Achieve Better Co-pointing, Pointing, and Collimation on your Science Targets.

Or: How to take better data with LBC.

LBC now operates in a fully binocular sense, so there is no longer any designation of "master" and "slave" sides or guiding of the mount as was done in the past. Effectively, the mount now tracks the RA,DEC coordinates you supply in an open-loop "TRACK mode" sense and the two LBCs "guide" by executing coma-free pointing offsets with their respective primary mirrors. You minimize the demands on the guide loop by making sure the mount correctly knows where it is pointing, ensuring the two sides are co-pointed, and collimating both mirrors. This can be achieved by modifying the normal sequencing of observing to allow for a correction of the pointing and co-pointing, along with the collimation, near your science target field.

First, select an easily identifiable (isolated, 10-12mag, fainter is OK as long as it is identifiable) source with proper motion corrected coordinates as the "co-point" target that you preset to in the co-point OB. This can be your science target itself if it is visible, or a star in the same field, or if your science target is significantly extended it could be a nearby field. And please do your queue observers a favor and make a finding chart identifying the pointing source you selected! Additional details are given below, but we start with a brief outline:

Prepare yourCoPoint.ob with the following parameters:
  • Use any source with known coordinates as the "target"
  • For collimation, DOFPIA assumes the following are true:
    • Collimation targets are unresolved point sources (stars)
    • Field is uncrowded (not much contamination from overlapping pupils)
    • Field is free of significant background gradients (large galaxies, nebulosity)
    • Example of a bad field for DOFPIA can be found below
  • Make a finding chart (5'x5' from DSS is fine) identifying the source
  • Read full CHIP2 ccd, the others are not needed
  • 0.0 focus offset
  • This is NOT a focus OB!
  • 1.0 sec exposure
  • Same PA as your science target
  • Use r-SLOAN (Red) and V-BESSEL (Blue), same as dofpia

Prepare yourscience.ob as before

Execute the observations in this order:
  • yourCoPoint.ob (interrupt the OB, you only need the preset executed)
  • dofpia, /X2 (the X2 doubles the exposure time to 32 seconds)
  • yourCoPoint.ob (allow it to complete and take the in-focus images)
  • run lbcrangebal on these images
  • yourscience.ob

Send a list of:
  • UTDate of observation
  • The ID and coordinates of your pointing reference
  • The timestamps of the images

Additional considerations

The Details:

On traditional (one-eyed) telescopes one does not have to worry about co-pointing at all, and the pointing and collimation are well understood such that one can often "just" observe. Our asymmetric telescope design and extremely fast optics makes the LBT more sensitive to thermal effects and relatively rapid drifts in collimation. Until these effects are better understood and modeled, the best approach to observing with LBC will be to locally correct the collimation, pointing, and co-pointing just before taking your science data. This is achieved by preparing a "co-point" OB that provides a suitable coordinate reference source near your science target as well as multiple stars usable for collimation.

Properly pointed and co-pointed, the telescope pointing control system (PCS) will accurately generate the tracking polynomials that drive the telescope mount to follow the field. However, our sensitivity to thermal changes means our pointing and co-pointing can drift off this just-corrected position, by up to several tens of arcseconds over the course of an hour or less. This would result in differential field rotation, where the sources in your science images follow a circular path with a radius equal to this drift off the position the mount is tracking. The rate the sources cover this circle is proportional to the change in parallactic angle over the exposure, so the sources will trail over an arc of length r*Theta (radius in arcsec times the change in parallactic angle in radians).

Guiding can correct this trailing if it happens slow enough. You are thus most sensitive to this positional drifting in the pointing and co-pointing when the parallactic angle changes the fastest (meridian crossings, especially at high elevation) as well as when the telescope is out of thermal equilibrium (just after opening of the dome, as well as changes in the weather). Correcting the pointing and co-pointing immediately before and very near to your science observations will minimize the guiding corrections that would be necessary and produce better images.

Co-Point OB

A suitable coordinate reference can be any easily identifiable, relatively isolated, source with accurate coordinates. Almost any star in the USNO-B1.0 catalog would do. In crowded fields it would be a good idea to provide a finding chart as well so that identification of the correct source is assured. Because the coordinate reference source is relatively bright, only a short exposure is needed for the pointing correction. This co-point OB should be set up with the parameters listed in the summary above.

Note that very crowded (i.e. Galactic plane) fields can be problematic for the collimation script as the out-of-focus pupils overlap. Select your fields carefully.


Send the yourCoPoint.ob but interrupt it because you only need to send the preset to the TCS and do not need to take the exposures with LBC yet. Collimation needs to be done before the pointing correction as the collimation process for LBC is not done in a pointing-free manner in order to preserve range of motion of the primary mirrors. Use the IDL procedure dofpia to iterate on the collimation using extra-focal pupil images obtained by driving the primary mirrors -0.8mm down (away from LBC). Note that DOFPIA assumes that your collimation field contains only uncrowded stars and the background is reasonably uniform. Fields containing any significant nebulosity or a large galaxy are unsuitable. Any nearby (within a degree or so) field can be used that satisfies DOFPIA's assumptions. Run dofpia with the /X2 switch to double the usual exposure time to 32 seconds (this will be made the default at some point). The script takes an exposures with both cameras and reads out only the top ~5 arcmin of chip 2, finds relatively isolated pupil images of sufficient signal-to-noise ratio and derives corrections to the wavefront. After waiting for the corrections to be applied, dofpia will check the corrections versus convergence criteria and iterate as necessary.

At the beginning of the night, and any other time you are out of thermal equilibrium dofpia will struggle to collimate the telescopes. Once things have settled down, dofpia will converge quickly at each new field, generally within two or three iterations.


Once collimated, re-send the yourfocus.ob and allow it to complete. An iraf script is available in the LBTtools/Observe package called "lbcrangebal" to do the optimal pointing and co-pointing corrections. This script takes as input the LBCB and LBCR filename time-stamps for the images taken in the co-point OB. It will read from the telescope control software two current pointing model parameters (IE and CA), and from the image headers the pixel coordinates of the mechanical rotator axis' intersection with the telescope focal plane for both cameras. Tip-tilt requests to the two primary mirrors and an offset to the pointing model are calculated in a manner that maximizes the available range of the primary mirror while simultaneously correcting any common offsets from the pointing model with an offset to the mount pointing model.

Science OB

At this point you have completed the collimation, pointing, and co-pointing so you are ready to submit your science OB. We recommend you design your science observing scripts to complete in 30-40 minutes so that dofpia can be run again, between the science OBs.

Additional Considerations
  • Run dofpia more frequently in poor conditions
  • Suitable star on the displaced tech chip?

Additional Work before Release?
  • Add links to other pages for details on dofpia and lbcrangebal

-- DavidThompson - 23 Sep 2011
  • Bad_DOFPIA_Field.png:

Example of a bad co-pointing field on which to run DOFPIA. The gradient in the background from the large galaxy goes from 35K to 3K ADU. At a somewhat harder stretch one can also see pupils from fainter stars overlapping most of the brighter pupils. Both of these cause problems for the DOFIPA routine. Please select co-pointing fields free of resolved sources and reasonably uncrowded, DOFPIA needs clean pupils and a flat background to correct the telescope collimation.
Topic revision: r12 - 29 Oct 2013, DavidThompson
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback