Adaptive Coronagraphy Using a Digital Micromirror Array Oregon State University Department of Physics by Brad Hermens Advisor: Dr. William Hetherington June 6, 2014
Abstract Coronagraphs have been used on telescopes for many years to block out bright sources in the sky to allow nearby objects to be resolved. A digital micromirror array (DMA) can be used in the focal plane of a telescope to serve as an adaptive coronagraph. The DMA can be configured to block out user defined sources after the telescope has been pointed. In this test, a new Python library was written to control the DMA, and its capabilities were tested on LEDs lit at various intensities. The new library did not increase the speed capabilities of the DMA but it offers a way for other programs to interface with the device. The test system was able to block out a single large LED in order to resolve the image of 4 small LEDs lit at roughly 1/800 the intensity. 1
CONTENTS CONTENTS Contents 1 Introduction 3 2 Methods 3 2.1 Equipment................................... 3 2.2 Programming................................. 4 2.3 Test Setup................................... 4 2.4 Data Collection................................ 7 3 Results 7 3.1 Programming................................. 7 3.2 Masking Results and Image Quality..................... 8 4 Discussion 10 5 Conclusion 11 2
1 Introduction A coronagraph is an attachment for a telescope that blocks out the light from bright objects so that nearby dim objects can be resolved. It was originally invented to block out the light of the sun to observe the corona, but can be used to block out any bright object. The coronagraph usually consists of a circular disk that is placed in the focal plane of the telescope where the bright object is in focus [4]. This masks the bright object and prevents its light from traveling any further through the optical system. A proper sized disk must be chosen and the telescope correctly aligned in order for this masking technique to work. The purpose of this research is to explore the viability of using a digital micromirror array as an adaptive mask. The micromirror array can quickly and easily be configured, removing the need to swap out masking disks. This could allow the user to change the mask while viewing or block out many objects at once. 2 Methods 2.1 Equipment The Meade LX200 12 inch telescope will be used for all imaging experiments. This telescope uses a cassegrain reflector which consists of a parabolic rear mirror and a hyperbolic front mirror. The rear parabolic mirror, also known as the primary mirror, has a large circular hole in the center. Light is collected from the front of the telescope, reflected off the primary mirror to the secondary mirror, and brought to a focal plane behind the primary mirror. The light path is shown in Figure 1. A focus knob on the back of the telescope allows the user to adjust the position of the primary mirror which will change the location of the focal plane. Figure 1: The light path of a cassegrain reflector. [2] The digital micromirror array (DMA) is a 1024 by 768 grid of tiny mirrors, each of which can be independently moved into one of three positions. The mirrors rest in a flat position and can also be tilted 12 up or down. The tilt direction is at 45 with reference to the face of the mirror array [5]. Figure 2 shows this tilt direction with respect to the face of the DMA and Figure 3 shows a close up of the individual mirrors. The DMA used for 3
this experiment measures.55 inches on the diagonal. The mirror array is attached to a control board which communicates with a computer using a USB connection. Mirror Tilt Direction DMA Face Figure 2: The micromirrors tilt in the direction of the arrows Figure 3: A closeup of the individual mirrors [1] The charge coupled device (CCD) used for all detection is the SBIG STF-8300M. The CCD is an array of photosensitive capacitors. These capacitors charge depending on the duration and intensity of light they are exposed to [3]. The charges can then be read by a computer to produce an image. 2.2 Programming The DMA comes with a simple program that uses monochrome bitmap images to set the direction of all the mirrors. Any black pixel on the bitmap represents a mirror that will be tilted towards the CCD, a white pixel is a mirror that will reflect light away. This program works well for simple low speed tasks but does not allow any sort of interfacing with other programs. The program provided with the DMA only works on the Windows XP operating system. The DMA program comes with two libraries for interfacing with the mirror array, a dynamic link library (DLL) for sending low level commands to the mirror array as well as a higher level ActiveX library. A wrapper, a thin layer of code that translates one interface to another, was written for the DLL to port the functionality to the Python programming language. This wrapper will allow other programs to be written that can interface with the DMA. The wrapper will be tested for ability to control the DMA and for execution speed. 2.3 Test Setup The entire experiment was set up on a cart so it can easily be moved out of the lab to be tested in different locations. Two optics breadboards were placed on top of a cart, one 4
laying flat and the other affixed on the edge. The telescope is positioned at one end of the flat breadboard and pointed out from the cart. Using a spirit level, the telescope was oriented to be parallel with the flat breadboard. Figure 6 is an image of the test system. The base of the telescope is butted against two vertical stops that are screwed into the breadboard and is surrounded with magnetic bases to hold it in place(figure 4). The telescope was set to the zero mark on the base and locked into position (Figure 5). Since the telescope was the most difficult to move, all other components of the optical system are moved to accommodate the telescopes position. Once the telescope was locked in place, only the focus would need to be adjusted for the remainder of the experiment. Figure 4: The back of the telescope. Stops are used to assure the base of the telescope is straight and the magnetic bases are used to hold it in place. Figure 5: Rotational axis of the telescope aligned with the base The DMA must be positioned at the center of the telescopes focal plane. To do this, the telescope field of view must first be mapped out. An LED was placed on the wall at the far end of the room and the telescope was pointed at it. The focus on the telescope must be adjusted so the image appears in the plane where the DMA will be mounted. With the image of the LED in focus, the LED is then moved left, right, up, and down while markings were made on the wall when the image left the field of view. The LED is then placed in the center of all the makings which is the center of the field of view for the telescope. With the LED in the center of the field of view, its image will be projected directly behind the telescope and served as a reference point for mounting the DMA. 5
Figure 6: A photo of the experimental setup The DMA was mounted at a 45 angle so the mirrors will tilt vertically. This allows the user to direct light 24 up or down. The LED used previously to find the field of view was replaced with a red laser spot projected against the wall to remove the need to affix a LED to the wall. The image of the laser spot is used to center the DMA in the image plane of the telescope. Once the DMA was in place, a lens and the CCD were added to the system. The purpose of the lens was to focus an image of the DMA surface on the CCD. The distance between the DMA and the CCD can not be much more than 50cm due to the size of the optical breadboard and the placement of the telescope. Since the mirror is at an angle with respect to the lens and CCD, a lens with a long focal length must be chosen to maximize depth of field. To find the lens with the largest focal length that will work in a 50cm space, the thin lens equation was used (equation 1). 1 S 1 + 1 S 2 = 1 f (1) S 1 and S 2 are the respective object and image distances from the lens and f is it s focal length. To get a real image, the object must be placed some distance away from the lens that is greater than the focal length. For a lens to work with the given distance, there must be a solution for equation 1 where the sum of S 1 and S 2 is equal or less than the distance between the DMA and the CCD. The maximum focal length that can be used must be 1/4 the length between the CCD and the DMA. The lens used in the system had a focal length of 120mm which is slightly less than the maximum focal length of 125mm. Once a lens was selected, the lens equation was used to find a rough starting point to focus the image. To get a clear image to focus, a checker pattern was set on the mirror 6
array. An image of a bitmap and the DMA showing the bitmap are shown in figures 7 and 8. The fine edges of the pattern allow a sharp image of the mirror surface to be focused on the CCD. Once the mirror array is in focus, the image from the telescope can be adjusted to be in same plane as the mirror. Figure 7: A bitmap used to set the state of the micromirror array. Figure 8: The pattern as displayed on the DMA. Four LEDs are placed on the wall in the center of the telescopes field of view. The CCD was configured to take successive pictures and display them on the monitor of a computer. Since the CCD was already focused on the surface of the mirror array, adjusting the focus on the telescope will brought the LEDs into focus on the mirror surface and made them visible in the pictures captured on the computer. 2.4 Data Collection The Python wrapper itself will be a form of data and it s capabilities will determine the abilities of the masking system. Masking data for this experiment consists of bitmap images captured from the CCD. The images will be of LEDs placed on the wall lit at various intensities and results of blocking them out with the DMA. 3 Results 3.1 Programming The Python library produced for the experiment added several capabilities that were not available with the original software. The new library allows the user to create lists that can set the state of the mirrors instead of relying only on bitmap images. This new 7
program also allows others the option to directly control the mirror array using their own custom programs. The library was modeled after the ActixeX library provided with the DMA and control board. An attempt was made to use the ActiveX library directly with no success. The source code for both libraries was provided upon request from Digital Light Innovations and provided useful insight on how to write a functional Python library. This new library makes use of the DLL provided to give the user functions that are slightly easier to use. The library allows the user to send raw data to the mirror array rather than being forced to use a monochrome bitmap. It can also be used in other programs to control the DMA for any sort of automated process that may be needed. The library allows the user to change the state of the DMA roughly 30 times per second. Although this is only 0.1% of the speed the mirror array is capable of, it is fast enough for coronagraphy. 3.2 Masking Results and Image Quality Figures 9 and 10 show 4 LEDs in focus emitting light at various intensities. The system was focused with all of the LEDs lit at the same intensity and every mirror oriented to reflect all light to the CCD. The best focus obtained is shown in figure 9. The intensity of the LEDs was reduced for the image displayed in figure 10, with two of the lights lit at an intensity much less than the others to test whether the finer details of the lights could be distinguished. The square shape of the LEDs can be seen in the two dim LEDs on the top of figure 10. 8
Figure 9: An image of 4 LEDs, all illuminated at the same intensity Figure 10: The same LEDs but the top ones have had their intensities significantly reduced. To determine the effectiveness of the masking capabilities, a large bright white LED was affixed next to the four LEDs from the previous section. The large LED is roughly 800 times brighter than the 4 small LEDS. A piece of black paper was placed between the bright LED and the dim ones to prevent reflections off solder joints around the small LEDs. The paper is oriented in a way that does not obscure the view of the LEDs through the telescope. The top image on figure 11 is a picture taken with all mirrors oriented to reflect light towards the CCD. In the bottom picture, light from the bright LED has been directed away from the CCD. The sharp line across the bottom image is the where the mirrors have been oriented in the opposite direction. The four dim LEDs are faintly visible in the bottom image. 9
Brad Hermens Thesis Ph 403 Figure 11: The top image shows a bright LED next to 4 dim LEDs. The 4 dim LEDs are not visible in the top image. In the bottom image the bright LED has bee blocked out by the DMA. The 4 dim LED are now visible due to the ability to take a longer exposure. 4 Discussion The results show that the DMA can be controlled by custom software and is capable of serving as an adaptable coronagraph. There was no noticeable effect on image quality due to diffraction from the DMA or effects from the lens. The DMA does not completely block out the large bright LED when the mirrors are directed away from the CCD. A faint image of the large LED can be seen at the bottom of figure 11. The ghost image of the large bright LED measures to be approximately the same intensity as the small LEDs. A significant amount of scattered light from the bright LED can be see despite the mask. The exact source of this is unknown, but it may be due to reflection off of a protective layer of glass covering the DMA surface. The mirror array limits the resolution of the image that can be captured as well as the minimum area that can be blocked. In figure 12, individual mirrors can be seen reflecting light from the LED. This resolution limit is directly proportional to the distance between the DMA and the telescope. The DMA in the test setup only reflects a small portion of the telescopes field of view back to the CCD. The DMA will need to be positioned much closer to the telescope in order to make a system that allows the telescope to move freely, which will reduce the image resolution significantly. 10
Figure 12: The individual mirrors can be seen near the edges of the light sources. 5 Conclusion A library was successfully created for the DMA. The test system successfully blocked out a bright light source allowing nearby dim lights to be resolved. This shows that the DMA is a viable option as a mask in coronagraphy. Future projects will need to significantly reduce the size of the system so it can be mounted on the back of the telescope. This will require new tests to determine if any depth of field or diffraction effects cause image degradation. Once a smaller system is created, it can be taken out of the lab and tested on stars and planets in the night sky. 11
References [1] Dmdchip.gif (GIF image, 512375 pixels). URL http://escience.anu.edu.au/lecture/cg/display/image/dmdchip.gif [2] File:Cassegrain-Telescope.svg. URL http://en.wikipedia.org/wiki/file:cassegrain-telescope.svg [3] James R. Janesick. Scientific Charge-coupled Devices. SPIE Press, January 2001. ISBN 9780819436986. [4] Marc Kuchner and Wesley Traub. Kuchner & traub, TPF coronagraph, August 2001. URL http://iopscience.iop.org/0004-637x/570/2/900/fulltext/ [5] Texas Instruments. DMD 101 - Introduction to DMD technology, June 2009. Rev. A. 12