III. A basic observation with PUMA

Before the run, check the different connexions showed in fig. II.3.1.
The CS100 with PUMA.
The FP with the CS100 controller.
The Sun station ESTOPA with PUMA.
The connexion to the power of PUMA and CS100.
The dry Nitrogen with the FP.
Then to put the power on, you may:

- Check that the red switch on the rack (see fig. II.3.1) is on.
When it is PUMA is moving the filter wheel to the position 0, the  calibration lamps wheel to the first lamp and put the FP out the optical path.

III.1  Observation strategy
A basic observation with PUMA is mainly composed of 2 parts: the calibration lamp acquisition and the object acquisition itself the object acquisition itself. The calibration is crucial to give the wavelength origin building the phase map. From the quality of this phase map it will depend the quality of the observations. Of course due to the fact that PUMA is using a CCD detector, it is necessary to go through the standard procedure of bias and flat fields. The flat field should be done for each filter used during the night.
 
III.2  A session PUMA
III.2.1 Description
The acquisition system is operating through an Xwin environment. From the Linux PC ALPHA in the account observa (login: observa
password: xxxxx), you click on the icon "Puma", it appears the main dialog window (Figure III.2.1.1) and a small window where you have to write the number of your session like this: p###, where ### is the number of the current observation, then return (Figure III.2.1.1bis).  It will also create a sub directory p###, in /home/observa/puma/work where will be saved your observations.
When you are a PUMA session you have to save the parameters first by clicking "escribir el archivo de parametros" even if you did fill out the file.
 


 

Figure III.2.1.1
 
 

Figure III.2.1.1bis

From the panel showed in Figure III.2.1.1 you can control everything, we are going to detail the different options:
 
  • Parametros generales : With this option you will fill up the general parameters of your observation. The figure   III.3.1.2 shows the window.
        By clicking "escribir" you can write parameters on 
        a file cip###.
Figure III.2.1.2

 
 
  • reductor focal : This option give you access to all the mobil parts of PUMA. There are 4 sections: "Espejo", "lamparas", "filtros", "Etalon". See figure III.3.1.3.
    • Espejo : by clicking on "Adentro" or "Fuera" you will put the mirror in or out of the light path.
    • lamparas : here you will control the calibration lamp. It is wise to click on "inicializa" to do a reset before moving the lamp wheel. You can choose between Neon, Helium or Hydrogen lamps. You will switch the lamps on or off by clicking on "Apaga" and "Enciende".
    •  filtros : this section allows you to change the position filter. Here too it is good to do a reset with "inicializa".

    • Etalon : You can remove or put the FP in the light path by click on "Adentro" or "Fuera". "Configuracion" allows you to configure the FP for a calibration or an observation. "Barrido del Etalon" initiates the scanning of the FP and "Reset" is doing a reset of the CS100.
Figure III.2.1.3
 

 
  • Fabry Perot : This option allows you to fill FP parameters. The QG constant is 10.74 and the order at Ha is 355. Those 2 parameters should not be modified. You may after put the different wavelength: Calibration wavelength, Nebula wavelength, the wavelength you want to scan and the systemic velocity of your object.  You may specifiy the number of channel you want to scan (Numeros de  canales) and then click on "Calculo de los parametros". See Figure III.3.1.4.
Figure III.2.1.4.

 
 
  • CCD : From this window you have the control on CCD parameters. You can choose the binning you want, the field you want to use: all of it or binning you want, the field you want to use: all of it or just a part of it. See Figure III.3.1.5.
  • For a calibration cube (cie) the gain (ganancia) should be 1 and the time exposure 0.1s. For a science cube (cio) the gain could be 1 or 2.
Figure III.2.1.5.

 
  • The Octagon window is showing all orders than you are giving to PUMA and the answers. If the message "time out" appears, repeat the operation.
Figure III.2.1.6

 
 

III.2.2 Parallelism of the FP
To observe it is necessary to ensure that the 2 plates of the interferometer are perfectly parallel. This proceduinterferometer are perfectly parallel. This procedure is done trough the CS100 controller. The procedure of parallelization of the FP plates should be done by the technical staff or the resident astronomer present just before the observing run.
Some typical values of the CS100 controller are given in Tables III.2.2.1.and III.2.2.2. The final tuning of the parallelism will be done from the control room trough PUMA Xwin dialog box.
 
 
 
  Fine 
  Coarse 
  Balance 
 X 
4.50
-0.2
4.04
Y
4.50
0.0
4.58
Z
1.94
3
3
3.88
 
     Gain 
     Time Ct 
 X 
32
25ms
Y
32
25ms
Z
32
25ms
Table III.2.2.1
Table III.2.2.2

NOTE: When the humidity is larger than 90%, the servostabilized control could not work in order to hold the parallelism. DO NOT USE THE INTERFEROMETER UNDER THIS CIRCUMSTANCE.

III.3  Visualisation with CIGALE
To visualizisation with CIGALE To visualize an image, you have to run the window  "CIGALE" from the main PUMA window. It then will appear an Xterm terminal where you can run the different command of CIGALE reduction package. With civis you can visualize a cube, or a channel. Figure III.3 shows an example. Writing the command civis cie (or cio), you will be able to visualize the observation currentlly in the memory of ALPHA. Writing civis ciep###, you will visualize any cube you previoully saved in Cigale format.
By clicking on the image loaded after writing a civis command with the mouse left button, you change the low cut and the high cut
of the image with the minimum and the maximum on the pixel you clicked. With the up and down arrows you can change the
cuts. With the left and right arrows you increase or decrease the channel number.
 
 
  • With civis ze=+1>With civis  ciepXXX you can visualize a cube like shown in figure III.3. This figure is showing an example of calibration ring.
Figure III.3
 

III.4 Running an observation

Once the technical staff has set the parallelism, you may begin the observation. You first should focus of the telescope and center the field of PUMA.
 
  Before doing the observation, verify the general parameters, the focal reducer parameters, the FP parameters and the CCD parameters.
You may font size=+1>You may begin by doing a calibration exposition. To do this, move the mirror in  (espejo adentro), choose the desired calibration lamp and switch it on. Then choose the time exposure. Click on "cie" on the main PUMA window and put the first channel: 0 and the last one: 47. Click "exponer cubo" to begin the exposition.
 
  • IMPORTANT: at the end of the scan, you have to save the data cube clicking on "Salvar cubo", the program DOES NOT SAVE OBSERVATION AUTOMATICALLY. You can save either your files in FITS or cigale format. See figure III.4.1
Figure III.4.1

Usually it is wise to do 2 calibration exposures, one before the science exposition and the other after in order to check eventual shifts on FP settings. But due to the long time taken for an exposition (exposition itself, CCD reading, images transfer to ALPHA...) and the great number of channels to be done (48), only 1 calibration exposition could be made (before or after the science observation).
The procedure for a science exposition is the same, except that you may change integration time and select "cio" on the main PUMA window.

A complete package of data reduction has been developed by  J. Boulesteix from Marseilles Observatory (France). ADHOC package can be transfer from the web.

You will find a log sheet of PUMA observation here .





 
  body>