Closing the loop between AESOP and Metrology
by Joe Liske on 2022-08-17
One of the more complex functions of 4MOST is the process of positioning all fibre tips exactly on the objects to be observed for every observed field. To accomplish this all the fibres are backlit and their positions are measured very precisely by four cameras mounted on the M2 spider of the VISTA telescope during telescope and instrument setup. Deviations from the desired positions are then sent to the fibre positioner(called AESOP) as corrections. This process is repeated until (essentially) all fibres are in the correct position.
Roland Winkler reports: “After months of development time, the Instrument Control Software is now able to control AESOP’s spines in closed loop with the Metrology system. In the AIP integration hall four Metrology cameras were mounted inside the MetCam Test Frame. This was connected to the Focal Surface Alignment Stand, which in turn contained AESOP.
Getting all electronics and software components ready and working with each other was a joint achievement by AAO, LSW, AIP and ESO. The team had to overcome a large number of obstacles, of which some had been foreseen but many others were unexpected, such as AESOP electronics failures and computer network problems. Through the diligent support of Gerard Zins (ESO), the background software within the ESO software framework was completely rewritten in python to be able to quickly iterate on functionality and for ease of development.
Prior to closing the control loop, we were able to prove the capability of the Metrology system to measure spots with a precision of better than 3 µm. This precision is currently limited by residual errors in the optical model of the test stand (as the flat mirrors have quite a substantial surface profile), and by seeing in the lab. With better quality optics in the telescope, the first problem will go away while we expect roughly the same impact of seeing in the telescope.
Currently, a complete fibre re-configuration takes around 3 minutes for 7 AESOP iterations. The requirement for the combined system is 2 minutes, with a goal of 1 minute. We are confident to be able to reach the 2 minute requirement, as small amounts of measurement performance can be traded for large amounts of time savings during the coordinate transformation step.”
This is a major step forward for 4MOST system testing in Potsdam. While there is still work remaining, it is very encouraging that the system is now fundamentally working and showing promising performance. Congratulations to all involved!