After we realised that a CA05 input block had dropped out in IF1, we switched to recording DAS2 with channels 5→8 at about 22:23 UTC. It actually dropped out at 21:09:28, but CA05 was left in the tied array, meaning the data might not be very useful.