In areadetector hdf writer if capture fails, then close() always fails

[stan-dot]

line 138 in areadetector/hdfwriter.py is when the error shows up, during running a plan with a ttrm2

[coretl]

A general problem is that if something goes wrong with the caput callback on Acquire or Capture, the awaiting it: https://github.com/bluesky/ophyd-async/blob/b153c5013f3ebb3f0d1697e96a33ed0746c95117/src/ophyd_async/epics/areadetector/writers/hdf_writer.py#L138 will cause an error. This will cause unstage to error.

We need to investigate whether this is just noise, and we should make a better error message, or if it stops anything working