Filter Changer Module Work

Due to the discovery by CTIO engineer Patricio Schurter that the DECam Filter Control Module (FCM) contains LEDs that are shining brightly, a recovery path for the FCM was laid out.  It is still unclear how these LEDs got through the builder tests and the tests at Fermilab, but it is clear that these LEDs will be masked. Today four of the six DECam filters were removed from the FCM and the remaining two will be removed tomorrow.  Here are some photos of the activities (photos courtesy Ann Elliott).

Here is the telescope at the North position.

Very carefully the filters are removed.

A scissor lift is rented to provide access to the Filter Changer Module.

The filter is lowered to the ground.  Success!

Testing Guide Mode: AUTO

The DECam guiding has been shown to work if you select your guide stars.  However, the guide software has been written with something more sophisticated in mind--it will select guide stars automatically, with no interaction from the observer. 

Tonight we started testing this "auto" mode.  We first loaded the "starfinder" software in a focus and alignment  (F&A) CCD to see where the starfinder stars are.  This screenshot (courtesy Aaron R.) shows the results.  The red circles designate the starfinder stars, and the donuts are the stars in an F&A CCD. The difference between the starfinder stars and a star in a F&A CCD is approximately 20 arcseconds. 

 We then started the guide software in auto mode and with a region of interest 100 x 100 pixels.  There were no stars in the guide CCDs.  We then started increased the region of interest to 300 x 300 pixels, and guide stars were seen in all four guide CCDs!  *success*  this screenshot (courtesy Ann E.) shows the four stars.

The TSCINTERFACE seemed to be sending offset/guiding commands to the telescope control system (TCS)--we could see these offset commands appear in the TCSINTERFACE console.  However, we do not know if these actually reached the TCS.  We have our doubts--this is because the RA centroid and DEC centroid did not converge about ~0 arcsec (see bottom right hand corner).  Instead, as clearly seen, the centroids are getting worse with time.

Then the clouds rolled in, and the humidity spiked.  We closed.  We'll be doing more guider tests tomorrow night, weather permitting.  We need to understand (1.) how to make the starfinder stars match closer to a star in a Guide CCD and (2.) why the RA and DEC centroid values are not converging around zero.

Commissioning Week Overview: Alistair Walker

Report for Week 4 of Commissioning September 21-October 4
October 5, Alistair Walker (DECam Instrument Scientist)
=================================
A.  Here are the goals for this week as written in last week's report:

1.  Do a pointing map, actually probably one before and one after establishing a static look-up table (LUT) for the hexapod.

2.  Continue testing and debugging the guider. 

3.  Quantify tracking rate quality, slew to track times, etc. 

4.  Test and integrate the active optics system (donut + BCam), including the static LUT.

5.  Juan Estrada will investigate the low full well of the two half-F&A CCDs.

6.  As a minority activity, in suitable conditions take "pre-SV" science and technical observations.


B.  Here is what we accomplished this week, addressing the above goals first.

1.  A pointing map was done on the night of October 3.  The hexapod was set to the nominal (zenith) position for this test.   Additional confirmatory measurements were made the following night for some areas iof sky to give a finer grid.  There is some evidence for hysteresis at the 0.5 mm level.

2.  The short visit of Santiago Serrano was very successful.  The guider was successfully made to work in "self" mode and some initial tuning done.  

3.  The tracking, slew to track etc tuning and evaluation remains to be done.

4.   There has been much work on the F&A analysis code and preparing the temperature, filter offsets and position LUT.

5.  Attempts by Juan Estrada to repair the low full well of two halves of two focus and alignment CCDs were not successful.

6.  An improved WCS was installed as from the night of October 3,

7.  Sets of calibration frames (flats, zeros) were obtained regularly,

8.  Morning twilight exposures were taken to test the function of the protection diodes and to ascertain the level at which they should be set trip.   The trip level was subsequently revised,

9.   Data were taken to investigate scattered light as the telescope is moved to within 5 deg. of the Full Moon.

10.  Some science data were taken.

11.  Many improvements and bug fixes made to SISPI.


C.  Issues

1.  Early in the week SISPI developed poor reliability for the loading-phase.  Once loaded it was OK, although the need to refresh GUIs due to memory leak is annoying.    Cleanup of stale processes and files, and some attention to interactions between the Labview and the python codes have improved the situation.


D.  Goals for Next Week

1.  On-sky is short, after Monday night the telescope is off-sky in order to remove the filter change module to zap the LEDs.  

2.  Priority over the next three nights is installing the LUT, and doing more guiding tests. Some of this work should be done as part of "observing".

3.  Evaluate the pointing and TCS performance as soon as the new pointing model is installed.

Focus, focus, focus

Focusing a camera attached to a telescope has always been a challenge.  Changing night time weather conditions, such as temperature fluctuations, contribute to the difficulty of keeping stars sharp and in focus.  Dr. Kevin Reil and Dr. Aaron Roodman are two DECam commissioning team members working hard on the focus and alignment (F&A) system for DECam.  First light for the F&A occurred two weeks ago, but the system requires more work and fine tuning.

(Image credit: Dark Energy Survey Collaboration)

This section of a DECam image shows donuts (out-of-focus stars) in the bottom left CCD, but stars in good focus in the neighboring CCDs. The CCD in the bottom left corner is one of the 2k x 2k focus and alignment CCDs.  It is defocussed on purpose by 1.5mm and the resulting donuts will give an indication on how to best focus the camera.

Last night, images much more than 1.5mm out-of-focus were taken.  The top panel of this next plot shows stars +12mm out of focus.  The resulting donuts are enormous!  The F&A team used their code to model these donuts, and a fit of their (simplified) donut model to these images is shown in the bottom panel.  The purpose of using such a large donut, is to make sure that the strange shape of the central obstruction is being well modeled.  Images were also taken at zenith and at a large zenith distance, to see if there are any aberrations from mis-collimation that may effect the location of the central obstruction.

(Image credit: Dr. Aaron Roodman)

Perfecting the modeling of the central obstruction is not a simple task.  As can be easily seen in the +12mm out of focus images, there are quite a few different features that need to be taken into account.

First DECam Guiding

In a previous blog post (The DECam guider), I wrote about the first stages of testing the new DECam guide system.  Progress was made quickly at first, but then slowed down.  Eventually, the guide system expert, Santiago from Spain, hopped on an airplane, and 50 hours later (nightmare travel), he arrived on Tololo.  But Santiago didn't sleep when he arrived on Tololo--he went right to work with the rest of the commissioning team.  Here are two of the more obscure issues that were encountered, and luckily, eventually fixed.

PROBLEM:  In "self" mode (when the guider automatically is supposed to select the guide star), no stars were identified.

SOLUTION:  It was discovered that there was a desynchronization between the guider readout and the CCD shutter.  A guider exposure of 300 ms was taken, but the CCD shutter would open only after this time was done.  With the CCD integrating during the guider readout, it is no surprise that no guide stars showed up!  The guide exposure time was increased to 1000 ms, and now, finally guide stars were seen.

PROBLEM: Although guide stars are selected, the guider was still not working.

SOLUTION:  The TCSINTERFACE command relays information to the telescope control system (TCS).  This value needs to be a float value, with no special characters.  Unfortunately, a RA=%5.2f in the code was adding a space!  This was a quick fix -- but only once identified.

This image shows the interface of the DECam guider--here one region of interest of one (of the four) guide CCDs is shown.  The "best" guide star is marked in green, and second guide star is also in the region of interest, to further help with accurate guiding. 

The guider is not working 100% yet, but it is exciting that the first guided images with DECam have been achieved.

Linear DECam CCDs

Dear DECamers,

Today Jiangang H. (Fermilab) and I have been running CCD linearity and full-well tests for the DECam ccds with beautiful results (graphics and statistics courtesy of work by Jiangang Hao - Fermilab).

Linearity and full-well analysis for the left amplifier of DECam CCD N4 , showing inverse gain and full-well depth.

Linearity and full well analysis for the right amplifier of DECam CCD N4, showing inverse gain and full-well depth.

Analysis for the remaining CCDs is on-going, 

a la belle etoile,

David James/Jiangang Hao

CTIO/Fermilab

DECam Flat Fields

Dear DECamers, 

After a very busy, very eventful 5-nights on the mountain this week, with Andrea's expert help for our non-NOAO visitors, we've made a lot of progress this week; not least among our successes, are the psuedo-final settings for the flat-field projector (shown below) settings.

DECam will be calibrated by the DECal system. Designed and constructed by our friends at Texas A&M University (Darren Depoy, Jennifer Marshall, Jean-Philippe Rheault and formerly Jason Wise), and with me acting as CTIO's local "expert", the system consists of three main components:

1. An LED projector cone
2. A photodiode (and thermocouple) 
3. A fibre projection unit

One of DECal calibration units, which total four in number, prior to its installation 

of the top ring of the Blanco 4m telescope. The large black cone houses the seven LED 

lamps that shine onto the flat-field screen inside the dome. The medium-sized black-coloured 

box houses a photodiode and thermocouple assembly, while the small black unit (on the right) is 

the fibre projection unit. 

Each of the four LED projection units consists of seven LEDs, which for are designed for flat-field and internal calibration purposes, are powered contemporaneously. The emitting light level of each LED can be adjusted independently, allowing us to set the flux level of flat-field data in each filter per unit exposure time. The constraining parameters for our "ideal" exposures are for a mean flux of about 15k ADUs above BIAS across the array in a minimum exposure time of 10-seconds. 

Through empirical experimentation, afternoon dome flat-flats are now routinely obtained using the following parameters: 

           LED#                 Peak                   LED 
                                Wavelength             Power

• LED1:        6500 Angstroms             5% 
• LED2:        3650 Angstroms         100% 
• LED3:        7800 Angstroms             5% 
• LED4:        9700 Angstroms            8% 
• LED5:      10300 Angstroms          75% 
• LED6:        9050 Angstroms            8% 
• LED7:        4700 Angstroms          15% 

         Filter:       Exp              ADUs

                          Time        above BIAS

•     u'       15-secs         ~ 13k 
•     g'       30-secs         ~ 14k 
•     r'       10-secs         ~ 17k 
•     i'        22-secs        ~ 17k 
•     z'       10-secs         ~ 19k 
•     Y       10-secs         ~ 17k 

 

For those of you with access to the archives, a good set of flat-field calibration data are to be found in files #136529 -- 136584. 

a la belle etoile, 

David James,                        

djj@ctio.noao.edu

Dark Energy Camera Calibration Team [DECal],

CTIO Staff Astronomer.