This morning during the maintenance I re-measured the OLTF of the mechanical loop of the MC when using the error signal coming from the arms. The delay is still there (see entry 46076). The origin is not understood.