Preliminary Fabry-Perot Manual, Using Arcon CCD controllers and User Interface S. Heathcote and R. A. Schommer Feb 1995 Arcon with the Rutger's Fabry-Perot at the 1.5m 9 August 1994 _______________________________________________ _____________ Currently the Rutger's Fabry Perot must still be controlled via the LSI-11 software normaly used with the VEB controllers. However, we have set things up so that control of the instrument is fully integrated into the Arcon IRAF interface software. For instance the observe command will prompt for the desired central wavelength (or equivalently etalon spacing "Z"). These parameters will also be recorded in the image headers. This has been done by connecting the serial port on the LSI-11 (normaly used for the astronomers terminal) to the serial port on the Sun. The IRAF interface software then moves the RFP hardware by simulating a human user typing the neccesary commands. SEE BELOW, STARTING UP LSI-11 and INITIALIZING THE ARCON SOFTWARE for notes on startup procedures Special FP related commands ___________________________ Most of the observing commands are identical to those used for direct imaging except that you will be prompted for FP related parameters such as the central wavelength or etalon spacing ("Z") in addition to the filter position and telescope focus. Also some hardware is moved automaticaly according to the type of exposure taken. For exposures of type object, sflat and dflat, all the lamps are turned off and the internal comp mirror is removed from the beam. For exposures of type comp, you are prompted for the comp source required and the comp mirror is inserted into the beam. For exposures of type pflat the quartz lamp is switched on and the comparison mirror is set up as for comp. The following commands can be used to setup the hardware without taking an exposure: fpinit - Initialise cs100 and FP controller fpstatus - List status of RFP setcomp - Set comparison optics and lamps setfilter - Set filter positions and focus setview - Move field viewer in/out setwave - Set central wavelength instrument - Sets everything but is cumbersome The obspars and instrpars psets look like: fp> lpar obspars ccdtype = "object" Exposure type npics = 1 Number of exposures to take picture = 22 Picture number of first exposure exposure = 0. Exposure time title = "" Title of picture nfexpo = 7 Number of focus exposures freference = 5000 focus value fdelta = 50 Focus increment (autopicnum = yes) Generate picture number automatically ?\n\n# PO (setcomp = "auto") Query and set comp. mirror and lamp\n\n# SELEC (setwave = "wavelength") Query and set wavelength or Z value (setfilters = "three") Query and set filters? (filtype = "instrument") Type of filters to use\n\n# SETTING FOCUS FOR E (setfocus = no) Query and set focus? (foctype = "manual") Type of focus to use (temperature = 300.) Telescope temperature (basefocus = INDEF) Focus base value (reftemp = 300.) Telescope temperature for base focus value (tfrcoefs = "") Coeficients of Temperature-Focus relationship\n (refis = "middle") Reference is first, middle or last exposure? (focmode = "manual") Focus mode (shtype = "detector") Shift type (fra_offset = 0.) Focus offset in RA (fdec_offset = 20.) Focus offset in Declination (fnrows = 30) Focus number of rows to reverse shift\n (mode = "ql") Most of these parameters are exactly as for direct imaging. setcomp controls the automatic movement of the comparison optics. setwave controls whether to prompt for the central wavelength, the "Z" value or neither. If setfilters is set to "three" then you will be prompted for the name of the filter in the RFP's internal filter bolt. This information gets recorded in the image header but is not used for anything else. To disable this prompt set setfilters="none". fp> lpar instrpars compmirror = "direct" Comparison mirror position complamp = "tube" Comparison lamp\n wavelength = 5000. Desired central wavelength fpz = 0 Fabry-Perot Z value filter1 = "1" Filter in wheel one filter2 = "1" Filter in wheel two filter3 = "" Filter in fixed internal slide fpview = "in" Position for Field view mirror (out to observe) (fpwave0 = 5000.) Zero point of wavelength-Z relationship (fpdwdz = 0.1) Slope of wavelength-Z relationship\n (wheel1 = "") Filter info. pset for wheel one (wheel2 = "") Filter info. pset for wheel two\n (etalon = "") Etalon identification\n (instrname = "fpccd") Instrument name (mtrctlr = "lsi11") Motor controller for filter wheel (comptype = "internal") use external or internal comp lamp (mode = "ql") Almost all these parameters are prompted for as needed. You will want to set the correct values of fpwave0 and fpdwdz, used in calcualting the Z value required for a given central wavelength, from time to time Note the normal LSI-11 software commands can be used to move the hardware by prefixing them with lsi e.g fp> lsi fplampon will turn on the quartz lamp. Helpful Fabry-Perot Observing Hints ---------------------------------- Take a calibration ring approximately every hour during the night. This is particularly inportant during the start of the night, when the temperature changes (drops) quickly, and it is important to monitor the wavelength zeropoint drift. Take a set of calibration rings during one afternoon (at least) to obtain the full wavelength solution. At least 3 different spectral lines should be used, and 3-4 different wavelenght settings per line measured. Over less than 70 A range, a linear wavelength fit is adequate, but a quadratic improves the fit, and is nessary to find a good solution over more than 100A calibration series. See the ducmentation below from the IRAF reduction task. Flatfields are best taken with dome flats. Twilights do not work well, because of the fine struction of emission and abortption in the sky. A dome flat need not be taken at every wavelength, the change is slow over that scale. But every 5-10A is certainly necessary. One strategy we have used is to take 3 flats (for CR removel) every 3-4A, and thus for a typcial galaxy tuen of 10-15A we would have 3-5 different flats. Alternatively, flats at every wavelength could be taken, and a running median (of 3) could be used to provide a CR-cleaned flatfield. TYPICAL NIGHTTIME PROBLEMS -------------------------- For a variety of reasons, the system will sometimes hang while asking the parementers for a typical observation. The commands stop being sent to the LSI-11 in the arsh window, and the user interface window simply stops responding. The first attempt to recover should be to use ctrl-C, which exits the observe task. Then several "flpr" commands should be sent to clean up the various IRAF nasties. Then the observe command can be attempted again. If the above fails, the one_button warm restart available on the root arcon menu should be executed. After this, and at all such restarts, 3 things should be checked. 1) "setdet force+" -- needed to reset gain, binning and region of interest parameters to the user choices. The default (re-)startup values are likely incorrect. 2) check the imdir and reset if necessary 3) check the working directory and change if necessary (you will be restarted in your home directory, not in any previous subdirectory). NEW OPTICS-CCD COMBINATIONS FOR THE RFP --------------------------------------- Since August 1994 we have been using Arcon controlled TEk CCDs as teh detector for the Rutgers Imaging Fabry-Perot. These CCDs have 24micron pixels, low readout noise (typically 3-4.5 e-), and excellent quantum efficiency and response flatness. The software interface for the Arcon users allows the standard RFP spectrograph control to be done via an IRAF parameter file, which make the instrument use very transparent. A similar interface is in use on our 4m RC spectrograph + Tek (F/Schmidt camera) + Blue Air Schmidt and Loral 3k. Because the Tek2k and Tek1k have large image areas we have also provided a longer focal length camera, of particular use on the 1.5m. Below are the telescope/focus/camera options currently in use. Telescope 135mm (f/2.0) camera 200mm (f/4) lens 4m f/7.76 0.36 arcsec/pixel 0.24 arcsec/pixel 1.5m f/13.5 0.54 arcsec/pixel 0.36 arcsec/pix 1.5m f/7.5 0.97 arcsec/pixel 0.65 arcsec/pix The preferred options are the 135m camera for the 4m, and the 200mm camera for the 1.5m (at f/7.5). With the 200mm lens the field projects onto approximately 750 pixels at the detector. Either the TEk1k or Tek2k chip can (and have been) be used. The choice is sometimes driven by other scheduling constraints. Currently (Feb 95) the Tek2k in Arcon 3.6 offers slightly lower read noise (however, it is in heavy demand for 4m instruments). The FOV remains determined by the 1-inch focal plane aperture of the instrument. On the 4m, this gives about 160 arcsec of useable field, while on the 1.5m the FOV is 440 arcsec at f/7.5 and 245 arcsec at f/13.5. It should be noted that the 4m and 1.5m at f/13.5 do provide significant numbers of nights with sub-arcsecond seeing. Due to optics problems the 1.5m at f/7.5 never produces images as small as 1 arcsec, and 1.5 arcsec is typical. Reseting the LSI-11 Software: _____________________________ IMPORTANT NOTE: it is not neccesary to reboot the LSI-11 (a slow painfull procedure) each time you reload the Arcon software. If communications to the FP controller get hung up: Set the A/B switch to "A" then on the pc type * ^t ^c task ..... aborted kernel> ^r * initcs100 * initfp Switch back to "B" when all is happy again. If the LSI-11 realy gets in a tizzy it may need to be re-bootstrapped. This can be done by switching control back to the pc (A/B switch to A) and following the usual startup procedure: kernel kernel> 0 0! 000002 @77773000g $db 1173 CACHE + MEM TEST anwer "y" to the first question. it takes a very long time for the LSI to reboot during much of which nothing will appear to be happening. Be patient!. Eventualy you will see the message: System now connected to user task and ready for input. * Make sure all is well again and then set the A/B switch back to "B". Starting up the LSI-11 program: _______________________________ This will normaly be done by observer support. The serial line to the LSI-11 is connected via an "A/B switcher". This is located in the grey box in the 1.5-m computer room (ask observer support to show you where this is). With the switch in the A position the LSI-11 is connected to the "observers PC" in the console room. In the B position it is connected to port B (ttyb) on ctioa2. Set the switch to A so that you can bootstrap the LSI-11 from the PC. Follow the normal LSI-11 bootstrap procedure. When answering the questions specify that you are using the fabry-perot, the "new" shutter filter bolts, do have a camac crate, and are using a pc-emulator. Once you have bootstrapped succesfully initialise the cs100 (initcs100) and RFP controller (initfp). Next check normal opperation of the RFP: switch on/off the various comparison sources; move the comparison optics and viewer; check that the etalon spacing can be changed. Once you are happy that the RFP is working normaly set the switch to the B position. Initialising the IRAF Interface software: _________________________________________ Loading the Arcon software follows the normal procedure; the fpccd package should be used for this instrument (manual real soon now, mean while use the direct imaging manual). List the instrpars pset: cf> lpar instrpars the last three entries should be: (instrna= fpccd) Instrument name (mtrctlr= lsi11) Motor controller for filter wheel (comptyp= internal) use external or internal comp lamp Now run the command fpgag. This downloads modifications to the FORTH software so that it doesn't insist on printing out the RFP status screen for every command entered. This must be run each time the LSI-11 is rebootstrapped. Having run this command you MUST bye out of fpccd and renter: fp> fpgag .... lots of gobbeldygook fp> bye ar> fpccd Reseting the IRAF interface and Arcon Software ______________________________________________ Malfunctions of the system might also be due to problems in the IRAF interface software or in the muxnex communications software. To clear this follow the warm start procedure given in the manual. Since this is a fast procedure, try this before you rebootstrap the LSI-11 unless it is obvious to you that the LSI-11 is out to lunch. _______________________________________________________________________________ Configuring RFP package for 4m or 1.5m telescopes 8 August 94 ________________________________________________ __________ There are slight differences between the hardware used with the RFP at the 4.0-m and 1.5-m telescopes. Hence, it is necessary to configure the RFP package accordingly. This involves changing the default value and option list for some parameters in obspars and instrpars. The option list is changed to prevent curious users from shooting themselves in the foot. To make life easier, there are two versions of each pset file in the fpccd directory and a links are used to select the appropriate ones. To configure for 4.Om (while logged in as root or other deity): %> cd /xp/iraf/fpccd %> rm obspars.par %> ln -s obspars4m.par obspars.par %> rm instrpars.par %> ls -s instrpars4m instrpars.par To configure for 1.5m: %> cd /xp/iraf/fpccd %> rm obspars.par %> ln -s obspars60.par obspars.par %> rm instrpars.par %> ls -s instrpars60 instrpars.par You should unlearn instrpars and obspars having made this change. One day soon the software will be made smart enough to do this itself based on which telescope is in use. NOTES on REDUCTIONS from the IRAF Fabry-Package *** Notes added by R. A. Schommer, Feb 1995 FABRY (July87) fabry FABRY (July87) Notes on the Fabry Perot Reduction Package George Jacoby July 1987 1. Introduction The Rutgers imaging Fabry Perot (RFP), is an piezo-electric controlled etalon imaging system which effectively produces very narrow-band images. The wavelength of the device is controlled by adjusting the etalon spacing, or gap (associated with the variable "z"). 2. Wavelength Calibration The light from the telescope converging beam must pass through the etalon as parallel light, and so re-imaging optics is also provided. However, images arising away from the collimator optical axis pass through the etalon at some angle which increases with radius. The effect is to reduce the wavelength of light which is passed at that radius, and so "rings" of constant wavelength are introduced according to the equation lambda(R) = Lambda(0) * Cos (arctan (R/C)) where C is a constant to be determined. Wavelength calibration is determined by observing a lamp having emission lines of know wavelength (He-Ne-Ar, for example) at a variety of etalon spacings, and for a variety of lines from the reference lamp. The ring diameters (or radii), and the etalon gaps are used to derive the constants A and B of the dispersion solution: Lambda(0) = A + Bz and where the constant C is defined above. Generally the constant, B, is very stable over a run, and C is reasonably stable, as well. The zero point, A, however can drift throughout the night, and so reference lamp observations are recommended during the night. ** Note, as mentioned in the main text, a quadratic fit to the wavelength solution is recommended, particularly if 100A or more is to be fit. Thus in the above terminology, Lambda(0) = A + Bz + Cz*z is required. Versions of such a wavelnght fitter are available, although theya re fairly simple to construct and users may chose to implement this off_line at their home institutions. -1- FABRY (July87) fabry FABRY (July87) 3. Positional Corrections Because the images of the object can drift positionally even when auto-guiding, the object frames must be shifted to a uniform and congruent spatial grid. Thereafter the object frames can be built into an image cube. Any postional shifts have to be remembered since the wavelength in each frame depends on its radius from the optical axis. 4. Other Corrections Additional corrections of each frame in the cube are required to account for flux normalization (due to atmospheric extinction or clouds), and sky subtraction. 5. Velocity Maps Once these corrections have been applied, a velocity map can be created by fitting a Gaussian to the spectrum at each point in the spatial sample. The intensity along the wavelength axis is extracted at each pixel, and the wavelength is computed using the constants B and C above, plus a revised A derived from recent calibration lamp exposures. The radius of the pixel is calculated, with corrections for shifts relative to the center of the ring found for the calibration lamp. The Gaussian is used as an approximation to the instrumental profile. If the instrument actually resolves structure in the emission, then the derived center may be erroneous. The line center is used to compute a velocity for the emission of the pixel and placed into an output velocity map. If the signal is adequate, a velocity dispersion parameter (the sigma of the Gaussian in km/s) is also placed into a second map. The continuum and peak intensity derived from the Gaussian fits are 2 more maps which are generated. Four additional maps, the error in each of the four parameters as derived from the formal fitting procedure, are also created. The basic steps required after the images have been initially processed through the flat-field and cosmic ray removal stages, are outlined below. Note that for the ring parameter program to work well, the ring images should not be flat fielded. 6. Summary of Steps Required 1. RINGPARS -- find the ring parameteres: x-center, y-center, and radius 2. FITRING -- fit the ring parameters to the dispersion equation -2- FABRY (July87) fabry FABRY (July87) 3. ICNTR -- find the centers of stars used to define the shifts 4. MKSHIFT -- use the centers to create a script for IMSHIFT 5. MKCUBE -- combine the shifted images into a single cube 6. ICUR -- Not supported on SUN version, use jdisplay instead 7. FINDSKY -- using coordinates for sky, estimate the sky in each frame 8. NORMALIZE - enter the normalization factors for each frame 9. RINGPARS -- measure parameters for the recent reference lamp frames 10. ZEROPT -- compute the wavelength zero points using recent lamp frames 11. VELOCITY - generate velocity map and spectral plots 7. Known Bugs/Problem Areas The MKSHIFT routine simply adds an "s" to the image name to create an output name to save the user from thinking about more names. If the input files have been generated with FILES, they may have ".imh" in their names, and the "s" will be appended to create image.imhs.imh kinds of files. This should be fixed. The etalon zero point may have some thermal drift, and so the ZEROPT program needs to interpolate in time if more than one lamp observation was taken. This can be entered by hand, at present, using HEDIT. The plot code is pretty slow. I can't see that much can be done about that. The fitting code isn't all that swift either, and it may be possible to improve soemthing, but I haven't looked at that at all. There is also a different fitting routine which I haven't considered. *** Note that we currently recommend NOT using the velocity fitting code in the IRAF package. It does not appear to fit reasonable data sets very well, and often produces no fit on a perfectly acceptable line profile. Contact Ted Williams at Rutgers or Bob Schommer at CTIO for different versions of the line fitting routine.