Text Size:
Smaller Text Normal Text Larger Text


Follow us on Twitter


The FIBGUIDER functionality needs to be tested on Technical and Service nights.
Please send us (PS, JT) your feedback.

In order to test the system, you really need to use only a few commands.

In the testing phase, it will be possible to activate/deactivate the fibguider through a simple sequencer command.

  • To activate the fibguider system: setup-fibguider-on
  • To deactivate and return to old system: setup-fibguider-off

Once the system is running, you really need to use only a few commands.

    In case the guider is not centering properly, for whatever reason, and you want to open the guide loop, i.e. to not give telescope offsets:
      fibguide-config guider=disabled

    To close the loop again, e.g. for the next star, use
      fibguide-config guider=enabled

    In case you want to accept the suggested fiber position, i.e. the blue cross, as the new reference position, i.e. green cross, use

    In case you want to update the green-cross position manually, use
      fibpos-man <fibernr>

In case the system is doing a good job, and the telescope offsets (see green talker messages) are fine, then you just do nothing at all.


The fibguider uses the target's position on the StanCam fibmovie images to tweak the telescope position. In principle the fibguider works with and without the main guiding system, but the movie and guiding frequency is too low to do proper guiding on its own. So in its normal operation, the fibguider provides a secondary guiding loop, that corrects the main NOT guiding system.
Secondly, the fibguider speeds up target acquisition, as a high-gain telescope correction is done on the first image of an acquisition movie sequence.

Once the fibguider system has been activated, every fibmovie image will be analyzed by the fibguider and the results overlayed on the DS9 image. The initial acquisition of the target in direct StanCam imaging mode remains the same, but once the pickoff mirror has moved to FIES (aka split position) and the first fibmovie image arrives, the fibguider will be working the following way:

  • 1'st image: fibguider will find the brightest target in the FOV and apply one single offset to bring the target onto the fiber.
  • 2-11'th images: The fibguider will be working with a very low gain, meaning that any offsets are hardly corrected for. The reason for this becomes clear(er) below, but basically at this moment of time the fiber position is yet undetermined and could be wrong with some few pixels. Setting a low gain allows the user to correct for this offset using the orange cursor keys without the fibguider pulling the star away again. It may be possible already at this early stage to redetermine the fiber position manually (see below).
  • 11'th image onwards: The fibguider parameters are set to the defaults (see below) which gives a well-behaved guiding. The fibguider will also after every incoming image make an analysis of the last 10 images to determine the fiber center. The suggested fiber center will not be applied automatic, only displayed on the DS9 image. It is up to the user to decide whether to use the computed center or not. The command to use for adopting the computed fiber center as the new reference is fibpos-update

Similar to quickspec, a configuration file is used to set certain parameters of the fibguider routine. The sequencer script is fibguide-config, and called without any parameters will list the current config, for instance
tc=60, gain=0.5, guider=enabled, analyzer=enabled, outermask=big hicut=auto

  • tc: Time constant in seconds used in the low-pass filter applied to the offset. Keep at recommended value of 60 seconds for now. Be aware that no actual telescope offsets are applied faster than this number.
  • gain: The percentage of the offset value that is actually transmitted to the TCS for teloffset. 0.5 seems a good value.
  • outermask: Circle on the DS9 image which limits the analysis region, useful in case the image got more than one object visible. Values can be big,small or disabled.
  • hicut: The hicut value on the DS9 image can be set automatic based on the image analysis. auto or manual
  • analyzer: Always keep enabled! Disabling will disable image display!
  • guider: Apply the offsets to the TCS, or run the fibguider in 'open loop' mode. There are circumstances in which the fibguider will not produce useful results, such as having multiple objects within the circle described by the outermask, in which case the only sensible thing to do is to disable the fibguider (fibguide-config guider=disabled) and observe with normal autoguiding only.

If the above mentioned fiber-position determination routine suggests a reasonable improvement for the fiber position, you can accept the suggestion by simply using fibpos-update. The fiber-position determination routine can also be invoked on the command line by running the sequencer script fibpos-compute [filename] [fiber] [no_images_to_analyze] which by being supplied with the above parameters will run in an interactive mode, bringing up a plot of the fiber position analysis. And then the fiber position can be set totally manually, through an imexam session of the loaded image, by executing fibpos-man [fiber]. In case the fiber positions are completely messed up, some approximate default values can be set by executing fibpos-snafu

The fibguider will make several region plots onto the DS9 display which need to be understood:

  • Green Cross: Reference fiber position. Offsets are calculate and applied with respect to this point. Hence, this cross should always indicate the actual center of the fiber.
  • Green Circle: The outermask circle, outside which the image analysis will not take place.
  • Blue Cross: Calculated fiber position, based on last 10 images. Blue cross will not be present until after the 11'th fibmovie image. From this moment onwards, the cross will update after every incoming image. Issuing the fibpos-update command will convert the last computed value to new reference value.
  • Red X: The calculated centroid of the object, based on the PyGuide rountine. A new centroid is calculated for each incoming image. In case no centroid can be calculated (too faint, too cloudy, too much down in fiber), no X will be shown.

Keep an eye on the Talker. All important information about the fibguider is reported there.


  • You know the fibguiding is doing well when all markings fall on top of each other and that that point corresponds to the true fiber center as seen on the DS9.
  • When the computed fiber position (blue cross) consistently are offset with respect to the reference position (green cross), and this new position seems to better reflect the fiber center, it's time to update the fiber position with fibpos-update.
  • Be patient. The system is designed to react slow and conservative. Last thing we want is actually pulling the star away from the fiber.

Back to top Last modified: 27-Nov-2013