreme Y end of the guide patrol field and try to keep the system from incorrectly dialing in coma.
- Parallactic mode acquire preset to a nearby bright star, adjusting IE/CA as per Andrew's instructions
IE = -75
CA = +70
- Active preset to the same star (position mode), on-axis for the WFS to collimate the telescope: converges to ~ 150 nm.
- Active preset using a far off-axis star. Sequence of Zernikes (in nm) sent to PSF for each iteration of active optics follows:
|| last iteration
These totals represent the error that we converged to. GCS reported a RMS WFE of ~ 150 nm here.
- Stopping GCS, Doug is installing his reconstructor, and restarting...
|| last iteration
Interesting, it diverged rather than converging. So the RMS WFE reported here by GCS never converged, unlike above.
- Sending an on-axis preset again with the reconstructor still installed.
Looks awful at the start because of the previous off-axis experiment, the star has a big "fox tail" of coma.
|| last iteration
Starting to converge steadily, however, but it is still pretty bad, then diverged badly.
- Stopping GCS, restoring the old reconstructor, and restarting
Now converging nicely, so we've restored function.
Active optics values for X, Y, Rx and Ry came back to their pre-experiment values.
Now if Doug can make his reconstructor register low aberrations for the first off-axis measurement above, and not diverge from there, we'll have something that will probably work.
That was instructive, perhaps we'll have an iteration on the new reconstructor later this run.
M3 Photometric Calibration Field
- going to two fields in M3 with secondary SDSS griz photometric standards. Problem is that the seeing is starting to degrade, about 1.6-arcsec FWHM. We also have to use off-axis stars, so we still have a problem with the WFS when off axis. It is not helped by being at 25° elevation, but the goal is to get a high(er) airmass measurement to derive extinction terms. This may not be fun...
- Starting in the "1444" (east) field.
Running imphot.cal for this field, also taking u despite not having standard measures here. Can at least use griz. Starting at airmass 1.85
- g: 2x30s and 2x10s + r: 2x30s + 2x10s
- u: 60s + i: 2x30s
- u: 60s + z: 2x30s
- going to the "1605" (west) field and repeating the sequence above.
Seeing has degraded to 2-arcsec, and we're losing our guide star in the bright moonlight every now and then during the worst puff. No sign of clouds in the all-sky camera or visually.
- seeing about 2-arcsec and the collimation is giving us lots of coma, so we had to abort the sequence during the u+1 set. We note that the temperature has started to rise (now 2.2°C) and the humidity has dropped to 5%. Wind still at 10m/s out of the SW.
- active preset to BS9142 as an on-axis WFS star to see if we can't clean up the collimation and check the seeing at higher elevation (80°) as opposed to 29° at the M3 field.
Getting reasonable convergence, the images started out pretty comatic and messed up. Now 0.8-arcsec FWHM and 153nm RMS wavefront error.
Feige 34 - Spectrophotometric Standard Star
- Going to Feige 34, using a relatively close guide star to avoid the coma problem as much as we can.
First acquisition much better, 0.6 arcsec FWHM, and quicly got 275nm RMS wavefront error after the first iteration, 90nm a couple after. This is good.
Using the LS5x60x1.2 slit mask, same drill as before for acquisition. 30s for the slit view, then 5s for the star (star was sat'd). Using 1Kx1K ROI to reduce readout overhead.
Offsetxy 2.58 0 REL, then a 5s thru the slit image to confirm. Followed by 1s, dx=+0.123 (1 pixel) and another 1s snap
- Dual Grating Mode
weird filter glitch on red, got confused as to where it was.
- Starting with a 60s snapshot, Airmass=1.03. Got 18K and 15K ADU in red and blue, respectively
- Running a 180s exposure in both channels - seeing 0.7 FWHM and 95nm RMS WFS error. Pretty nice here... Peak counts of 56K and 48K ADU, red and blue, respectively. Probably want 120-150sec exposures.
- Dual Prism Mode
First run, 30sec - both saturated (mildly, just a few pixels)
Second run, 10sec - blue 57KADU, Red 23KADU
Calib run - Blue 5x10s, Red 5x15s. Both are readout-time limited so there is only marginal deadtime.
Seeing during the sequence was 0.6-0.7 arcsec FWHM and 80-90nm RMS WF error
- Slit function - imaging thru slit 1s, then remove slit and image field 1s. Using a 1Kx1K ROI for lower overhead. Guider seeing is 0.65 arcsec FWHM and 100nm RMS WF error.
Oops, Feige 34 is too bright for that... My bad.
- Red-only grating mode - 120sec and GG495 filter. Seeing 0.65 arcsec FWHM, 130nm RMS WF error.
- Blue-only grating mode - 120 sec and Clear filter. Seeing 0.8 arcsec FWHM, 92nm RMS WF error
- Blue-only prism mode - 10sec and Clear filter. 58K ADU, Running a 3x10s sequence. Seeing 0.7arcsec FWHM, 92nm RMS WF error
- Red-only prism mode - 10s and GG495 filter. 15K ADU peak. Followed with 1x30s image, Seeing 0.6-0.7 arcsec FWHM, 70-90nm RMS WF error
Got 62K ADU in 30s, taking 20s to finish sequence, for 37K ADU at peak. Very nice.
Hz44 Spectrophotometric Standard Star
- Sent preset with nearer-by guide star.
Acquisition snap at the start was way off - the guide star (if it was the target) was way near the top. Going to a bright (M5) pointing star.
- Pointing check with M5_1332
IE = -75 (unchanged)
CA = +50
- Active on-axis preset to BS9152 - initial pointing was OK, a little low in vertical, but otherwise it is converging quick. Seeing 0.5-0.7 arcsec, getting 100nm RMS WF error.
- Resending the Hz44 preset - verified again with the target selection tool - much better this time. First WFS image was 616nm RMS WF error, started converging
- Taking a quick snapshot after RMS WF error dropped to ~200nm. Oops, still had the slit in for the thru-slit sky image, took it out and reshot . This star has high PM, so want to confirm field first given our previous pointing problems.
- Confirmatory shot after OFFSETXY dx=7.63 dy=0, only one step into the slit
- Dual Grating Mode
- 60s shot - noticed the WF error is not very stable, seen RMS between 100 and 300nm. Seeing ~0.7arcsec FWHM - got ~8K ADU peak per channel.
- 300s shot - Good signal, but seeing started to fuzz a little towards the end.
J1148+5252 - a challenging z-band target acquisition problem
- Target is a Z=6.4 QSO discovered by Fan et al. 2003 as an i-band drop out that is only detectable in the SDSS z band. Source brightness is z=20mag. This presents a challenge for our slit acquisition procedure: we have a faint target visible only in one of the lower-efficiency imaging bands at the far end of our spectral coverage. If we can acquire the target, we'll try a red-only prism and slit-less prism image to see what we get. We'll continue to use the 1.2-arcsec long slit mask.
- Preset failed - no star found. Going to a nearby brighter (R=12) star on-axis active preset to tweak collimation and check pointing. Preset went OK, a bit fuzzy on-target at start with ~1000nm RMS WF error on iteration 1, next iteration was 200nm RMS. That helps.
- Using instead PA=90 and a brighter near-axis guide star. The original choice is seen in a better image as a galaxy! Thanks DPOSS.
Got it, seeing ~0.7 arcsec. first WFS image 190nm RMS WF error.
- Taking a 60s z image (this is an i drop-out object) - of course in all the excitement (and not using the scripting) I forgot the IMCS lock-on first. Oops. Shooting again (that's why we're here - work out the process, find the pitfalls, rinse, repeat...)
- Oh yeah, we nailed it, and seeing is about 0.6-arcsec, too. Beauty. Only 1.35 arcsec E of the slit. Confirmed using the Fan et al. 2003 finding chart.
- Rotator fault! NOOOO! Rotator Error GUI shows a "SERDES Error" Geno attempting a reset.
- No luck, looking it up in IssueTrak
. Nothing relevant found, calling Dan Cox at 0947. Dan directed Geno to IssueTrak
- Dan talked Geno through the fault recovery. Resending the preset... faults immediately. Geno back on the phone with Dan and is preparing to head upstairs to swap a fiber. I've seen this fix employed on other LDG rotator faults. This fiber has been swapped on every run where we've used the LDG rotator while observing or moving during engineering.
- Resent the preset at Geno's request (he's on level 5). Executed but failed to find a guide star. Geno on his way back down to the control room.
- Sending pointing acquire preset, star M5_0441 (not to far from where we were). Way the hell off, not even on screen, but we can see its spikes.
IE = -66
CA = +89
- Re-sending the J1148 active preset... failed because of probe move would be out of range (huh?) My bad, grabbed the wrong acq file info when cutting and pasting.
- On-axis active preset to BS9152 to set collimation... Not a bad move, it was kind of fuzzy at the start. Converging quickly.
- Re-sending preset after collimation tweak up... much better. Seeing 0.6 arcsec before first WFS image, first WFS images 420nm RMS, second is
- converged at 70nm RMS, IMCS locked, taking 60s field finder image in the z_sdss filter.
QSO at the same place, 1.35 in X offset, 0 in Y (like before the fault). Confirmation shot thru slit at 1040: Nailed it dead center of the slit!!!
- Reconfigured as in red-only Prism mode (blue is a waste at z=6.4), locked-on IMCS, and started image at 1044, 600sec.
0.6-arcsec FWHM seeing on guider, WFS showing 55nm RMS WF error at best moments. Guiding steadily...
- 120s snap with the 1.2-arcsec slit
We solidly detect Ly-a and the Si lines in 600s with S/N=30/pixel (rough estimate) in the peak, but we could confirm that it was a QSO in a 2-minute integration with a whopping signal in Ly-alpha - looks like about S/N=5-6 per pixel in the line peak. Cool.
- Just for grins, taking a 300s slitless
spectrum through the grating of the field... Yeah, you can see that the QSO is "not like the others", having only a tiny nub of a spectrum compared to the stars and galaxies. This is going to be a fun observing mode
- Setup red-only grating mode, shooting at 600s for a start.
- An OFFSETXY then a PRESET failed because the GCS is grabbing the bright background instead of the star (the "stupid" version of the brightness veto in which it does not recognize that the "star" it found has a FWHM of 100s of pixels). The problem is exacerbated by the low setting Moon filling the chamber with moonlight. Grumble. I thought that problem had been fixed. This kind of messed up our last observation (an offset along the slit to make a delta-spectrum as the final acquisition mode for bright sky/faint object conditions). Nothing to be done, there is no manual override of the star finding algorithm when it is being stupid.
GD140 - Spectrophotometric Standard Star
- Mr. Moon is denying us further high-Z fun, so we're going to observe yet another standard star closer to home.
- a long slow slew, starting with a 10s field snap, then putting in the long slit. 1Kx1K ROI.
- Acquired with 2.95" X, 0Y offset. Confirming shot shows we are in the slit dead center.
- Dual Grating Mode - 180s in 1.2-arcsec slit. - 6K and 8K ADU peak in red and blue, respectively.
- Dual Prism Mode - start with 10s snap, then 180s. Getting 28K and 13K ADU in blue and red, respectively
- Red-only Prism Mode - 3x180s, useful for the deep QSO comparison extraction (rough flux calibration in far red).
M3 griz photometric field - higher elevation this time
- Pointing to M3 field B (1605 on the finder). El=76° this time, so images should be much better than before. big WF error (1900nm) on getting there. Oof.
- thought better of it and instead putting in an on-axis active preset of BS9158 to set the collimation on-axis and then trying again.
- A little slow to converge on both sides, and slews are now quite slow (AZ drive running at half-speed out of concern for the fatigued bogies).
running the imphot.cal script. Guider seeing is 0.8-arcsec, WFS is getting 200nm WF errors
- Started the end biases while the dome is dark. 10x3Kx3K and 10x8x3K in both channels using the doBias meta-script (which contains execMODS calls executing bias3K.cal and bias8K.cal).
- while taking biases, Geno is closing up
- 14 Feb 2011