CFHT, Instruments, Spectroscopy, Fabry-Perot, Observing Procedures.

Canada-France-Hawaii Telescope

User's Manual for Scanning Fabry-Perot Spectrography

R. Arsenault, P. Martin
Version 3.2
December, 1998

3 Observing Procedures

3.1 Getting Started on the Sky

Sign On

To have the proper Pegasus environement set up, you must login on the X-terminal in the control room on the account: mosfp. Ask your support astronomer for the password. This activates your Pegasus session and a top menubar will appear on the screen. The buttons displayed below are the same that you will see on the menubar. As for the real Pegasus session, you can click on these buttons, and a window will pop-up, a process quite similar to your Pegasus session.

Startup Checks

Upon login and before to proceed with various operations or calibrations have a check at the various status displays: disk space and CPU load. You should start your run with a large disk space. If the display indicates larger than 90 % full, it is your responsibility to clean up your directory. Use the FILES button for this purpose. Sometimes other accounts (for other instruments) might take up disk space. If you removed all unnecessary files and the disk still appear quite full contact the software group. If during acquisition a process appears extremely slow, have a look at the CPU load. You might have a runaway process. Before to take exposures make sure the CCD raster is defined according to your needs.


MOS or OSIS are focal reducers and this implies that 2 focii have to be checked: the matching of the telescope focus with the entrance focal plane of MOS or OSIS, and the matching of the MOS or OSIS camera focus with the detector. The latter is set up by CFHT's staff prior to the visiting astronomer arrival. The (or "OSIS-FP") button in the Pegasus menubar contains a field that controls the camera focus. The value varies from 1 to 5000 and the units are 1.25 micron.

It is the task of the visiting astronomer to focus the telescope. The value changes from night to night (even during a single night) or after large change in telescope pointing and therefore the visiting astronomer must know how to perform this task.

First ask the T.O. to point the telescope on a 5 to 8 magnitude star close to your field, (the appropriate star brightness depends on your interference filter-etalon combination). Move the telescope focus to 1950 (good guess to start with; the MOS camera focus has been set up by the technical staff). Localize your focus star on the CCD (normal exposure) and then select an appropriate focus raster in the FRASTER form (100x100 pixels is usually adequate; remember DO NOT BIN THE CCD). Use the form, and select: focus analytical. This will display the stellar image in SAOimage, and automatically measure FWHM and other parameters (see the button to determine which ones). Move slowly the telescope focus by pushing the appropriate buttons on the handpaddle. You can either do a careful focussing in this way (steps of 5 units), or a rough estimate (steps of 15-20 units) and then use to obtain a finer focus. Be prepared to repeat this operation after large movements of the telescope (more than 30 degrees).

There is a slight difference between focussing with MOS/FP and OSIS/FP. For both instruments, CFHT technical staff focusses the camera in the afternoon preceding your run. Then the visiting astronomer focusses the telescope to get a sharp image on the detector. This procedure is valid for MOS/FP only. For OSIS/FP, there is the additional constraint of the guide probe (for fast guiding) which implies that the telescope focus, once set without a filter in the entrace focal plane, must not be changed once a filter is inserted. The only choice is then to change the camera focus in the "OSIS/FP" window. This causes a slight de-collimation on the Fabry-Perot etalon, but calculations show that the loss in spectral resolution is minimal.

For FOCAM/FP on the AOB (Adaptive Optics Bonnette) the technical staff setup the AOB WFS (WaveFront Sensor) so that when the Adaptive Optics correction is active, the image is focussed on the detector. There is no need for the observer to run a focus sequence for the telescope when the AOB is used.

Focus Changes

Remember to adjust the telescope focus for different filter thicknesses. The visiting observer must calibrate the focus change for each of the filters, and manually change the telescope focus at each change. Do not change the MOS camera focus to adjust for different filters. However, it is appropriate to do so for OSIS/FP. The formula below is valid only for MOS/FP.

General Rule: Insert (thicker) filter ==> Increase Cass focus value

Focus change: DelF = (n-1)* t/n

where DelF is the focus change (in mm), n is the refraction index of BK7 and t is the filter thickness. The cassegrain upper-end focus value unit is 22 micron.

CCD Control

The operation of the CCD is described in detail in the FOCAM User's Manual@, please refer to this manual for further information. The basic operations, Bias/Dark/Flat/Object/Comparison exposures are easily performed from the form. After selecting ACCEPT on that form, always wait 20-30 seconds for the small camera window or exposure status window to appear on the right of the screen. At the end of an exposure check in the Feedback window that the message 'successful exposure' has been written, and keep track of the exposure number for your observing log, and later transfer to tape.

Before starting your calibrations or sky exposures make sure you have selected the appropriate subraster and binning factor ( button).

3.2 How to Scan

Before to explain the various way of scanning offered by the software, the observers should be aware of the effects of differential refraction when using MOS/FP or OSIS/FP. Atmospheric refraction depends on the zenithal angle and occurs for all instruments. However there is one reason why it is more critical here. Both for MOS/FP, OSIS/FP and FOCAM/FP (AOB) exposures can be very long. One scans many channels and it is not unusual to spend more than 5 hours on a same object. During this time, differential refraction may cause the objects to drift on the detector. To avoid this make sure that the guiding is done in a similar color as your observations. The Cassegrain bonnette offer the following choice of filters:

3  RG610    3 mm    > 600 nm
4  OG515    3 mm    > 500 nm
5  GG420    1 mm    > 400 nm
6  BG12     4 mm    320 nm to 500 nm + 1% red leak

There are 3 possible ways of scanning offered by Pegasus: manual, full and partial.

The first one is the most basic but requires the osbervers to do more homework. One simply writes in the Pegasus "Fabry-Perot" form the BCV (Binary Control Value) that will be sent to the CS100 (Fabry-Perot Controller). The BCV is proportional to the spacing between the FP mirrors and ranges from -2048 to +2047, which covers 1.7 interference order at Halpha.

The second method is used when a full CCD field coverage or full spectral order is needed. The user provides only the number of channels. The number of interferograms or channels to be taken is determined by the finesse of the etalon. It must be 2.2 multiplied by the finesse. This criterion (the Nyquist criterion) insures that the spectral profile is adequately sampled. The options for a, so-called, full scanning is with overlap or without. With overlap means that the last channel provides the same spectral information as the first one, only one interference order higher. This helps to see if exactly one interference order has been scanned. The option "no overlap" means that the next to last channel, if taken, would be the first plus one interference order. The last option offers a sequential scanning mode from 1 to N, while the interlaced option proceeds as such: 1,7,13,19...2,8,14,20...3,9,15,21.....N. The latter method insure a "pseudo" monitoring of the sky transparency by scanning repeatedly the full interference order.

Given a number of channels m=2.2xN, there are 2 possibilities where one might not have to scan m channels:

- the science object is limited in extend (spatially),
- the field of view is completely covered by a fraction of a full interference order.

The first case is clear, a small quasi-point source, distant galaxy, or a close spiral galaxy arm is observed. A few channels, well chosen are sufficient. The technique consists in setting the FP at the BCV value that brings an interference ring in spatial correspondance with the object. Then a few channels are scanned. The Pegasus partial scanning algorithm carry this task for you. The partial scanning option fulfills this purpose. However, a careful calibration must be made. The interference order of the etalon, the calibration wavelength and the wavelength one wishes to scan must be accurately known. The BCV value requested in the form is the setting for which a calibration ring appears on the center of the detector or field. Actually it can be anywhere on the detector, but this needs to be calibrated at the beginning of the night (with the help of your support astronomer). The following formula relates the required BCV to the calibration wavelength, the wavelength observed and the calibration BCV. It is the formula used in the Pegasus partial scanning algorithm. You must set the Fabry-Perot controller (CS100) Z channel voltage to zero. This insure that the etalon is at the nominal settings (the spacing for which the interference order is the one specified by the manufacturer).

The second case occurs when the field of view is very small or the etalon interference order very low. Only the very central portion of the ring pattern can be seen. Therefore, only a few channels are required to construct the phase map. In this case also less than 2.2xN channels are required. This also assumes that one does not need the complete spectral coverage offered by the etalon, or the free spectral range. If the spectral features of the objects under study are of the order of the free spectral range of the etalon, then a full scan of 2.2xN channels is required.

To conclude, a careful evaluation of your observing need is in order. The spatial coverage and spectral coverage of your data set are not independent.

Be aware that no error message will appear on the Pegasus session if the CS100 loop opens. You should routinely check it up when you go to the 5th floor (led on the CS100).

3.3 Calibrations

The following calibrations are needed:

A set of at least 10 bias frames at the beginning of the night, plus several biases taken during the night,
Dome or Sky (dawn) flat fields, at least 10 per night,
Spectral calibration with a monochromatic lamp, with the same scanning procedure than for the sky. A light level of about 10,000 - 20,000 electrons per pixel is recommended.

Two Neon calibration lamps are installed in the Gumball device. One of them is coupled with a filter to isolate a line at 6598.953A, which is usually used for Halpha work. This configuration eliminates the need for a calibration filter in MOS or OSIS filter slides. For more redshifted sources, Neon has another line at 6717A. If you need other calibration lines or lamps for your program, contact your support astronomer well in advance of your run.

To carry out a calibration scan you must select the exposure type "c" (for comparison) in the form. The calibration lamp will be switched on when the exposure starts. The lamp or lamps selected in the form will be switched on at that time. Refer to the calibration system page (Gumball) for further detilas concerning the utilisation of the calibration spectral sources.

Since the spectral bandwidth of this instrument can be very narrow (0.5 A) the flux level will be low for flat fields. Besides, there is a clear preference among "fabry-perotists" to obtain a scan of flats (same number of channels as for the science observations). This technique insures that one does not correct the science frames with a single flat corrupted by an eventual spectral features in the white lamps. Therefore, sky flat might be simply too time consuming for fabry-perot observations. We advise the use of a series of white lamps located on the f/8 upper end. They can be controlled from the 4th floor auxiliary control room for dome flats. Ask your support astronomer or telescope operator for the details.

Because of non negligible phase effects in the multilayer coatings on the Fabry-Perot plates, the interference order P at a wavelength W cannot be accurately computed from the known values of P' at wavelength W' by the usual formula: PW=P'W'. It is thus strongly advised, when only a partial scanning (i.e. less than one interference order) is sought for, to use a calibration wavelength as close as possible to the spectral region actually scanned.

Return to the Table of Contents
Please send comments to: