Yeah – I remember seeing that when the report was originally published. It notes that both the priming switches and transfer switches could have been knocked during or after impact, but it’s certainly odd that two fuel switches that should have been on were off and two adjacent fuel switches that should have been off were on.
Makes you wonder if he turned the prime switches on in response to the low fuel caution, and believed that the subsequent warnings were simply because the supply tanks hadn’t replenished? Particularly as at several points after the initial caution, the caution cleared itself.
But the overall arrangement does seem very manual and error prone. If I’m reading the report correctly, there’s nothing to remind you to turn a pump back on once it’s resubmerged in fuel, and no immediate warning if, when you turn off an unsubmerged pump, the other pump isn’t running. Seems like it’d be quite easy to end up with neither pump running.
During flight the pilot has to routinely respond to warnings that pumps need to be turned off, but the low fuel caution requires you to turn them back on. All of which needs to be done whilst flying a helicopter in the dark.
The other thing that struck me is that the manufacturers went to the effort of installing NVM to record the last 31 state changes from the warning unit, but didn’t take the extra step of adding any sort of timestamp to the record. Even if completely unsynchronised with a real clock, just knowing the number of seconds between the warnings would make the NVM massively more useful.