1. There are problems with the right side of the COSMOS detector (problems with the electronics). Occasionally, that part of the frame is unusable. There was an attempt to fix it by re-connecting the cables and restarting the controller, but it did not help, just took 30min of time. Example image: n1.64615.fits. Also, there is a strong materess-like pattern at +-15 counts visible in nearly all images (check any slit image or bias).

2. There were several situations with the stuck readout, which required us to restart all the software. It takes a few minutes each time. And we lost several exposures (standard stars) because of that. The last exposure of the night failed for that reason but it was too late to retake it.

3. The alignment procedure takes forever (30-50min for a long slit) and often does not converge. The worst situation is a bright standard star that saturates in the alignment images. There must be a procedure to observe such targets (at least aligning on a particular pixel rather than using the IRAF imexam which does not work).

4. There is a strange drop in flux in the last exposure (HIP38095, n1.64648): either there is a condensation on the glass somewhere or something else (wrong) is going on.