In early 2014, during the first run of Semester 2014A, our team discovered that stars imaged with the Gemini South Adaptive Optics
Imager (GSAOI) were elongated especially at
the edges of the field, yielding poor performance (up to 250 milliarcseconds). One of
the issues uncovered was an incorrect procedure used for saving Zernike coefficients
(that control the figure of the Gemini South
8.1-meter primary mirror (M1)). Figure 1 illustrates the image elongation issue.
Once the M1 model was fixed, we still had
a somewhat intermediary situation, where
elongation was present, but to a lesser extent. After a few simple tests, we determined
that the LGSWFS system was the problem,
since running GeMS with only the Deformable Mirror (DM) conjugated to 0 km (DM0),
used essentially in a ground layer adaptive
optics mode, showed no signs of the elongation seen while running in a full MCAO
mode with DM at an altitude of 9 km (DM9;
Figure 2).
This indicated that we had something odd
coming from the LGSWFS that was being
offloaded to the DM when conjugated to an
altitude of 9 km (DM9). During two further
runs in 2014 (May and June), we acquired
some science data using a software patch
that removed a semi-static shape on the
DM9. As several optical phenomena could
cause this issue, the AO team decided to
shutdown the GeMS system in concert with
the general telescope shutdown and proceed with a more thorough investigation
inside Canopus — the AO bench that is the
heart of GeMS.
To maximize the effectiveness of this shutdown, we approached previous GeMS AO
team members François Rigaut, Benoit
Neichel, and Marcos van Dam, all of whom
agreed to assi 7BF