We had four lock losses from nominal low noise state this afternoon. At least three of them seem to be caused by a suddend excursion in DARM correction, which leads to a saturation of the ESD. The reason is yet unclear.