search

bluesky / ophyd-async

Hardware abstraction for bluesky written using asyncio
https://blueskyproject.io/ophyd-async
BSD 3-Clause "New" or "Revised" License
11 stars 24 forks source link

Kickoff/Complete/Collect crashes with an attribute error when using internal trigger on standard det #603

Open jwlodek opened 3 weeks ago

jwlodek commented 3 weeks ago

When using a standard detector in internal trigger mode, trying to use kickoff/complete/collect, will crash with an attribute error because self._fly_start is never set because of this conditional: https://github.com/bluesky/ophyd-async/blob/e291396be1ae795ad153f7c8242243f462cc30e5/src/ophyd_async/core/_detector.py#L331

Is there a reason we need this conditional? Everything works fine if I just remove it. Using kickoff/complete/collect with internal trigger is something I want to do because we want to be able to re-use the same plan for both the everything is hardware triggered experiment, and the everything is timestamped experiment.

Steps To Reproduce

Steps to reproduce the behavior:

  1. Create standard detector.
  2. Prepare in internal trigger mode.
  3. Run kickoff/complete/collect
  4. See traceback.

Acceptance Criteria

ZohebShaikh commented 2 weeks ago

This change was made in response to the comment here: https://github.com/bluesky/ophyd-async/pull/568#discussion_r1756674086. @coretl may have more context on the necessity of this requirement.

coretl commented 2 weeks ago

We want to be able to parallelize the slow operations of our flyscan:

But actually we might want to do the same flyscan 10 times at different energies, so we decided that we should be able to arm for multiple runs at the start:

We decided that we should put the arm in prepare for this, but actually I'm beginning to think we should put the arm back in kickoff after all. Then the first detector kickoff of the 10 will arm for 10x flyscans, and subsequent ones will not have to do anything.

@ZohebShaikh I'm not sure if we explored that as an option?