 | |||
|
| Home > Astronomy > LIRIS > |
LIRIS Calibrations
Bias Darks Imaging Sky Flats Imaging Dome Flats Polarimetric Sky Flats Spectroscopic Lamp Flats (for zJ Grism) Spectroscopic Dome Flats (for HK or zJ Grisms) Exposure Times for Spectroscopic Flats Arcs BiasThe bias level of LIRIS is unstable with time, as is the case for most near-IR detectors, and a master BIAS therefore probably will not reflect a later state of the detector. Therefore a bias image (pre-read) is taken and subtracted automatically for every image taken. Further calibration with extra biases is therefore not needed.If you do want extra biases, first move all blanks into the light path to eliminate stray light: SYS> lblanks then to take a bias: SYS> bias liris or to get N BIASes: SYS> multbias liris <N> (Top of page) DarksThe dark current of LIRIS is very small. It is not stable with time, so the detector might be in a slightly different equilibrium state during later exposures than it was at the beginning of the night. A master DARK therefore probably does not reflect a later state of the detector. Separate dark calibrations are usually not necessary since the dark contribution is automatically subtracted from the data during the sky subtraction. This is considered superior to extra calibrations, since the correction is determined from the data itself. A master dark can serve as a very good bad pixel mask, since bad pixels usually saturate quickly in detectors.If you do want extra dark calibrations, take at least 30 darks immediately before and after your observation, and possibly in between if your exposure sequence is long. Reject the first 5 darks from each series since they are not representative. If you want take darks, move all blanks into the light path to eliminate stray light SYS> lblanks then to take a DARK exposure: SYS> dark liris or to take N DARKs: SYS> multdark liris <N> <int time> At the end of the night one can be taking the darks in more automized way. (Top of page) Imaging Sky FlatsIt is essential to take flat fields since the detector is not homogeneously sensitive and shows some cosmetic features and variations of up to 20% across the field. Most of these are corrected well by flatfields, others disappear later on with the sky background subtraction. The flat field itself is very stable with time, e.g. to within a few percent over months.Skyflats are superior to domeflats since they preserve the natural light path, and do not suffer from possibly enhanced fringing effects coming from the flatfield lamps. You can start taking flatfields in narrow band filters shortly before sunset, and in broadband filters a few minutes later. Since the read-out time of LIRIS is short (2 sec) you can easily afford to take a sufficiently large number of flats, such as 20, per filter in order to obtain an optimal S/N ratio. Select an exposure time such that you reach about 20000 counts in the image (post-read minus pre-read). Although ideally one should separately reduce (or discard) the first and second flat exposures (because of the detector reset anomaly), getting 15000-20000 ADU will minimise the DRA and allow one to get the pixel to pixel sensitivity. To avoid stars ending up on the same detector pixels, there are two options:
(Top of page) Imaging Dome FlatsIf it is not possible to take sky flats, dome flats may be used instead. However domeflats can have unwanted spectral characteristics due to the lamp and/or the reflected light from the painting in the dome, so skyflats should always be your first choice.Ask the telescope operator (OSA) to point the telescope to a suitable position. To take dome flats, first open the mirror petals: SYS> petals open Switch on the flat field lamps, which are mounted on the top-ring of the telescope and illuminate the upper part of the dome. You can ask the OSS or OSA to show you how. To switch the weak lamp on: SYS> fflamp 1 on Leave the four additional, much stronger flat field lamps turned off. Even in narrow-band filters LIRIS is too sensitive for even one of the brighter lamps. However, for example z, brg and He1 filters, you should also switch on the second top-ring lamp: SYS> fflamp 2 on Some narrowband filters need even more lamps on (see table below). Then to take N exposures (see table below for orientative exposure times and lamp information): SYS> multflat liris <N> <int time> "dome flat" When done, turn off the lamps: SYS> fflamp all off and finally close the mirror petals: SYS> petals close
Note that, in particular for the K-band filters around 2µm, the illumination level mostly depends on the dome temperature rather than the lamp brightness, so exposures times can vary. (Top of page) Polarimetric Sky FlatsThese must be taken as low on the sky as possible since the twilight sky is highly polarised at larger elevations. Ask the OSA to point away from the Sun and to lower the telescope to about 20 degrees. Even though skyflats are preferable, domeflats can be a good alternative for the more complicated skyflats. See calibration page for imaging polarimetry using the half-wave plates. Or if you want to use the old method without half-wave plates: Take a series of domeflats, then turn the LIRIS turntable by 90 degrees (ask the OSA to do this), and then take another series of domeflats.(Top of page) Spectroscopic Lamp Flats for zJ GrismYou can use the tungsten (W) lamp that is integrated in the A&G box. Or you can also take dome flats (see below).See the Exposure Time Table In order to bring the light from the lamp to the detector we need to move a mirror into the light path, then include a neutral density filter, and switch the tungsten lamp on: SYS> agcomp SYS> compfilta i (with i=4 or 6 for ND=0.8 or ND=3.0) SYS> complamps w In the A&G box mimic the changes you have done are shown. This can also all be done instead from within the Instrument Control GUI window. In the A&G Observer menu tab, select "W" for the tungsten lamp, then click the gears symbol next to it to switch the lamp on (or off). In the Mirror Control section, select Acq/Comp to move the mirror in (or out). In the A&G menu, select the Main ND Filter you desire. In the A&G Observer window, (un-)select the "W" for the Wolfram lamp, then click the gears symbol next to it to switch the lamp on (off). In the Mirror Control section, select Acq/Comp to move the mirror in, or Out to remove it again. Now bring LIRIS into spectroscopic mode by inserting your desired grism and slit: SYS> lspec <grism name> <slit mask name> To take a spectroscopic flat: SYS> flat liris <int time> "W flat" or to take N flats: SYS> multflat liris <N> <int time> "W flat" Once you are done, switch off the lamp, remove the neutral filter and move the mirror out of the beam so that LIRIS can see the sky again: SYS> complamps off SYS> compfilta clear SYS> agmirror out (Top of page) Dome Flats for HK and K (and zJ) GrismsSpectroscopic flats in the HK or K grisms are better done as dome flats due to a broad absorption feature in the W flats. They can also be done for the zJ grism. Even though by observing the dome means looking through just a few metres of air, you will see the very prominent absorption lines due to water vapour, forming the natural separation between the H- and the Ks-bands.Take out the Acq/Comp mirror to allow light to pass into LIRIS from the dome: SYS> agmirror out and open the mirror petals: SYS> petals open Switch on all the dome flat lamps, but leave the main dome illumination itself switched off, since this features lots of emission lines: SYS> fflamp all on In the case of the widest (l10) slit, use only the two faintest top-ring lamps. The lamps can be switched on individually with the command: SYS> fflamp N on where N is the desired lamp (1 = 9W, 2 = 25W, 3 = 150W, 4 = 500W, 5 = 500W). Now bring LIRIS into spectroscopic mode by inserting your desired grism and slit: SYS> lspec <grism name> <slit mask name> To take N spectroscopic flats. SYS> multflat liris <N> <int time> "Specflat" When finished, the lamps can be switched off with: SYS> fflamp N off SYS> fflamp all off When done, close the mirror petals: SYS> petals close (Top of page) Exposure Times for Spectroscopic Flats
These should give you a level of about 20,000 counts (and thus minimises the detector reset anomaly effect). It is recommended to check the count level as these are only orientative values. (Top of page) ArcsIt is usually sufficient if arcs are taken at the beginning of the night, but you can take more before or after your target as well. However, the best wavelength calibration comes from the numerous sky lines themselves, since they are tied directly to each individual exposure.For wavelength calibration of LIRIS with arcs, there is an Argon lamp (lamp 1) and a Xenon lamp (lamp 2). First, bring LIRIS into spectroscopic mode by inserting your desired grism: and slit: SYS> lspec <grism name> <slit mask name> Then switch on one or both of the calibration lamps. If you take arcs in the HK bands, switch both lamps on since the number of lines is low. It takes a short while until the lamps are warmed up and you get the prompt back. SYS> llampon 1 SYS> llampon 2 Move the comparison mirror into the light path to direct the light to LIRIS: SYS> agcomp Then to take an arc spectrum: SYS> arc liris <int time> "Arc" or to take N arc spectra: SYS> multarc liris <N> <int time> "Arc" When done, switch the lamps off and move the mirror out again: SYS> llampoff all SYS> agmirror out
For the lr_zj grism two exposure times are recommended, since there is a small number of bright lines that saturate quickly. Spectral atlas. In text form Argon HK , Argon zJ , Xenon HK and Xenon zJ. (Top of page) | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Top | Back |
|
