Deliver Fly Scanning - Real Beamline

[gkalua]

Goal: Make Flyscanning work on a real beamline

• Speak to HS about Tuesday sessions
• Audit the current flyscanning on i13
• Speak to MX about how they do Flyscanning (I04, I03)
• Consolidate I13 plans with Visr plans
• Increase bit depth -(cannot be increased)
• Create issue for int16 work around
• Look at FastCS-Eiger-Odin detectors
• Look at F-E-O instantiation

[gkalua]

Spoken to HS about planning and auditing. Meetup scheduled for the 26th of May.

[gkalua]

• Visited i13 on 26-06-2026
• Encountered the issue with pmac parenting; was fixed by changing pmactrajectory... class from a device.
• Turns out the workspace version of ophyd-async was outdated
• The plan is to do as much testing as possible over the next three weeks (shutdown)

The to-do list for the period is as follows;

• Fix a panda bug (that has now disappeared, without recreating)
• Fix any bugs that come up
• Confirm that the Eiger-Odin detector can work with fastcs and with ophyd,
• Try the fastcs panda IOC on the beamline
• Consolidate the plan to be similar to the visr plan
• Saw that running a 20KHz scan breaches the bit depth limit - hence a suggestion to tweak it, need to understand why it is 16 atm.
• Might be 16 as a panda default, a work around can be to create multiple rows when it exceeds 16.
• Minimum frame duration seems to have increased.

[gkalua]

Asked TC about the panda default, he confimed our suspicions and the work-around. Changes to be made in the panda code to allow it to make a new row when the no. of repeats exeeds the limit.

Asked TC about the proper way to instantiate the FastCS-Eiger-Odin detector, have to ask SS, but researching what was, what is, and what should be is a good place to start. i07 and i15-1 might have the newest changes in their 'dodals'

Did quick research to understand the fuss behind FastCS-Eiger-Odin. Essentially, the eiger is a detector (handling its control), the odin does the file writing for the eiger. They are both first layer abstracted in fastcs, then in ophyd-async another layer of hardware abstraction occurs.

The eiger dectector within ophyd-async inherits from standard detector and has od: Odin.OdinIO as a class parameter. od is also used in the class method.

In essence this structure allows the eiger and odin to act as one virtual detector handling two parts of one process.