The collimator is left at about 30mm opening (7000 steps from the closed position).
We used the primitive telnet interface to the controller. The GUI program seems to be confused. At the "negative end switch", corresponding to the collimator closed, the program decides to flip the direction of motion and instead of opening, starts closing.
A copy of this log entry has been emailed to: gorbenko@jlab.org, romanip@jlab.org