The TRIPPEL Spectrograph A User Guide

Transcription

1 The TRIPPEL Spectrograph A User Guide April 8, 2006

2 Contents 1 Introduction 3 2 Overview 4 3 Spectrograph properties Wavelength tuning Filters Slit holder Spectrograph control Stability Image rotation Preparing a spectrograph setup Mounting and focusing spectrum cameras on ports A and B Mounting and focusing a spectrum camera on port C Focusing the telescope on the spectrum camera Chromatic effects Slit-jaw camera Simultaneous observations of two wavelengths Parallel imaging Observations Flats and darks Geometric distortions

3 5.3 Reductions Correlation tracker and adaptive optics Scanning Fine-tuning slit position Where to find it 21 7 Warnings 21 2

4 Figure 1: The spectrograph with the grating box to the right. 1 Introduction The TRI-Port Polarimetric Echelle-Littrow (TRIPPEL) spectrograph at the SST saw first light on June 6, This document describes the TRIPPEL spectrograph and its use with the general observer in mind. Thus details about the inner workings and how it is set up are skipped. It is assumed that everything is ready except the science cameras. It is also assumed that the spectroscopist is accustomed with the procedures for imaging observations at the SST. The stress here is on the differences to imaging procedures. Many things mentioned here may soon be outdated as experience is being gathered and we learn more about the instrument. Obviously transient facts are marked NOW:. Direct feedback to dan at astro.su.se. 3

5 2 Overview TRIPPEL is situated in the spectrograph room which is adjacent to the imaging ( URSIES ) room. The AO platform is rotated 180 degrees to feed the spectrograph with light. (General observers do not have to do this beam rotation.) The light beam goes through a hole in the wall. Observers can access the spectrograph room from the URSIES room via a somewhat larger hole. Watch your head! And be careful not to bump into the two Dalsa cameras and their cables. NOW: The light beam encounters a beamsplitter cube which currently and nominally transmits 90% of the light and sends the rest to the wavefront sensor and correlation tracker camera. The current beamsplitter performance degrades below 500 nm it is not well fitted for Ca H&K observations. TRIPPEL is built on two optical tables which are connected but have different heights. The slitbox is on the first higher table where also the WFS and the correlation tracker setups are built, as well as any slit-jaw setup or other imaging cameras. TRIPPEL is a Littrow spectrograph using a 79 grooves/mm echelle grating with a blaze of degrees. The slit has a width of 25 µm and is etched on a glassplate with a reflecting layer of coated silver. The slit plate is mounted on the turnable slit holder which consists one wall of the slit box. The slit plate is interchangeable. The slit box also contains a corrector lens and mirrors that divert the outgoing beam to the exit ports. The grating is mounted on a rotational stage in the grating box. The Littrow lens a doublet with 1500 mm focal length is on a translation stage in front of it. The lens should always be at a fixed position determined by the optical design. It is not used for focusing the spectrograph that is done by moving the spectrum cameras. The lens is only to be moved while checking the focus position of a spectrum camera. The grating is slightly tilted to the vertical so that the image of the slit the spectrum falls on a plane slightly above the slit. The three exit ports are called A (to the left when facing the sun), B (to the right when facing the sun), and C (central port on the top of the slit box). Currently only ports A and B are operated. Spectropolarimetry will be possible via port A in the future. Outside the exit ports the observer can mount cameras of the same sort and on the same kind of mounts as when doing imaging at the SST. (With the exception for port C for which there is a special camera mount that fits only Megaplus II cameras.) For ports A and B, the only difference is that the beam is higher so the track has to be placed on an aluminium block to get to the proper height. Since there is no predisperser, all spectral orders fall on top of each other. To separate orders 4

6 filters are used. The spectrograph has an off-plane design which causes keystone and smile deformation of the spectra. (There is also some image distortion from the corrector lens.) Thus the spectral lines and the dispersion direction will be slightly curved and not exactly orthogonal. 3 Spectrograph properties NOW: Will soon show some plots. Until then some theoretical numbers. Bandpass (limited by slit width) is around 1.25 km/s throughout the observable wavelength range. (This number is simply the product of the dispersion and the slit width.) Thus the theoretical effective spectral resolution is around , a more refined analysis shows that it should be at least Wavelength window with a Megaplus 1.6: around 0.9 nm at 400 nm, 1.8 nm at 800 nm. Dispersion is around 15 mm/nm at 400 nm, 7 mm/nm at 800 nm. The spatial scale should be close to the usual scale given by the telescope and its reimaging optics since the Littrow system gives 1:1 imaging. Thus the relevant numbers can be found here. A Megaplus camera at standard gain (0 db gain for old cameras with control box which is equal to +6dB gain for new or i cameras) saturates at about 290 ms exposure time in the 630 nm continuum at solar disk centre. 3.1 Wavelength tuning The grating can be rotated to allow the desired wavelengths to the exit ports. The spectrograph should always operate near its blaze angle. NOW: Calibrations and routines for quick wavelength tuning are in the pipeline. In the meantime, here are some numbers from which wavelengths should be possible to find with moderate precision. 5

7 Infall angle Newport remark α [deg] angle [deg] th order back towards slit Grating surface parallel to optical axis Blaze angle Port lambda [nm] order Newport [deg] A all B A A B B A A A A The grating constant is 79 grooves/mm. The focal length is 1500 mm. Note that the cameras can be moved sidewise so there will never be a calibration table or program with 0.01 deg or even 0.1 deg precision. The observer must always be prepared to identify lines using a spectral atlas. NOW: There is an IDL program to help finding and combining wavelengths. Start IDL on the server machine (number six) and give the command littslideplot. Necessary routines in idl/litt and idl/atlas. Note that it is always possible to observe a wavelength at another order, away from the blaze. There is a definite price to pay in that efficiency goes down, but this could make more wavelength combinations possible. 3.2 Filters There is a number of filters suitable for order sorting. At shorter wavelengths the requirements are such that they must be custom ordered. At red or NIR wavelengths one may use standard filters that manufacturers have in stock, if the wavelengths are suitable. Custom-ordered filters take several months to manufacture. NOW: We will prepare some notes to help with filter specifications. 6

8 Here is a list of the filters. Transmission curves are available through the internet version of the manual./bf NOW: not yet! There are of course suitable slit-jaw filters among the imaging filters too, but the ones listed here were ordered specifically for this purpose and are stored together with the spectrograph filters. As can be gleaned from the pass-bands, the requirements are easier at longer wavelengths and some filter could be useful both for order-sorting and slit-jaw imaging. CWL/FWHM approx. useful Manufacturer, remarks. designated [nm] range (untilted) A=Andover, B=Barr slit-jaw filter [nm] / B, Ca II H / B, [Mg I line, in steep gradient A /1 and A / / B, C I A 532.1/ / B, Fe Doppler / A, Na D, 25 mm diameter / B, Fe II mag / B, Fe magn., telluric A / / A, H-α / B, Fe Doppler A / / A, 25 mm diameter / A / B, O I / A / A, no good central plateau / A, Ca II / A, Ca II / The filters can be used at slightly shorter wavelengths by tilting them. Transmission plots are stored with the filters. Please put the filters back in their boxes which are marked spec or slitjaw and put the boxes back in the green drawer. Otherwise imaging people may take them and destroy them. 3.3 Slit holder Figure 2 shows the slit holder. It can rotate along a vertical axis which passes through the slit. Thus the reflection from the slit plate can be directed towards the desired direction without changing the slit position. The central part of the holder can rotate in the plane of the reflecting slit plate to change the slit orientation. The reflective slit plate is interchangeable. 7

9 Figure 2: The slit holder. A: Unscrew this to allow rotation of slit holder. B: Clamps fixing the slit rotation. C: Screw for adjusting slit rotation. D: Screws for adjusting upper and lower slit jaw. E: Rods for focusing device. F: Clamps for reflecting slit plate. 8

10 Figure 3: Newport control. 3.4 Spectrograph control The Littrow lens movement and the grating rotation is controlled via a Newport control unit placed on the shelf above the spectrograph. Keep this turned off when all adjustments have been made. Figure 3 shows the control unit. There are buttons and displays on the unit for manual control. Unit 1 is the lens, unit 2 is the grating. Note that the motors must be turned on with the buttons to right of the display. After turning on the unit, hit the home buttons for calibration. There should be a manual on the top of the unit, but operations are rather intuitive. A blinking E in the corner of the display indicates an error report which can be accessed via the MENU button and arrow keys. Error reports are given rather liberally even for such things as trying to move a motor while it was turned off. Thus most are not very interesting. The Newport unit can be controlled via a computer interface which disables manual control when active. Start the interface by clicking the SPECT icon on the desktop control panel. If the control unit is turned off or not connected there will be a complaint. The stages will be calibrated at startup and go to their preset starting positions. The lens can be set to three predetermined positions of which one is the focus position where it should ALWAYS be when observing. The grating can be controlled either by typing in a rotation angle or by clicking on arrow buttons. NOW: The spectrograph control is connected to royac19. 9

11 3.5 Stability The spectrograph is not perfectly stable. Long test runs show small movements of the spectral lines. Further work may identify the causes and come up with remedies in the form of stabilising hardware or introducing reference spectra. 3.6 Image rotation The slit will be rotating relative to the sun. The only way to counteract this is by scanning to make the effective area studied broader. Consider running the telescope in FLIP mode in the afternoons to decrease the rotation. 4 Preparing a spectrograph setup 4.1 Mounting and focusing spectrum cameras on ports A and B 1. Choose your camera. Megaplus 1.6 is probably the best choice because of its higher frame rate. If a longer spatial or spectral coverage is really needed, a 4.2 could be used. Note that these cameras show surprisingly different fringing properties. Notably bad are the 1.6 cameras XII, XIII, and XIV which show a strong cross-like pattern, see Fig. 4. Avoid these cameras. 2. Select your filter and mount it in front of the camera. Sometimes putting tension on the filter by taping it or tilting it to a certain angle may cause fringing. 3. Place the camera on a mount and slide it onto the track. Start it in focus mode. 4. Turn the grating to the approximate wavelength of your target line. For most angles, you will see no spectrum because the filter is blocking the light. 5. Focus the camera roughly by sliding it along the track while looking at the spectral lines. By now it is most convenient to have the computer monitor that is connected to the switch moved to a place where it can be seen from besides the spectrograph. 6. Adjust the height and the tilt of the camera. Illuminate the slit with sunlight. Carefully slide the focusing device (found in a marked container on the slit box the device is shown in Fig. 5) over these rods and make sure that it touches the slit plate. The focusing device has an air slit of 100 µm width. Thus the spectrum camera will show a very narrow spectrum. Adjust the height of the spectrum camera to get that centred. Use shims, etc, as for an imaging setup. Adjust the tilt at the same time so that the dispersion is as parallel 10

12 Figure 4: Bad fringes from camera XIV (June 2004). 11

13 Figure 5: Both sides of the focusing device. as possible with the CCD rows. Be careful to do this adjustment with the camera close to focus. The outgoing beam has an upward tilt and the mirrors are not perfectly mounted so the lightbeam is not exactly parallel to the track. 7. The spectrum camera is focused using a method similar to that used for focusing imaging camera in that it involves finding the focus position as the mean of two well-defined outof-focus positions. Instead of measuring intensities, eye estimates of sharpness are used. Instead of moving the camera between two positions straddling the focus, the Littrow lens is moved between to preset positions. The slit should be illuminated. Around the slit there are three rods. Carefully slide the focusing device in place and make sure that it touches the slit plate and that it does not get stuck on the rods. Focus on the narrow spectrum by sliding the camera along the track which should have a ruler taped to it. For each position the Littrow lens is moved between its + and - positions. note which of the two positions that give the sharpest spectrum edges. Use the Zoom display option and set the contrast high limit to about 20 counts. The spectrum will then look wider in the vertical direction when out of foucs. When the focus is reached, the + and the - positions give equally sharp spectral edges (not spectral lines). With the settings recommended here it i easy to look for the spectrum being equally wide at the + and - positions. The method is surprisingly fast and efficient and allows the focus to be determined with 1/2 mm precision. Do not forget to bring the Littrow lens back to its focus position by pressing the Focus button on the spectrograph GUI! The button will turn green. Being two observers, one in the spectrograph room and one in the observing room, with an intercom connection makes focusing somewhat faster. Note that is not possible to focus on the slit-jaw edges! They are situated several mm 12

14 behind the slit. 8. Once the camera is securely fastened to its track, it should be straylight protected by building a house of black paper over it. 9. If needed, the slit should be turned to have the spectral lines close to vertical on the CCD (they are slightly curved so this cannot be exact): Loosen the screws of the clamps holding the slit plate. Then turn the screw that rotates the slit until it reaches the desired orientation. 10. Adjust the slit jaws. There should be no more light than necessary being let into the spectrograph. It may be useful to shade the upper and/or lower part of the spectrum CCD to have some control on straylight levels. 11. The camera is in place and everything can now be controlled via computer. Finetune the grating. Then turn off the Newport control unit. If a new wavelength is chosen, the camera must be refocused. 4.2 Mounting and focusing a spectrum camera on port C The C port on top of the spectrograph has a custom-made camera holder. This only allows the small Megaplus II cameras. Once the camera is fastened to the holder, the filters can be switch by removing the plate beneath the camera (fastened by two screws). Withdraw the plate and put the filter in place, then put the plate back. NOW: The Megaplus II cameras have a dark level problem. To reduce the impact of this, loosen the filterplate and slide in a strip of shim plastic so that the field of view of the camera is reduced slightly on one edge. The edge to be covered is the one opposite to the side where the cable contact in the camera house is situated. When mounted on port C this side is the one closest to the wall. Adjust the camera holder so that the dispersion direction is horizontal, etc. Then fasten it with four screws. The focusing is then easy because there is a screw to move the camera vertically and a scale to read its position. 4.3 Focusing the telescope on the spectrum camera This procedure is equal to cofocusing the wavefront sensor with the slit position. This is usually done once when the spectrograph has been set up. It is not necessary otherwise unless there has 13

15 been some changes of the setup that affects the light path to the wavefront sensor or the wavefront sensor itself has been moved. Follow the first part of this procedure to check the focus, but do not start moving the wavefront sensor until having consulted the SST staff. 1. Do the normal wavefront sensor and control matrix calibration for the AO system. Then put in the clear-glass ND filters used when focusing the CCD cameras (if they are not already there). 2. By adjusting X-offset voltages, move the 40 µm (AO-) pinhole exactly to the middle of the slit sideways (vertical position is irrelevant). 3. Enable (unclick disable ) tip-tilt correction on the AO GUI such that the pinhole will not move on the slit when closing the loop. 4. Close the AO loop on the pinhole. (The pinhole will of course not be exactly at the centre of the subimages.) Repeat the previous steps if necessary in order to keep the pinhole very well centered on the slit. 5. With the AO loop closed, check the focus of the pinhole on the SPECTRUM camera in the vertical direction. Again use the - Focus and + Focus positions. When the height of the spectrum is equally large in these two positions, then the wavefront sensor is in the right position. If this is the case, you do not have to do anything. (This is more difficult than the spectrum-camera focusing procedure.) 6. If the wavefront sensor is not in the right position, consult the staff first. If given a go for refocusing, you have to move the wavefront sensor in order to improve focus. Open the AO loop, move the wavefront sensor a few mm in one direction, close the AO loop again and see if the focus of the spectrum camera improved or became worse (such that you know in what direction to move depending on which of the two - Focus and + Focus positions look best. This is because we did not yet have the possibility to check which way the wavefront sensor should be moved when e.g the + Focus position gives the sharpest image). Repeat the procedure with smaller movements until you feel confident that the wavefront sensor is within 0.5 mm of the correct focus. 7. Check the wavefront-sensor optical alignment (also pupil illumination), adjust if necessary, repeat the AO calibration and lock up again on the pinhole. Verify that the focus on the pinhole is still OK. If not, adjust. 4.4 Chromatic effects The reimaging triplet (on the AO platform) has a focus curve that is slightly wavelength dependent. This is a problem for spectroscopy because here we need to have the image formed on 14

16 Figure 6: Theoretical focus curve of the reimaging triplet lens. the slit. In imaging every camera is focused individually. Thus a slit-wfs cofocus found at one wavelength will not be exactly valid at another wavelength. Figure 6 shows a plot of the theoretical focus curve. Problems can be expected when co-observing lines in the blue and in the red, and when going into the IR. 4.5 Slit-jaw camera The slit-jaw camera is used to check the position of the slit. It should have a filter close in wavelength to the spectrum wavelength one reason for this is the differential atmospheric dispersion. One slit-jaw approach is to run the camera at the same exposure time as the spectrum camera in order to give exactly the same view. It must then be precisely synched with the spectrum camera. Alternatively the exposure time can be short for optimal image quality and possibly image reconstruction from several exposures. Then the stored timing of the exposures has to be used to approximately pair spectrum frames with slit-jaw images. We use a 240-mm Rodenstock lens for imaging the slit. Typically one wants to demagnify the image somewhat which means that the distance from the slit to the lens should be slightly more than twice the focal length. Make sure that the slit-jaw view covers all of the spatial view of the spectrum camera. This lens is not perfectly centred in its holder. Experts may want to take this into account in the setup. 15

17 Figure 7: A slit-jaw setup with two cameras. In this setup, the reflected beam is split by a cube just behind the 240-mm reimaging lens (in the black holder). The slitplate can be rotated around a vertical axis which coincides with the slit. Use this to direct the beam. Loosen the two protruding screws on both sides of the slit holder. Then just turn the plate and fasten the screws. The smile distortion causes the spectral lines in the cameras to be tilted (and curved) relative to the vertical. This causes the slit to deviate from the vertical when it is rotated to get lines and dispersion as orthogonal as possible in the spectrum cameras. It does not matter very much for the spectra, but anyone who wants the slit to be nicely vertical in the slit-jaw camera field of view will have to tilt that camera. Choose a filter with a CWL close to the wavelength of the spectrum. If the same exposure time is to be used, a lot of neutral blocking is needed (unless the filter is very narrow). Set the slit-jaw camera to the lowest available gain to minimise the blocking requirement. It is convenient to have rotating polaroids so that exposure levels can be finetuned with the spectrum camera. But ND filters and polaroids introduce reflections and straylight which can be quite problematic. Otherwise follow the instructions for mounting and focusing imaging cameras. This means that the camera should be focused using a pinhole while the AO is running. Do not focus on the slit. The slit will be sharp if the slit-jaw wavelength has a focus that is close to that of the wavelength where slit-wfs cofocusing was made (see Fig.6). Otherwise the slit may be noticably out of focus. Note that if the setup is demagnifying, the sharpness depth will be decreased and the 16

18 situation may look worse than it really is. Still the aim must be to get the sharpest possible images of the sun, not of the slit. Figure 7 shows an example of a slit-jaw setup. 4.6 Simultaneous observations of two wavelengths TRIPPEL allows observations of three different wavelengths simultaneously. The grating rotation sets what wavelengths are possible to observe in the different ports. Obviously not all combinations are possible. There is some stretch allowed in that the cameras can be moved sideways. There are also other problems: Currently a master camera can only have one slave. Thus only two cameras can be synched. This is a definite problem when scanning, because only one camera communicates with the tracker. The solution is to run the third and fourth cameras with external trigger enable. This will synchronise exposures and if all cameras run save all mode this will ensure that frame selection does not mix images up. Refer to the camera manual for more information. Differential refraction causes the slit to sample different parts on the sun at different wavelengths. The SST can in principle correct for this by tilting its field mirror, but there is not yet a system for doing this. Some wavelength combinations will suffer from longitudinal colour in the reimaging lens that causes different focus positions. 4.7 Parallel imaging The observer is free to do anything with the slit-jaw beam, but that will always show the slit (and dust specks that can be hard to remove). There is a possibility of putting a dichroic beamsplitter in front of the beam to give access to blue light. It is in principle possible to put a cube just before the WFS. This requires refocusing of the WFS and changing its filter combinations, however. In principle, though, all available photons should go to the spectrograph. 17

19 5 Observations 5.1 Flats and darks Make them in the usual way. But take flatfields often. MUCH more often than in imaging work. The flats must be acquired very close in time to the observations. Do not hesitate to take a break in the observations and make flat fields even if the seeing is good. Without flats the spectra are worthless. 5.2 Geometric distortions To rectify the spectra, the spectral lines can be used in the spatial direction. For the other direction direction, we have a device that places a grid of 0.5-mm spaced lines in front of the slit giving rise to a set of lines following the dispersion direction on the spectral frame. This is called the gg device. Place it at the slit and make a flatfield exposure. NOW: The current gg device is provisional. The glass plate with the etched lines is glued to a piece of pink plastic that can be slid on the three rods around the slit in a similar way to the focus device. As always be careful when working close to the slit. 5.3 Reductions Flat fields are constructed by computing a mean spectrum for each flatfield exposure, then dividing each spectral row with this mean. If the spectra are curved and not orthogonal, this becomes complicated. We lack a good set of routines for flatfielding with the current image distortions (smile and keystone). This may be made available later. NOW: An old set of routines (history: dan rouppe dettori dan) has been brushed up and is available on royac17. Use flat setup as usual. then give the command spectflat which has the same interface as flat. Be careful to define the image area to include only the illuminated part of the CCD. The spectflat command is defined by an alias in the.cshrc of obs. Code is in /home/obs/spec. Use this for a quick check. It is recommended to look at reduced data as soon as possible preferably the very same day. Are there heavy, time-dependent fringes? Are there enough flatfields? Have there been wavelength shifts? 18

20 Figure 8: Wavefront sensor and correlation tracker. 5.4 Correlation tracker and adaptive optics The correlation tracker hardware is set up on the small table in front of the spectrograph as shown in Fig. 8. Except for an extra folding cube everything is like the imaging setup. Operations are identical except for some coordinate changes. This should not concern general observers, but here they are. The calibration procedure takes care of the tracker orientation. What must be manually set is the signs for the drift corrections that are being sent to the turret computer. Under the menu: system - drift correction, set sign X to 1 and sign Y to 1. Remember to press save and restart the program. On the image table the sign of the x coordinate changes. The wavefront sensor is also set up on the small table. Operations are identical to those on the imaging table except that the orientations of the electrode response wavefronts that are displayed after a control matrix run are different. The illumination levels will be different from those on the imaging table. 19

21 5.5 Scanning The slit is narrow (0.11 arcsec). Seeing movements, image rotation, and, when two separate wavelengths are simultaneously observed, differential refraction all calls for some scanning procedure unless purely statistical studies are enough. The scanning method available involves sending offsets to the tracker system and having it introduce the appropriate tilt on the mirror. To use scanning simply enable the Spectrograph Step in the options menu. The following values are read from the configuration file defining operation. (An easier to use interface is on the way) SCAN HOST The correlation tracker computer. royac15 at the time of writing. SCAN PORT The port, this should always be SCAN STEPX The step size in the x direction (arcseconds). SCAN STEPY The step size in the y direction (arcseconds). SCAN STEPS The number of steps to make. When running in all data taking modes the system will scan make a single step in x and y at the end of eachh data taking period. It will repeat this SCAN STEPS times before going back to the starting position. Note the following: Set the repeat counter to 0 for frame selection. Scanning will change the AO target. There is an upper limit to the scanning range. 5.6 Fine-tuning slit position. Depending on how precise you want to be, it can be quite hard to get the slit exactly where you want it using the telescope handpaddle. To allow fine tuning of the slit position an interface is available on all camera computers for adjusting the slit position using the same mechanism as scanning. To start the interface (shown in figure 9) simply click on the slit icon on toolbar (when logged in as obs.) 20

22 Figure 9: Interface for fine tuning slit position). To use the system one should first choose a good correlation tracker target and lock on it. Then by using this interface you can use the tracker (which is much more precis than the turret handpaddle) to fine tune the slit position. The smallest step size is 0.11 (the slit width) and 0.5, 1 for medium and large respectively. Note that the AO target will change as you do this. 6 Where to find it Spectrograph parts are stored in green drawers at the entrance to the spectrograph room. There you find the filters. They are all marked spec on their boxes. Please put them back in the boxes and put the boxes back in the green drawer after use. The focus device is in a box top of the slit box. The gg device should be close by. 7 Warnings An observer should not move anything in the path from the telescope to the wavefront sensor or the spectrograph. An exception may be a dichroic beamsplitter which could in principle be mounted before the first cube, because this will change the optical distance to the WFS and the slit with the same amount. If this distance is changed, then the WFS and the slit will have to be cofocused again. The cofocusing is crucial for getting spectra with high spatial resolution. Do not move any parts of the spectrograph except the spectrum cameras. You may also adjust 21

23 the slit plate rotation, the slit rotation, and the slit-jaws. Be very careful when working with these adjustments. Do not touch the reflecting slit plate. 22