LBTI
Detailed startup guide (Vidhya, Feb. 2012)
For DX side in AO Engineering mode (with pictures!)
Reconstructor parameters
WFS Schematic Diagram
Archived instructions for older software versions
Contents
I. Initial Setup
Initialize WFS
Initialize DM
Start AOS (if using)
Find the star (nudge the telescope)
IIa. AOS Mode
Prerequisites
Safe operation
Close and open loop using the AOS
IIb. AO Engineering Mode
Prerequisites
Safe operation
Close and open loop using the AO Engineering mode
III. Removing Static Aberrations
IV. Troubleshooting
I. Initial Startup
This step only initializes all the components. A more detailed description of the setup when using the Engineering Mode or AOS is in step II.
WFS Software
- ssh -XC sxwunit@lbti-sxwfs (dxwunit@lbti-dxwfs)
- startAO.py &
- This brings up a the WFS software main control window
- If it does not, the WFS software must be started using the command "w_start"
- The software can be shut down using the command "w_stop"
- Open "WFS HW GUI"
- Move the bayside stages to their nominal positions
- Choose CCD39 (WFS) in WFS Hardware GUI.
- If the status reports 'STATE_READY', Select '1x1' binning.
- Once it reports 'STATE_OPERATING,' choose 'Live View' to see display.
- Click Equalize Quadrants to automatically set appropriate bias levels for the 4 CCD quadrants
- Choose CCD47 (view finder camera) in WFS Hardware GUI. Make sure Status reports 'STATE_READY'.
- If the status reports 'STATE_READY', select '4x4' binning.
- Once it reports 'STATE_OPERATING,' choose 'Live View' to see display.
- If changing binning, you must power cycle the ccd47 tray power using lbti-web/power
- Click Equalize Quadrants.
- Make sure other essential components are up and running
- Look for a green check at the top of the pages for:
- tip/tilt; filter wheel 1 & 2; camera lens; ADC (on DX only)
- Open "WFS Arbitrator GUI"
- From the drop down Power menu, select "WFS_complete_(without_ccd47)" and click "Operate"
Secondary Mirror Startup
- ssh -XC AOeng@adsecsx (adsecdx)
- adsceng &
- Open "AdSec Mirror GUI."
- This shows the mirror status (position, current, commands)
- Open "AdSecControl GUI"
- If the status is SHELL RIP, and the box is yellow, the DM must be flattened before proceeding.
- Make sure telescope is above 30 degrees elevation.
- Click "SetFlatAO"
- In the Focal Station tab, select "Bent Gregorian Center"
Start AOS (if using)
- from lbtieng account on any obs machine
- AOSGUI left & (AOSGUI right &)
- Click the 'command' button to open the AOS Commands window
- Set Magn. equal to the approximate V band magnitude of the star, if it's not correct already.
- Make sure AO Mode is ACE-AO.
Configuring the ACE Table
The AOS chooses reconstructors, filters, etc based on an ACE table. If you need to edit this file, it's located in $ADOPT_ROOT/calib/wfs/LBTISX(DX)/ao/table_LBTI_ACE-AO.txt
Finding the Star (Nudging the Telescope)
- Adjust filter wheel 1 (transmits to CCD39, reflects to CCD47)
- For star finding with CCD47, set to least transmissive
- Adjust filter wheel 2 (transmits to CCD47)
- For star finding, set to most transmissive
- Use IE and CA telescope offset commands to move the star onto LMIRCam/NOMIC. This should also put the star on the WFS if the baysides are in their nominal positions.
- Use Steve's nifty GUI (must run on obs1) to send pointing offsets and spiral.
>> ssh -XC LBTIeng@obs1
>> cd /scratch/sallanson/pointnear
>> ./pngui &
- Select your side (left or right)
- PTPADDLE: send offsets to one or both sides.
- Checking "both" will send the offset to both sides at once by moving the telescope mount. If "both" is unchecked, the primary is moved instead. Moving the primary is a slow process - expcect a delay of several seconds.
- Use the "Option" check box to use CCD47 or LMIRCam coordinates. The current IE and CA offet values are listed here. You can also adjust
- PTSPIRAL: automatic spiral search on one or both sides
- Again, spiraling on both sides moves the telescope mount, which is much faster than moving the primary on one side alone. If you do not see the star on either side, start by spiraling on both sides.
- The FOV of the CCD47 is ~20", so set the spiral pitch to 7-10". Set the delay to at least 3 seconds for "both" and at least 5 seconds for single sided.
- When you've found the star, his "Pause".
- "OK! Reset" will accept the current location. This will set the current location as the center for a new spiral.
- A cheatsheet with bayside, IE/CA, and AZ/EL for the WFS and LMIRCam can be found on the home page (LBTIOrientation under Operational Guides).
Verify your side!
Make sure that your WFS HW GUI, your AdSec GUIs, and your AOS GUI (if using) all report the same side (right or left).IIa. AOS Mode (AOSGUI version)
Prerequisites:
- Need an image somewhere on CCD47 when the bayside stages are in their nominal ("sweet spot") positions. Otherwise, the AOS will fail. If absolutely necessary, you can change the sweet spot by editing the file fov.txt under $ADOPT_ROOT/calib/wfs/...
- Need reasonable image quality. If the image is too aberrated, the loop can't close. Ask the TO to fix static offsets if needed.
- AOSGUI is up and running
Safe operation:
-
Open or pause the loop if you see a "safe skip" on the AdSec Control GUI (the safe skip box will turn red)
- Do not allow actuator stroke of >10um to build up on the mirror, or it will RIP. You can monitor this on the AdSec Mirror GUI 'commands' display. This is especially important during closed loop nodding. You can't move the bayside x/y more than 2 cts at a time without exceeding the maximum stroke. You must wait for the offload to the primary before making another move.
-
To avoid a messy loop break, open the loop whenever the counts per subaperture (shown on the CCD39 viewer) fall too low. When using the maximum number of modes (DX=200 SX=400), the loop can become unstable below 100-150 cts.
Closing the Loop
- Go to the AOS Command GUI.
- For the following 3 commands, check command progress at the bottom of the window. The log is displayed at the bottom of the main AOS window. In the AOS commands window, the black box to the right of any button should read "Success" after the command is completed.
- If any command fails, try once or twice more. If still failing, try to repeat previous command.
- Click PresetAO
- This command gets darks for the CCD39 and CCD47.
- Click Acquire Ref
- This command centers the star on the CCD47 sweet spot, sets the filters, takes new darks for new frame rate, and removes static aberrations.
- Click StartAO
- This command optimizes the AO gain and closes the loop.
Opening the loop
- To completely stop the loop, click the StopAO button on the AOS Commands window. You'll have to use StartAO to close the loop.
- To only temporarily stop, use Pause. Use Resume to close the loop again immediately.
IIb. AO Engineering Mode
Use this mode when the AOS is not functional.
Prerequisites:
- Need an image with reasonable quality somewhere on CCD47. If the image is too aberrated, the loop can't close. Ask the TO to fix static offsets if needed.
- The image must be nearly centered on the CCD39. Move the bayside stages (from the WFS HW GUI) as needed.
Safe operation:
-
Open or pause the loop if you see a "safe skip" on the AdSec Control GUI (the safe skip box will turn red)
- Do not allow actuator stroke of >10um to build up on the mirror, or it will RIP. You can monitor this on the AdSec Mirror GUI 'commands' display. This is especially important during closed loop nodding. You can't move the bayside x/y more than 2 cts at a time without exceeding the maximum stroke. You must wait for the offload to the primary before making another move.
- To avoid a messy loop break, open the loop whenever the counts per subaperture (shown on the CCD39 viewer) fall too low. When using the maximum number of modes (DX=200 SX=400), the loop can become unstable below 100-150 cts.
Closing the loop
- In the WFS HW GUI
- Double check CCD39 framerate.
- Choose FW1 so the counts per px (shown in the CCD39 viewer) are between several hundred and several thousand.
- Adjust the camera lens to its nominal position.
- In the CCD39 viewer, click "Cloop Pupils" to show the pupil masks. Make sure that the pupil images are approximately aligned with the mask. Adjust camera lens as necessary, but ensure it does not come to close to the end of its travel. If it does, see Optical Alignment Troubleshooting.
- In the CCD39 viewer, make sure the Realtime Pupils are running, either by clicking the check box or verifying Diam, CX, and CY are populated with sensible values.
- In the AdSec Control GUI
- Go to the Reconstructor tab
- Choose the modal basis from the drop down: KL_v7 on DX ; KL_v9 on SX
- Pick the reconstructor (ReconstructorParameters.pdf at the top of this page)
- If you're closing for the first time and there are significant aberrations, start with 10 mode.
- Use a higher order recon if the pupils are fairly even or you've already flattened using a 10 mode recon.
- Hit Apply (Say 'yes' to the pupil mismatch error which will probably appear)
- Check that the pupils listed in the CCD39 viewer match those used for the reconstructor. If not, from the command line, use change_binning.py 1 (or 2, 3, 4) to load the appropriate pupils.
- Click the "Set zero gain" button
- In the WFS Control GUI
- In Loop Params box, set Binning, Loop Frequency, and Modulation. Hit Apply.
- In the Loop Control box, click "Close"
- Enable Camera Lens tracking. (Only if static aberrations are minor. Otherwise, postpone this until you've flattened the mirror at a very low high order gain.)
- In the AdSec Control GUI
- Shape Control -> Enable LO Offload
- Reconstructor Control -> TT gain to 0.05, stepping up to 0.1-0.2.
- Enable HO Offloads under Shape Control
- If there are a lot of static aberrations
- Turn HO gain up for brief periods of time until pupils more evenly illuminated.
- Turn HO gain to 0.05-0.10, ensuring HO offsets are enabled under Shape Control.
- Enable Cameral Lens Tracking if not already on.
- Optimizing gain
- Automatic: Use the Optimize Gain button on the WFS Control GUI
- Manual:
- Tip tilt: increase until oscillations appear. The DC level of the pupils will oscillate/slosh back and forth. Decrease gain 0.1-0.15 below this level.
- Higher Order: increase until oscillations appear. Stripes or ripples will appear. Decrease gain 0.1-0.15 below this level.
Opening the loop
- In WFS Control GUI, click Open.
III. Removing Static Aberrations
Apply a static offset to the WFS slopes (ie: a static aberration). These scripts are run from the WFS computer.
Use a live image on LMIRCam to view the effect.
Apply a single aberration (tested Nov 2011)
- >> cd $ADOPT_SOURCE/scripts
- >> python set_slopenull.py bin basis intmat_tracknum mode amplitude
- eg: python set_slopenull.py 1 KL_v9 20120315_123404 2 10e-9
- The amplitude is absolute. To remove an aberration, send a command with amplitude 0.
- You can't apply more than one mode at a time.
Apply multiple aberrations (untested)
- >> cd $ADOPT_ROOT/bin
- edit set_slopenull_LBTIDX.py and change the following variables:
- modes: which modes to apply
- coeff: the amplitude of those modes in [m]
- modalbasis: the KL_v# used for your reconstructor
- im_tracknum: the name of the reconstructor file you're using (eg. 20120315_123404)
- bin: the CCD39 binning associated with your reconstructor
- >> set_slopenull_LBTIDX.py (from anywhere on the WFS computer)
DX needs approx +70nm of astigmatism (mode 2) in Nov 2011
SX needs approx +10nm of trefoil mode 5 and +10nm of trefoil mode 6 in Nov 2011
If you open the loop, you may need to reapply the offset when you close again. Check the CCD39 and LMIRCam images.
SX needs approx +10nm of trefoil mode 5 and +10nm of trefoil mode 6 in Nov 2011
If you open the loop, you may need to reapply the offset when you close again. Check the CCD39 and LMIRCam images.
IV. Troubleshooting
Main Troubleshooting Page
AO Operation
This includes: Recovering from shell RIP, restarting WFS hardware processes, coping with frozen CCDs during closed loop, etc.
AOS Troubleshooting
AOS-specific issues.
/groups/lbti/search/index.rss?tag=hotlist/groups/lbti/search/?tag=hotWhat’s HotHotListHot!?tag=hot1/groups/lbti/sidebar/HotListAdded tag - hotphinzPhil Hinz2010-10-15 20:37:53+00:002010-10-15 20:37:53addTag3phinzPhil Hinz2010-10-15 13:18:40+00:002010-10-15 13:18:40updated2First createdphinzPhil Hinz2010-10-15 13:05:39+00:002010-10-15 13:05:39created1weblog2010-10-15T13:18:40+00:00groups/lbti/weblog/b808eFalseFirst Night Results/groups/lbti/weblog/b808e/First_Night_Results.htmlPhil Hinz3 updatesFirst Night Results
It was an eventful first night on the LBT. Network connections failed, hardware refused to turn on, all the typical problems of a new instrument. Aft...Falsephinz2010-10-15T13:18:40+00:00hot/groups/lbti/search/index.rss?sort=modifiedDate&kind=all&sortDirection=reverse&excludePages=wiki/welcomelist/groups/lbti/search/?sort=modifiedDate&kind=all&sortDirection=reverse&excludePages=wiki/welcomeRecent ChangesRecentChangesListUpdates?sort=modifiedDate&kind=all&sortDirection=reverse&excludePages=wiki/welcome0/groups/lbti/sidebar/RecentChangesListmodifiedDateallRecent ChangesRecentChangesListUpdateswiki/welcomeNo recent changes.reverse5searchlist/groups/lbti/calendar/Upcoming EventsUpcomingEventsListEvents1Getting events…
Comments