pysmurf-controller: Update sodetlib version to v0.6.2#976
pysmurf-controller: Update sodetlib version to v0.6.2#976BrianJKoopman merged 3 commits intomainfrom
Conversation
New sodetlib version fixes bug in time constant analysis for detectors near saturation, and provides additional configuration parameters to control amplifier biasing.
|
@BrianJKoopman seems like simonsobs/sodetlib#494 caused this test image build to fail: [ocs-pysmurf-agent 5/9] RUN pip3 install -e . |
Can you try dropping the I'm pushing towards the PyPI installation for this, so we'll start installing from there soon. I just opened simonsobs/sodetlib#497. |
BrianJKoopman
left a comment
There was a problem hiding this comment.
This looks good to me. Let me know how you want to handle the one comment I made about the -e flag.
Also, is there any effect on the 'user' end for this agent? Config changes needed or anything?
New sodetlib version fixes bug in time constant analysis for detectors near saturation, and provides additional configuration parameters to control amplifier biasing.
Description
fit_tmin is again available as a parameter that can be specified. If left as None, dynamic tmin calculation is used.
Dynamic tmin calculation rewritten to rely on the stored polarity of a detector's response, rather than abs(signal)
For amplifier bias parameters,
amp_{amp}_gate_volt_{min/max}are added to the dev cfg and referenced in the amplifier biasing routines.Motivation and Context
Issues with time constant fitting of near-saturated detectors shown here:
https://docs.google.com/presentation/d/1oGefvf0J7R-OqJHX4xY_ArVm5rNWToZq-wb7sOqg1t8/edit?slide=id.g3b710008148_0_20#slide=id.g3b710008148_0_20
How Has This Been Tested?
Kyohei and Daniel have re-run bias step analyses and saw the spuriously-low time constants disappear, with results similar to the old fixed tmin=1.5 ms results.
Types of changes
Checklist: