Technical report: Focusing the DK-1.54m telescope (step-by-step user manual/guide)

FOCUSING THE DK-1.54m TELESCOPE WITH DFOSC IMAGING CCD A manual to relatively quickly focus the telescope with DFOSC camera Version history / change log: 20230701: Ver. 1 (Hinse). Created document using LibreOffice (*odt file). 20251002: Ver. 2 (Hinse, Liu & Salinas San Martín). Introduction This little manual should help you to focus the telescope using the DFOSC camera in a step-by-step guide. It is assumed that you know the TCS (Telescope Control System running on the Windows PC) and ARS (Astro Robotic System running on the Linux 1 PC). The DFOSC (Danish Faint Object Spectrograph and Camera) camera https://www.eso.org/public/sweden/teles-instr/lasilla/danish154/dfosc/ is a CCD camera mounted on the back of the Danish 1.54m telescope. In the photo, it's the silvercolored metallic cube with a fan, kept cold using liquid nitrogen stored in the silver cylinder (ESO/OHB technicians refill that cylinder/dewar with liquid nitrogen every day cooling the CCD down to around -120 °C). The camera captures sky images using a series of filters (U, B, V, etc.), which are selected by rotating a filter wheel positioned in front of the CCD. At the back end of the telescope, light is reflected through multiple mirrors and converges at the focal plane. To get sharp images, the CCD must sit exactly at this focal point—this ensures the point spread function (PSF) is as small as possible (perfect focus). However, the focus can shift due to a few factors: i) dome/environmental temperature changes: even a 1 °C change can slightly deform the telescope structure, altering the length of optical path and thus moving the focal plane and ii) filter changes: light of different wavelengths have different focal lengths (chromatic aberration). Page 1 / 14 The good news is that the filter-induced shift in focal length is automatically corrected by the TCS system. When you select a new filter in the TCS (or from within the ARS), it slightly adjusts the focus as well. But temperature-related focus changes are not corrected automatically, so you'll need to manually recalibrate the focus as environmental conditions change. The focus procedure will be done by following the iterative “adapting-sampling” method. The focusing can be done at any temperature and using a bright star close to the zenith. Chose a suitable bright star by clicking the “wrench” symbol in the TCS. Bright stars are always visible as they are pre-selected at all all Ras (00 – 24 hrs). Focus will be done in the Bessel U-filter. This is because in U filter the bright star appears dimmer and thus will in most cases not saturate the CCD. This is important to determine a good focus. Usually, a good exposure time is 0.1 seconds. Opting for very short exposure times (0.001 s) might result in failure for the DFOSC camera to acquire a useful image. There are two methods to focus the telescope: i) uniform sampling (traditional or brute-force focusing which is used in the Odin software when focusing the telescope with the LI / EMCCD cameras) and ii) adaptive sampling, which is more time-efficient achieving the same focus precision with only about 10% as many images. Method 1: Uniform sampling In this method, you take images at regular intervals, for example every 0.05 mm, from 37.5 mm to 39.5 mm. This results in a total of 41 images. As you go through this range: • At shorter focus lengths, the star looks like a donut or blurry circle. • It becomes increasingly compact as you approach the best focus. Then it becomes fuzzy again as you move past it. You find the best focus by comparing all the images and identifying where the star appears most compact. The drawback is that many of these images are unnecessary. You don’t need high-precision sampling far from the best focus, so time and effort are wasted. Page 2 / 14 Method 2: Adaptive Sampling (Recommended) This method is more efficient. Instead of taking all images at once, you take one image at a time and narrow in on the best focus step-by-step. Here’s how it works: • Take three images at equal intervals. • Identify the two images where the star appears more compact. The best focus lies between them. • Take a new image at the midpoint between these two. • Repeat the process: in every round, always compare three neighbouring points and narrow the interval. Each time you do this, you halve the step size and improve the focus estimate. After 6 images, the precision is 0.0625 mm. After 7 images, it reaches 0.03125 mm, which is more precise than what the eye can distinguish under normal seeing conditions. For comparison, using Method 1 to reach this level of precision would take about 40 images (0.05mm) or 60 images (0.03mm) - about 10 times more effort. Page 3 / 14 Here is the overall procedure: 1. Point to a bright star. 2. Change to U filter, CCD binning to 1x1, window the CCD (!!), set exposure to 0.1 s. 3. Follow this manual and determine a best focus value in millimeters. 4. Record the temperature from the TCS (Meteo Data). 5. Report your findings to the night log on the wiki page. In the following we describe the iterative focusing technique. It is based on the assumption that the focus changes symmetrically around the best in-focus value. We have tested this and confirm that the assumption is correct based on one focusing test experiment at temparature 15.2C (24. Sep. 2025). In the first iterative level (level 1 = Round 1) you create three images at three different focus settings. Iteration Level 1 (Round 1) Image1 @ 37.500 mm Image2 @ 38.500 mm Image3 @ 39.500 mm You will then load the three images into DS9 and adjust the zoom and scale for each image so they appear identical. Your job is then to decide – from visual inspection -- which two images are the two best focus images. For example you find Image 2 and Image 3 to be best in focus. That means the best-focus is somewhere in between focus setting for Image 2 and Image 3 (including the end points). Then you find the (half) focus in between image 2 and 3 and take a new image (image 4) with the new (halved) focus value set in TCS. At this point you are in the iteration level 2. New images are written in bold face. Iteration Level 2 (Round 2) Image2 @ 38.500 mm Image4 @ 39.000 mm Image3 @ 39.500 mm See the following diagram / figure. Page 4 / 14 The iteration procedure continues down to level 4 / round 4 (a total of 6 images) or level 5 / round 5 (a total of 7 images). Remember that the best-focus is a function of temperature in the dome / La Silla mountain. We judge that it will be hard to estimate / detect any improvement in focus by eye at level 5 / round 5. At level 4 / round 4 it might already be hard to judge between the three stellar point-spreadfunctions (PSFs) as displayed in DS9. The 1x1 binning will help to increase spatial resolution. STEP-BY-STEP MANUAL Step 01: In TCS (“Movable Carriage”): Position to DFOSC imaging position (click “park” = 300 mm button). Step 02: In TCS (“CORRECTIONS”): Point to a bright star near/close to zenith. Click the wrench icon. Slew the telescope to the bright star. Step 03: In ARS (“Setting”): Chose 1x1 binning (remember to click “set binning”). Step 04: In ARS (“Setting”): Window the CCD to only read out a central square area with the bright star located near or around the middle. The CCD dimension is 2148 pixels along x-axis and 2064 along the y-axis. The center is thus at (x,y) = (1124, 1032). We aim to window the CCD to a 400 pixel x 400 pixel sub window around the center. X begin = 924. X Size = 400. Y begin = 932. Y Size = 400. (The X and Y begin values come from: half the X or Y CCD axis dimension – 200 pixels). Click “set” and double check in ARS for desired values. Double check binning. Step 05: Page 5 / 14 In ARS (“Setting”): Set the exposure time to 0.1 seconds. Step 06: In ARS (“Setting”): Chose a sensible filename. For example FocusU For each image the ARS software will append a number. For example the first image will have the filename FocusU_000001.fits Step 07: In ARS (“Setting”): Set the image type to “Light”. Step 08: In ARS (“Setting”): Choose the U filter (Filter wheel A: empty. Filter wheel B: U). Step 09: In TCS (“Main Focus”): Set the “ABS” focus to 37.500 mm and click “POSIT” Step 10: In ARS: Take an exposure. Focus image 1 (400 x 400 pixels) is stored in /data/YYYYMMDD Page 6 / 14 Step 11: In TCS (“Main Focus”): Set the “ABS” focus to 38.500 mm and click “POSIT” Step 12: In ARS: Take an exposure. Focus image 2 is stored in /data/YYYYMMDD Step 13: In TCS (“Main Focus”): Set the “ABS” focus to 39.500 mm and click “POSIT” Step 14: In ARS: Take an exposure. Focus image 3 is stored in /data/YYYYMMDD At this stage you have now generated three images at different focus values. You should now repeat the above steps and generate image 4, image 5, image 6 and eventually image 7. To give you an idea what to expect we will display our results as obtained from a 45 minute test experiment on 24. September 2025. To open the images in DS9 you need to open a command-line terminal on the DFOSC Linux 1 PC and issue the command: dk154@dk154:/data/YYYYMMDD$ ds9 FocusU_000001.fits FocusU_000002.fits FocusU_000003.fits where YYYYMMDD is today’s date. Our results look like as follows: Page 7 / 14 Above screenshot of DS9 (Linux 1 PC running ARS) shows the bright star at three different focus settings at iteration level 1. Focus Iteration Level 1 / Round 1 Left Image Middle Image Right Image Image 1 (37.50 mm) Image 2 (38.50 mm) Image 3 (39.50 mm) From visual inspection the best focus is between Image 2 and Image 3 Page 8 / 14 Above screenshot of DS9 (Linux 1 PC running ARS) shows the bright star at three different focus settings at iteration level 2. Focus Iteration Level 2 / Round 2 Left Image Middle Image Right Image Image 2 (38.50 mm) Image 4 (39.00 mm) Image 3 (39.50 mm) From visual inspection the best focus is between Image 2 and Image 4 Page 9 / 14