PNS MANUAL: Focussing the Spectrograph

Focussing the spectrograph

The PN.S has a mask which can be flipped into the focal plane. It consists of a regular grid of holes, with a small slit in the center of the field. By illuminating the mask with an arc lamp, focus exposures can be taken so that each arm may be focussed onto the mask. Later on, at night, the telescope is also focussed onto this same focal plane.

Focus adjustments are made with three capstans on the mount of each CCD. The procedure described below first finds a focus in the center of the field, and then corrects for tilt of the CCD with respect to the image plane.

Usually the CCDs will be close to focus after being put on the PNS by the ING team.

Focussing probably has to be re-done after any filter change! For a list of focus offsets between filters, see here.

It is simplest to focus one arm at a time. It may take several hours to complete this task!

  1. Window the CCDs. Fast readout speed may be used for focussing.
  2. Make sure image viewer (ximtool or whatever) is viewing the whole chip: set stdimage=imt4096
  3. Note the initial positions of the capstans (they are gradated, but it is not possible to get an absolute reading from them as their full range is more than one turn).
  4. Loosen the locking screws if required
  5. For initial starting point, manually set the capstans to positions such that the mount is evenly spaced away from the instrument. A poor starting position can result in an inability to focus the spectrograph (because of finding local minima?)
  6. Put in the mask and open the shutter: issue CALIB command on PNS console.
  7. Issue AGCOMP on the A&G console.
  8. Take an arc exposure (ca. 3 minutes): issue ARC   PNSL   (or PNSR) 180 on DAS console, then switch on lamps (COMPLAMPS   CUAR+CUNE on A&G console). Switch them off (COMPLAMPS   OFF) before end of exposure.
  9. Now do a focus sequence: turn all capstans by the same amount and take new exposures. Measure the fwhm of a bright arc line in the images (using the "j" key in imexam). Avoid the corners, where the image quality degrades a bit. Best measure the spectra from spots diagonally inwards from the corners (3 in x, 3 in y) and near the center. As holes in the mask may have different sizes, be consistent about which holes to measure. Record the measured fwhms and corresponding capstan turns. Typical focus steps are 0.1 - 0.2 turns. Note it may be desirable to adapt this procedure to using the tungsten lamp since the exposure time would be much shorter.
  10. Once the best average focus has been found, look for systematic changes in focus that may indicate tilt. Tilt causes a systematic degradation of focus away from a line through the center of the field. Decide which direction to correct.
    • A tilt about a vertical (i.e. spatial direction) line can be corrected by turning capstan 1 and 2 clockwise, and 3 twice as far counterclockwise.
    • A tilt about the horizontal (dispersion) axis can be corrected for by turning 1 and 2 equal amounts in opposite directions, and leaving 3 alone.
    Typical steps for tilt corrections are 0.3 turns.
  11. Perform a sequence of measurements at different tilts, until the focus is as uniform as possible.
  12. Finally it is a good idea to tweak the focus again, without changing the tilt. FWHM should be close to 2.5 pixels over most of the field when the focus is good.
  13. Now repeat for the other arm!
Once the focus and tilt are properly corrected, lock down the dewars. This is done by winding back the locking nuts, then gently winding down the screws with finger pressure only until they contact the black flange (visually check this). Then wind down the locking nuts to eliminate the possibility of the screws loosening.


It seems that filter changes affect the focussing. Here are the offsets determined for the A -> B change, i.e. the changes required if the spectrograph was in focus with A and you change to B.

All turns counter-clockwise
L1 0.925 -> 0.45
L2 0.875 -> 0.40
L3 0.475 -> 0.00

R1 0.15 -> 0.725
R2 0.85 -> 0.425
R3 0.75 -> 0.32

On 23/04/2012 we used the following offsets to change from AB0 to A0
Offsets (- is counter-clockwise):

     AB0          applied      A0
   (absolute)     offset    (absolute)
L1  0.10          -0.10       0.00
L2  0.05          -0.10       0.95
L3  0.25          -0.10       0.15

R1  0.80          -0.10       0.70
R2  0.80          -0.10       0.70
R3  0.70          -0.10       0.60

PSF is about ~2.9 pixels (a-imexam, moffat profile)

Here are the offsets for the AB -> B change
0.5 turn counter-clockwise
We thank an anonymous author

About this Document

We need your input to keep these notes up-to-date. Please add to this document, or create new ones which can be linked from it. Send your input by email. Last updated 28-03-2006

Documents of this background colour are part of a set linked from the manual