Panda grid scans: Use the internal Eiger clock for time-based triggers

Our PandA grid scans currently work like this - for each row:
1. Trigger sent to Eiger once PandA PCOMP tells us we have entered the start of the grid
2. PandA uses its own clock to send a trigger to the Eiger at `frequency=exposure time + eiger deadtime`
3. Repeat step 2 for `N=number of boxes per row`

It would be more robust to use the Eiger's internal clock for steps 2 and 3 instead, and it may shed some light on #831 . This would mean the PandA still uses PCOMP at the start of each row, but only sends a trigger to the Eiger. The Eiger then handles its triggers for that row automatically - since it knows its exposure and deadtime, and we can tell it the number of triggers per row

## Acceptance Criteria
- Sucessfully used Eiger uses internal triggering for PandA grid scans
- PandA position data from this scan looks correct